Article

THE DIAGNOSIS: Apocrine Hidrocystoma

Histology for all 3 lesions demonstrated similar cystic structures lined by a dual layer of epithelial cells, with the outermost layer composed of flattened myoepithelial cells and the inner layer composed of cells with apocrine features (Figure 1). Based on these findings, a diagnosis of apocrine hidrocystoma was made. The patient underwent successful surgical excision shortly thereafter without recurrence at follow-up 1 year later. 

Apocrine hidrocystomas are rare benign cystic lesions that are considered to be adenomatous proliferations of apocrine glands. They typically manifest as solitary asymptomatic lesions measuring 3 to 15 mm.¹ They tend to appear on the face, usually in the periorbital region, but also have been described on the neck, scalp, trunk, arms, and legs.^2-4 Multiple apocrine hidrocystomas can be a marker of 2 rare inherited disorders: Gorlin-Goltz syndrome and Schopf-Schulz-Passarge syndrome.⁵ Apocrine hidrocystomas may be flesh colored or may have a blue, black, or brown appearance due to the Tyndall effect, in which light with shorter wavelengths is scattered by the contents of the lesions.² Histologically, apocrine hidrocystomas are cysts lined by a dual layer of epithelial cells. The inner layer is composed of cells with apocrine features, and the outer layer is composed of flattened myoepithelial cells. Due to their range of colors and predilection for sun-exposed surfaces, apocrine hidrocystomas may be mistaken for various malignant neoplasms, including melanoma.^6,7

The differential diagnosis for our patient included agminated blue nevi, melanoma, pigmented basal cell carcinoma (BCC), and seborrheic keratosis. A blue nevus is a dermal melanocytic lesion that manifests as a well-demarcated, blue to blue-black papule that typically appears on the face, scalp, arms, legs, lower back, and buttocks. Although there are several histologic subtypes, the common blue nevus usually manifests as a solitary lesion measuring less than 1 cm, often developing during childhood to young adulthood.⁸ Histologically, common blue nevi are characterized by a dermal proliferation of deeply pigmented bipolar spindled melanocytes embedded in thickened collagen bundles, often with scattered epithelioid melanophages, and no conspicuous mitotic activity (Figure 2).⁹ There are other types of blue nevi, including cellular blue nevi, which tend to be larger and manifest commonly on the buttocks and sacrococcygeal region in early adulthood.⁹ Histologically, cellular blue nevi contain oval to spindled melanocytes with scattered melanophages forming a well-demarcated nodule typically in the reticular dermis. There may be bulbous extension into the subcutaneous adipose tissue. Occasional mitoses may be seen.^9,10 Melanoma can arise from common or cellular blue nevi, though it more frequently occurs with cellular blue nevi. Other subtypes of blue nevi have been described, including the sclerosing, plaque-type, combined, hypomelanotic/amelanotic, and pigmented epithelioid melanocytoma.¹¹ However, they typically have features of the common blue nevus or cellular blue nevus, such as oval/spindle cell morphology, some degree of melanin, and biphasic architecture, but are classified according to their dominant histologic characteristics. 

Given the location of our patient’s lesions on the scalp and his extensive history of sun exposure, malignancy was high in the differential. Multiple synchronous primary melanomas including nodular melanoma, blue nevus–like metastatic melanoma, and metastatic melanoma were considered. The leg and the scalp have the highest reported incidence of cutaneous metastases of melanoma, with many cases presenting as dermal or subcutaneous nodules and eruptive blue nevus–like papules, similar to our patient’s clinical presentation.^12,13 Nodular melanoma (NM) is one of 4 major types of melanoma, accounting for approximately 15% to 30% of cases in the United States.¹⁴ Nodular melanoma typically manifests as a smooth, raised, symmetric, well-circumscribed lesion with variable pigmentation, from very dark to amelanotic. Histologically, NM is defined as a dermal mass, either in isolation or with an epidermal component, not to exceed 3 rete ridges beyond the dermal component.¹⁵ Tumor cells have a high cell density with pleomorphism, usually with atypical epithelioid cells with vesicular nuclei and irregular cytoplasm, and occasionally spindle cells (Figure 2).¹⁶ Mitoses and necrosis are frequent. Scalp location independently is responsible for worse survival, both overall and melanoma specific.¹⁷ Nodular melanoma tends to have greater Breslow thickness at diagnosis than other melanoma subtypes and often carries a worse prognosis. 

Malignant melanomas that develop from or in conjunction with or bear histologic resemblance to blue nevi are termed blue nevus–like melanoma or blue nevus–associated melanoma. These malignancies are exceedingly rare, accounting for only 0.3% of melanomas in one Turkey-based multicenter study.¹⁸ The histologic criteria for diagnosing blue nevus–like melanoma are poorly defined, and terminology of these lesions has led to some debate in naming conventions.¹⁹ Nevertheless, unlike blue nevus, blue nevus–like melanoma demonstrates histologic features of malignancy, including pleomorphism, prominent nucleoli, mitotic activity, vascular invasion, and potential necrosis.¹⁰ The lack of an inflammatory infiltrate, surrounding fibrosis, junctional activity, and pre-existing nevus can help distinguish cutaneous melanoma metastases from primary nodular melanoma. Immunohistochemical stains such as S100, Melan-A/MART1, or SOX-10 can help confirm melanocytic lineage.¹² 

Pigmented BCC is a clinical and histologic variant of BCC characterized by increased melanin pigmentation due to melanocytes admixed with tumor cells. Dermoscopically, the pigment can have a maple leaf–like appearance with spoke-wheel areas, in-focus dots, and concentric structures at the dermoepidermal junction, which is more characteristic of superficial and infiltrating BCC.²⁰ In nodular BCC, the pigment occurs as blue-gray ovoid nests and globules in deeper layers of the dermis.²⁰ 

Seborrheic keratoses (SKs) can vary widely in clinical appearance, with pigmentation ranging from flesh colored to yellow to brown to black. Melanoacanthomas are acanthotic SKs that are highly pigmented due to intermixed epidermal melanocytes and subepidermal melanophages.²¹ Dermoscopy can help distinguish cutaneous malignancies from SKs, which often demonstrate fissures and ridges, comedolike openings, and milialike cysts. Biopsy sometimes is required to assess for malignancy, as was the case in our patient. The classic histologic features of SKs include acanthosis, papillomatosis, and hyperkeratosis.²² 

This case highlights the need to consider apocrine hidrocystoma, along with malignancy, in the differential diagnosis of pigmented cystic masses of the face and scalp. Because apocrine hidrocystomas are benign, they do not need to be treated but often are surgically excised for cosmesis or complete histopathologic examination. Destruction via electrodessication, carbon dioxide ablation, trichloroacetic acid chemical ablation, botulinum toxin injection, and anticholinergic creams sometimes is used, especially for cosmetic treatment of multiple small lesions.⁵ Our patient was treated with surgical excision with no evidence of recurrence on follow-up 1 year later. 

THE DIAGNOSIS: Apocrine Hidrocystoma

Histology for all 3 lesions demonstrated similar cystic structures lined by a dual layer of epithelial cells, with the outermost layer composed of flattened myoepithelial cells and the inner layer composed of cells with apocrine features (Figure 1). Based on these findings, a diagnosis of apocrine hidrocystoma was made. The patient underwent successful surgical excision shortly thereafter without recurrence at follow-up 1 year later. 

Apocrine hidrocystomas are rare benign cystic lesions that are considered to be adenomatous proliferations of apocrine glands. They typically manifest as solitary asymptomatic lesions measuring 3 to 15 mm.¹ They tend to appear on the face, usually in the periorbital region, but also have been described on the neck, scalp, trunk, arms, and legs.^2-4 Multiple apocrine hidrocystomas can be a marker of 2 rare inherited disorders: Gorlin-Goltz syndrome and Schopf-Schulz-Passarge syndrome.⁵ Apocrine hidrocystomas may be flesh colored or may have a blue, black, or brown appearance due to the Tyndall effect, in which light with shorter wavelengths is scattered by the contents of the lesions.² Histologically, apocrine hidrocystomas are cysts lined by a dual layer of epithelial cells. The inner layer is composed of cells with apocrine features, and the outer layer is composed of flattened myoepithelial cells. Due to their range of colors and predilection for sun-exposed surfaces, apocrine hidrocystomas may be mistaken for various malignant neoplasms, including melanoma.^6,7

The differential diagnosis for our patient included agminated blue nevi, melanoma, pigmented basal cell carcinoma (BCC), and seborrheic keratosis. A blue nevus is a dermal melanocytic lesion that manifests as a well-demarcated, blue to blue-black papule that typically appears on the face, scalp, arms, legs, lower back, and buttocks. Although there are several histologic subtypes, the common blue nevus usually manifests as a solitary lesion measuring less than 1 cm, often developing during childhood to young adulthood.⁸ Histologically, common blue nevi are characterized by a dermal proliferation of deeply pigmented bipolar spindled melanocytes embedded in thickened collagen bundles, often with scattered epithelioid melanophages, and no conspicuous mitotic activity (Figure 2).⁹ There are other types of blue nevi, including cellular blue nevi, which tend to be larger and manifest commonly on the buttocks and sacrococcygeal region in early adulthood.⁹ Histologically, cellular blue nevi contain oval to spindled melanocytes with scattered melanophages forming a well-demarcated nodule typically in the reticular dermis. There may be bulbous extension into the subcutaneous adipose tissue. Occasional mitoses may be seen.^9,10 Melanoma can arise from common or cellular blue nevi, though it more frequently occurs with cellular blue nevi. Other subtypes of blue nevi have been described, including the sclerosing, plaque-type, combined, hypomelanotic/amelanotic, and pigmented epithelioid melanocytoma.¹¹ However, they typically have features of the common blue nevus or cellular blue nevus, such as oval/spindle cell morphology, some degree of melanin, and biphasic architecture, but are classified according to their dominant histologic characteristics. 

Given the location of our patient’s lesions on the scalp and his extensive history of sun exposure, malignancy was high in the differential. Multiple synchronous primary melanomas including nodular melanoma, blue nevus–like metastatic melanoma, and metastatic melanoma were considered. The leg and the scalp have the highest reported incidence of cutaneous metastases of melanoma, with many cases presenting as dermal or subcutaneous nodules and eruptive blue nevus–like papules, similar to our patient’s clinical presentation.^12,13 Nodular melanoma (NM) is one of 4 major types of melanoma, accounting for approximately 15% to 30% of cases in the United States.¹⁴ Nodular melanoma typically manifests as a smooth, raised, symmetric, well-circumscribed lesion with variable pigmentation, from very dark to amelanotic. Histologically, NM is defined as a dermal mass, either in isolation or with an epidermal component, not to exceed 3 rete ridges beyond the dermal component.¹⁵ Tumor cells have a high cell density with pleomorphism, usually with atypical epithelioid cells with vesicular nuclei and irregular cytoplasm, and occasionally spindle cells (Figure 2).¹⁶ Mitoses and necrosis are frequent. Scalp location independently is responsible for worse survival, both overall and melanoma specific.¹⁷ Nodular melanoma tends to have greater Breslow thickness at diagnosis than other melanoma subtypes and often carries a worse prognosis. 

Malignant melanomas that develop from or in conjunction with or bear histologic resemblance to blue nevi are termed blue nevus–like melanoma or blue nevus–associated melanoma. These malignancies are exceedingly rare, accounting for only 0.3% of melanomas in one Turkey-based multicenter study.¹⁸ The histologic criteria for diagnosing blue nevus–like melanoma are poorly defined, and terminology of these lesions has led to some debate in naming conventions.¹⁹ Nevertheless, unlike blue nevus, blue nevus–like melanoma demonstrates histologic features of malignancy, including pleomorphism, prominent nucleoli, mitotic activity, vascular invasion, and potential necrosis.¹⁰ The lack of an inflammatory infiltrate, surrounding fibrosis, junctional activity, and pre-existing nevus can help distinguish cutaneous melanoma metastases from primary nodular melanoma. Immunohistochemical stains such as S100, Melan-A/MART1, or SOX-10 can help confirm melanocytic lineage.¹² 

Pigmented BCC is a clinical and histologic variant of BCC characterized by increased melanin pigmentation due to melanocytes admixed with tumor cells. Dermoscopically, the pigment can have a maple leaf–like appearance with spoke-wheel areas, in-focus dots, and concentric structures at the dermoepidermal junction, which is more characteristic of superficial and infiltrating BCC.²⁰ In nodular BCC, the pigment occurs as blue-gray ovoid nests and globules in deeper layers of the dermis.²⁰ 

Seborrheic keratoses (SKs) can vary widely in clinical appearance, with pigmentation ranging from flesh colored to yellow to brown to black. Melanoacanthomas are acanthotic SKs that are highly pigmented due to intermixed epidermal melanocytes and subepidermal melanophages.²¹ Dermoscopy can help distinguish cutaneous malignancies from SKs, which often demonstrate fissures and ridges, comedolike openings, and milialike cysts. Biopsy sometimes is required to assess for malignancy, as was the case in our patient. The classic histologic features of SKs include acanthosis, papillomatosis, and hyperkeratosis.²² 

This case highlights the need to consider apocrine hidrocystoma, along with malignancy, in the differential diagnosis of pigmented cystic masses of the face and scalp. Because apocrine hidrocystomas are benign, they do not need to be treated but often are surgically excised for cosmesis or complete histopathologic examination. Destruction via electrodessication, carbon dioxide ablation, trichloroacetic acid chemical ablation, botulinum toxin injection, and anticholinergic creams sometimes is used, especially for cosmetic treatment of multiple small lesions.⁵ Our patient was treated with surgical excision with no evidence of recurrence on follow-up 1 year later. 

THE DIAGNOSIS: Apocrine Hidrocystoma

Histology for all 3 lesions demonstrated similar cystic structures lined by a dual layer of epithelial cells, with the outermost layer composed of flattened myoepithelial cells and the inner layer composed of cells with apocrine features (Figure 1). Based on these findings, a diagnosis of apocrine hidrocystoma was made. The patient underwent successful surgical excision shortly thereafter without recurrence at follow-up 1 year later. 

Apocrine hidrocystomas are rare benign cystic lesions that are considered to be adenomatous proliferations of apocrine glands. They typically manifest as solitary asymptomatic lesions measuring 3 to 15 mm.¹ They tend to appear on the face, usually in the periorbital region, but also have been described on the neck, scalp, trunk, arms, and legs.^2-4 Multiple apocrine hidrocystomas can be a marker of 2 rare inherited disorders: Gorlin-Goltz syndrome and Schopf-Schulz-Passarge syndrome.⁵ Apocrine hidrocystomas may be flesh colored or may have a blue, black, or brown appearance due to the Tyndall effect, in which light with shorter wavelengths is scattered by the contents of the lesions.² Histologically, apocrine hidrocystomas are cysts lined by a dual layer of epithelial cells. The inner layer is composed of cells with apocrine features, and the outer layer is composed of flattened myoepithelial cells. Due to their range of colors and predilection for sun-exposed surfaces, apocrine hidrocystomas may be mistaken for various malignant neoplasms, including melanoma.^6,7

The differential diagnosis for our patient included agminated blue nevi, melanoma, pigmented basal cell carcinoma (BCC), and seborrheic keratosis. A blue nevus is a dermal melanocytic lesion that manifests as a well-demarcated, blue to blue-black papule that typically appears on the face, scalp, arms, legs, lower back, and buttocks. Although there are several histologic subtypes, the common blue nevus usually manifests as a solitary lesion measuring less than 1 cm, often developing during childhood to young adulthood.⁸ Histologically, common blue nevi are characterized by a dermal proliferation of deeply pigmented bipolar spindled melanocytes embedded in thickened collagen bundles, often with scattered epithelioid melanophages, and no conspicuous mitotic activity (Figure 2).⁹ There are other types of blue nevi, including cellular blue nevi, which tend to be larger and manifest commonly on the buttocks and sacrococcygeal region in early adulthood.⁹ Histologically, cellular blue nevi contain oval to spindled melanocytes with scattered melanophages forming a well-demarcated nodule typically in the reticular dermis. There may be bulbous extension into the subcutaneous adipose tissue. Occasional mitoses may be seen.^9,10 Melanoma can arise from common or cellular blue nevi, though it more frequently occurs with cellular blue nevi. Other subtypes of blue nevi have been described, including the sclerosing, plaque-type, combined, hypomelanotic/amelanotic, and pigmented epithelioid melanocytoma.¹¹ However, they typically have features of the common blue nevus or cellular blue nevus, such as oval/spindle cell morphology, some degree of melanin, and biphasic architecture, but are classified according to their dominant histologic characteristics. 

Given the location of our patient’s lesions on the scalp and his extensive history of sun exposure, malignancy was high in the differential. Multiple synchronous primary melanomas including nodular melanoma, blue nevus–like metastatic melanoma, and metastatic melanoma were considered. The leg and the scalp have the highest reported incidence of cutaneous metastases of melanoma, with many cases presenting as dermal or subcutaneous nodules and eruptive blue nevus–like papules, similar to our patient’s clinical presentation.^12,13 Nodular melanoma (NM) is one of 4 major types of melanoma, accounting for approximately 15% to 30% of cases in the United States.¹⁴ Nodular melanoma typically manifests as a smooth, raised, symmetric, well-circumscribed lesion with variable pigmentation, from very dark to amelanotic. Histologically, NM is defined as a dermal mass, either in isolation or with an epidermal component, not to exceed 3 rete ridges beyond the dermal component.¹⁵ Tumor cells have a high cell density with pleomorphism, usually with atypical epithelioid cells with vesicular nuclei and irregular cytoplasm, and occasionally spindle cells (Figure 2).¹⁶ Mitoses and necrosis are frequent. Scalp location independently is responsible for worse survival, both overall and melanoma specific.¹⁷ Nodular melanoma tends to have greater Breslow thickness at diagnosis than other melanoma subtypes and often carries a worse prognosis. 

Malignant melanomas that develop from or in conjunction with or bear histologic resemblance to blue nevi are termed blue nevus–like melanoma or blue nevus–associated melanoma. These malignancies are exceedingly rare, accounting for only 0.3% of melanomas in one Turkey-based multicenter study.¹⁸ The histologic criteria for diagnosing blue nevus–like melanoma are poorly defined, and terminology of these lesions has led to some debate in naming conventions.¹⁹ Nevertheless, unlike blue nevus, blue nevus–like melanoma demonstrates histologic features of malignancy, including pleomorphism, prominent nucleoli, mitotic activity, vascular invasion, and potential necrosis.¹⁰ The lack of an inflammatory infiltrate, surrounding fibrosis, junctional activity, and pre-existing nevus can help distinguish cutaneous melanoma metastases from primary nodular melanoma. Immunohistochemical stains such as S100, Melan-A/MART1, or SOX-10 can help confirm melanocytic lineage.¹² 

Pigmented BCC is a clinical and histologic variant of BCC characterized by increased melanin pigmentation due to melanocytes admixed with tumor cells. Dermoscopically, the pigment can have a maple leaf–like appearance with spoke-wheel areas, in-focus dots, and concentric structures at the dermoepidermal junction, which is more characteristic of superficial and infiltrating BCC.²⁰ In nodular BCC, the pigment occurs as blue-gray ovoid nests and globules in deeper layers of the dermis.²⁰ 

Seborrheic keratoses (SKs) can vary widely in clinical appearance, with pigmentation ranging from flesh colored to yellow to brown to black. Melanoacanthomas are acanthotic SKs that are highly pigmented due to intermixed epidermal melanocytes and subepidermal melanophages.²¹ Dermoscopy can help distinguish cutaneous malignancies from SKs, which often demonstrate fissures and ridges, comedolike openings, and milialike cysts. Biopsy sometimes is required to assess for malignancy, as was the case in our patient. The classic histologic features of SKs include acanthosis, papillomatosis, and hyperkeratosis.²² 

This case highlights the need to consider apocrine hidrocystoma, along with malignancy, in the differential diagnosis of pigmented cystic masses of the face and scalp. Because apocrine hidrocystomas are benign, they do not need to be treated but often are surgically excised for cosmesis or complete histopathologic examination. Destruction via electrodessication, carbon dioxide ablation, trichloroacetic acid chemical ablation, botulinum toxin injection, and anticholinergic creams sometimes is used, especially for cosmetic treatment of multiple small lesions.⁵ Our patient was treated with surgical excision with no evidence of recurrence on follow-up 1 year later. 

To the Editor:

A 34-year-old Brazilian woman presented to the dermatology department with pruritic lesions on the neck and chest that had been present since adolescence. She reported a family history of Darier disease in her father. Physical examination revealed erythematous follicular papules on the neck, inframammary region, and abdomen (Figure 1A), as well as longitudinal bandlike leukonychia and distal nail splits on the fingernails (Figure 1B). Histopathology of a lesion on the back revealed compact hyperkeratosis and parakeratosis above an acantholytic cleft accompanied by dyskeratotic keratinocytes, including some corps ronds and grains, which supported the clinical impression of Darier disease (Figure 2). The typical clinical presentation along with the family history and histopathology confirmed the diagnosis. After therapeutic failure with topical corticosteroids and oral antibiotics for 3 months, low-dose oral naltrexone (4.5 mg/d) as monotherapy noticeably improved the lesions and pruritus within 2 months, with near-complete regression at 6 months, achieving disease stability (Figures 1C and 1D). The patient remained stable with no recurrence after 1 year of follow-up.

Darier disease is an autosomal-dominant genodermatosis caused by a mutation in the ATP2A2 gene, which encodes the sarco/endoplasmic reticulum calcium ATPase, leading to defective intracellular calcium signaling and alterations in epidermal adhesion and keratinization.¹ Darier disease typically begins in adolescence and is aggravated by exposure to heat and friction. It is characterized by seborrheic distribution of painful and pruritic red-brown keratotic papules. Nail manifestations include longitudinal ridges—erythronychia and/or leukonychia—and grooves that end in a V-shaped notch. The differential diagnosis includes Hailey-Hailey disease, psoriasis, and pityriasis rubra pilaris.^1,2 The diagnosis is clinical and is confirmed by histopathology, which reveals suprabasal cleavage, acantholytic dyskeratosis, corps ronds, and grains. Treatment options are limited and include corticosteroids, oral and/or topical antibiotics, and systemic retinoids.²

 

Oral naltrexone has been used in Darier disease based on its observed effectiveness in Hailey-Hailey disease, considering the histopathologic similarities and alterations in calcium homeostasis in both conditions. Low-dose oral naltrexone (1-5 mg/d) increases the expression of opioid receptors (δ, μ, κ), enhancing its immunomodulatory and antinociceptive effects. The δ opioid receptor regulates the expression of desmoglein, improving epidermal differentiation and wound healing.³ Activation of the δ and μ receptors increases intracellular calcium through the inositol phosphate pathway, which contributes to calcium homeostasis.⁴ Naltrexone blocks the nonopioid toll-like receptor 4 found in keratinocytes and macrophages, exerting an anti-inflammatory effect by reducing proinflammatory cytokines.³ Adverse events associated with low-dose naltrexone are minimal, mostly mild, and often related to sleep disorders^3,5; however, patients should undergo screening for prior opioid dependence, recent opioid usage, and signs of opioid withdrawal before initiating naltrexone treatment.⁵

Boehmer et al⁶ used naltrexone (4.5 mg/d) and oral magnesium (200 mg/d) in 6 patients with inconsistent results, except for 1 case that concurrently used acitretin (25 mg/d) with satisfactory improvement. Pessoa et al⁷ added naltrexone (4.5 mg/d) to oral isotretinoin (0.5 mg/kg/d) in 1 patient, resulting in notable improvement of lesions within 3 months. 

In our patient with Darier disease, low-dose naltrexone demonstrated a substantial response as monotherapy after 2 months of treatment and nearly complete regression of lesions within 6 months, with no reported side effects after 1 year of follow-up. The use of low-dose naltrexone could be a promising and safe treatment option as monotherapy or in combination with conventional therapy for Darier disease; however, further studies are needed.

To the Editor:

A 34-year-old Brazilian woman presented to the dermatology department with pruritic lesions on the neck and chest that had been present since adolescence. She reported a family history of Darier disease in her father. Physical examination revealed erythematous follicular papules on the neck, inframammary region, and abdomen (Figure 1A), as well as longitudinal bandlike leukonychia and distal nail splits on the fingernails (Figure 1B). Histopathology of a lesion on the back revealed compact hyperkeratosis and parakeratosis above an acantholytic cleft accompanied by dyskeratotic keratinocytes, including some corps ronds and grains, which supported the clinical impression of Darier disease (Figure 2). The typical clinical presentation along with the family history and histopathology confirmed the diagnosis. After therapeutic failure with topical corticosteroids and oral antibiotics for 3 months, low-dose oral naltrexone (4.5 mg/d) as monotherapy noticeably improved the lesions and pruritus within 2 months, with near-complete regression at 6 months, achieving disease stability (Figures 1C and 1D). The patient remained stable with no recurrence after 1 year of follow-up.

Darier disease is an autosomal-dominant genodermatosis caused by a mutation in the ATP2A2 gene, which encodes the sarco/endoplasmic reticulum calcium ATPase, leading to defective intracellular calcium signaling and alterations in epidermal adhesion and keratinization.¹ Darier disease typically begins in adolescence and is aggravated by exposure to heat and friction. It is characterized by seborrheic distribution of painful and pruritic red-brown keratotic papules. Nail manifestations include longitudinal ridges—erythronychia and/or leukonychia—and grooves that end in a V-shaped notch. The differential diagnosis includes Hailey-Hailey disease, psoriasis, and pityriasis rubra pilaris.^1,2 The diagnosis is clinical and is confirmed by histopathology, which reveals suprabasal cleavage, acantholytic dyskeratosis, corps ronds, and grains. Treatment options are limited and include corticosteroids, oral and/or topical antibiotics, and systemic retinoids.²

 

Oral naltrexone has been used in Darier disease based on its observed effectiveness in Hailey-Hailey disease, considering the histopathologic similarities and alterations in calcium homeostasis in both conditions. Low-dose oral naltrexone (1-5 mg/d) increases the expression of opioid receptors (δ, μ, κ), enhancing its immunomodulatory and antinociceptive effects. The δ opioid receptor regulates the expression of desmoglein, improving epidermal differentiation and wound healing.³ Activation of the δ and μ receptors increases intracellular calcium through the inositol phosphate pathway, which contributes to calcium homeostasis.⁴ Naltrexone blocks the nonopioid toll-like receptor 4 found in keratinocytes and macrophages, exerting an anti-inflammatory effect by reducing proinflammatory cytokines.³ Adverse events associated with low-dose naltrexone are minimal, mostly mild, and often related to sleep disorders^3,5; however, patients should undergo screening for prior opioid dependence, recent opioid usage, and signs of opioid withdrawal before initiating naltrexone treatment.⁵

Boehmer et al⁶ used naltrexone (4.5 mg/d) and oral magnesium (200 mg/d) in 6 patients with inconsistent results, except for 1 case that concurrently used acitretin (25 mg/d) with satisfactory improvement. Pessoa et al⁷ added naltrexone (4.5 mg/d) to oral isotretinoin (0.5 mg/kg/d) in 1 patient, resulting in notable improvement of lesions within 3 months. 

In our patient with Darier disease, low-dose naltrexone demonstrated a substantial response as monotherapy after 2 months of treatment and nearly complete regression of lesions within 6 months, with no reported side effects after 1 year of follow-up. The use of low-dose naltrexone could be a promising and safe treatment option as monotherapy or in combination with conventional therapy for Darier disease; however, further studies are needed.

To the Editor:

A 34-year-old Brazilian woman presented to the dermatology department with pruritic lesions on the neck and chest that had been present since adolescence. She reported a family history of Darier disease in her father. Physical examination revealed erythematous follicular papules on the neck, inframammary region, and abdomen (Figure 1A), as well as longitudinal bandlike leukonychia and distal nail splits on the fingernails (Figure 1B). Histopathology of a lesion on the back revealed compact hyperkeratosis and parakeratosis above an acantholytic cleft accompanied by dyskeratotic keratinocytes, including some corps ronds and grains, which supported the clinical impression of Darier disease (Figure 2). The typical clinical presentation along with the family history and histopathology confirmed the diagnosis. After therapeutic failure with topical corticosteroids and oral antibiotics for 3 months, low-dose oral naltrexone (4.5 mg/d) as monotherapy noticeably improved the lesions and pruritus within 2 months, with near-complete regression at 6 months, achieving disease stability (Figures 1C and 1D). The patient remained stable with no recurrence after 1 year of follow-up.

Darier disease is an autosomal-dominant genodermatosis caused by a mutation in the ATP2A2 gene, which encodes the sarco/endoplasmic reticulum calcium ATPase, leading to defective intracellular calcium signaling and alterations in epidermal adhesion and keratinization.¹ Darier disease typically begins in adolescence and is aggravated by exposure to heat and friction. It is characterized by seborrheic distribution of painful and pruritic red-brown keratotic papules. Nail manifestations include longitudinal ridges—erythronychia and/or leukonychia—and grooves that end in a V-shaped notch. The differential diagnosis includes Hailey-Hailey disease, psoriasis, and pityriasis rubra pilaris.^1,2 The diagnosis is clinical and is confirmed by histopathology, which reveals suprabasal cleavage, acantholytic dyskeratosis, corps ronds, and grains. Treatment options are limited and include corticosteroids, oral and/or topical antibiotics, and systemic retinoids.²

 

Oral naltrexone has been used in Darier disease based on its observed effectiveness in Hailey-Hailey disease, considering the histopathologic similarities and alterations in calcium homeostasis in both conditions. Low-dose oral naltrexone (1-5 mg/d) increases the expression of opioid receptors (δ, μ, κ), enhancing its immunomodulatory and antinociceptive effects. The δ opioid receptor regulates the expression of desmoglein, improving epidermal differentiation and wound healing.³ Activation of the δ and μ receptors increases intracellular calcium through the inositol phosphate pathway, which contributes to calcium homeostasis.⁴ Naltrexone blocks the nonopioid toll-like receptor 4 found in keratinocytes and macrophages, exerting an anti-inflammatory effect by reducing proinflammatory cytokines.³ Adverse events associated with low-dose naltrexone are minimal, mostly mild, and often related to sleep disorders^3,5; however, patients should undergo screening for prior opioid dependence, recent opioid usage, and signs of opioid withdrawal before initiating naltrexone treatment.⁵

Boehmer et al⁶ used naltrexone (4.5 mg/d) and oral magnesium (200 mg/d) in 6 patients with inconsistent results, except for 1 case that concurrently used acitretin (25 mg/d) with satisfactory improvement. Pessoa et al⁷ added naltrexone (4.5 mg/d) to oral isotretinoin (0.5 mg/kg/d) in 1 patient, resulting in notable improvement of lesions within 3 months. 

In our patient with Darier disease, low-dose naltrexone demonstrated a substantial response as monotherapy after 2 months of treatment and nearly complete regression of lesions within 6 months, with no reported side effects after 1 year of follow-up. The use of low-dose naltrexone could be a promising and safe treatment option as monotherapy or in combination with conventional therapy for Darier disease; however, further studies are needed.

Peer-review, evidence-based, detailed gene/drug clinical practice guidelines suggest that genetic variations can impact how individuals metabolize medications, which is sometimes included in medication prescribing information.^1-3 Pharmacogenomic testing identifies genetic markers so medication selection and dosing can be tailored to each individual by identifying whether a specific medication is likely to be safe and effective prior to prescribing.⁴

Pharmacogenomics can be a valuable tool for personalizing medicine but has had suboptimal implementation since its discovery. The US Department of Veterans Affairs (VA) health care system reviewed the implementation of the Pharmacogenomic Testing for Veterans (PHASER) program. This review identified clinician barriers pre- and post-PHASER program implementation; staffing issues, competing clinical priorities, and inadequate PHASER program resources were the most frequently reported barriers to implementation of pharmacogenomic testing.⁵

Another evaluation of the implementation of the PHASER program that surveyed VA patients found that patients could be separated into 3 groups. Acceptors of pharmacogenomic testing emphasized potential health benefits of testing. Patients that declined testing often cited concerns for genetic information affecting insurance coverage, being misused, or being susceptible to data breach. The third group—identified as contemplators—reported the need for clinician outreach to impact their decision on whether or not to receive pharmacogenomic testing.⁶ These studies suggest that removing barriers by providing ample pharmacogenomics resources to clinicians, in addition to detailed training on how to offer and follow up with patients regarding pharmacogenomic testing, is crucial to successful implementation of the PHASER program.

PHASER

In 2019, the VA began working with Sanford Health to establish the PHASER program and offer pharmacogenomic testing. PHASER has since expanded to 25 VA medical centers, including the VA Central Ohio Healthcare System (VACOHCS).^7,8 Pharmacogenomic testing through PHASER is conducted using a standardized laboratory panel that includes 12 different medication classes.⁹ The drug classes include certain anti-infective, anticoagulant, antiplatelet, cardiovascular, cholesterol, gastrointestinal, mental health, neurological, oncology, pain, transplant, and other miscellaneous medications. Medications are correlated to each class and assessed for therapeutic impacts based on gene panel results.

Clinical recommendations for medication-gene interactions can range from monitoring for increased risk of adverse effects or therapeutic failure to recommending avoiding a medication. For example, patients who test positive for the HLA-B gene have significantly increased risk of hypersensitivity to abacavir, an HIV treatment.¹⁰

Similarly, patients who cannot adequately metabolize cytochrome P450 2C19 should consider avoiding clopidogrel as they are unlikely to convert clopidogrel to its active prodrug, which reduces its effectiveness.¹¹ Pharmacists can play a critical role educating patients about pharmacogenomic testing, especially within hematology and oncology.¹² Patients can benefit from this testing even if they are not currently taking medications with known concerns as they could be prescribed in the future. The SLCO1B1 gene-drug test, for example, can identify risk for statin-associated muscle symptoms.¹³

Clinical pharmacist practitioners (CPPs) can increase access to genetic testing because they interact with patients in a variety of settings and can order this laboratory test.^12,14 Recent research has demonstrated that most VA patients carry ≥ 1 genetic variant that may influence medication decisions and that half of veterans are prescribed a medication with known gene-drug interactions.¹⁵ CPP ordering of pharmacogenomic tests at the VACOHCS outpatient clinic was evaluated through collection of baseline data from March 8, 2023, to September 8, 2023. A goal was identified to increase orders by 50% for a patient care quality improvement initiative and use CPPs to increase access to pharmacogenomic testing. The purpose of this quality improvement initiative was to expand access to pharmacogenomic testing through process implementation and improvement within CPP-led clinic settings.

Gap Analysis

Lean Six Sigma A3 methodology was used to identify ways to increase the use of pharmacogenomic testing for veterans at VACOHCS and develop an improved process for increased ordering of pharmacogenomic testing. Lean Six Sigma A3 methodology is a stepwise approach to process improvement that helps identify gaps in efficiency, sustainable changes, and eliminate waste.¹⁶ Baseline data were collected from March 8, 2023, to September 8, 2023, to determine the frequency of CPPs ordering pharmacogenomic laboratory panels during clinic appointments. The ordering of pharmacogenomic panels was monitored by the VACOHCS PHASER coordinator.

CPPs were surveyed to identify perceived barriers to PHASER implementation. A gap analysis was conducted using Lean Six Sigma A3 methodology. Gap analyses use lean tools such as a Fishbone Diagram to illustrate and identify the gap between current state and ideal state. (Figure 1).The following barriers were identified: lack of clinician education materials, lack of a standardized patient screening process, time constraints on patient education and ordering, higher priority clinical needs, forgetting to order, lack of comfort with pharmacogenomics ordering and education, lack of support for the initiative, and increased workload and burnout. Among these perceived barriers, higher priority clinical needs, forgetting to order, and time constraints ranked highest in importance among CPPs. 

In line with Lean Six Sigma A3 methodology, several tests of change were used to improve pharmacogenomic testing ordering. These changes focused on increasing patient and clinician awareness, facilitating discussion, educating clinicians, and simplifying documentation to ease time constraints. Several strategies were employed postimplementation (Figure 2). Prefilled templates simplified documentation. These templates helped identify patients without pharmacogenomic testing, provided reminders, and saved documentation time during visits. CPPs also received training and materials on PHASER ordering and documentation within encounter notes. Additionally, patient-directed advertisements were displayed in CPP examination rooms to help inspire and facilitate discussion between veterans and CPPs.

Process Improvement Data

The quality improvement project goal was to increase PHASER orders by 50% after 3 months. PHASER orders increased from 87 at baseline (March 8, 2023, to September 8, 2023) to 196 during the intervention (November 16, 2023, to February 16, 2024), a 125% increase. Changes were consistent and sustained with 65 orders the first month, 67 orders the second month, and 64 orders the third month.

Discussion

Using Lean Six Sigma A3 methodology for a quality improvement process to increase PHASER orders by CPPs revealed barriers and guided potential solutions to overcome these barriers. Interventions included additional CPP training and ordering, tools for easier identification of potential patients, documentation best practices, patient-directed advertisements to facilitate conversations. These interventions required about 8 hours for preparation, distribution, development, and interpretation of surveys, education, and documentation materials. The financial impact of these interventions was already included in allotted office materials budgeted and provided. Additional funding was not needed to provide patient-directed advertisements or education materials. The VACOHCS pharmacogenomics CPP discusses PHASER test results with patients at a separate appointment.

Future directions include educating other CPPs to assist in discussing results with veterans. Overall, the changes implemented to improve the PHASER ordering process were low effort and exemplify the ease of streamlining future initiatives, allowing for sustained optimal implementation of pharmacogenomic testing.

Conclusions

A quality improvement initiative resulted in increased PHASER orders and a clearly defined process, allowing for a continued increase and sustained support. Perceived barriers were identified, and the changes implemented were often low effort but exhibited a sustained impact. The insights gleaned from this process will shape future process development initiatives and continue to sustain pharmacogenomic testing ordering by CPPs. This process will be extended to other VACOHCS clinical departments to further support increased access to pharmacogenomic testing, reduce medication trial and error, and reduce hospitalizations from adverse effects for veterans.

Peer-review, evidence-based, detailed gene/drug clinical practice guidelines suggest that genetic variations can impact how individuals metabolize medications, which is sometimes included in medication prescribing information.^1-3 Pharmacogenomic testing identifies genetic markers so medication selection and dosing can be tailored to each individual by identifying whether a specific medication is likely to be safe and effective prior to prescribing.⁴

Pharmacogenomics can be a valuable tool for personalizing medicine but has had suboptimal implementation since its discovery. The US Department of Veterans Affairs (VA) health care system reviewed the implementation of the Pharmacogenomic Testing for Veterans (PHASER) program. This review identified clinician barriers pre- and post-PHASER program implementation; staffing issues, competing clinical priorities, and inadequate PHASER program resources were the most frequently reported barriers to implementation of pharmacogenomic testing.⁵

Another evaluation of the implementation of the PHASER program that surveyed VA patients found that patients could be separated into 3 groups. Acceptors of pharmacogenomic testing emphasized potential health benefits of testing. Patients that declined testing often cited concerns for genetic information affecting insurance coverage, being misused, or being susceptible to data breach. The third group—identified as contemplators—reported the need for clinician outreach to impact their decision on whether or not to receive pharmacogenomic testing.⁶ These studies suggest that removing barriers by providing ample pharmacogenomics resources to clinicians, in addition to detailed training on how to offer and follow up with patients regarding pharmacogenomic testing, is crucial to successful implementation of the PHASER program.

PHASER

In 2019, the VA began working with Sanford Health to establish the PHASER program and offer pharmacogenomic testing. PHASER has since expanded to 25 VA medical centers, including the VA Central Ohio Healthcare System (VACOHCS).^7,8 Pharmacogenomic testing through PHASER is conducted using a standardized laboratory panel that includes 12 different medication classes.⁹ The drug classes include certain anti-infective, anticoagulant, antiplatelet, cardiovascular, cholesterol, gastrointestinal, mental health, neurological, oncology, pain, transplant, and other miscellaneous medications. Medications are correlated to each class and assessed for therapeutic impacts based on gene panel results.

Clinical recommendations for medication-gene interactions can range from monitoring for increased risk of adverse effects or therapeutic failure to recommending avoiding a medication. For example, patients who test positive for the HLA-B gene have significantly increased risk of hypersensitivity to abacavir, an HIV treatment.¹⁰

Similarly, patients who cannot adequately metabolize cytochrome P450 2C19 should consider avoiding clopidogrel as they are unlikely to convert clopidogrel to its active prodrug, which reduces its effectiveness.¹¹ Pharmacists can play a critical role educating patients about pharmacogenomic testing, especially within hematology and oncology.¹² Patients can benefit from this testing even if they are not currently taking medications with known concerns as they could be prescribed in the future. The SLCO1B1 gene-drug test, for example, can identify risk for statin-associated muscle symptoms.¹³

Clinical pharmacist practitioners (CPPs) can increase access to genetic testing because they interact with patients in a variety of settings and can order this laboratory test.^12,14 Recent research has demonstrated that most VA patients carry ≥ 1 genetic variant that may influence medication decisions and that half of veterans are prescribed a medication with known gene-drug interactions.¹⁵ CPP ordering of pharmacogenomic tests at the VACOHCS outpatient clinic was evaluated through collection of baseline data from March 8, 2023, to September 8, 2023. A goal was identified to increase orders by 50% for a patient care quality improvement initiative and use CPPs to increase access to pharmacogenomic testing. The purpose of this quality improvement initiative was to expand access to pharmacogenomic testing through process implementation and improvement within CPP-led clinic settings.

Gap Analysis

Lean Six Sigma A3 methodology was used to identify ways to increase the use of pharmacogenomic testing for veterans at VACOHCS and develop an improved process for increased ordering of pharmacogenomic testing. Lean Six Sigma A3 methodology is a stepwise approach to process improvement that helps identify gaps in efficiency, sustainable changes, and eliminate waste.¹⁶ Baseline data were collected from March 8, 2023, to September 8, 2023, to determine the frequency of CPPs ordering pharmacogenomic laboratory panels during clinic appointments. The ordering of pharmacogenomic panels was monitored by the VACOHCS PHASER coordinator.

CPPs were surveyed to identify perceived barriers to PHASER implementation. A gap analysis was conducted using Lean Six Sigma A3 methodology. Gap analyses use lean tools such as a Fishbone Diagram to illustrate and identify the gap between current state and ideal state. (Figure 1).The following barriers were identified: lack of clinician education materials, lack of a standardized patient screening process, time constraints on patient education and ordering, higher priority clinical needs, forgetting to order, lack of comfort with pharmacogenomics ordering and education, lack of support for the initiative, and increased workload and burnout. Among these perceived barriers, higher priority clinical needs, forgetting to order, and time constraints ranked highest in importance among CPPs. 

In line with Lean Six Sigma A3 methodology, several tests of change were used to improve pharmacogenomic testing ordering. These changes focused on increasing patient and clinician awareness, facilitating discussion, educating clinicians, and simplifying documentation to ease time constraints. Several strategies were employed postimplementation (Figure 2). Prefilled templates simplified documentation. These templates helped identify patients without pharmacogenomic testing, provided reminders, and saved documentation time during visits. CPPs also received training and materials on PHASER ordering and documentation within encounter notes. Additionally, patient-directed advertisements were displayed in CPP examination rooms to help inspire and facilitate discussion between veterans and CPPs.

Process Improvement Data

The quality improvement project goal was to increase PHASER orders by 50% after 3 months. PHASER orders increased from 87 at baseline (March 8, 2023, to September 8, 2023) to 196 during the intervention (November 16, 2023, to February 16, 2024), a 125% increase. Changes were consistent and sustained with 65 orders the first month, 67 orders the second month, and 64 orders the third month.

Discussion

Using Lean Six Sigma A3 methodology for a quality improvement process to increase PHASER orders by CPPs revealed barriers and guided potential solutions to overcome these barriers. Interventions included additional CPP training and ordering, tools for easier identification of potential patients, documentation best practices, patient-directed advertisements to facilitate conversations. These interventions required about 8 hours for preparation, distribution, development, and interpretation of surveys, education, and documentation materials. The financial impact of these interventions was already included in allotted office materials budgeted and provided. Additional funding was not needed to provide patient-directed advertisements or education materials. The VACOHCS pharmacogenomics CPP discusses PHASER test results with patients at a separate appointment.

Future directions include educating other CPPs to assist in discussing results with veterans. Overall, the changes implemented to improve the PHASER ordering process were low effort and exemplify the ease of streamlining future initiatives, allowing for sustained optimal implementation of pharmacogenomic testing.

Conclusions

A quality improvement initiative resulted in increased PHASER orders and a clearly defined process, allowing for a continued increase and sustained support. Perceived barriers were identified, and the changes implemented were often low effort but exhibited a sustained impact. The insights gleaned from this process will shape future process development initiatives and continue to sustain pharmacogenomic testing ordering by CPPs. This process will be extended to other VACOHCS clinical departments to further support increased access to pharmacogenomic testing, reduce medication trial and error, and reduce hospitalizations from adverse effects for veterans.

Peer-review, evidence-based, detailed gene/drug clinical practice guidelines suggest that genetic variations can impact how individuals metabolize medications, which is sometimes included in medication prescribing information.^1-3 Pharmacogenomic testing identifies genetic markers so medication selection and dosing can be tailored to each individual by identifying whether a specific medication is likely to be safe and effective prior to prescribing.⁴

Pharmacogenomics can be a valuable tool for personalizing medicine but has had suboptimal implementation since its discovery. The US Department of Veterans Affairs (VA) health care system reviewed the implementation of the Pharmacogenomic Testing for Veterans (PHASER) program. This review identified clinician barriers pre- and post-PHASER program implementation; staffing issues, competing clinical priorities, and inadequate PHASER program resources were the most frequently reported barriers to implementation of pharmacogenomic testing.⁵

Another evaluation of the implementation of the PHASER program that surveyed VA patients found that patients could be separated into 3 groups. Acceptors of pharmacogenomic testing emphasized potential health benefits of testing. Patients that declined testing often cited concerns for genetic information affecting insurance coverage, being misused, or being susceptible to data breach. The third group—identified as contemplators—reported the need for clinician outreach to impact their decision on whether or not to receive pharmacogenomic testing.⁶ These studies suggest that removing barriers by providing ample pharmacogenomics resources to clinicians, in addition to detailed training on how to offer and follow up with patients regarding pharmacogenomic testing, is crucial to successful implementation of the PHASER program.

PHASER

In 2019, the VA began working with Sanford Health to establish the PHASER program and offer pharmacogenomic testing. PHASER has since expanded to 25 VA medical centers, including the VA Central Ohio Healthcare System (VACOHCS).^7,8 Pharmacogenomic testing through PHASER is conducted using a standardized laboratory panel that includes 12 different medication classes.⁹ The drug classes include certain anti-infective, anticoagulant, antiplatelet, cardiovascular, cholesterol, gastrointestinal, mental health, neurological, oncology, pain, transplant, and other miscellaneous medications. Medications are correlated to each class and assessed for therapeutic impacts based on gene panel results.

Clinical recommendations for medication-gene interactions can range from monitoring for increased risk of adverse effects or therapeutic failure to recommending avoiding a medication. For example, patients who test positive for the HLA-B gene have significantly increased risk of hypersensitivity to abacavir, an HIV treatment.¹⁰

Similarly, patients who cannot adequately metabolize cytochrome P450 2C19 should consider avoiding clopidogrel as they are unlikely to convert clopidogrel to its active prodrug, which reduces its effectiveness.¹¹ Pharmacists can play a critical role educating patients about pharmacogenomic testing, especially within hematology and oncology.¹² Patients can benefit from this testing even if they are not currently taking medications with known concerns as they could be prescribed in the future. The SLCO1B1 gene-drug test, for example, can identify risk for statin-associated muscle symptoms.¹³

Clinical pharmacist practitioners (CPPs) can increase access to genetic testing because they interact with patients in a variety of settings and can order this laboratory test.^12,14 Recent research has demonstrated that most VA patients carry ≥ 1 genetic variant that may influence medication decisions and that half of veterans are prescribed a medication with known gene-drug interactions.¹⁵ CPP ordering of pharmacogenomic tests at the VACOHCS outpatient clinic was evaluated through collection of baseline data from March 8, 2023, to September 8, 2023. A goal was identified to increase orders by 50% for a patient care quality improvement initiative and use CPPs to increase access to pharmacogenomic testing. The purpose of this quality improvement initiative was to expand access to pharmacogenomic testing through process implementation and improvement within CPP-led clinic settings.

Gap Analysis

Lean Six Sigma A3 methodology was used to identify ways to increase the use of pharmacogenomic testing for veterans at VACOHCS and develop an improved process for increased ordering of pharmacogenomic testing. Lean Six Sigma A3 methodology is a stepwise approach to process improvement that helps identify gaps in efficiency, sustainable changes, and eliminate waste.¹⁶ Baseline data were collected from March 8, 2023, to September 8, 2023, to determine the frequency of CPPs ordering pharmacogenomic laboratory panels during clinic appointments. The ordering of pharmacogenomic panels was monitored by the VACOHCS PHASER coordinator.

CPPs were surveyed to identify perceived barriers to PHASER implementation. A gap analysis was conducted using Lean Six Sigma A3 methodology. Gap analyses use lean tools such as a Fishbone Diagram to illustrate and identify the gap between current state and ideal state. (Figure 1).The following barriers were identified: lack of clinician education materials, lack of a standardized patient screening process, time constraints on patient education and ordering, higher priority clinical needs, forgetting to order, lack of comfort with pharmacogenomics ordering and education, lack of support for the initiative, and increased workload and burnout. Among these perceived barriers, higher priority clinical needs, forgetting to order, and time constraints ranked highest in importance among CPPs. 

In line with Lean Six Sigma A3 methodology, several tests of change were used to improve pharmacogenomic testing ordering. These changes focused on increasing patient and clinician awareness, facilitating discussion, educating clinicians, and simplifying documentation to ease time constraints. Several strategies were employed postimplementation (Figure 2). Prefilled templates simplified documentation. These templates helped identify patients without pharmacogenomic testing, provided reminders, and saved documentation time during visits. CPPs also received training and materials on PHASER ordering and documentation within encounter notes. Additionally, patient-directed advertisements were displayed in CPP examination rooms to help inspire and facilitate discussion between veterans and CPPs.

Process Improvement Data

The quality improvement project goal was to increase PHASER orders by 50% after 3 months. PHASER orders increased from 87 at baseline (March 8, 2023, to September 8, 2023) to 196 during the intervention (November 16, 2023, to February 16, 2024), a 125% increase. Changes were consistent and sustained with 65 orders the first month, 67 orders the second month, and 64 orders the third month.

Discussion

Using Lean Six Sigma A3 methodology for a quality improvement process to increase PHASER orders by CPPs revealed barriers and guided potential solutions to overcome these barriers. Interventions included additional CPP training and ordering, tools for easier identification of potential patients, documentation best practices, patient-directed advertisements to facilitate conversations. These interventions required about 8 hours for preparation, distribution, development, and interpretation of surveys, education, and documentation materials. The financial impact of these interventions was already included in allotted office materials budgeted and provided. Additional funding was not needed to provide patient-directed advertisements or education materials. The VACOHCS pharmacogenomics CPP discusses PHASER test results with patients at a separate appointment.

Future directions include educating other CPPs to assist in discussing results with veterans. Overall, the changes implemented to improve the PHASER ordering process were low effort and exemplify the ease of streamlining future initiatives, allowing for sustained optimal implementation of pharmacogenomic testing.

Conclusions

A quality improvement initiative resulted in increased PHASER orders and a clearly defined process, allowing for a continued increase and sustained support. Perceived barriers were identified, and the changes implemented were often low effort but exhibited a sustained impact. The insights gleaned from this process will shape future process development initiatives and continue to sustain pharmacogenomic testing ordering by CPPs. This process will be extended to other VACOHCS clinical departments to further support increased access to pharmacogenomic testing, reduce medication trial and error, and reduce hospitalizations from adverse effects for veterans.

Statins are widely used to reduce low-density lipoprotein (LDL) and non-high-density lipoprotein (HDL) levels for the prevention of atherosclerotic cardiovascular disease (ASCVD).¹ However, despite maximally tolerated statin therapy, many patients may not reach their LDL and non-HDL goals. Some patients may experience adverse events (AEs), particularly muscle-related AEs, which can limit the use of these medications.

The 2022 American College of Cardiology (ACC) expert consensus pathway recommends a goal LDL of < 55 mg/dL in very high-risk patients, defined as those with a history of multiple major ASCVD events or 1 major ASCVD event and multiple high-risk conditions.² Major ASCVD events include acute coronary syndrome within 12 months, history of myocardial infarction (MI) or ischemic stroke, and symptomatic peripheral arterial disease (ie, claudication with ankle-brachial index < 0.85 or previous revascularization or amputation). Factors for being considered high risk include age > 65 years, heterozygous familial hypercholesterolemia, history of prior coronary artery bypass surgery or percutaneous coronary intervention outside the major ASCVD events, diabetes, hypertension, chronic kidney disease (CKD) (estimated glomerular filtration rate [eGFR] 15-59 mL/min/1.73 m²), current smoking, persistently elevated LDL cholesterol (LDL-C) levels despite maximally tolerated statin therapy and ezetimibe, and history of congestive heart failure.² For these patients, statin therapy alone may not achieve LDL goal. 

The ACC recommends ezetimibe as the initial nonstatin therapy in patients who are not at their goal LDL.² Ezetimibe works by inhibiting Niemann-Pick C1-Like 1 protein, which causes reduced cholesterol absorption in the small intestine.^2,3 Previous studies have shown the benefit of ezetimibe for LDL reduction and ASCVD prevention.^4-7 The 2015 IMPROVE-IT study found the combination of simvastatin and ezetimibe resulted in a significantly lower risk of cardiovascular events than simvastatin monotherapy. IMPROVE-IT also reported a further clinical benefit when lower LDL targets (ie, < 55 mg/dL) are achieved, which aligns with the expert consensus pathway recommendations for a lower LDL goal for very high-risk patients.^2,5

The RACING trial found that treatment with a moderate-intensity statin and ezetimibe was noninferior to treatment with a high-intensity statin for the primary outcome of occurrence of cardiovascular death, major cardiovascular events, or nonfatal stroke within 3 years. The combination of moderate-intensity statin and ezetimibe achieved lower LDL-C levels and lower incidence of drug intolerance compared to high intensity statin monotherapy.⁶ The SHARP-CKD study assessed major atherosclerotic events in patients with CKD who had no history of MI or coronary revascularization. The study found that lowering LDL-C with the combination of simvastatin plus ezetimibe safely reduces the risk of major atherosclerotic events in a wide range of patients with CKD.⁷

Lastly, the 2019 EWTOPIA 75 study found that ezetimibe noted a statistically significant reduction in the incidence of the composite of sudden cardiac death, MI, coronary revascularization, or stroke compared to placebo. Ezetimibe showed benefits in preventing ASCVD events independently of statin therapy.⁸ These clinical trials provided evidence for the efficacy of ezetimibe for secondary or primary prevention of ASCVD, patients with CKD, and patients who are not at their LDL goal despite maximally tolerated statin therapy.

Reductions in LDL levels with ezetimibe are reported to be 15% to 19% for monotherapy and 13% to 25% when used in combination with a statin.⁴ Given that the ACC now recommends lower LDL goals, patients may need additional lowering despite taking maximally tolerated statin therapy.² Additionally, the package insert for ezetimibe reports increased area under the curve (AUC) values of ezetimibe and its metabolites in patients with severe renal disease. It is anticipated that ezetimibe may show an increased reduction of LDL and non-HDL, but there may also be an increased risk for muscle-related AEs.³

This quality-assurance quality improvement project investigated the use of ezetimibe in patients with CKD to determine whether there is further LDL and non-HDL reduction in this patient population. It sought to determine the LDL and non-HDL percentage reduction in patients with and without CKD at the Wilkes-Barre Veterans Affairs Medical Center (WBVAMC) and whether there is an increased risk for muscle-related AEs. Determining the percentage reduction of LDL and non-HDL within this population can help increase use of ezetimibe in patients not at their LDL or non-HDL goal or for those patients unable to tolerate statin therapy.

Methods

This single-center retrospective chart review investigated patients prescribed ezetimibe by a patient aligned care team (PACT) pharmacist at WBVAMC between September 1, 2021, and September 1, 2023. This project was determined to be nonresearch by the Veterans Integrated Service Network 4 multisite institutional review board. Patients were excluded from the review if they started taking ezetimibe outside of the prespecified time frame, if ezetimibe was initiated by a non-WBVAMC PACT pharmacist, or if there was no follow-up lipid panel obtained within 6 months of initiation of ezetimibe.

The primary outcomes were to determine the percentage mean change in LDL and non-HDL reduction and the incidence of muscle-related AEs after initiation of ezetimibe in patients without CKD. The secondary outcomes were to determine the percentage mean change in LDL and non-HDL levels and the incidence of muscle-related AEs after initiation of ezetimibe in patients with CKD. For this study, CKD was defined as a patient having an eGFR 15 to 60 ml/min/1.73 m². Non-HDL is the combination of LDL-C and very LDL-C and represents all potentially atherogenic particles. The 2022 Expert Consensus Pathway included non-HDL goals in addition to LDL goals.² Non-HDL cholesterol levels can be used for patients with elevated triglycerides where LDL levels may not be as accurate. To account for instances of elevated triglycerides, this study assessed changes in both LDL and non-HDL levels.

Data were collected from the US Department of Veterans Affairs (VA) Computerized Patient Record System (CPRS) and recorded in a spreadsheet. Collected data included age, sex, race, concomitant cholesterol-lowering medications (statin, proprotein convertase subtilisin/kexin type 9 [PCSK9] inhibitor, bempedoic acid, fish oil, niacin, bile acid sequestrants, and fibrates), baseline lipid panel, lipid panel within 6 months of ezetimibe initiation, and eGFR level. If the patient’s LDL or non-HDL levels worsened on the follow-up lipid panel, their baseline LDL and non-HDL levels were used to calculate the percentage reduction; thus, the percentage reduction would be 0%. This strategy was used in prior research, notably the IMPROVE-IT and SHARP-CKD trials. 

Ezetimibe 5 mg once daily was used in this study based on a 2008 VA study that evaluated the use of ezetimibe 5 mg vs ezetimibe 10 mg and the percentage reduction of LDL with each dose. The study found no significant difference between the 5 mg and 10 mg dose.⁹ Most patients included in this study received the 5 mg dose.

Results

This retrospective chart review consisted of 173 patients, 137 (79.2%) without CKD and 36 (20.8%) with CKD at baseline. The mean age was 69.6 years, 155 (89.6%) patients were male, and 18 (10.4%) were female. There were 164 concomitant medications, including 115 patients prescribed a statin and 38 patients prescribed fish oil (Table 1).

Patients without CKD had mean reductions in LDL levels of 23.5% and non-HDL levels of 21.7% (Figure). Patients who had an increase in LDL and non-HDL levels were excluded to control for potential confounding factors such as dietary changes, discontinuation of ezetimibe therapy, nonadherence to ezetimibe, and medication changes that impacted follow-up laboratory tests such as discontinuation of a statin. Fifteen patients experienced an increase in LDL or non-HDL levels. After excluding these patients, those without CKD had a mean reduction in LDL levels of 28.0% and non-HDL levels of 25.5%. Nineteen (13.9%) patients without CKD experienced a muscle-related AE (Table 2). One patient discontinued ezetimibe and statin use following a Lyme disease diagnosis due to concerns over potential muscle-related AEs. 

Patients with CKD had a mean reduction in LDL and non-HDL levels of 27.0% and 24.8%, respectively. Patients with an increase in LDL or non-HDL levels were also excluded to help control for potential confounding factors. After excluding 4 patients with increased LDL and non-HDL levels, the mean reduction in LDL and non-HDL levels was 30.5% and 27.5%, respectively. Five (13.9%) patients with CKD experienced muscle-related AEs thought to be due to ezetimibe. Other AEs (eg, urticaria, diarrhea, reflux, dizziness, headache, upset stomach) were reported that led to discontinuation of ezetimibe, but only muscle-related AEs were analyzed.

Discussion

This retrospective chart review found larger reductions in LDL and non-HDL levels for patients with CKD than reported in the literature.⁴ Based on the findings that indicate a greater cholesterol reduction with ezetimibe, the results suggest an underutilization of ezetimibe in clinical practice, which may be due to clinicians favoring statin therapy and overlooking ezetimibe as a viable option based on recommendation in earlier guidelines. The 2022 guidelines transitioned from a statin focus to a focus on LDL targets and goals.²

According to the ACC, there is evidence to support a direct relationship between LDL-C levels, atherosclerosis progression, and ASCVD event risk.² Absolute LDL-C level reduction is directly associated with ASCVD risk reduction which supports the LDL hypothesis. There appears to be no specific LDL-C level below which benefit ceases.² This suggests that lower LDL-C targets (< 55 mg/dL) should be used when clinically indicated. Many patients are either unable to reach their goal LDL levels with statin monotherapy or are unable to tolerate statin therapy at higher doses, which may require additional pharmacotherapy to reach goal LDL-C. The ACC expert consensus pathway recommends ezetimibe as the initial add-on treatment to statins.² The RACING trial showed the benefit of adding ezetimibe to a moderate-intensity statin vs increasing to a high-intensity statin dose. This trial found patients had lower LDL levels and lower rates of intolerances, which further supports ezetimibe use.⁶

This quality improvement project assessed LDL and non-HDL level reduction in patients with CKD. As anticipated, there was greater reduction in LDL and non-HDL levels seen in patients with CKD. The SHARP-CKD trial also found reductions in LDL levels with ezetimibe in patients with CKD.⁷ Given the reduction in LDL and non-HDL levels with ezetimibe in patients with or without CKD, add-on therapy of ezetimibe should be recommended for patients who do not achieve their LDL goals with statin therapy or for patients who intolerant to statin therapy. 

The ezetimibe package insert reports myalgias incidence to be < 5% in patients and research has shown up to a 20% incidence of muscle-related AEs with statin therapy.^3,10 Based on the package information reporting increased AUC values of ezetimibe and its metabolites in patients with severe renal disease, it was anticipated there may be an increased risk of muscle-related AEs in patients with CKD.³ However, this study found the same incidence of muscle-related AEs in patients with and without CKD. Previous research on statin-intolerant patients found the incidence of muscle-related AEs with ezetimibe to be 23.0% and 28.8%.^11,12 This increased incidence of muscle-related AEs may be the result of including patients with a history of statin intolerance. Collectively, data from clinical trials and this study indicate that patients with prior intolerances to statins appear to have a higher likelihood of developing a muscle-related AEs with ezetimibe.^11,12 Clinicians and patients should be educated on the potential for these AEs and be aware that the likelihood may be greater if there is a history of statin intolerance. To our knowledge, this was the first study to evaluate muscle-related AEs with ezetimibe in patients with and without CKD.

Limitations

This retrospective chart review was performed over a prespecified period and only patients initiated on ezetimibe by a PACT pharmacist were included. This study did not assess the percentage of LDL reduction in patients on concomitant statins vs those who were not on concomitant statins. The study only included 173 patients. Additionally, the study was primarily composed of White men and may not be representative of other populations. In addition, veterans may not be representative of the general population given their high comorbidity burden and other exposures. Some reported muscle-related AEs associated with ezetimibe may be attributed to the nocebo effect.

Conclusions

The results of this retrospective chart review suggest there may be a larger mean reduction in LDL and non-HDL levels seen with ezetimibe therapy than reported within the literature. There was a larger mean reduction in LDL and non-HDL levels in patients with CKD than in patients without CKD. Additionally, there were the same rates of muscle-related AEs with ezetimibe therapy in patients with and without CKD. The rates of muscle-related AEs with ezetimibe therapy were higher than reported in the medication’s package insert, but lower than reported in literature that included statin-intolerant patients. These results indicate there may be a benefit to an increase in use of ezetimibe in clinical practice due to its increased effectiveness and safety in patients with and without CKD. Ultimately, this can help patients achieve their LDL goals as recommended by ACC clinical practice guidelines.

Statins are widely used to reduce low-density lipoprotein (LDL) and non-high-density lipoprotein (HDL) levels for the prevention of atherosclerotic cardiovascular disease (ASCVD).¹ However, despite maximally tolerated statin therapy, many patients may not reach their LDL and non-HDL goals. Some patients may experience adverse events (AEs), particularly muscle-related AEs, which can limit the use of these medications.

The 2022 American College of Cardiology (ACC) expert consensus pathway recommends a goal LDL of < 55 mg/dL in very high-risk patients, defined as those with a history of multiple major ASCVD events or 1 major ASCVD event and multiple high-risk conditions.² Major ASCVD events include acute coronary syndrome within 12 months, history of myocardial infarction (MI) or ischemic stroke, and symptomatic peripheral arterial disease (ie, claudication with ankle-brachial index < 0.85 or previous revascularization or amputation). Factors for being considered high risk include age > 65 years, heterozygous familial hypercholesterolemia, history of prior coronary artery bypass surgery or percutaneous coronary intervention outside the major ASCVD events, diabetes, hypertension, chronic kidney disease (CKD) (estimated glomerular filtration rate [eGFR] 15-59 mL/min/1.73 m²), current smoking, persistently elevated LDL cholesterol (LDL-C) levels despite maximally tolerated statin therapy and ezetimibe, and history of congestive heart failure.² For these patients, statin therapy alone may not achieve LDL goal. 

The ACC recommends ezetimibe as the initial nonstatin therapy in patients who are not at their goal LDL.² Ezetimibe works by inhibiting Niemann-Pick C1-Like 1 protein, which causes reduced cholesterol absorption in the small intestine.^2,3 Previous studies have shown the benefit of ezetimibe for LDL reduction and ASCVD prevention.^4-7 The 2015 IMPROVE-IT study found the combination of simvastatin and ezetimibe resulted in a significantly lower risk of cardiovascular events than simvastatin monotherapy. IMPROVE-IT also reported a further clinical benefit when lower LDL targets (ie, < 55 mg/dL) are achieved, which aligns with the expert consensus pathway recommendations for a lower LDL goal for very high-risk patients.^2,5

The RACING trial found that treatment with a moderate-intensity statin and ezetimibe was noninferior to treatment with a high-intensity statin for the primary outcome of occurrence of cardiovascular death, major cardiovascular events, or nonfatal stroke within 3 years. The combination of moderate-intensity statin and ezetimibe achieved lower LDL-C levels and lower incidence of drug intolerance compared to high intensity statin monotherapy.⁶ The SHARP-CKD study assessed major atherosclerotic events in patients with CKD who had no history of MI or coronary revascularization. The study found that lowering LDL-C with the combination of simvastatin plus ezetimibe safely reduces the risk of major atherosclerotic events in a wide range of patients with CKD.⁷

Lastly, the 2019 EWTOPIA 75 study found that ezetimibe noted a statistically significant reduction in the incidence of the composite of sudden cardiac death, MI, coronary revascularization, or stroke compared to placebo. Ezetimibe showed benefits in preventing ASCVD events independently of statin therapy.⁸ These clinical trials provided evidence for the efficacy of ezetimibe for secondary or primary prevention of ASCVD, patients with CKD, and patients who are not at their LDL goal despite maximally tolerated statin therapy.

Reductions in LDL levels with ezetimibe are reported to be 15% to 19% for monotherapy and 13% to 25% when used in combination with a statin.⁴ Given that the ACC now recommends lower LDL goals, patients may need additional lowering despite taking maximally tolerated statin therapy.² Additionally, the package insert for ezetimibe reports increased area under the curve (AUC) values of ezetimibe and its metabolites in patients with severe renal disease. It is anticipated that ezetimibe may show an increased reduction of LDL and non-HDL, but there may also be an increased risk for muscle-related AEs.³

This quality-assurance quality improvement project investigated the use of ezetimibe in patients with CKD to determine whether there is further LDL and non-HDL reduction in this patient population. It sought to determine the LDL and non-HDL percentage reduction in patients with and without CKD at the Wilkes-Barre Veterans Affairs Medical Center (WBVAMC) and whether there is an increased risk for muscle-related AEs. Determining the percentage reduction of LDL and non-HDL within this population can help increase use of ezetimibe in patients not at their LDL or non-HDL goal or for those patients unable to tolerate statin therapy.

Methods

This single-center retrospective chart review investigated patients prescribed ezetimibe by a patient aligned care team (PACT) pharmacist at WBVAMC between September 1, 2021, and September 1, 2023. This project was determined to be nonresearch by the Veterans Integrated Service Network 4 multisite institutional review board. Patients were excluded from the review if they started taking ezetimibe outside of the prespecified time frame, if ezetimibe was initiated by a non-WBVAMC PACT pharmacist, or if there was no follow-up lipid panel obtained within 6 months of initiation of ezetimibe.

The primary outcomes were to determine the percentage mean change in LDL and non-HDL reduction and the incidence of muscle-related AEs after initiation of ezetimibe in patients without CKD. The secondary outcomes were to determine the percentage mean change in LDL and non-HDL levels and the incidence of muscle-related AEs after initiation of ezetimibe in patients with CKD. For this study, CKD was defined as a patient having an eGFR 15 to 60 ml/min/1.73 m². Non-HDL is the combination of LDL-C and very LDL-C and represents all potentially atherogenic particles. The 2022 Expert Consensus Pathway included non-HDL goals in addition to LDL goals.² Non-HDL cholesterol levels can be used for patients with elevated triglycerides where LDL levels may not be as accurate. To account for instances of elevated triglycerides, this study assessed changes in both LDL and non-HDL levels.

Data were collected from the US Department of Veterans Affairs (VA) Computerized Patient Record System (CPRS) and recorded in a spreadsheet. Collected data included age, sex, race, concomitant cholesterol-lowering medications (statin, proprotein convertase subtilisin/kexin type 9 [PCSK9] inhibitor, bempedoic acid, fish oil, niacin, bile acid sequestrants, and fibrates), baseline lipid panel, lipid panel within 6 months of ezetimibe initiation, and eGFR level. If the patient’s LDL or non-HDL levels worsened on the follow-up lipid panel, their baseline LDL and non-HDL levels were used to calculate the percentage reduction; thus, the percentage reduction would be 0%. This strategy was used in prior research, notably the IMPROVE-IT and SHARP-CKD trials. 

Ezetimibe 5 mg once daily was used in this study based on a 2008 VA study that evaluated the use of ezetimibe 5 mg vs ezetimibe 10 mg and the percentage reduction of LDL with each dose. The study found no significant difference between the 5 mg and 10 mg dose.⁹ Most patients included in this study received the 5 mg dose.

Results

This retrospective chart review consisted of 173 patients, 137 (79.2%) without CKD and 36 (20.8%) with CKD at baseline. The mean age was 69.6 years, 155 (89.6%) patients were male, and 18 (10.4%) were female. There were 164 concomitant medications, including 115 patients prescribed a statin and 38 patients prescribed fish oil (Table 1).

Patients without CKD had mean reductions in LDL levels of 23.5% and non-HDL levels of 21.7% (Figure). Patients who had an increase in LDL and non-HDL levels were excluded to control for potential confounding factors such as dietary changes, discontinuation of ezetimibe therapy, nonadherence to ezetimibe, and medication changes that impacted follow-up laboratory tests such as discontinuation of a statin. Fifteen patients experienced an increase in LDL or non-HDL levels. After excluding these patients, those without CKD had a mean reduction in LDL levels of 28.0% and non-HDL levels of 25.5%. Nineteen (13.9%) patients without CKD experienced a muscle-related AE (Table 2). One patient discontinued ezetimibe and statin use following a Lyme disease diagnosis due to concerns over potential muscle-related AEs. 

Patients with CKD had a mean reduction in LDL and non-HDL levels of 27.0% and 24.8%, respectively. Patients with an increase in LDL or non-HDL levels were also excluded to help control for potential confounding factors. After excluding 4 patients with increased LDL and non-HDL levels, the mean reduction in LDL and non-HDL levels was 30.5% and 27.5%, respectively. Five (13.9%) patients with CKD experienced muscle-related AEs thought to be due to ezetimibe. Other AEs (eg, urticaria, diarrhea, reflux, dizziness, headache, upset stomach) were reported that led to discontinuation of ezetimibe, but only muscle-related AEs were analyzed.

Discussion

This retrospective chart review found larger reductions in LDL and non-HDL levels for patients with CKD than reported in the literature.⁴ Based on the findings that indicate a greater cholesterol reduction with ezetimibe, the results suggest an underutilization of ezetimibe in clinical practice, which may be due to clinicians favoring statin therapy and overlooking ezetimibe as a viable option based on recommendation in earlier guidelines. The 2022 guidelines transitioned from a statin focus to a focus on LDL targets and goals.²

According to the ACC, there is evidence to support a direct relationship between LDL-C levels, atherosclerosis progression, and ASCVD event risk.² Absolute LDL-C level reduction is directly associated with ASCVD risk reduction which supports the LDL hypothesis. There appears to be no specific LDL-C level below which benefit ceases.² This suggests that lower LDL-C targets (< 55 mg/dL) should be used when clinically indicated. Many patients are either unable to reach their goal LDL levels with statin monotherapy or are unable to tolerate statin therapy at higher doses, which may require additional pharmacotherapy to reach goal LDL-C. The ACC expert consensus pathway recommends ezetimibe as the initial add-on treatment to statins.² The RACING trial showed the benefit of adding ezetimibe to a moderate-intensity statin vs increasing to a high-intensity statin dose. This trial found patients had lower LDL levels and lower rates of intolerances, which further supports ezetimibe use.⁶

This quality improvement project assessed LDL and non-HDL level reduction in patients with CKD. As anticipated, there was greater reduction in LDL and non-HDL levels seen in patients with CKD. The SHARP-CKD trial also found reductions in LDL levels with ezetimibe in patients with CKD.⁷ Given the reduction in LDL and non-HDL levels with ezetimibe in patients with or without CKD, add-on therapy of ezetimibe should be recommended for patients who do not achieve their LDL goals with statin therapy or for patients who intolerant to statin therapy. 

The ezetimibe package insert reports myalgias incidence to be < 5% in patients and research has shown up to a 20% incidence of muscle-related AEs with statin therapy.^3,10 Based on the package information reporting increased AUC values of ezetimibe and its metabolites in patients with severe renal disease, it was anticipated there may be an increased risk of muscle-related AEs in patients with CKD.³ However, this study found the same incidence of muscle-related AEs in patients with and without CKD. Previous research on statin-intolerant patients found the incidence of muscle-related AEs with ezetimibe to be 23.0% and 28.8%.^11,12 This increased incidence of muscle-related AEs may be the result of including patients with a history of statin intolerance. Collectively, data from clinical trials and this study indicate that patients with prior intolerances to statins appear to have a higher likelihood of developing a muscle-related AEs with ezetimibe.^11,12 Clinicians and patients should be educated on the potential for these AEs and be aware that the likelihood may be greater if there is a history of statin intolerance. To our knowledge, this was the first study to evaluate muscle-related AEs with ezetimibe in patients with and without CKD.

Limitations

This retrospective chart review was performed over a prespecified period and only patients initiated on ezetimibe by a PACT pharmacist were included. This study did not assess the percentage of LDL reduction in patients on concomitant statins vs those who were not on concomitant statins. The study only included 173 patients. Additionally, the study was primarily composed of White men and may not be representative of other populations. In addition, veterans may not be representative of the general population given their high comorbidity burden and other exposures. Some reported muscle-related AEs associated with ezetimibe may be attributed to the nocebo effect.

Conclusions

The results of this retrospective chart review suggest there may be a larger mean reduction in LDL and non-HDL levels seen with ezetimibe therapy than reported within the literature. There was a larger mean reduction in LDL and non-HDL levels in patients with CKD than in patients without CKD. Additionally, there were the same rates of muscle-related AEs with ezetimibe therapy in patients with and without CKD. The rates of muscle-related AEs with ezetimibe therapy were higher than reported in the medication’s package insert, but lower than reported in literature that included statin-intolerant patients. These results indicate there may be a benefit to an increase in use of ezetimibe in clinical practice due to its increased effectiveness and safety in patients with and without CKD. Ultimately, this can help patients achieve their LDL goals as recommended by ACC clinical practice guidelines.

Statins are widely used to reduce low-density lipoprotein (LDL) and non-high-density lipoprotein (HDL) levels for the prevention of atherosclerotic cardiovascular disease (ASCVD).¹ However, despite maximally tolerated statin therapy, many patients may not reach their LDL and non-HDL goals. Some patients may experience adverse events (AEs), particularly muscle-related AEs, which can limit the use of these medications.

The 2022 American College of Cardiology (ACC) expert consensus pathway recommends a goal LDL of < 55 mg/dL in very high-risk patients, defined as those with a history of multiple major ASCVD events or 1 major ASCVD event and multiple high-risk conditions.² Major ASCVD events include acute coronary syndrome within 12 months, history of myocardial infarction (MI) or ischemic stroke, and symptomatic peripheral arterial disease (ie, claudication with ankle-brachial index < 0.85 or previous revascularization or amputation). Factors for being considered high risk include age > 65 years, heterozygous familial hypercholesterolemia, history of prior coronary artery bypass surgery or percutaneous coronary intervention outside the major ASCVD events, diabetes, hypertension, chronic kidney disease (CKD) (estimated glomerular filtration rate [eGFR] 15-59 mL/min/1.73 m²), current smoking, persistently elevated LDL cholesterol (LDL-C) levels despite maximally tolerated statin therapy and ezetimibe, and history of congestive heart failure.² For these patients, statin therapy alone may not achieve LDL goal. 

The ACC recommends ezetimibe as the initial nonstatin therapy in patients who are not at their goal LDL.² Ezetimibe works by inhibiting Niemann-Pick C1-Like 1 protein, which causes reduced cholesterol absorption in the small intestine.^2,3 Previous studies have shown the benefit of ezetimibe for LDL reduction and ASCVD prevention.^4-7 The 2015 IMPROVE-IT study found the combination of simvastatin and ezetimibe resulted in a significantly lower risk of cardiovascular events than simvastatin monotherapy. IMPROVE-IT also reported a further clinical benefit when lower LDL targets (ie, < 55 mg/dL) are achieved, which aligns with the expert consensus pathway recommendations for a lower LDL goal for very high-risk patients.^2,5

The RACING trial found that treatment with a moderate-intensity statin and ezetimibe was noninferior to treatment with a high-intensity statin for the primary outcome of occurrence of cardiovascular death, major cardiovascular events, or nonfatal stroke within 3 years. The combination of moderate-intensity statin and ezetimibe achieved lower LDL-C levels and lower incidence of drug intolerance compared to high intensity statin monotherapy.⁶ The SHARP-CKD study assessed major atherosclerotic events in patients with CKD who had no history of MI or coronary revascularization. The study found that lowering LDL-C with the combination of simvastatin plus ezetimibe safely reduces the risk of major atherosclerotic events in a wide range of patients with CKD.⁷

Lastly, the 2019 EWTOPIA 75 study found that ezetimibe noted a statistically significant reduction in the incidence of the composite of sudden cardiac death, MI, coronary revascularization, or stroke compared to placebo. Ezetimibe showed benefits in preventing ASCVD events independently of statin therapy.⁸ These clinical trials provided evidence for the efficacy of ezetimibe for secondary or primary prevention of ASCVD, patients with CKD, and patients who are not at their LDL goal despite maximally tolerated statin therapy.

Reductions in LDL levels with ezetimibe are reported to be 15% to 19% for monotherapy and 13% to 25% when used in combination with a statin.⁴ Given that the ACC now recommends lower LDL goals, patients may need additional lowering despite taking maximally tolerated statin therapy.² Additionally, the package insert for ezetimibe reports increased area under the curve (AUC) values of ezetimibe and its metabolites in patients with severe renal disease. It is anticipated that ezetimibe may show an increased reduction of LDL and non-HDL, but there may also be an increased risk for muscle-related AEs.³

This quality-assurance quality improvement project investigated the use of ezetimibe in patients with CKD to determine whether there is further LDL and non-HDL reduction in this patient population. It sought to determine the LDL and non-HDL percentage reduction in patients with and without CKD at the Wilkes-Barre Veterans Affairs Medical Center (WBVAMC) and whether there is an increased risk for muscle-related AEs. Determining the percentage reduction of LDL and non-HDL within this population can help increase use of ezetimibe in patients not at their LDL or non-HDL goal or for those patients unable to tolerate statin therapy.

Methods

This single-center retrospective chart review investigated patients prescribed ezetimibe by a patient aligned care team (PACT) pharmacist at WBVAMC between September 1, 2021, and September 1, 2023. This project was determined to be nonresearch by the Veterans Integrated Service Network 4 multisite institutional review board. Patients were excluded from the review if they started taking ezetimibe outside of the prespecified time frame, if ezetimibe was initiated by a non-WBVAMC PACT pharmacist, or if there was no follow-up lipid panel obtained within 6 months of initiation of ezetimibe.

The primary outcomes were to determine the percentage mean change in LDL and non-HDL reduction and the incidence of muscle-related AEs after initiation of ezetimibe in patients without CKD. The secondary outcomes were to determine the percentage mean change in LDL and non-HDL levels and the incidence of muscle-related AEs after initiation of ezetimibe in patients with CKD. For this study, CKD was defined as a patient having an eGFR 15 to 60 ml/min/1.73 m². Non-HDL is the combination of LDL-C and very LDL-C and represents all potentially atherogenic particles. The 2022 Expert Consensus Pathway included non-HDL goals in addition to LDL goals.² Non-HDL cholesterol levels can be used for patients with elevated triglycerides where LDL levels may not be as accurate. To account for instances of elevated triglycerides, this study assessed changes in both LDL and non-HDL levels.

Data were collected from the US Department of Veterans Affairs (VA) Computerized Patient Record System (CPRS) and recorded in a spreadsheet. Collected data included age, sex, race, concomitant cholesterol-lowering medications (statin, proprotein convertase subtilisin/kexin type 9 [PCSK9] inhibitor, bempedoic acid, fish oil, niacin, bile acid sequestrants, and fibrates), baseline lipid panel, lipid panel within 6 months of ezetimibe initiation, and eGFR level. If the patient’s LDL or non-HDL levels worsened on the follow-up lipid panel, their baseline LDL and non-HDL levels were used to calculate the percentage reduction; thus, the percentage reduction would be 0%. This strategy was used in prior research, notably the IMPROVE-IT and SHARP-CKD trials. 

Ezetimibe 5 mg once daily was used in this study based on a 2008 VA study that evaluated the use of ezetimibe 5 mg vs ezetimibe 10 mg and the percentage reduction of LDL with each dose. The study found no significant difference between the 5 mg and 10 mg dose.⁹ Most patients included in this study received the 5 mg dose.

Results

This retrospective chart review consisted of 173 patients, 137 (79.2%) without CKD and 36 (20.8%) with CKD at baseline. The mean age was 69.6 years, 155 (89.6%) patients were male, and 18 (10.4%) were female. There were 164 concomitant medications, including 115 patients prescribed a statin and 38 patients prescribed fish oil (Table 1).

Patients without CKD had mean reductions in LDL levels of 23.5% and non-HDL levels of 21.7% (Figure). Patients who had an increase in LDL and non-HDL levels were excluded to control for potential confounding factors such as dietary changes, discontinuation of ezetimibe therapy, nonadherence to ezetimibe, and medication changes that impacted follow-up laboratory tests such as discontinuation of a statin. Fifteen patients experienced an increase in LDL or non-HDL levels. After excluding these patients, those without CKD had a mean reduction in LDL levels of 28.0% and non-HDL levels of 25.5%. Nineteen (13.9%) patients without CKD experienced a muscle-related AE (Table 2). One patient discontinued ezetimibe and statin use following a Lyme disease diagnosis due to concerns over potential muscle-related AEs. 

Patients with CKD had a mean reduction in LDL and non-HDL levels of 27.0% and 24.8%, respectively. Patients with an increase in LDL or non-HDL levels were also excluded to help control for potential confounding factors. After excluding 4 patients with increased LDL and non-HDL levels, the mean reduction in LDL and non-HDL levels was 30.5% and 27.5%, respectively. Five (13.9%) patients with CKD experienced muscle-related AEs thought to be due to ezetimibe. Other AEs (eg, urticaria, diarrhea, reflux, dizziness, headache, upset stomach) were reported that led to discontinuation of ezetimibe, but only muscle-related AEs were analyzed.

Discussion

This retrospective chart review found larger reductions in LDL and non-HDL levels for patients with CKD than reported in the literature.⁴ Based on the findings that indicate a greater cholesterol reduction with ezetimibe, the results suggest an underutilization of ezetimibe in clinical practice, which may be due to clinicians favoring statin therapy and overlooking ezetimibe as a viable option based on recommendation in earlier guidelines. The 2022 guidelines transitioned from a statin focus to a focus on LDL targets and goals.²

According to the ACC, there is evidence to support a direct relationship between LDL-C levels, atherosclerosis progression, and ASCVD event risk.² Absolute LDL-C level reduction is directly associated with ASCVD risk reduction which supports the LDL hypothesis. There appears to be no specific LDL-C level below which benefit ceases.² This suggests that lower LDL-C targets (< 55 mg/dL) should be used when clinically indicated. Many patients are either unable to reach their goal LDL levels with statin monotherapy or are unable to tolerate statin therapy at higher doses, which may require additional pharmacotherapy to reach goal LDL-C. The ACC expert consensus pathway recommends ezetimibe as the initial add-on treatment to statins.² The RACING trial showed the benefit of adding ezetimibe to a moderate-intensity statin vs increasing to a high-intensity statin dose. This trial found patients had lower LDL levels and lower rates of intolerances, which further supports ezetimibe use.⁶

This quality improvement project assessed LDL and non-HDL level reduction in patients with CKD. As anticipated, there was greater reduction in LDL and non-HDL levels seen in patients with CKD. The SHARP-CKD trial also found reductions in LDL levels with ezetimibe in patients with CKD.⁷ Given the reduction in LDL and non-HDL levels with ezetimibe in patients with or without CKD, add-on therapy of ezetimibe should be recommended for patients who do not achieve their LDL goals with statin therapy or for patients who intolerant to statin therapy. 

The ezetimibe package insert reports myalgias incidence to be < 5% in patients and research has shown up to a 20% incidence of muscle-related AEs with statin therapy.^3,10 Based on the package information reporting increased AUC values of ezetimibe and its metabolites in patients with severe renal disease, it was anticipated there may be an increased risk of muscle-related AEs in patients with CKD.³ However, this study found the same incidence of muscle-related AEs in patients with and without CKD. Previous research on statin-intolerant patients found the incidence of muscle-related AEs with ezetimibe to be 23.0% and 28.8%.^11,12 This increased incidence of muscle-related AEs may be the result of including patients with a history of statin intolerance. Collectively, data from clinical trials and this study indicate that patients with prior intolerances to statins appear to have a higher likelihood of developing a muscle-related AEs with ezetimibe.^11,12 Clinicians and patients should be educated on the potential for these AEs and be aware that the likelihood may be greater if there is a history of statin intolerance. To our knowledge, this was the first study to evaluate muscle-related AEs with ezetimibe in patients with and without CKD.

Limitations

This retrospective chart review was performed over a prespecified period and only patients initiated on ezetimibe by a PACT pharmacist were included. This study did not assess the percentage of LDL reduction in patients on concomitant statins vs those who were not on concomitant statins. The study only included 173 patients. Additionally, the study was primarily composed of White men and may not be representative of other populations. In addition, veterans may not be representative of the general population given their high comorbidity burden and other exposures. Some reported muscle-related AEs associated with ezetimibe may be attributed to the nocebo effect.

Conclusions

The results of this retrospective chart review suggest there may be a larger mean reduction in LDL and non-HDL levels seen with ezetimibe therapy than reported within the literature. There was a larger mean reduction in LDL and non-HDL levels in patients with CKD than in patients without CKD. Additionally, there were the same rates of muscle-related AEs with ezetimibe therapy in patients with and without CKD. The rates of muscle-related AEs with ezetimibe therapy were higher than reported in the medication’s package insert, but lower than reported in literature that included statin-intolerant patients. These results indicate there may be a benefit to an increase in use of ezetimibe in clinical practice due to its increased effectiveness and safety in patients with and without CKD. Ultimately, this can help patients achieve their LDL goals as recommended by ACC clinical practice guidelines.

Artificial intelligence (AI) is rapidly evolving, with large language models (LLMs) marking a significant milestone in processing and generating human-like responses to natural language prompts. However, this advancement only signals the beginning of a more profound transformation in AI capabilities. The development of AI agents represents a new paradigm at the forefront of this evolution.

BACKGROUND

AI agents represent a leap forward from traditional LLM applications. While definitions may vary slightly among technology developers, the core concept remains: these agents are autonomous software entities designed to interact with their environment, make independent decisions, and execute tasks based on predefined goals.^1-3 What sets AI agents apart is their combination of sophisticated components within structured architectures. At their core, AI agents incorporate an LLM for response generation, which is augmented by a suite of tools to optimize workflow and complete tasks, memory capabilities for personalized interactions, and autonomous reasoning. This combination allows AI agents to plan, create subtasks, gather information, and learn iteratively from their own experiences or other AI agents.

The true potential of this technology becomes apparent when multiple AI agents collaborate within multiagent AI systems. This concept introduces a new level of flexibility and capability in tackling complex tasks. Autogen, CrewAI, and LangChain offer various agent network configurations, including hierarchical, sequential, conditional, or even parallel task execution.^4-6 This adaptability opens up a world of possibilities across various industries, but perhaps nowhere is the potential impact more exciting and profound than in health care.

AI agents in health care present an opportunity to revolutionize patient care, streamline administrative processes, and support complex clinical decision-making. This review examines 3 scenarios that illustrate the impact of AI agents in health care: a hypothetical sepsis management system, chronic disease management, and hospital patient flow optimization. This article will provide a detailed look at the technical implementation challenges, including the integration with existing health care IT systems, data privacy considerations, and the crucial role of explainable AI in maintaining trust and transparency.

It is challenging to implement AI agents in health care. Concerns include ensuring data quality and mitigating bias, seamlessly integrating these systems into existing clinical workflows, and navigating the complex ethical considerations that arise when deploying autonomous systems in health care. The integration with Internet of Things (IoT) devices for real-time patient data monitoring and the development of more sophisticated natural language interfaces to enhance future human-AI collaboration.

The adoption of AI agents in health care is only beginning, and it promises to be transformative. As AI continues to evolve, a comprehensive understanding of its applications, limitations, and ethical considerations is essential. This report provides a comprehensive overview of the current state, potential applications, and future directions of AI agents in health care, offering insights valuable to researchers, clinicians, and policymakers.

MultiAgent AI architecture

Sepsis Management

Despite advancements in broad-spectrum antibiotics, imaging, and life support systems, mortality rates associated with sepsis remain high. The complexity of optimizing care in clinical settings has hindered progress in managing sepsis. Previous attempts to develop predictive sepsis models have proven challenging.⁷ This report proposes a multiagent AI system designed to enhance comprehensive patient monitoring and care through coordinated AI-driven interventions.

Data Collection and Integration Agent. Powered by a controlled vocabulary to specify all data, the primary function for the data collection and integration agent is to clean, transform, and organize patient data from structured and unstructured sources. This agent prepares succinct summaries of consultant notes and formats data for human and machine consumption. All numerical data are presented graphically, including relevant historical data trends. The agent also digitally captures all orders in a structured format using a specified controlled vocabulary. This structured data feed supports the output of other agents, including documentation, treatment planning, and risk stratification, while also supplying the data structures for future training.

Diagnostic Agent. Critical illness is characterized by multiple abnormalities across a wide array of tests, ranging from plain chest X-ray, computed tomography (CT), blood cell composition, plasma chemistry, and microscopic evaluation of specimens. Additionally, life support parameters provide insights into disease severity and can inform management recommendations. These data offer a wide array of visual and numerical data to be used as input for computation, recommendation, and further training. For example, to evaluate fluid overload on chest X-rays or tissue histopathology slides, an AI agent can leverage deep learning models such as convolutional neural networks and vision transformers to analyze images like radiographs and histopathology slides.^8,9 Recurrent neural networks or transformer models process sequential data like time-series vital signs. The agent also implements ensemble methods that combine multiple machine learning algorithms to enhance diagnostic accuracy.

Risk Stratification Agent. This assesses severity and predicts potential outcomes. Morbidity and mortality risks are calculated using an established scoring system and individualized based on the history of other agents’ conditional patients. These are presented graphically, with major risk factors highlighted for explainability. 

Treatment Recommendation Agent. Using a reinforcement learning framework supplemented by up-to-date clinical guidelines, this system leverages historical data structured with standardized vocabulary to analyze patients with similar clinical features. Training is also conducted on the patient’s physiological data. All recommendations are presented via a dedicated user interface in a readable format, along with recommendations for editable, orderable items, references, and full-text snippets from previous research. Stop rules end computing if confidence in recommendations is too broad or no clear pathway can be computed with certainty, prompting human mitigation.

Resource Management Agent. This agent coordinates hospital resources using constraint programming techniques for optimal resource allocation, uses queueing theory models to predict and manage patient flow, and implements genetic algorithms for complex scheduling problems.^10,11

Monitoring and Alert Agent. By tracking patients’ progress and alerting staff to changes, this agent uses anomaly detection algorithms to identify unusual patterns in patient data and implement time-series forecasting models, such as autoregressive integrated moving average and prophet, to predict future patient states. The agent also uses stream-processing techniques for real-time data analysis.^12,13

Documentation and Reporting Agent. This agent maintains comprehensive medical records and generates reports. It employs advanced natural language processing techniques for automated report generation, uses advanced LLMs fine-tuned on medical corpora for narrative creation, and implements information-retrieval techniques to efficiently query patient records.

CLINICAL CASE STUDIES

To illustrate the functionality of a multiagent system, this report examines its application for managing sepsis. The data collection and integration agent continuously aggregates patient data from various sources, normalizing and timestamping it for consistent processing. The diagnostic agent analyzes this integrated data in real time, applying sepsis criteria and utilizing a deep learning model trained on a large sepsis dataset to detect subtle patterns.

The risk stratification agent calculates severity scores, such as the Sepsis-related Organ Failure Assessment (SOFA), quick SOFA (qSOFA), and Acute Physiology and Chronic Health Evaluation II, upon detecting a possible sepsis case.¹⁴ It predicts the likelihood of specific outcomes and estimates the potential trajectory of the patient’s condition for the next 24 to 48 hours. Based on this assessment, the treatment recommendation agent suggests an initial treatment plan, including appropriate antibiotics, fluid resuscitation protocols, and vasopressor recommendations, recommendations when indicated.

Concurrently, the resource management agent checks the availability of necessary resources and prioritizes allocation based on the severity. The monitoring agent tracks the patient’s response to interventions in real time, alerting the care team to any concerning changes or lack of expected improvement. Throughout this process, the documentation agent ensures that all actions, responses, and outcomes are meticulously recorded in a structured format and generates real-time updates for the patient’s electronic health record (EHR) and preparing summary reports for handoffs between care teams.

Administrative Workflow Support

Modern health care operations are resource-intensive, requiring coordination of advanced imaging, procedures, laboratory testing, and professional consultations.¹⁵ AI-powered health care administrative workflow systems are revolutionizing how medical facilities coordinate patient care. For patients with chronic cough, these systems seamlessly integrate scheduling, imaging, diagnostics, and follow-up care into a cohesive process that reduces administrative burden while improving patient outcomes. Through an intuitive interface and automated assistance, health care practitioners (HCPs) can track patient progress from initial consultation through diagnosis and treatment.

The process begins when an HCP enters a patient into the system, which triggers an automated CT scan scheduling system. The system considers factors like urgency, facility availability, and patient preferences to suggest optimal appointment times. Once imaging is complete, AI agents analyze the radiology reports, extract key findings, and generate structured summaries that highlight critical information such as “mild bronchial wall thickening with patchy ground-glass opacities” or “findings consistent with chronic bronchitis.”

Based on these findings, the system automatically generates evidence-based recommendations for follow-up care, such as pulmonology consultations or follow-up imaging in 3 months. These recommendations are presented to the ordering clinician, along with suggested appointment slots for specialist consultations. The system then manages the coordination of multiple appointments, ensuring each step in the patient’s care plan is properly sequenced and scheduled.

The entire process is monitored through a comprehensive dashboard that provides real-time updates on patient status, appointment schedules, and clinical recommendations. HCPs can track which patients require immediate attention, view upcoming appointments, and monitor the progress of ongoing care plans.

Multiagent AI Operation Optimization

Hospitals are complex entities that must function at different scales and respond in an agile, timely manner at all hours, deploying staff at various positions.¹⁶ A system of AI agents can receive signals from sensors monitoring foot traffic in the emergency department and trauma unit, as well as the availability of operating room staff, equipment, and intensive care unit beds. Smart sensors enable this monitoring through IoT networks. These networks benefit from advances in adaptive and consensus networking algorithms, along with recent advances in bioengineering and biocomputing.¹⁷

For example, in the case of imaging for suspected abdominal obstruction, an AI agent tasked with scheduling CTs could time the patient’s arrival based on acuity. Another AI agent could alert staff transporting the patient to the CT appointment, with the next location contingent on a clinical decision to proceed to the operating room. Yet another AI agent could summarize radiology interpretations and alert the surgery and anesthesia teams to a potential case, while others could notify operating room staff of equipment needs or reserve a bed. In this paradigm, AI agents facilitate more precise and timely communication between multiple staff members.

TECHNICAL IMPLEMENTATION

Large Language Models

Each agent uses a different LLM optimized for its specific task. For example, the diagnostic agent uses an LLM pretrained on a large corpus of biomedical literature and fine-tuned on a dataset of confirmed sepsis cases and their presentations.¹⁸ It implements few-shot learning techniques to adapt to rare or atypical presentations. The treatment recommendation agent also uses an LLM, employing a retrieval-augmented generation approach to access the latest clinical guidelines during inference. The documentation agent uses another advanced language model, fine-tuned on a large corpus of high-quality medical documentation, implementing controlled text generation techniques and utilizing a separate smaller model for real-time error checking and correction.

Interagent Quality Control

Agents learn from their own experience and the experience of other agents. They are equipped with user-defined rule-based and model-based systems for quality assurance, with clear stopping rules for human involvement and mitigation.

Sophisticated quality control measures bolster the system’s reliability, including ensemble techniques for result comparison, redundancy for critical tasks, and automatic human review for disagreements above a certain threshold. Each agent provides a calibrated confidence score with its output, used to weigh inputs in downstream tasks and trigger additional checks for low-confidence outputs.

A dedicated quality control agent monitors output from all agents, employing both supervised and unsupervised anomaly detection techniques. Feedback loops allow agents to evaluate the quality and utility of information received from other agents. The system implements a multiarmed bandit approach to dynamically adjust the influence of different agents based on their performance and periodically retrains agent models using federated learning techniques.¹⁹

Electronic Health Record Integration

Seamless EHR integration is crucial for practical implementation. The system has secure application programming interface access to various EHR platforms, implements OAuth 2.0 for authentication, and use HTTPS with perfect forward secrecy for all communications.²⁰ It works with HL7 FHIR to ensure interoperability and uses SNOMED CT for clinical terminology to ensure semantic interoperability across different EHRs.^21,22

The system implements a multilevel approval system for write-backs to EHRs, with different thresholds based on the information’s criticality. It uses digital signatures to ensure the integrity and nonrepudiation of AI-generated entries and implements blockchain technology to create an immutable and distributed ledger of all AI system actions.²³

Decision Transparency

To ensure transparency in decision-making processes, the system applies techniques (eg, local interpretable model-agnostic explanations and Shapley additive explanations) to provide insights into agent decision-making processes.^24-26 It provides customized visualizations for different stakeholders and allows users to explore alternative decision paths through what-if scenario modeling.²⁷

The system provides calibrated confidence indicators for each recommendation or decision, implementing a novel confidence calibration agent that continuously monitors and adjusts confidence scores based on observed outcomes.

Continuous Learning and Adaptation

The system employs several techniques to remain current with evolving medical knowledge. Federated learning includes information from diverse datasets across multiple institutions without compromising patient privacy.²⁸ A/B testing is used to safely deploy and compare new agent versions in controlled settings, implementing multiarmed bandit algorithms to efficiently explore new models while minimizing potential negative impacts. Human-in-the-loop learning and active learning techniques are used to incorporate feedback from HCPs and efficiently solicit expert input on the most informative data.²⁹

CLINICAL IMPLICATIONS

The implementation of multiagent AI systems in health care has several potential benefits: enhanced diagnostic accuracy, personalized treatment, improved efficiency, continuous monitoring, and resource optimization. A recent review of AI sepsis predictive models exhibited superior results to standard clinical scoring methods like qSOFA.³⁰ In oncology, such systems can result in more tailored treatments, enhancing outcomes.³¹ The implementation of an ambient dictation system can improve workflow and prevent HCP burnout.³²

ETHICAL CONSIDERATIONS AND AI OVERSIGHT

Integrating AI agents into health care raises significant ethical considerations that must be carefully addressed to ensure equitable and effective care delivery. One primary concern involves cultural and linguistic competency, as AI systems may struggle with cultural nuances, idioms, and context-specific communication patterns. This becomes particularly challenging in regions with diverse ethnic populations or immigrant communities, where medical terminology may not have direct translations and cultural beliefs significantly influence health care decisions. AI systems also may inherit and amplify existing biases in health care delivery, whether through HCP bias reflected in training data, patient bias affecting acceptance of AI-assisted care, or demographic underrepresentation during system development.

AI agents present unique opportunities for improving health care access and outcomes through community engagement, though such initiatives require thoughtful implementation. Predictive analytics can identify high-risk individuals within communities who may benefit from preventive care, while analysis of social determinants of health can enable more targeted interventions. However, these capabilities must be balanced with privacy concerns and the risk of surveillance, particularly in communities that distrust health care institutions. The potential for AI to bridge health care gaps must be weighed against the need to maintain cultural sensitivity and community trust.

The governance and oversight of health care AI systems requires a multistakeholder approach with clear lines of responsibility and accountability. This includes involvement from government health care agencies, professional medical associations, ethics boards, and independent auditors, all working together to establish and enforce standards while monitoring system performance and addressing potential biases. Health care organizations must maintain transparent policies about AI use, implement regular monitoring and evaluation protocols, and establish precise mechanisms for patient feedback and grievance resolution. Ongoing assessment and adjustment of these systems, informed by community feedback and outcomes data, will be crucial for their ethical implementation, ensuring that AI agents complement, rather than replace, human judgment and cultural sensitivity.

FUTURE DIRECTIONS

Despite the potential benefits, implementing multiagent AI systems in health care faces significant challenges that require careful consideration. Beyond the fundamental issues such as data quality and bias mitigation, health care organizations struggle with fragmented systems, inconsistent data formats, and varying quality. Technical infrastructure requirements are substantial, particularly in rural or underserved areas that lack robust networks and cybersecurity. HCPs already face significant cognitive load and time pressures, making integrating AI agents into existing workflows particularly challenging. There is also the critical issue of transparency and interpretability, as health care decisions require clear reasoning and accountability that many black-box AI systems struggle to provide.

The legal landscape introduces another layer of complexity, particularly regarding liability, consent, and privacy questions. When AI agents contribute to medical decisions, establishing clear lines of responsibility becomes crucial. There are also serious concerns about algorithmic fairness and the potential for AI systems to perpetuate or amplify existing inequities. The cost of implementation remains a significant barrier, requiring substantial investment in technology, training, and ongoing maintenance while ensuring resources are not diverted from direct patient care. Moreover, HCPs may resist adoption due to concerns about job security, loss of autonomy, or skepticism about AI capabilities while paradoxically facing risks of overreliance on AI systems that could lead to the degradation of human clinical skills.

Addressing these challenges requires a multifaceted approach that combines technical solutions with organizational and policy changes. Health care organizations must implement rigorous data validation processes and interoperability standards while developing hybrid models that balance sophisticated AI capabilities with interpretable techniques. Extensive research and iterative design processes, with direct input from HCPs, are essential for successful integration. Establishing independent ethics boards to oversee system development and deployment, conducting multicenter randomized controlled trials, and creating clear regulatory frameworks will ensure safe and effective implementation. Success will ultimately depend on ongoing collaboration between technology developers, HCPs, policymakers, and patients, maintaining a steady focus on improving patient care and outcomes while carefully navigating the complex challenges of AI integration in health care.^33-35

As multiagent AI systems in health care evolve, several exciting directions emerge. These include the integration of IoT and wearable devices, the development of more sophisticated natural language interfaces, and applying these systems to predictive maintenance of medical equipment.

CONCLUSIONS

The advent of multiagent AI systems in health care represents a paradigm shift in the approach to patient care, clinical decision making, and health care management. While these systems offer immense potential to transform health care delivery, their development and implementation must be guided by rigorous scientific validation, ethical considerations, and a patient-centered approach. The ultimate goal remains clear: harnessing the power of AI to improve patient outcomes, enhance the efficiency of health care delivery, and ultimately advance the health and well-being of patients.

Artificial intelligence (AI) is rapidly evolving, with large language models (LLMs) marking a significant milestone in processing and generating human-like responses to natural language prompts. However, this advancement only signals the beginning of a more profound transformation in AI capabilities. The development of AI agents represents a new paradigm at the forefront of this evolution.

BACKGROUND

AI agents represent a leap forward from traditional LLM applications. While definitions may vary slightly among technology developers, the core concept remains: these agents are autonomous software entities designed to interact with their environment, make independent decisions, and execute tasks based on predefined goals.^1-3 What sets AI agents apart is their combination of sophisticated components within structured architectures. At their core, AI agents incorporate an LLM for response generation, which is augmented by a suite of tools to optimize workflow and complete tasks, memory capabilities for personalized interactions, and autonomous reasoning. This combination allows AI agents to plan, create subtasks, gather information, and learn iteratively from their own experiences or other AI agents.

The true potential of this technology becomes apparent when multiple AI agents collaborate within multiagent AI systems. This concept introduces a new level of flexibility and capability in tackling complex tasks. Autogen, CrewAI, and LangChain offer various agent network configurations, including hierarchical, sequential, conditional, or even parallel task execution.^4-6 This adaptability opens up a world of possibilities across various industries, but perhaps nowhere is the potential impact more exciting and profound than in health care.

AI agents in health care present an opportunity to revolutionize patient care, streamline administrative processes, and support complex clinical decision-making. This review examines 3 scenarios that illustrate the impact of AI agents in health care: a hypothetical sepsis management system, chronic disease management, and hospital patient flow optimization. This article will provide a detailed look at the technical implementation challenges, including the integration with existing health care IT systems, data privacy considerations, and the crucial role of explainable AI in maintaining trust and transparency.

It is challenging to implement AI agents in health care. Concerns include ensuring data quality and mitigating bias, seamlessly integrating these systems into existing clinical workflows, and navigating the complex ethical considerations that arise when deploying autonomous systems in health care. The integration with Internet of Things (IoT) devices for real-time patient data monitoring and the development of more sophisticated natural language interfaces to enhance future human-AI collaboration.

The adoption of AI agents in health care is only beginning, and it promises to be transformative. As AI continues to evolve, a comprehensive understanding of its applications, limitations, and ethical considerations is essential. This report provides a comprehensive overview of the current state, potential applications, and future directions of AI agents in health care, offering insights valuable to researchers, clinicians, and policymakers.

MultiAgent AI architecture

Sepsis Management

Despite advancements in broad-spectrum antibiotics, imaging, and life support systems, mortality rates associated with sepsis remain high. The complexity of optimizing care in clinical settings has hindered progress in managing sepsis. Previous attempts to develop predictive sepsis models have proven challenging.⁷ This report proposes a multiagent AI system designed to enhance comprehensive patient monitoring and care through coordinated AI-driven interventions.

Data Collection and Integration Agent. Powered by a controlled vocabulary to specify all data, the primary function for the data collection and integration agent is to clean, transform, and organize patient data from structured and unstructured sources. This agent prepares succinct summaries of consultant notes and formats data for human and machine consumption. All numerical data are presented graphically, including relevant historical data trends. The agent also digitally captures all orders in a structured format using a specified controlled vocabulary. This structured data feed supports the output of other agents, including documentation, treatment planning, and risk stratification, while also supplying the data structures for future training.

Diagnostic Agent. Critical illness is characterized by multiple abnormalities across a wide array of tests, ranging from plain chest X-ray, computed tomography (CT), blood cell composition, plasma chemistry, and microscopic evaluation of specimens. Additionally, life support parameters provide insights into disease severity and can inform management recommendations. These data offer a wide array of visual and numerical data to be used as input for computation, recommendation, and further training. For example, to evaluate fluid overload on chest X-rays or tissue histopathology slides, an AI agent can leverage deep learning models such as convolutional neural networks and vision transformers to analyze images like radiographs and histopathology slides.^8,9 Recurrent neural networks or transformer models process sequential data like time-series vital signs. The agent also implements ensemble methods that combine multiple machine learning algorithms to enhance diagnostic accuracy.

Risk Stratification Agent. This assesses severity and predicts potential outcomes. Morbidity and mortality risks are calculated using an established scoring system and individualized based on the history of other agents’ conditional patients. These are presented graphically, with major risk factors highlighted for explainability. 

Treatment Recommendation Agent. Using a reinforcement learning framework supplemented by up-to-date clinical guidelines, this system leverages historical data structured with standardized vocabulary to analyze patients with similar clinical features. Training is also conducted on the patient’s physiological data. All recommendations are presented via a dedicated user interface in a readable format, along with recommendations for editable, orderable items, references, and full-text snippets from previous research. Stop rules end computing if confidence in recommendations is too broad or no clear pathway can be computed with certainty, prompting human mitigation.

Resource Management Agent. This agent coordinates hospital resources using constraint programming techniques for optimal resource allocation, uses queueing theory models to predict and manage patient flow, and implements genetic algorithms for complex scheduling problems.^10,11

Monitoring and Alert Agent. By tracking patients’ progress and alerting staff to changes, this agent uses anomaly detection algorithms to identify unusual patterns in patient data and implement time-series forecasting models, such as autoregressive integrated moving average and prophet, to predict future patient states. The agent also uses stream-processing techniques for real-time data analysis.^12,13

Documentation and Reporting Agent. This agent maintains comprehensive medical records and generates reports. It employs advanced natural language processing techniques for automated report generation, uses advanced LLMs fine-tuned on medical corpora for narrative creation, and implements information-retrieval techniques to efficiently query patient records.

CLINICAL CASE STUDIES

To illustrate the functionality of a multiagent system, this report examines its application for managing sepsis. The data collection and integration agent continuously aggregates patient data from various sources, normalizing and timestamping it for consistent processing. The diagnostic agent analyzes this integrated data in real time, applying sepsis criteria and utilizing a deep learning model trained on a large sepsis dataset to detect subtle patterns.

The risk stratification agent calculates severity scores, such as the Sepsis-related Organ Failure Assessment (SOFA), quick SOFA (qSOFA), and Acute Physiology and Chronic Health Evaluation II, upon detecting a possible sepsis case.¹⁴ It predicts the likelihood of specific outcomes and estimates the potential trajectory of the patient’s condition for the next 24 to 48 hours. Based on this assessment, the treatment recommendation agent suggests an initial treatment plan, including appropriate antibiotics, fluid resuscitation protocols, and vasopressor recommendations, recommendations when indicated.

Concurrently, the resource management agent checks the availability of necessary resources and prioritizes allocation based on the severity. The monitoring agent tracks the patient’s response to interventions in real time, alerting the care team to any concerning changes or lack of expected improvement. Throughout this process, the documentation agent ensures that all actions, responses, and outcomes are meticulously recorded in a structured format and generates real-time updates for the patient’s electronic health record (EHR) and preparing summary reports for handoffs between care teams.

Administrative Workflow Support

Modern health care operations are resource-intensive, requiring coordination of advanced imaging, procedures, laboratory testing, and professional consultations.¹⁵ AI-powered health care administrative workflow systems are revolutionizing how medical facilities coordinate patient care. For patients with chronic cough, these systems seamlessly integrate scheduling, imaging, diagnostics, and follow-up care into a cohesive process that reduces administrative burden while improving patient outcomes. Through an intuitive interface and automated assistance, health care practitioners (HCPs) can track patient progress from initial consultation through diagnosis and treatment.

The process begins when an HCP enters a patient into the system, which triggers an automated CT scan scheduling system. The system considers factors like urgency, facility availability, and patient preferences to suggest optimal appointment times. Once imaging is complete, AI agents analyze the radiology reports, extract key findings, and generate structured summaries that highlight critical information such as “mild bronchial wall thickening with patchy ground-glass opacities” or “findings consistent with chronic bronchitis.”

Based on these findings, the system automatically generates evidence-based recommendations for follow-up care, such as pulmonology consultations or follow-up imaging in 3 months. These recommendations are presented to the ordering clinician, along with suggested appointment slots for specialist consultations. The system then manages the coordination of multiple appointments, ensuring each step in the patient’s care plan is properly sequenced and scheduled.

The entire process is monitored through a comprehensive dashboard that provides real-time updates on patient status, appointment schedules, and clinical recommendations. HCPs can track which patients require immediate attention, view upcoming appointments, and monitor the progress of ongoing care plans.

Multiagent AI Operation Optimization

Hospitals are complex entities that must function at different scales and respond in an agile, timely manner at all hours, deploying staff at various positions.¹⁶ A system of AI agents can receive signals from sensors monitoring foot traffic in the emergency department and trauma unit, as well as the availability of operating room staff, equipment, and intensive care unit beds. Smart sensors enable this monitoring through IoT networks. These networks benefit from advances in adaptive and consensus networking algorithms, along with recent advances in bioengineering and biocomputing.¹⁷

For example, in the case of imaging for suspected abdominal obstruction, an AI agent tasked with scheduling CTs could time the patient’s arrival based on acuity. Another AI agent could alert staff transporting the patient to the CT appointment, with the next location contingent on a clinical decision to proceed to the operating room. Yet another AI agent could summarize radiology interpretations and alert the surgery and anesthesia teams to a potential case, while others could notify operating room staff of equipment needs or reserve a bed. In this paradigm, AI agents facilitate more precise and timely communication between multiple staff members.

TECHNICAL IMPLEMENTATION

Large Language Models

Each agent uses a different LLM optimized for its specific task. For example, the diagnostic agent uses an LLM pretrained on a large corpus of biomedical literature and fine-tuned on a dataset of confirmed sepsis cases and their presentations.¹⁸ It implements few-shot learning techniques to adapt to rare or atypical presentations. The treatment recommendation agent also uses an LLM, employing a retrieval-augmented generation approach to access the latest clinical guidelines during inference. The documentation agent uses another advanced language model, fine-tuned on a large corpus of high-quality medical documentation, implementing controlled text generation techniques and utilizing a separate smaller model for real-time error checking and correction.

Interagent Quality Control

Agents learn from their own experience and the experience of other agents. They are equipped with user-defined rule-based and model-based systems for quality assurance, with clear stopping rules for human involvement and mitigation.

Sophisticated quality control measures bolster the system’s reliability, including ensemble techniques for result comparison, redundancy for critical tasks, and automatic human review for disagreements above a certain threshold. Each agent provides a calibrated confidence score with its output, used to weigh inputs in downstream tasks and trigger additional checks for low-confidence outputs.

A dedicated quality control agent monitors output from all agents, employing both supervised and unsupervised anomaly detection techniques. Feedback loops allow agents to evaluate the quality and utility of information received from other agents. The system implements a multiarmed bandit approach to dynamically adjust the influence of different agents based on their performance and periodically retrains agent models using federated learning techniques.¹⁹

Electronic Health Record Integration

Seamless EHR integration is crucial for practical implementation. The system has secure application programming interface access to various EHR platforms, implements OAuth 2.0 for authentication, and use HTTPS with perfect forward secrecy for all communications.²⁰ It works with HL7 FHIR to ensure interoperability and uses SNOMED CT for clinical terminology to ensure semantic interoperability across different EHRs.^21,22

The system implements a multilevel approval system for write-backs to EHRs, with different thresholds based on the information’s criticality. It uses digital signatures to ensure the integrity and nonrepudiation of AI-generated entries and implements blockchain technology to create an immutable and distributed ledger of all AI system actions.²³

Decision Transparency

To ensure transparency in decision-making processes, the system applies techniques (eg, local interpretable model-agnostic explanations and Shapley additive explanations) to provide insights into agent decision-making processes.^24-26 It provides customized visualizations for different stakeholders and allows users to explore alternative decision paths through what-if scenario modeling.²⁷

The system provides calibrated confidence indicators for each recommendation or decision, implementing a novel confidence calibration agent that continuously monitors and adjusts confidence scores based on observed outcomes.

Continuous Learning and Adaptation

The system employs several techniques to remain current with evolving medical knowledge. Federated learning includes information from diverse datasets across multiple institutions without compromising patient privacy.²⁸ A/B testing is used to safely deploy and compare new agent versions in controlled settings, implementing multiarmed bandit algorithms to efficiently explore new models while minimizing potential negative impacts. Human-in-the-loop learning and active learning techniques are used to incorporate feedback from HCPs and efficiently solicit expert input on the most informative data.²⁹

CLINICAL IMPLICATIONS

The implementation of multiagent AI systems in health care has several potential benefits: enhanced diagnostic accuracy, personalized treatment, improved efficiency, continuous monitoring, and resource optimization. A recent review of AI sepsis predictive models exhibited superior results to standard clinical scoring methods like qSOFA.³⁰ In oncology, such systems can result in more tailored treatments, enhancing outcomes.³¹ The implementation of an ambient dictation system can improve workflow and prevent HCP burnout.³²

ETHICAL CONSIDERATIONS AND AI OVERSIGHT

Integrating AI agents into health care raises significant ethical considerations that must be carefully addressed to ensure equitable and effective care delivery. One primary concern involves cultural and linguistic competency, as AI systems may struggle with cultural nuances, idioms, and context-specific communication patterns. This becomes particularly challenging in regions with diverse ethnic populations or immigrant communities, where medical terminology may not have direct translations and cultural beliefs significantly influence health care decisions. AI systems also may inherit and amplify existing biases in health care delivery, whether through HCP bias reflected in training data, patient bias affecting acceptance of AI-assisted care, or demographic underrepresentation during system development.

AI agents present unique opportunities for improving health care access and outcomes through community engagement, though such initiatives require thoughtful implementation. Predictive analytics can identify high-risk individuals within communities who may benefit from preventive care, while analysis of social determinants of health can enable more targeted interventions. However, these capabilities must be balanced with privacy concerns and the risk of surveillance, particularly in communities that distrust health care institutions. The potential for AI to bridge health care gaps must be weighed against the need to maintain cultural sensitivity and community trust.

The governance and oversight of health care AI systems requires a multistakeholder approach with clear lines of responsibility and accountability. This includes involvement from government health care agencies, professional medical associations, ethics boards, and independent auditors, all working together to establish and enforce standards while monitoring system performance and addressing potential biases. Health care organizations must maintain transparent policies about AI use, implement regular monitoring and evaluation protocols, and establish precise mechanisms for patient feedback and grievance resolution. Ongoing assessment and adjustment of these systems, informed by community feedback and outcomes data, will be crucial for their ethical implementation, ensuring that AI agents complement, rather than replace, human judgment and cultural sensitivity.

FUTURE DIRECTIONS

Despite the potential benefits, implementing multiagent AI systems in health care faces significant challenges that require careful consideration. Beyond the fundamental issues such as data quality and bias mitigation, health care organizations struggle with fragmented systems, inconsistent data formats, and varying quality. Technical infrastructure requirements are substantial, particularly in rural or underserved areas that lack robust networks and cybersecurity. HCPs already face significant cognitive load and time pressures, making integrating AI agents into existing workflows particularly challenging. There is also the critical issue of transparency and interpretability, as health care decisions require clear reasoning and accountability that many black-box AI systems struggle to provide.

The legal landscape introduces another layer of complexity, particularly regarding liability, consent, and privacy questions. When AI agents contribute to medical decisions, establishing clear lines of responsibility becomes crucial. There are also serious concerns about algorithmic fairness and the potential for AI systems to perpetuate or amplify existing inequities. The cost of implementation remains a significant barrier, requiring substantial investment in technology, training, and ongoing maintenance while ensuring resources are not diverted from direct patient care. Moreover, HCPs may resist adoption due to concerns about job security, loss of autonomy, or skepticism about AI capabilities while paradoxically facing risks of overreliance on AI systems that could lead to the degradation of human clinical skills.

Addressing these challenges requires a multifaceted approach that combines technical solutions with organizational and policy changes. Health care organizations must implement rigorous data validation processes and interoperability standards while developing hybrid models that balance sophisticated AI capabilities with interpretable techniques. Extensive research and iterative design processes, with direct input from HCPs, are essential for successful integration. Establishing independent ethics boards to oversee system development and deployment, conducting multicenter randomized controlled trials, and creating clear regulatory frameworks will ensure safe and effective implementation. Success will ultimately depend on ongoing collaboration between technology developers, HCPs, policymakers, and patients, maintaining a steady focus on improving patient care and outcomes while carefully navigating the complex challenges of AI integration in health care.^33-35

As multiagent AI systems in health care evolve, several exciting directions emerge. These include the integration of IoT and wearable devices, the development of more sophisticated natural language interfaces, and applying these systems to predictive maintenance of medical equipment.

CONCLUSIONS

The advent of multiagent AI systems in health care represents a paradigm shift in the approach to patient care, clinical decision making, and health care management. While these systems offer immense potential to transform health care delivery, their development and implementation must be guided by rigorous scientific validation, ethical considerations, and a patient-centered approach. The ultimate goal remains clear: harnessing the power of AI to improve patient outcomes, enhance the efficiency of health care delivery, and ultimately advance the health and well-being of patients.

Artificial intelligence (AI) is rapidly evolving, with large language models (LLMs) marking a significant milestone in processing and generating human-like responses to natural language prompts. However, this advancement only signals the beginning of a more profound transformation in AI capabilities. The development of AI agents represents a new paradigm at the forefront of this evolution.

BACKGROUND

AI agents represent a leap forward from traditional LLM applications. While definitions may vary slightly among technology developers, the core concept remains: these agents are autonomous software entities designed to interact with their environment, make independent decisions, and execute tasks based on predefined goals.^1-3 What sets AI agents apart is their combination of sophisticated components within structured architectures. At their core, AI agents incorporate an LLM for response generation, which is augmented by a suite of tools to optimize workflow and complete tasks, memory capabilities for personalized interactions, and autonomous reasoning. This combination allows AI agents to plan, create subtasks, gather information, and learn iteratively from their own experiences or other AI agents.

The true potential of this technology becomes apparent when multiple AI agents collaborate within multiagent AI systems. This concept introduces a new level of flexibility and capability in tackling complex tasks. Autogen, CrewAI, and LangChain offer various agent network configurations, including hierarchical, sequential, conditional, or even parallel task execution.^4-6 This adaptability opens up a world of possibilities across various industries, but perhaps nowhere is the potential impact more exciting and profound than in health care.

AI agents in health care present an opportunity to revolutionize patient care, streamline administrative processes, and support complex clinical decision-making. This review examines 3 scenarios that illustrate the impact of AI agents in health care: a hypothetical sepsis management system, chronic disease management, and hospital patient flow optimization. This article will provide a detailed look at the technical implementation challenges, including the integration with existing health care IT systems, data privacy considerations, and the crucial role of explainable AI in maintaining trust and transparency.

It is challenging to implement AI agents in health care. Concerns include ensuring data quality and mitigating bias, seamlessly integrating these systems into existing clinical workflows, and navigating the complex ethical considerations that arise when deploying autonomous systems in health care. The integration with Internet of Things (IoT) devices for real-time patient data monitoring and the development of more sophisticated natural language interfaces to enhance future human-AI collaboration.

The adoption of AI agents in health care is only beginning, and it promises to be transformative. As AI continues to evolve, a comprehensive understanding of its applications, limitations, and ethical considerations is essential. This report provides a comprehensive overview of the current state, potential applications, and future directions of AI agents in health care, offering insights valuable to researchers, clinicians, and policymakers.

MultiAgent AI architecture

Sepsis Management

Despite advancements in broad-spectrum antibiotics, imaging, and life support systems, mortality rates associated with sepsis remain high. The complexity of optimizing care in clinical settings has hindered progress in managing sepsis. Previous attempts to develop predictive sepsis models have proven challenging.⁷ This report proposes a multiagent AI system designed to enhance comprehensive patient monitoring and care through coordinated AI-driven interventions.

Data Collection and Integration Agent. Powered by a controlled vocabulary to specify all data, the primary function for the data collection and integration agent is to clean, transform, and organize patient data from structured and unstructured sources. This agent prepares succinct summaries of consultant notes and formats data for human and machine consumption. All numerical data are presented graphically, including relevant historical data trends. The agent also digitally captures all orders in a structured format using a specified controlled vocabulary. This structured data feed supports the output of other agents, including documentation, treatment planning, and risk stratification, while also supplying the data structures for future training.

Diagnostic Agent. Critical illness is characterized by multiple abnormalities across a wide array of tests, ranging from plain chest X-ray, computed tomography (CT), blood cell composition, plasma chemistry, and microscopic evaluation of specimens. Additionally, life support parameters provide insights into disease severity and can inform management recommendations. These data offer a wide array of visual and numerical data to be used as input for computation, recommendation, and further training. For example, to evaluate fluid overload on chest X-rays or tissue histopathology slides, an AI agent can leverage deep learning models such as convolutional neural networks and vision transformers to analyze images like radiographs and histopathology slides.^8,9 Recurrent neural networks or transformer models process sequential data like time-series vital signs. The agent also implements ensemble methods that combine multiple machine learning algorithms to enhance diagnostic accuracy.

Risk Stratification Agent. This assesses severity and predicts potential outcomes. Morbidity and mortality risks are calculated using an established scoring system and individualized based on the history of other agents’ conditional patients. These are presented graphically, with major risk factors highlighted for explainability. 

Treatment Recommendation Agent. Using a reinforcement learning framework supplemented by up-to-date clinical guidelines, this system leverages historical data structured with standardized vocabulary to analyze patients with similar clinical features. Training is also conducted on the patient’s physiological data. All recommendations are presented via a dedicated user interface in a readable format, along with recommendations for editable, orderable items, references, and full-text snippets from previous research. Stop rules end computing if confidence in recommendations is too broad or no clear pathway can be computed with certainty, prompting human mitigation.

Resource Management Agent. This agent coordinates hospital resources using constraint programming techniques for optimal resource allocation, uses queueing theory models to predict and manage patient flow, and implements genetic algorithms for complex scheduling problems.^10,11

Monitoring and Alert Agent. By tracking patients’ progress and alerting staff to changes, this agent uses anomaly detection algorithms to identify unusual patterns in patient data and implement time-series forecasting models, such as autoregressive integrated moving average and prophet, to predict future patient states. The agent also uses stream-processing techniques for real-time data analysis.^12,13

Documentation and Reporting Agent. This agent maintains comprehensive medical records and generates reports. It employs advanced natural language processing techniques for automated report generation, uses advanced LLMs fine-tuned on medical corpora for narrative creation, and implements information-retrieval techniques to efficiently query patient records.

CLINICAL CASE STUDIES

To illustrate the functionality of a multiagent system, this report examines its application for managing sepsis. The data collection and integration agent continuously aggregates patient data from various sources, normalizing and timestamping it for consistent processing. The diagnostic agent analyzes this integrated data in real time, applying sepsis criteria and utilizing a deep learning model trained on a large sepsis dataset to detect subtle patterns.

The risk stratification agent calculates severity scores, such as the Sepsis-related Organ Failure Assessment (SOFA), quick SOFA (qSOFA), and Acute Physiology and Chronic Health Evaluation II, upon detecting a possible sepsis case.¹⁴ It predicts the likelihood of specific outcomes and estimates the potential trajectory of the patient’s condition for the next 24 to 48 hours. Based on this assessment, the treatment recommendation agent suggests an initial treatment plan, including appropriate antibiotics, fluid resuscitation protocols, and vasopressor recommendations, recommendations when indicated.

Concurrently, the resource management agent checks the availability of necessary resources and prioritizes allocation based on the severity. The monitoring agent tracks the patient’s response to interventions in real time, alerting the care team to any concerning changes or lack of expected improvement. Throughout this process, the documentation agent ensures that all actions, responses, and outcomes are meticulously recorded in a structured format and generates real-time updates for the patient’s electronic health record (EHR) and preparing summary reports for handoffs between care teams.

Administrative Workflow Support

Modern health care operations are resource-intensive, requiring coordination of advanced imaging, procedures, laboratory testing, and professional consultations.¹⁵ AI-powered health care administrative workflow systems are revolutionizing how medical facilities coordinate patient care. For patients with chronic cough, these systems seamlessly integrate scheduling, imaging, diagnostics, and follow-up care into a cohesive process that reduces administrative burden while improving patient outcomes. Through an intuitive interface and automated assistance, health care practitioners (HCPs) can track patient progress from initial consultation through diagnosis and treatment.

The process begins when an HCP enters a patient into the system, which triggers an automated CT scan scheduling system. The system considers factors like urgency, facility availability, and patient preferences to suggest optimal appointment times. Once imaging is complete, AI agents analyze the radiology reports, extract key findings, and generate structured summaries that highlight critical information such as “mild bronchial wall thickening with patchy ground-glass opacities” or “findings consistent with chronic bronchitis.”

Based on these findings, the system automatically generates evidence-based recommendations for follow-up care, such as pulmonology consultations or follow-up imaging in 3 months. These recommendations are presented to the ordering clinician, along with suggested appointment slots for specialist consultations. The system then manages the coordination of multiple appointments, ensuring each step in the patient’s care plan is properly sequenced and scheduled.

The entire process is monitored through a comprehensive dashboard that provides real-time updates on patient status, appointment schedules, and clinical recommendations. HCPs can track which patients require immediate attention, view upcoming appointments, and monitor the progress of ongoing care plans.

Multiagent AI Operation Optimization

Hospitals are complex entities that must function at different scales and respond in an agile, timely manner at all hours, deploying staff at various positions.¹⁶ A system of AI agents can receive signals from sensors monitoring foot traffic in the emergency department and trauma unit, as well as the availability of operating room staff, equipment, and intensive care unit beds. Smart sensors enable this monitoring through IoT networks. These networks benefit from advances in adaptive and consensus networking algorithms, along with recent advances in bioengineering and biocomputing.¹⁷

For example, in the case of imaging for suspected abdominal obstruction, an AI agent tasked with scheduling CTs could time the patient’s arrival based on acuity. Another AI agent could alert staff transporting the patient to the CT appointment, with the next location contingent on a clinical decision to proceed to the operating room. Yet another AI agent could summarize radiology interpretations and alert the surgery and anesthesia teams to a potential case, while others could notify operating room staff of equipment needs or reserve a bed. In this paradigm, AI agents facilitate more precise and timely communication between multiple staff members.

TECHNICAL IMPLEMENTATION

Large Language Models

Each agent uses a different LLM optimized for its specific task. For example, the diagnostic agent uses an LLM pretrained on a large corpus of biomedical literature and fine-tuned on a dataset of confirmed sepsis cases and their presentations.¹⁸ It implements few-shot learning techniques to adapt to rare or atypical presentations. The treatment recommendation agent also uses an LLM, employing a retrieval-augmented generation approach to access the latest clinical guidelines during inference. The documentation agent uses another advanced language model, fine-tuned on a large corpus of high-quality medical documentation, implementing controlled text generation techniques and utilizing a separate smaller model for real-time error checking and correction.

Interagent Quality Control

Agents learn from their own experience and the experience of other agents. They are equipped with user-defined rule-based and model-based systems for quality assurance, with clear stopping rules for human involvement and mitigation.

Sophisticated quality control measures bolster the system’s reliability, including ensemble techniques for result comparison, redundancy for critical tasks, and automatic human review for disagreements above a certain threshold. Each agent provides a calibrated confidence score with its output, used to weigh inputs in downstream tasks and trigger additional checks for low-confidence outputs.

A dedicated quality control agent monitors output from all agents, employing both supervised and unsupervised anomaly detection techniques. Feedback loops allow agents to evaluate the quality and utility of information received from other agents. The system implements a multiarmed bandit approach to dynamically adjust the influence of different agents based on their performance and periodically retrains agent models using federated learning techniques.¹⁹

Electronic Health Record Integration

Seamless EHR integration is crucial for practical implementation. The system has secure application programming interface access to various EHR platforms, implements OAuth 2.0 for authentication, and use HTTPS with perfect forward secrecy for all communications.²⁰ It works with HL7 FHIR to ensure interoperability and uses SNOMED CT for clinical terminology to ensure semantic interoperability across different EHRs.^21,22

The system implements a multilevel approval system for write-backs to EHRs, with different thresholds based on the information’s criticality. It uses digital signatures to ensure the integrity and nonrepudiation of AI-generated entries and implements blockchain technology to create an immutable and distributed ledger of all AI system actions.²³

Decision Transparency

To ensure transparency in decision-making processes, the system applies techniques (eg, local interpretable model-agnostic explanations and Shapley additive explanations) to provide insights into agent decision-making processes.^24-26 It provides customized visualizations for different stakeholders and allows users to explore alternative decision paths through what-if scenario modeling.²⁷

The system provides calibrated confidence indicators for each recommendation or decision, implementing a novel confidence calibration agent that continuously monitors and adjusts confidence scores based on observed outcomes.

Continuous Learning and Adaptation

The system employs several techniques to remain current with evolving medical knowledge. Federated learning includes information from diverse datasets across multiple institutions without compromising patient privacy.²⁸ A/B testing is used to safely deploy and compare new agent versions in controlled settings, implementing multiarmed bandit algorithms to efficiently explore new models while minimizing potential negative impacts. Human-in-the-loop learning and active learning techniques are used to incorporate feedback from HCPs and efficiently solicit expert input on the most informative data.²⁹

CLINICAL IMPLICATIONS

The implementation of multiagent AI systems in health care has several potential benefits: enhanced diagnostic accuracy, personalized treatment, improved efficiency, continuous monitoring, and resource optimization. A recent review of AI sepsis predictive models exhibited superior results to standard clinical scoring methods like qSOFA.³⁰ In oncology, such systems can result in more tailored treatments, enhancing outcomes.³¹ The implementation of an ambient dictation system can improve workflow and prevent HCP burnout.³²

ETHICAL CONSIDERATIONS AND AI OVERSIGHT

Integrating AI agents into health care raises significant ethical considerations that must be carefully addressed to ensure equitable and effective care delivery. One primary concern involves cultural and linguistic competency, as AI systems may struggle with cultural nuances, idioms, and context-specific communication patterns. This becomes particularly challenging in regions with diverse ethnic populations or immigrant communities, where medical terminology may not have direct translations and cultural beliefs significantly influence health care decisions. AI systems also may inherit and amplify existing biases in health care delivery, whether through HCP bias reflected in training data, patient bias affecting acceptance of AI-assisted care, or demographic underrepresentation during system development.

AI agents present unique opportunities for improving health care access and outcomes through community engagement, though such initiatives require thoughtful implementation. Predictive analytics can identify high-risk individuals within communities who may benefit from preventive care, while analysis of social determinants of health can enable more targeted interventions. However, these capabilities must be balanced with privacy concerns and the risk of surveillance, particularly in communities that distrust health care institutions. The potential for AI to bridge health care gaps must be weighed against the need to maintain cultural sensitivity and community trust.

The governance and oversight of health care AI systems requires a multistakeholder approach with clear lines of responsibility and accountability. This includes involvement from government health care agencies, professional medical associations, ethics boards, and independent auditors, all working together to establish and enforce standards while monitoring system performance and addressing potential biases. Health care organizations must maintain transparent policies about AI use, implement regular monitoring and evaluation protocols, and establish precise mechanisms for patient feedback and grievance resolution. Ongoing assessment and adjustment of these systems, informed by community feedback and outcomes data, will be crucial for their ethical implementation, ensuring that AI agents complement, rather than replace, human judgment and cultural sensitivity.

FUTURE DIRECTIONS

Despite the potential benefits, implementing multiagent AI systems in health care faces significant challenges that require careful consideration. Beyond the fundamental issues such as data quality and bias mitigation, health care organizations struggle with fragmented systems, inconsistent data formats, and varying quality. Technical infrastructure requirements are substantial, particularly in rural or underserved areas that lack robust networks and cybersecurity. HCPs already face significant cognitive load and time pressures, making integrating AI agents into existing workflows particularly challenging. There is also the critical issue of transparency and interpretability, as health care decisions require clear reasoning and accountability that many black-box AI systems struggle to provide.

The legal landscape introduces another layer of complexity, particularly regarding liability, consent, and privacy questions. When AI agents contribute to medical decisions, establishing clear lines of responsibility becomes crucial. There are also serious concerns about algorithmic fairness and the potential for AI systems to perpetuate or amplify existing inequities. The cost of implementation remains a significant barrier, requiring substantial investment in technology, training, and ongoing maintenance while ensuring resources are not diverted from direct patient care. Moreover, HCPs may resist adoption due to concerns about job security, loss of autonomy, or skepticism about AI capabilities while paradoxically facing risks of overreliance on AI systems that could lead to the degradation of human clinical skills.

Addressing these challenges requires a multifaceted approach that combines technical solutions with organizational and policy changes. Health care organizations must implement rigorous data validation processes and interoperability standards while developing hybrid models that balance sophisticated AI capabilities with interpretable techniques. Extensive research and iterative design processes, with direct input from HCPs, are essential for successful integration. Establishing independent ethics boards to oversee system development and deployment, conducting multicenter randomized controlled trials, and creating clear regulatory frameworks will ensure safe and effective implementation. Success will ultimately depend on ongoing collaboration between technology developers, HCPs, policymakers, and patients, maintaining a steady focus on improving patient care and outcomes while carefully navigating the complex challenges of AI integration in health care.^33-35

As multiagent AI systems in health care evolve, several exciting directions emerge. These include the integration of IoT and wearable devices, the development of more sophisticated natural language interfaces, and applying these systems to predictive maintenance of medical equipment.

CONCLUSIONS

The advent of multiagent AI systems in health care represents a paradigm shift in the approach to patient care, clinical decision making, and health care management. While these systems offer immense potential to transform health care delivery, their development and implementation must be guided by rigorous scientific validation, ethical considerations, and a patient-centered approach. The ultimate goal remains clear: harnessing the power of AI to improve patient outcomes, enhance the efficiency of health care delivery, and ultimately advance the health and well-being of patients.

DISCUSSION

A diagnosis of familial glomangiomatosis was made based on the clinical history and histopathologic findings from the punch biopsy. Glomus tumors are comprised of glomus cells, or undifferentiated smooth muscle cells responsible for thermoregulation.¹ Glomus tumors are classified into 3 categories: solid (predominantly glomus cells), glomangiomas (predominantly blood vessels), and glomangiomyomas (predominantly smooth muscle cells).² Glomangiomas, which comprise up to 20% of glomus tumors, typically present as bluish-purple, papular or nodular, hyperkeratotic lesions that are 2 to 10 mm in diameter.¹ These lesions are tender to palpation and pain may worsen with exposure to cold. Glomangiomas are associated with a classic triad of symptoms that include hypersensitivity, intermittent pain, and pinpoint pain, but patients rarely present with all 3.³

Glomangiomas tend to occur in areas rich with glomus bodies—the distal extremities—specifically the palms, wrists, forearms, feet, and subungual regions; visceral organ involvement including the GI tract is very rare.^1,4,5 About 38% to 68% of these lesions are hereditary or can be sporadic. If these lesions are hereditary, a patient is said to have familial glomangiomatosis. In familial glomangiomatosis, the glomulin gene is mutated in an autosomal dominant inheritance pattern with incomplete penetrance and variable expressivity. Inherited glomangiomas may present at birth or puberty similar to other vascular anomalies.⁴

Histopathology of glomangiomas shows rows of glomus cells (modified smooth muscle cells) surrounding distorted venous channels.^6,7 These lesions stain positive for CD34, vimentin, calponin, and α-smooth muscle actin, but are negative for desmin, S-100, and von Willebrand factor.^1,8 Although the patient’s medical history and physical examination are important in establishing the diagnosis, histopathology is confirmatory.

While the punch biopsy results were pending, a complete blood count (CBC) and fecal occult blood test (FOBT) were ordered due to concerns for blue rubber bleb nevus syndrome (BRBNS), a rare disorder with about 200 reported cases. Patients present with multiple blue to violaceous compressible nodules that feel rubbery in consistency and may be painful with compression. Lesions may be up to 5 cm in diameter and with time, the GI tract may also become involved.⁹ In the GI tract, the small bowel is the most common site of involvement and patients may present with severe iron deficiency anemia due to hemorrhage.¹⁰ Histopathologic features are nonspecific and have features of venous malformations but may include large, tortuous, dilated vessels with a single endothelial lining with possible smooth muscle in vessel walls or calcifications.¹¹ Due to concerns of BRBNS, laboratory studies (CBC and FOBT) were obtained but did not indicate the patient was experiencing a GI hemorrhage.

The differential diagnosis included Maffucci syndrome, also known as dyschondrodysplasia with hemangiomas, enchondromatosis with cavernous hemangiomas, or hemangiomatosis chondrodystrophic. Patients with Maffucci syndrome present with multiple enchondromas, soft tissue hemangiomas or lymphangiomas, and gliomas. These lesions tend to undergo malignant transformation from enchondromas to chondrosarcomas and hemangiomas to vascular sarcomas.¹² This diagnosis was less likely in the patient in this case as there were no concerns of skeletal involvement upon history and physical examination.

Lastly, Klippel-Trénaunay syndrome can be associated with similar cutaneous vascular manifestations.^13,14 This syndrome occurs due to somatic mutations altering angiogenesis during embryological development. This results in varicosities of superficial and deep venous systems, persistent embryonic veins, and valvular incompetence. However, these patients typically have capillary manifestations such as a flat, red, or purple port-wine stain present at birth and associated limb hypertrophy predominantly affecting a single lower limb.^15,16 The patient reported not having the lesions present at birth and because bilateral upper/lower extremity and trunk involvement is rare in this syndrome, a Klippel-Trénaunay syndrome diagnosis was unlikely even in the absence of biopsy results.

Treatment

Based on pathology results, the patient was diagnosed with familial glomangiomatosis and a discussion of treatment options ensued. Asymptomatic lesions can be periodically managed. In addition, there are several treatments for symptomatic lesions. Symptomatic lesions may be tender to palpation and or hypersensitive to temperature change (cold). Though they exhibit slow growth, they can invade surrounding tissues including nerve sheaths which can worsen pain.

Surgical resection, sclerotherapy, laser therapy, and electron beam radiation have been used on patients with symptomatic lesions.^8,17 Sclerotherapy involves introducing sterile solutions into a blood vessel’s lumen or into the vascular lesion itself to induce permanent endofibrosis and ablation.¹⁷ Hypertonic saline, sodium tetradecyl sulfate (STS), and absolute alcohol have been used to treat vascular anomalies as well as glomangiomas.¹⁷ Though case reports have noticed significant improvement in symptomatic lesions, sclerotherapy has been shown to be more effective in treating venous malformations than glomangiomas.^18,19

A long-pulsed 1064-nm neodymium-doped yttrium aluminum garnet (Nd:YAG) laser has also been effective in treating larger glomangiomas that would otherwise be difficult to excise.²⁰ The Nd:YAG laser has successfully treated lesions in patients with familial glomangiomatosis.^21,22

Our patient opted for sclerotherapy with STS on symptomatic lesions of the bilateral upper extremities and trunk. The patient reported moderate improvement of some lesions at a 4-week follow-up appointment and sclerotherapy with STS was repeated.

It is important to note that if a glomangioma is fully excised, the prognosis is favorable; however, recurrence after surgical excision is seen in 10% to 33% of cases.^23,24 Our patient had symptomatic lesions excised on the face, but they recurred. Glomangiomas confer a low risk of malignancy but some risk factors include lesions > 2 cm in size, deep lesions, muscle and/or bone invasion, and high mitotic activity.^17,25 If left untreated, high-risk glomangiomas can potentially be life-threatening due to growth, bleeding, or vital organ obstruction.²⁶

Primary Care Role

This patient was referred by his PCP assuming that these were symptomatic vascular lesions or telangiectasias (spider veins). Glomus cell tumors are classified as neurovascular neoplasms which may appear similar to vascular malformations or hemangiomas. ²⁷ PCPs serve an important role in performing cutaneous biopsies to increase patient access to dermatologic care, increase patient awareness of skin conditions including skin cancer, and to potentially diagnose a malignant lesion.²⁸ However, the PCP ultimately referred the patient to dermatology due to the number of growing, painful lesions. If the patient had a single lesion, it may have been appropriate to biopsy for diagnostic clarity.

A retrospective review found that the clinical diagnosis of glomus tumor showed concordance with histopathological diagnosis in 45.4% of cases. The most common alternate histopathological diagnoses were vascular tumors (25.9%) followed by other skin or soft tissue tumors like neuromas, leiomyomas, lipomas, or nevi.²⁹ Even if the PCP performed an initial biopsy with high clinical suspicion of a vascular malformation, some glomus cell tumors may be vascular tumors and vice versa.

Though the patient’s history was consistent with the classic triad of glomangiomas including hypersensitivity, intermittent pain, and pinpoint pain, histopathology was necessary to confirm the diagnosis. Given that these appeared to be similar to telangiectasias to the PCP, a rare condition like BRBNS was likely not considered upon initial presentation. Furthermore, the patient had a negative review of systems to include GI symptoms like melena or hematochezia. The PCP had no concern of GI hemorrhage as these lesions can involve the GI tract. If the patient were to endorse additional symptoms, a CBC to evaluate for anemia as well as a GI referral would be warranted.

CONCLUSIONS

This case exhibits the importance of differentiating glomus cell tumors from other more common vascular anomalies via a patient’s history and histopathological findings. Diagnosis and treatment may be difficult depending on the extent of lesions.

DISCUSSION

A diagnosis of familial glomangiomatosis was made based on the clinical history and histopathologic findings from the punch biopsy. Glomus tumors are comprised of glomus cells, or undifferentiated smooth muscle cells responsible for thermoregulation.¹ Glomus tumors are classified into 3 categories: solid (predominantly glomus cells), glomangiomas (predominantly blood vessels), and glomangiomyomas (predominantly smooth muscle cells).² Glomangiomas, which comprise up to 20% of glomus tumors, typically present as bluish-purple, papular or nodular, hyperkeratotic lesions that are 2 to 10 mm in diameter.¹ These lesions are tender to palpation and pain may worsen with exposure to cold. Glomangiomas are associated with a classic triad of symptoms that include hypersensitivity, intermittent pain, and pinpoint pain, but patients rarely present with all 3.³

Glomangiomas tend to occur in areas rich with glomus bodies—the distal extremities—specifically the palms, wrists, forearms, feet, and subungual regions; visceral organ involvement including the GI tract is very rare.^1,4,5 About 38% to 68% of these lesions are hereditary or can be sporadic. If these lesions are hereditary, a patient is said to have familial glomangiomatosis. In familial glomangiomatosis, the glomulin gene is mutated in an autosomal dominant inheritance pattern with incomplete penetrance and variable expressivity. Inherited glomangiomas may present at birth or puberty similar to other vascular anomalies.⁴

Histopathology of glomangiomas shows rows of glomus cells (modified smooth muscle cells) surrounding distorted venous channels.^6,7 These lesions stain positive for CD34, vimentin, calponin, and α-smooth muscle actin, but are negative for desmin, S-100, and von Willebrand factor.^1,8 Although the patient’s medical history and physical examination are important in establishing the diagnosis, histopathology is confirmatory.

While the punch biopsy results were pending, a complete blood count (CBC) and fecal occult blood test (FOBT) were ordered due to concerns for blue rubber bleb nevus syndrome (BRBNS), a rare disorder with about 200 reported cases. Patients present with multiple blue to violaceous compressible nodules that feel rubbery in consistency and may be painful with compression. Lesions may be up to 5 cm in diameter and with time, the GI tract may also become involved.⁹ In the GI tract, the small bowel is the most common site of involvement and patients may present with severe iron deficiency anemia due to hemorrhage.¹⁰ Histopathologic features are nonspecific and have features of venous malformations but may include large, tortuous, dilated vessels with a single endothelial lining with possible smooth muscle in vessel walls or calcifications.¹¹ Due to concerns of BRBNS, laboratory studies (CBC and FOBT) were obtained but did not indicate the patient was experiencing a GI hemorrhage.

The differential diagnosis included Maffucci syndrome, also known as dyschondrodysplasia with hemangiomas, enchondromatosis with cavernous hemangiomas, or hemangiomatosis chondrodystrophic. Patients with Maffucci syndrome present with multiple enchondromas, soft tissue hemangiomas or lymphangiomas, and gliomas. These lesions tend to undergo malignant transformation from enchondromas to chondrosarcomas and hemangiomas to vascular sarcomas.¹² This diagnosis was less likely in the patient in this case as there were no concerns of skeletal involvement upon history and physical examination.

Lastly, Klippel-Trénaunay syndrome can be associated with similar cutaneous vascular manifestations.^13,14 This syndrome occurs due to somatic mutations altering angiogenesis during embryological development. This results in varicosities of superficial and deep venous systems, persistent embryonic veins, and valvular incompetence. However, these patients typically have capillary manifestations such as a flat, red, or purple port-wine stain present at birth and associated limb hypertrophy predominantly affecting a single lower limb.^15,16 The patient reported not having the lesions present at birth and because bilateral upper/lower extremity and trunk involvement is rare in this syndrome, a Klippel-Trénaunay syndrome diagnosis was unlikely even in the absence of biopsy results.

Treatment

Based on pathology results, the patient was diagnosed with familial glomangiomatosis and a discussion of treatment options ensued. Asymptomatic lesions can be periodically managed. In addition, there are several treatments for symptomatic lesions. Symptomatic lesions may be tender to palpation and or hypersensitive to temperature change (cold). Though they exhibit slow growth, they can invade surrounding tissues including nerve sheaths which can worsen pain.

Surgical resection, sclerotherapy, laser therapy, and electron beam radiation have been used on patients with symptomatic lesions.^8,17 Sclerotherapy involves introducing sterile solutions into a blood vessel’s lumen or into the vascular lesion itself to induce permanent endofibrosis and ablation.¹⁷ Hypertonic saline, sodium tetradecyl sulfate (STS), and absolute alcohol have been used to treat vascular anomalies as well as glomangiomas.¹⁷ Though case reports have noticed significant improvement in symptomatic lesions, sclerotherapy has been shown to be more effective in treating venous malformations than glomangiomas.^18,19

A long-pulsed 1064-nm neodymium-doped yttrium aluminum garnet (Nd:YAG) laser has also been effective in treating larger glomangiomas that would otherwise be difficult to excise.²⁰ The Nd:YAG laser has successfully treated lesions in patients with familial glomangiomatosis.^21,22

Our patient opted for sclerotherapy with STS on symptomatic lesions of the bilateral upper extremities and trunk. The patient reported moderate improvement of some lesions at a 4-week follow-up appointment and sclerotherapy with STS was repeated.

It is important to note that if a glomangioma is fully excised, the prognosis is favorable; however, recurrence after surgical excision is seen in 10% to 33% of cases.^23,24 Our patient had symptomatic lesions excised on the face, but they recurred. Glomangiomas confer a low risk of malignancy but some risk factors include lesions > 2 cm in size, deep lesions, muscle and/or bone invasion, and high mitotic activity.^17,25 If left untreated, high-risk glomangiomas can potentially be life-threatening due to growth, bleeding, or vital organ obstruction.²⁶

Primary Care Role

This patient was referred by his PCP assuming that these were symptomatic vascular lesions or telangiectasias (spider veins). Glomus cell tumors are classified as neurovascular neoplasms which may appear similar to vascular malformations or hemangiomas. ²⁷ PCPs serve an important role in performing cutaneous biopsies to increase patient access to dermatologic care, increase patient awareness of skin conditions including skin cancer, and to potentially diagnose a malignant lesion.²⁸ However, the PCP ultimately referred the patient to dermatology due to the number of growing, painful lesions. If the patient had a single lesion, it may have been appropriate to biopsy for diagnostic clarity.

A retrospective review found that the clinical diagnosis of glomus tumor showed concordance with histopathological diagnosis in 45.4% of cases. The most common alternate histopathological diagnoses were vascular tumors (25.9%) followed by other skin or soft tissue tumors like neuromas, leiomyomas, lipomas, or nevi.²⁹ Even if the PCP performed an initial biopsy with high clinical suspicion of a vascular malformation, some glomus cell tumors may be vascular tumors and vice versa.

Though the patient’s history was consistent with the classic triad of glomangiomas including hypersensitivity, intermittent pain, and pinpoint pain, histopathology was necessary to confirm the diagnosis. Given that these appeared to be similar to telangiectasias to the PCP, a rare condition like BRBNS was likely not considered upon initial presentation. Furthermore, the patient had a negative review of systems to include GI symptoms like melena or hematochezia. The PCP had no concern of GI hemorrhage as these lesions can involve the GI tract. If the patient were to endorse additional symptoms, a CBC to evaluate for anemia as well as a GI referral would be warranted.

CONCLUSIONS

This case exhibits the importance of differentiating glomus cell tumors from other more common vascular anomalies via a patient’s history and histopathological findings. Diagnosis and treatment may be difficult depending on the extent of lesions.

DISCUSSION

A diagnosis of familial glomangiomatosis was made based on the clinical history and histopathologic findings from the punch biopsy. Glomus tumors are comprised of glomus cells, or undifferentiated smooth muscle cells responsible for thermoregulation.¹ Glomus tumors are classified into 3 categories: solid (predominantly glomus cells), glomangiomas (predominantly blood vessels), and glomangiomyomas (predominantly smooth muscle cells).² Glomangiomas, which comprise up to 20% of glomus tumors, typically present as bluish-purple, papular or nodular, hyperkeratotic lesions that are 2 to 10 mm in diameter.¹ These lesions are tender to palpation and pain may worsen with exposure to cold. Glomangiomas are associated with a classic triad of symptoms that include hypersensitivity, intermittent pain, and pinpoint pain, but patients rarely present with all 3.³

Glomangiomas tend to occur in areas rich with glomus bodies—the distal extremities—specifically the palms, wrists, forearms, feet, and subungual regions; visceral organ involvement including the GI tract is very rare.^1,4,5 About 38% to 68% of these lesions are hereditary or can be sporadic. If these lesions are hereditary, a patient is said to have familial glomangiomatosis. In familial glomangiomatosis, the glomulin gene is mutated in an autosomal dominant inheritance pattern with incomplete penetrance and variable expressivity. Inherited glomangiomas may present at birth or puberty similar to other vascular anomalies.⁴

Histopathology of glomangiomas shows rows of glomus cells (modified smooth muscle cells) surrounding distorted venous channels.^6,7 These lesions stain positive for CD34, vimentin, calponin, and α-smooth muscle actin, but are negative for desmin, S-100, and von Willebrand factor.^1,8 Although the patient’s medical history and physical examination are important in establishing the diagnosis, histopathology is confirmatory.

While the punch biopsy results were pending, a complete blood count (CBC) and fecal occult blood test (FOBT) were ordered due to concerns for blue rubber bleb nevus syndrome (BRBNS), a rare disorder with about 200 reported cases. Patients present with multiple blue to violaceous compressible nodules that feel rubbery in consistency and may be painful with compression. Lesions may be up to 5 cm in diameter and with time, the GI tract may also become involved.⁹ In the GI tract, the small bowel is the most common site of involvement and patients may present with severe iron deficiency anemia due to hemorrhage.¹⁰ Histopathologic features are nonspecific and have features of venous malformations but may include large, tortuous, dilated vessels with a single endothelial lining with possible smooth muscle in vessel walls or calcifications.¹¹ Due to concerns of BRBNS, laboratory studies (CBC and FOBT) were obtained but did not indicate the patient was experiencing a GI hemorrhage.

The differential diagnosis included Maffucci syndrome, also known as dyschondrodysplasia with hemangiomas, enchondromatosis with cavernous hemangiomas, or hemangiomatosis chondrodystrophic. Patients with Maffucci syndrome present with multiple enchondromas, soft tissue hemangiomas or lymphangiomas, and gliomas. These lesions tend to undergo malignant transformation from enchondromas to chondrosarcomas and hemangiomas to vascular sarcomas.¹² This diagnosis was less likely in the patient in this case as there were no concerns of skeletal involvement upon history and physical examination.

Lastly, Klippel-Trénaunay syndrome can be associated with similar cutaneous vascular manifestations.^13,14 This syndrome occurs due to somatic mutations altering angiogenesis during embryological development. This results in varicosities of superficial and deep venous systems, persistent embryonic veins, and valvular incompetence. However, these patients typically have capillary manifestations such as a flat, red, or purple port-wine stain present at birth and associated limb hypertrophy predominantly affecting a single lower limb.^15,16 The patient reported not having the lesions present at birth and because bilateral upper/lower extremity and trunk involvement is rare in this syndrome, a Klippel-Trénaunay syndrome diagnosis was unlikely even in the absence of biopsy results.

Treatment

Based on pathology results, the patient was diagnosed with familial glomangiomatosis and a discussion of treatment options ensued. Asymptomatic lesions can be periodically managed. In addition, there are several treatments for symptomatic lesions. Symptomatic lesions may be tender to palpation and or hypersensitive to temperature change (cold). Though they exhibit slow growth, they can invade surrounding tissues including nerve sheaths which can worsen pain.

Surgical resection, sclerotherapy, laser therapy, and electron beam radiation have been used on patients with symptomatic lesions.^8,17 Sclerotherapy involves introducing sterile solutions into a blood vessel’s lumen or into the vascular lesion itself to induce permanent endofibrosis and ablation.¹⁷ Hypertonic saline, sodium tetradecyl sulfate (STS), and absolute alcohol have been used to treat vascular anomalies as well as glomangiomas.¹⁷ Though case reports have noticed significant improvement in symptomatic lesions, sclerotherapy has been shown to be more effective in treating venous malformations than glomangiomas.^18,19

A long-pulsed 1064-nm neodymium-doped yttrium aluminum garnet (Nd:YAG) laser has also been effective in treating larger glomangiomas that would otherwise be difficult to excise.²⁰ The Nd:YAG laser has successfully treated lesions in patients with familial glomangiomatosis.^21,22

Our patient opted for sclerotherapy with STS on symptomatic lesions of the bilateral upper extremities and trunk. The patient reported moderate improvement of some lesions at a 4-week follow-up appointment and sclerotherapy with STS was repeated.

It is important to note that if a glomangioma is fully excised, the prognosis is favorable; however, recurrence after surgical excision is seen in 10% to 33% of cases.^23,24 Our patient had symptomatic lesions excised on the face, but they recurred. Glomangiomas confer a low risk of malignancy but some risk factors include lesions > 2 cm in size, deep lesions, muscle and/or bone invasion, and high mitotic activity.^17,25 If left untreated, high-risk glomangiomas can potentially be life-threatening due to growth, bleeding, or vital organ obstruction.²⁶

Primary Care Role

This patient was referred by his PCP assuming that these were symptomatic vascular lesions or telangiectasias (spider veins). Glomus cell tumors are classified as neurovascular neoplasms which may appear similar to vascular malformations or hemangiomas. ²⁷ PCPs serve an important role in performing cutaneous biopsies to increase patient access to dermatologic care, increase patient awareness of skin conditions including skin cancer, and to potentially diagnose a malignant lesion.²⁸ However, the PCP ultimately referred the patient to dermatology due to the number of growing, painful lesions. If the patient had a single lesion, it may have been appropriate to biopsy for diagnostic clarity.

A retrospective review found that the clinical diagnosis of glomus tumor showed concordance with histopathological diagnosis in 45.4% of cases. The most common alternate histopathological diagnoses were vascular tumors (25.9%) followed by other skin or soft tissue tumors like neuromas, leiomyomas, lipomas, or nevi.²⁹ Even if the PCP performed an initial biopsy with high clinical suspicion of a vascular malformation, some glomus cell tumors may be vascular tumors and vice versa.

Though the patient’s history was consistent with the classic triad of glomangiomas including hypersensitivity, intermittent pain, and pinpoint pain, histopathology was necessary to confirm the diagnosis. Given that these appeared to be similar to telangiectasias to the PCP, a rare condition like BRBNS was likely not considered upon initial presentation. Furthermore, the patient had a negative review of systems to include GI symptoms like melena or hematochezia. The PCP had no concern of GI hemorrhage as these lesions can involve the GI tract. If the patient were to endorse additional symptoms, a CBC to evaluate for anemia as well as a GI referral would be warranted.

CONCLUSIONS

This case exhibits the importance of differentiating glomus cell tumors from other more common vascular anomalies via a patient’s history and histopathological findings. Diagnosis and treatment may be difficult depending on the extent of lesions.

Comprehensive genomic profiling (CGP) is becoming progressively common and appropriate as the array of molecular targets expands. However, most hospital laboratories in the United States do not perform CGP assays in-house; instead, these tests are sent to reference laboratories. As evidenced by Inal et al, only a minority of guideline-indicated molecular testing is performed.¹

The workload associated with referral testing is a barrier to increased use of such tests; streamlined processes in pathology might increase molecular test use. At 6 high-complexity US Department of Veterans Affairs (VA) medical centers (VAMCs) (Manhattan, Los Angeles, San Diego, Denver, Kansas City, and Salisbury, Maryland) ranging from 150 to 750 beds, a consult process for anatomic pathology molecular testing has increased test utilization, appropriateness of orders, standardization of reporting, and efficiency of care. This report comprehensively describes and maps the anatomic pathology molecular testing consult process at a VAMC. We present areas of inefficiency and a target state process map that incorporates best practices.

MOLECULAR TESTING CONSULT PROCESS

At the Kansas City VAMC (KCVAMC), a consult process for anatomic pathology molecular testing was introduced in 2021. Prior to this, requesting anatomic pathology molecular testing was not standardized. A variety of opportunities and methods were used for requests (eg, phone, page, Teams message, email, Computerized Patient Record System alert; or in-person during tumor board, an office meeting, or in passing). Requests were not documented in a standardized way, resulting in duplicate requests. Testing status and updates were documented outside the medical record, so requests for status updates (via various opportunities and methods) were common and redundant. Data from the year preceding consult implementation and the year following consult implementation have demonstrated increased test utilization, appropriateness of orders, standardization of reporting, and efficiency of care.

Consult Request

The precision oncology testing process starts with a health care practitioner (HCP) request on behalf of any physician or advanced practice registered nurse. It can be placed by any health care employee and directed to a designated employee in the pathology department. The request is ultimately reviewed by a pathologist (Figure 1). At KCVAMC, this request comes in the form of a consult in the electronic health record (EHR) from the ordering HCP to a pathologist. The KCVAMC pathology consult form was previously published with a discussion of the rationale for this process as opposed to a laboratory order process.² This consult form ensures ordering HCPs supply all necessary information for the pathologist to approve the request and order the test without needing to, in most cases, contact the ordering HCP for clarification or additional information. The form asks the ordering HCP to specify which test is being requested and why. Within the Veterans Health Administration (VHA) there are local and national contracts with many laboratories with hundreds of precision oncology tests to choose from. Consulting with a pathologist is necessary to determine which test is most appropriate.

The precision oncology consult form cannot be submitted without completing all required fields. It also contains indications for the test the ordering HCP selects to minimize unintentionally inappropriate orders. The form asks which tissue the requestor expects the test to be performed on. The requestor must provide contact information for the originating institution when the tissue was collected outside the VHA. The consult form also asks whether another anatomic site is accessible and could be biopsied without unacceptable risk or impracticality, should all previously collected tissue be insufficient. For CGP requests, this allows the pathologist to determine the appropriateness of liquid biopsy without having to reach out to the ordering HCP or wait for the question to be addressed at a tumor board. When a companion diagnostic is available for a test, the ordering HCP is asked which drug will be used so that the most appropriate assay is chosen.

Consult Review

Pathology service involvement begins with pathologist review of the consult form to ensure that the correct test is indicated. Depending on the resources and preferences at a site, consults can be directed to and reviewed by the pathologist associated with the corresponding pathology specimen or to a single pathologist or group of pathologists charged with attending to consults.

The patient’s EHR is reviewed to verify that the test has not already been performed and to determine which tissue to review. Previous surgical pathology reports are examined to assess whether sufficient tissue is available for testing, which may be determined without the need for direct slide examination. Pathologists often use wording such as “rare cells” or in some cases specify that there are not enough lesional cells for ancillary testing. In biopsy reports, the percentage of tissue occupied by lesional cells or the greatest linear length of tumor cells is often documented. As for quality, pathologists may note that a specimen is largely necrotic, and gross descriptions will indicate if a specimen was compromised for molecular analysis by exposure to fixatives such as Bouin’s solution, B-5, or decalcifying agents that contain strong acids.

Tissue Retrieval

If, after such evaluation, the test is indicated and there is tissue that could be sufficient for testing, retrieval of the tissue is pursued. For in-house cases, the pathologist reviews the corresponding surgical pathology report to determine which blocks and slides to pull from the archives. In the cancer checklist, some pathologists specify the best block for subsequent ancillary studies. From the final diagnosis and gross description, the pathologist can determine which blocks are most likely to contain lesional tissue. These slides are retrieved from the archives.

For cases collected at an outside institution (other VHA facility or non-VHA facility/institution), the outside institution must be contacted to retrieve the needed slides and blocks. The phone numbers, fax numbers, email addresses, and mailing addresses for outside institutions are housed in an electronic file and are specific to the point of contact for such requests. Maintaining a record of contacts increases efficiency of the overall process; gathering contact information and successfully requesting tissue often involves multiple automated answering systems, misdirected calls, and failed attempts.

Tissue Review

After retrieving in-house tissue, the pathologist can proceed directly to slide review. For outside cases, the case must first be accessioned so that after review of the slides the pathologist can issue a report to confirm the outside diagnosis. In reviewing the slides, the pathologist looks to see that the diagnosis is correct, that there is a sufficient number of lesional cells in a section, that the lesional cells are of a sufficient concentration in a section, or subsection of the section that could be dissected, and that the cells are viable. Depending on the requested assay and the familiarity of the pathologist with that assay, the pathologist may need to look up the technical requirements of the assay and capabilities of the testing company. Assays vary in sensitivity and require differing amounts and concentrations of tumor. Some companies will dissect tissue, others will not.

If there is sufficient tissue in the material reviewed, the corresponding blocks are retrieved from in-house archives or requests are placed for outside blocks or unstained slides. If there was not enough tissue for testing, the same process is repeated to retrieve and evaluate any other specimens the patient may have. If there are no other specimens to review, this is simply communicated to the ordering HCP via the consult. If the patient is a candidate for liquid biopsy—ie, current specimens are of insufficient quality and/or quantity and a new tissue sample cannot be obtained due to unacceptable risk or impracticality—the order is placed at this time.

Tissue Transport and Testing

Unstained slides need to be cut unless blocks are sent. Slides, blocks, reports, and requisition forms are packaged for transport. An accession number is created for the precision oncology molecular laboratory test in the clinical laboratory section of the EHR system. The clinical laboratory accession number provides a way of tracking sendout testing status. The case is accessioned just prior to placement in the mail so that when an accession number appears in the EHR, the ordering HCP knows the case has been sent out. When results are received, the clinical laboratory accession is completed and a comment is added to indicate where in the EHR to find the report or, when applicable, notes that testing failed.

RESULT REPORTING

When a result becomes available, the report file is downloaded from the vendor portal. This full report is securely transmitted to the ordering HCP. The file is then scanned into the EHR. Additionally, salient findings from the report are abstracted by the pathologist for inclusion as a supplement to the anatomic pathology case. This step ensures that this information travels with the anatomic pathology report if the patient’s care is transferred elsewhere. Templates are used to ensure essential data is captured based on the type of test. The template reminds the pathologist to comment on things such as variants that may represent clonal hematopoiesis, variants that may be germline, and variants that qualify a patient for germline testing. Even with the template, the pathologist must spend significant time reviewing the chart for things such as personal cancer history, other medical history, other masses on imaging, family history, previous surgical pathology reports, and previous molecular testing.

If results are suboptimal, recommendations for repeat testing are made based on the consult response to the question of repeat biopsy feasibility and review of previous pathology reports. The final consult report is added as a consult note, the consult is completed, and the original vendor report file is associated with the consult note in the EHR.

Ancillary Testing Technician

Due to chronic KCVAMC understaffing in the clerical office, gross room, and histology, most of the consult tasks are performed by a pathologist. In an ideal scenario, the pathology staff would divide its time between a pathologist and another dedicated laboratory position, such as an ancillary testing technician (ATT). The ATT can assume responsibilities that do not require the expertise of a pathologist (Figure 2). In such a process, the only steps that would require a pathologist would be review of requests and slides and completion of the interpretive report. All other steps could be accomplished by someone who lacks certifications, laboratory experience, or postsecondary education.

The ATT can receive the requests and retrieve slides and blocks. After slides have been reviewed by a pathologist, the pathologist can inform the ATT which slides or blocks testing will be performed on, provide any additional necessary information for completing the order, and answer any questions. For send-out tests, this allows the ATT to independently complete online portal forms and all other physical requirements prior to delivery of the slides and blocks to specimen processors in the laboratory.

ATTs can keep the ordering HCPs informed of status and be identified as the point of contact for all status inquiries. ATTs can receive results and get outside reports scanned into the EHR. Finally, ATTs can use pathologistdesigned templates to transpose information from outside reports such that a provisional report is prepared and a pathologist does not spend time duplicating information from the outside report. The pathologist can then complete the report with information requiring medical judgment that enhances care.

Optimal Pathologist Involvement

Only 3 steps in the process (request review, tissue review, and completion of an interpretive report) require a pathologist, which are necessary for optimal care and to address barriers to precision oncology.3 While the laboratory may consume only 5% of a health system budget, optimal laboratory use could prevent as much as 30% of avoidable costs.4 These estimates are widely recognized and addressed by campaigns such as Choosing Wisely, as well as programming of alerts and hard stops in EHR systems to reduce duplicate or otherwise inappropriate orders. The tests associated with precision oncology, such as CGP assays, require more nuanced consideration that is best achieved through pathology consultation. In vetting requests for such tests, the pathologist needs information that ordering HCPs do not routinely provide when ordering other tests. A consult asking for such information allows an ordering HCP to efficiently convey this information without having to call the laboratory to circumvent a hard stop.

Regardless of whether a formal electronic consult is used, pathologists must be involved in the review of requests. Creation of an original in-house report also provides an opportunity for pathologists to offer their expertise and maximize the contribution of pathology to patient care. If outside (other VHA facility or non-VHA facility/institution) reports are simply scanned into the EHR without review and issuance of an interpretive report by an in-house pathologist, then an interpretation by a pathologist with access to the patient’s complete chart is never provided. Testing companies are not provided with every patient diagnosis, so in patients with multiple neoplastic conditions, a report may seem to indicate that a detected mutation is from 1 tumor when it is actually from another. Even when all known diagnoses are considered, a variant may be detected that the medical record could reveal to indicate a new diagnosis.

Variation in reporting between companies necessitates pathologist review to standardize care. Some companies indicate which variants may represent clonal hematopoiesis, while others will simply list the pathogenic variants. An oncologist who sees a high volume of hematolymphoid neoplasia may recognize which variants may represent clonal hematopoiesis, but others may not. Reports from the same company may vary, and their interpretation often requires a pathologist's expertise. For example, even if a sample meets the technical requirements for analysis, the report may indicate that the quality or quantity of DNA has reduced the sensitivity for genomic alteration detection. A pathologist would know how to use this information in deciding how to proceed. In a situation where quantity was the issue, the pathologist may know there is additional tissue that could be sent for testing. If quality is the issue, the pathologist may know that additional blocks from the same case likely have the same quality of DNA and would also be unsuitable for testing.

Pathologist input is necessary for precision oncology testing. Some tasks that would ideally be completed by a molecular pathologist (eg, creation of reports to indicate which variants may represent clonal hematopoiesis of indeterminate potential) may be sufficiently completed by a pathologist without fellowship training in molecular pathology.

There are about 15,000 full-time pathologists in the US.⁴ In the 20 years since molecular genetic pathology was formally recognized as a specialty, there have been < 500 pathologists who have pursued fellowship training in this specialty.5 With the inundation of molecular variants uncovered by routine next-generation sequencing (NGS), there are too few fellowship-trained molecular pathologists to provide all such aforementioned input; it is incumbent on surgical pathologists in general to take on such responsibilities.

Consult Implementation Data

These results support the feasibility and effectiveness of the consult process. Prior to consult implementation, many requests were not compliant with VHA National Precision Oncology Program (NPOP) testing guidelines. Since enactment of the consult, > 90% of requests have been in compliance. In the year preceding the consult (January 2020 to December 2021), 55 of 211 (26.1%) metastatic lung and prostate cancers samples eligible for NGS were tested and 126 (59.7%) NGS vendor reports were scanned into the EHR. The mean time from metastasis to NGS result was 151 days. In the year following enactment of the consult (January 2021 to December 2022), 168 of 224 (75.0%) of metastatic lung and prostate cancers eligible for NGS were tested and all 224 NGS vendor reports were scanned into the EHR. The mean time from metastasis to NGS result was 83 days. These data indicate that the practices recommended increase test use, appropriateness of orders, standardization of reporting, and efficiency of care.

CONCLUSIONS

Processing precision oncology testing requires substantial work for pathology departments. Laboratory workforce shortages and ever-expanding indications necessitate additional study of pathology processes to manage increasing workload and maintain the highest quality of cancer care through maximal efficiency and the development of appropriate staffing models. The use of a consult for anatomic pathology molecular testing is one process that can increase test use, appropriateness of orders, standardization of reporting, and efficiency of care. This report provides a comprehensive description and mapping of the process, highlights best practices, identifies inefficiencies, and provides a description and mapping of a target state.

Comprehensive genomic profiling (CGP) is becoming progressively common and appropriate as the array of molecular targets expands. However, most hospital laboratories in the United States do not perform CGP assays in-house; instead, these tests are sent to reference laboratories. As evidenced by Inal et al, only a minority of guideline-indicated molecular testing is performed.¹

The workload associated with referral testing is a barrier to increased use of such tests; streamlined processes in pathology might increase molecular test use. At 6 high-complexity US Department of Veterans Affairs (VA) medical centers (VAMCs) (Manhattan, Los Angeles, San Diego, Denver, Kansas City, and Salisbury, Maryland) ranging from 150 to 750 beds, a consult process for anatomic pathology molecular testing has increased test utilization, appropriateness of orders, standardization of reporting, and efficiency of care. This report comprehensively describes and maps the anatomic pathology molecular testing consult process at a VAMC. We present areas of inefficiency and a target state process map that incorporates best practices.

MOLECULAR TESTING CONSULT PROCESS

At the Kansas City VAMC (KCVAMC), a consult process for anatomic pathology molecular testing was introduced in 2021. Prior to this, requesting anatomic pathology molecular testing was not standardized. A variety of opportunities and methods were used for requests (eg, phone, page, Teams message, email, Computerized Patient Record System alert; or in-person during tumor board, an office meeting, or in passing). Requests were not documented in a standardized way, resulting in duplicate requests. Testing status and updates were documented outside the medical record, so requests for status updates (via various opportunities and methods) were common and redundant. Data from the year preceding consult implementation and the year following consult implementation have demonstrated increased test utilization, appropriateness of orders, standardization of reporting, and efficiency of care.

Consult Request

The precision oncology testing process starts with a health care practitioner (HCP) request on behalf of any physician or advanced practice registered nurse. It can be placed by any health care employee and directed to a designated employee in the pathology department. The request is ultimately reviewed by a pathologist (Figure 1). At KCVAMC, this request comes in the form of a consult in the electronic health record (EHR) from the ordering HCP to a pathologist. The KCVAMC pathology consult form was previously published with a discussion of the rationale for this process as opposed to a laboratory order process.² This consult form ensures ordering HCPs supply all necessary information for the pathologist to approve the request and order the test without needing to, in most cases, contact the ordering HCP for clarification or additional information. The form asks the ordering HCP to specify which test is being requested and why. Within the Veterans Health Administration (VHA) there are local and national contracts with many laboratories with hundreds of precision oncology tests to choose from. Consulting with a pathologist is necessary to determine which test is most appropriate.

The precision oncology consult form cannot be submitted without completing all required fields. It also contains indications for the test the ordering HCP selects to minimize unintentionally inappropriate orders. The form asks which tissue the requestor expects the test to be performed on. The requestor must provide contact information for the originating institution when the tissue was collected outside the VHA. The consult form also asks whether another anatomic site is accessible and could be biopsied without unacceptable risk or impracticality, should all previously collected tissue be insufficient. For CGP requests, this allows the pathologist to determine the appropriateness of liquid biopsy without having to reach out to the ordering HCP or wait for the question to be addressed at a tumor board. When a companion diagnostic is available for a test, the ordering HCP is asked which drug will be used so that the most appropriate assay is chosen.

Consult Review

Pathology service involvement begins with pathologist review of the consult form to ensure that the correct test is indicated. Depending on the resources and preferences at a site, consults can be directed to and reviewed by the pathologist associated with the corresponding pathology specimen or to a single pathologist or group of pathologists charged with attending to consults.

The patient’s EHR is reviewed to verify that the test has not already been performed and to determine which tissue to review. Previous surgical pathology reports are examined to assess whether sufficient tissue is available for testing, which may be determined without the need for direct slide examination. Pathologists often use wording such as “rare cells” or in some cases specify that there are not enough lesional cells for ancillary testing. In biopsy reports, the percentage of tissue occupied by lesional cells or the greatest linear length of tumor cells is often documented. As for quality, pathologists may note that a specimen is largely necrotic, and gross descriptions will indicate if a specimen was compromised for molecular analysis by exposure to fixatives such as Bouin’s solution, B-5, or decalcifying agents that contain strong acids.

Tissue Retrieval

If, after such evaluation, the test is indicated and there is tissue that could be sufficient for testing, retrieval of the tissue is pursued. For in-house cases, the pathologist reviews the corresponding surgical pathology report to determine which blocks and slides to pull from the archives. In the cancer checklist, some pathologists specify the best block for subsequent ancillary studies. From the final diagnosis and gross description, the pathologist can determine which blocks are most likely to contain lesional tissue. These slides are retrieved from the archives.

For cases collected at an outside institution (other VHA facility or non-VHA facility/institution), the outside institution must be contacted to retrieve the needed slides and blocks. The phone numbers, fax numbers, email addresses, and mailing addresses for outside institutions are housed in an electronic file and are specific to the point of contact for such requests. Maintaining a record of contacts increases efficiency of the overall process; gathering contact information and successfully requesting tissue often involves multiple automated answering systems, misdirected calls, and failed attempts.

Tissue Review

After retrieving in-house tissue, the pathologist can proceed directly to slide review. For outside cases, the case must first be accessioned so that after review of the slides the pathologist can issue a report to confirm the outside diagnosis. In reviewing the slides, the pathologist looks to see that the diagnosis is correct, that there is a sufficient number of lesional cells in a section, that the lesional cells are of a sufficient concentration in a section, or subsection of the section that could be dissected, and that the cells are viable. Depending on the requested assay and the familiarity of the pathologist with that assay, the pathologist may need to look up the technical requirements of the assay and capabilities of the testing company. Assays vary in sensitivity and require differing amounts and concentrations of tumor. Some companies will dissect tissue, others will not.

If there is sufficient tissue in the material reviewed, the corresponding blocks are retrieved from in-house archives or requests are placed for outside blocks or unstained slides. If there was not enough tissue for testing, the same process is repeated to retrieve and evaluate any other specimens the patient may have. If there are no other specimens to review, this is simply communicated to the ordering HCP via the consult. If the patient is a candidate for liquid biopsy—ie, current specimens are of insufficient quality and/or quantity and a new tissue sample cannot be obtained due to unacceptable risk or impracticality—the order is placed at this time.

Tissue Transport and Testing

Unstained slides need to be cut unless blocks are sent. Slides, blocks, reports, and requisition forms are packaged for transport. An accession number is created for the precision oncology molecular laboratory test in the clinical laboratory section of the EHR system. The clinical laboratory accession number provides a way of tracking sendout testing status. The case is accessioned just prior to placement in the mail so that when an accession number appears in the EHR, the ordering HCP knows the case has been sent out. When results are received, the clinical laboratory accession is completed and a comment is added to indicate where in the EHR to find the report or, when applicable, notes that testing failed.

RESULT REPORTING

When a result becomes available, the report file is downloaded from the vendor portal. This full report is securely transmitted to the ordering HCP. The file is then scanned into the EHR. Additionally, salient findings from the report are abstracted by the pathologist for inclusion as a supplement to the anatomic pathology case. This step ensures that this information travels with the anatomic pathology report if the patient’s care is transferred elsewhere. Templates are used to ensure essential data is captured based on the type of test. The template reminds the pathologist to comment on things such as variants that may represent clonal hematopoiesis, variants that may be germline, and variants that qualify a patient for germline testing. Even with the template, the pathologist must spend significant time reviewing the chart for things such as personal cancer history, other medical history, other masses on imaging, family history, previous surgical pathology reports, and previous molecular testing.

If results are suboptimal, recommendations for repeat testing are made based on the consult response to the question of repeat biopsy feasibility and review of previous pathology reports. The final consult report is added as a consult note, the consult is completed, and the original vendor report file is associated with the consult note in the EHR.

Ancillary Testing Technician

Due to chronic KCVAMC understaffing in the clerical office, gross room, and histology, most of the consult tasks are performed by a pathologist. In an ideal scenario, the pathology staff would divide its time between a pathologist and another dedicated laboratory position, such as an ancillary testing technician (ATT). The ATT can assume responsibilities that do not require the expertise of a pathologist (Figure 2). In such a process, the only steps that would require a pathologist would be review of requests and slides and completion of the interpretive report. All other steps could be accomplished by someone who lacks certifications, laboratory experience, or postsecondary education.

The ATT can receive the requests and retrieve slides and blocks. After slides have been reviewed by a pathologist, the pathologist can inform the ATT which slides or blocks testing will be performed on, provide any additional necessary information for completing the order, and answer any questions. For send-out tests, this allows the ATT to independently complete online portal forms and all other physical requirements prior to delivery of the slides and blocks to specimen processors in the laboratory.

ATTs can keep the ordering HCPs informed of status and be identified as the point of contact for all status inquiries. ATTs can receive results and get outside reports scanned into the EHR. Finally, ATTs can use pathologistdesigned templates to transpose information from outside reports such that a provisional report is prepared and a pathologist does not spend time duplicating information from the outside report. The pathologist can then complete the report with information requiring medical judgment that enhances care.

Optimal Pathologist Involvement

Only 3 steps in the process (request review, tissue review, and completion of an interpretive report) require a pathologist, which are necessary for optimal care and to address barriers to precision oncology.3 While the laboratory may consume only 5% of a health system budget, optimal laboratory use could prevent as much as 30% of avoidable costs.4 These estimates are widely recognized and addressed by campaigns such as Choosing Wisely, as well as programming of alerts and hard stops in EHR systems to reduce duplicate or otherwise inappropriate orders. The tests associated with precision oncology, such as CGP assays, require more nuanced consideration that is best achieved through pathology consultation. In vetting requests for such tests, the pathologist needs information that ordering HCPs do not routinely provide when ordering other tests. A consult asking for such information allows an ordering HCP to efficiently convey this information without having to call the laboratory to circumvent a hard stop.

Regardless of whether a formal electronic consult is used, pathologists must be involved in the review of requests. Creation of an original in-house report also provides an opportunity for pathologists to offer their expertise and maximize the contribution of pathology to patient care. If outside (other VHA facility or non-VHA facility/institution) reports are simply scanned into the EHR without review and issuance of an interpretive report by an in-house pathologist, then an interpretation by a pathologist with access to the patient’s complete chart is never provided. Testing companies are not provided with every patient diagnosis, so in patients with multiple neoplastic conditions, a report may seem to indicate that a detected mutation is from 1 tumor when it is actually from another. Even when all known diagnoses are considered, a variant may be detected that the medical record could reveal to indicate a new diagnosis.

Variation in reporting between companies necessitates pathologist review to standardize care. Some companies indicate which variants may represent clonal hematopoiesis, while others will simply list the pathogenic variants. An oncologist who sees a high volume of hematolymphoid neoplasia may recognize which variants may represent clonal hematopoiesis, but others may not. Reports from the same company may vary, and their interpretation often requires a pathologist's expertise. For example, even if a sample meets the technical requirements for analysis, the report may indicate that the quality or quantity of DNA has reduced the sensitivity for genomic alteration detection. A pathologist would know how to use this information in deciding how to proceed. In a situation where quantity was the issue, the pathologist may know there is additional tissue that could be sent for testing. If quality is the issue, the pathologist may know that additional blocks from the same case likely have the same quality of DNA and would also be unsuitable for testing.

Pathologist input is necessary for precision oncology testing. Some tasks that would ideally be completed by a molecular pathologist (eg, creation of reports to indicate which variants may represent clonal hematopoiesis of indeterminate potential) may be sufficiently completed by a pathologist without fellowship training in molecular pathology.

There are about 15,000 full-time pathologists in the US.⁴ In the 20 years since molecular genetic pathology was formally recognized as a specialty, there have been < 500 pathologists who have pursued fellowship training in this specialty.5 With the inundation of molecular variants uncovered by routine next-generation sequencing (NGS), there are too few fellowship-trained molecular pathologists to provide all such aforementioned input; it is incumbent on surgical pathologists in general to take on such responsibilities.

Consult Implementation Data

These results support the feasibility and effectiveness of the consult process. Prior to consult implementation, many requests were not compliant with VHA National Precision Oncology Program (NPOP) testing guidelines. Since enactment of the consult, > 90% of requests have been in compliance. In the year preceding the consult (January 2020 to December 2021), 55 of 211 (26.1%) metastatic lung and prostate cancers samples eligible for NGS were tested and 126 (59.7%) NGS vendor reports were scanned into the EHR. The mean time from metastasis to NGS result was 151 days. In the year following enactment of the consult (January 2021 to December 2022), 168 of 224 (75.0%) of metastatic lung and prostate cancers eligible for NGS were tested and all 224 NGS vendor reports were scanned into the EHR. The mean time from metastasis to NGS result was 83 days. These data indicate that the practices recommended increase test use, appropriateness of orders, standardization of reporting, and efficiency of care.

CONCLUSIONS

Processing precision oncology testing requires substantial work for pathology departments. Laboratory workforce shortages and ever-expanding indications necessitate additional study of pathology processes to manage increasing workload and maintain the highest quality of cancer care through maximal efficiency and the development of appropriate staffing models. The use of a consult for anatomic pathology molecular testing is one process that can increase test use, appropriateness of orders, standardization of reporting, and efficiency of care. This report provides a comprehensive description and mapping of the process, highlights best practices, identifies inefficiencies, and provides a description and mapping of a target state.

Comprehensive genomic profiling (CGP) is becoming progressively common and appropriate as the array of molecular targets expands. However, most hospital laboratories in the United States do not perform CGP assays in-house; instead, these tests are sent to reference laboratories. As evidenced by Inal et al, only a minority of guideline-indicated molecular testing is performed.¹

The workload associated with referral testing is a barrier to increased use of such tests; streamlined processes in pathology might increase molecular test use. At 6 high-complexity US Department of Veterans Affairs (VA) medical centers (VAMCs) (Manhattan, Los Angeles, San Diego, Denver, Kansas City, and Salisbury, Maryland) ranging from 150 to 750 beds, a consult process for anatomic pathology molecular testing has increased test utilization, appropriateness of orders, standardization of reporting, and efficiency of care. This report comprehensively describes and maps the anatomic pathology molecular testing consult process at a VAMC. We present areas of inefficiency and a target state process map that incorporates best practices.

MOLECULAR TESTING CONSULT PROCESS

At the Kansas City VAMC (KCVAMC), a consult process for anatomic pathology molecular testing was introduced in 2021. Prior to this, requesting anatomic pathology molecular testing was not standardized. A variety of opportunities and methods were used for requests (eg, phone, page, Teams message, email, Computerized Patient Record System alert; or in-person during tumor board, an office meeting, or in passing). Requests were not documented in a standardized way, resulting in duplicate requests. Testing status and updates were documented outside the medical record, so requests for status updates (via various opportunities and methods) were common and redundant. Data from the year preceding consult implementation and the year following consult implementation have demonstrated increased test utilization, appropriateness of orders, standardization of reporting, and efficiency of care.

Consult Request

The precision oncology testing process starts with a health care practitioner (HCP) request on behalf of any physician or advanced practice registered nurse. It can be placed by any health care employee and directed to a designated employee in the pathology department. The request is ultimately reviewed by a pathologist (Figure 1). At KCVAMC, this request comes in the form of a consult in the electronic health record (EHR) from the ordering HCP to a pathologist. The KCVAMC pathology consult form was previously published with a discussion of the rationale for this process as opposed to a laboratory order process.² This consult form ensures ordering HCPs supply all necessary information for the pathologist to approve the request and order the test without needing to, in most cases, contact the ordering HCP for clarification or additional information. The form asks the ordering HCP to specify which test is being requested and why. Within the Veterans Health Administration (VHA) there are local and national contracts with many laboratories with hundreds of precision oncology tests to choose from. Consulting with a pathologist is necessary to determine which test is most appropriate.

The precision oncology consult form cannot be submitted without completing all required fields. It also contains indications for the test the ordering HCP selects to minimize unintentionally inappropriate orders. The form asks which tissue the requestor expects the test to be performed on. The requestor must provide contact information for the originating institution when the tissue was collected outside the VHA. The consult form also asks whether another anatomic site is accessible and could be biopsied without unacceptable risk or impracticality, should all previously collected tissue be insufficient. For CGP requests, this allows the pathologist to determine the appropriateness of liquid biopsy without having to reach out to the ordering HCP or wait for the question to be addressed at a tumor board. When a companion diagnostic is available for a test, the ordering HCP is asked which drug will be used so that the most appropriate assay is chosen.

Consult Review

Pathology service involvement begins with pathologist review of the consult form to ensure that the correct test is indicated. Depending on the resources and preferences at a site, consults can be directed to and reviewed by the pathologist associated with the corresponding pathology specimen or to a single pathologist or group of pathologists charged with attending to consults.

The patient’s EHR is reviewed to verify that the test has not already been performed and to determine which tissue to review. Previous surgical pathology reports are examined to assess whether sufficient tissue is available for testing, which may be determined without the need for direct slide examination. Pathologists often use wording such as “rare cells” or in some cases specify that there are not enough lesional cells for ancillary testing. In biopsy reports, the percentage of tissue occupied by lesional cells or the greatest linear length of tumor cells is often documented. As for quality, pathologists may note that a specimen is largely necrotic, and gross descriptions will indicate if a specimen was compromised for molecular analysis by exposure to fixatives such as Bouin’s solution, B-5, or decalcifying agents that contain strong acids.

Tissue Retrieval

If, after such evaluation, the test is indicated and there is tissue that could be sufficient for testing, retrieval of the tissue is pursued. For in-house cases, the pathologist reviews the corresponding surgical pathology report to determine which blocks and slides to pull from the archives. In the cancer checklist, some pathologists specify the best block for subsequent ancillary studies. From the final diagnosis and gross description, the pathologist can determine which blocks are most likely to contain lesional tissue. These slides are retrieved from the archives.

For cases collected at an outside institution (other VHA facility or non-VHA facility/institution), the outside institution must be contacted to retrieve the needed slides and blocks. The phone numbers, fax numbers, email addresses, and mailing addresses for outside institutions are housed in an electronic file and are specific to the point of contact for such requests. Maintaining a record of contacts increases efficiency of the overall process; gathering contact information and successfully requesting tissue often involves multiple automated answering systems, misdirected calls, and failed attempts.

Tissue Review

After retrieving in-house tissue, the pathologist can proceed directly to slide review. For outside cases, the case must first be accessioned so that after review of the slides the pathologist can issue a report to confirm the outside diagnosis. In reviewing the slides, the pathologist looks to see that the diagnosis is correct, that there is a sufficient number of lesional cells in a section, that the lesional cells are of a sufficient concentration in a section, or subsection of the section that could be dissected, and that the cells are viable. Depending on the requested assay and the familiarity of the pathologist with that assay, the pathologist may need to look up the technical requirements of the assay and capabilities of the testing company. Assays vary in sensitivity and require differing amounts and concentrations of tumor. Some companies will dissect tissue, others will not.

If there is sufficient tissue in the material reviewed, the corresponding blocks are retrieved from in-house archives or requests are placed for outside blocks or unstained slides. If there was not enough tissue for testing, the same process is repeated to retrieve and evaluate any other specimens the patient may have. If there are no other specimens to review, this is simply communicated to the ordering HCP via the consult. If the patient is a candidate for liquid biopsy—ie, current specimens are of insufficient quality and/or quantity and a new tissue sample cannot be obtained due to unacceptable risk or impracticality—the order is placed at this time.

Tissue Transport and Testing

Unstained slides need to be cut unless blocks are sent. Slides, blocks, reports, and requisition forms are packaged for transport. An accession number is created for the precision oncology molecular laboratory test in the clinical laboratory section of the EHR system. The clinical laboratory accession number provides a way of tracking sendout testing status. The case is accessioned just prior to placement in the mail so that when an accession number appears in the EHR, the ordering HCP knows the case has been sent out. When results are received, the clinical laboratory accession is completed and a comment is added to indicate where in the EHR to find the report or, when applicable, notes that testing failed.

RESULT REPORTING

When a result becomes available, the report file is downloaded from the vendor portal. This full report is securely transmitted to the ordering HCP. The file is then scanned into the EHR. Additionally, salient findings from the report are abstracted by the pathologist for inclusion as a supplement to the anatomic pathology case. This step ensures that this information travels with the anatomic pathology report if the patient’s care is transferred elsewhere. Templates are used to ensure essential data is captured based on the type of test. The template reminds the pathologist to comment on things such as variants that may represent clonal hematopoiesis, variants that may be germline, and variants that qualify a patient for germline testing. Even with the template, the pathologist must spend significant time reviewing the chart for things such as personal cancer history, other medical history, other masses on imaging, family history, previous surgical pathology reports, and previous molecular testing.

If results are suboptimal, recommendations for repeat testing are made based on the consult response to the question of repeat biopsy feasibility and review of previous pathology reports. The final consult report is added as a consult note, the consult is completed, and the original vendor report file is associated with the consult note in the EHR.

Ancillary Testing Technician

Due to chronic KCVAMC understaffing in the clerical office, gross room, and histology, most of the consult tasks are performed by a pathologist. In an ideal scenario, the pathology staff would divide its time between a pathologist and another dedicated laboratory position, such as an ancillary testing technician (ATT). The ATT can assume responsibilities that do not require the expertise of a pathologist (Figure 2). In such a process, the only steps that would require a pathologist would be review of requests and slides and completion of the interpretive report. All other steps could be accomplished by someone who lacks certifications, laboratory experience, or postsecondary education.

The ATT can receive the requests and retrieve slides and blocks. After slides have been reviewed by a pathologist, the pathologist can inform the ATT which slides or blocks testing will be performed on, provide any additional necessary information for completing the order, and answer any questions. For send-out tests, this allows the ATT to independently complete online portal forms and all other physical requirements prior to delivery of the slides and blocks to specimen processors in the laboratory.

ATTs can keep the ordering HCPs informed of status and be identified as the point of contact for all status inquiries. ATTs can receive results and get outside reports scanned into the EHR. Finally, ATTs can use pathologistdesigned templates to transpose information from outside reports such that a provisional report is prepared and a pathologist does not spend time duplicating information from the outside report. The pathologist can then complete the report with information requiring medical judgment that enhances care.

Optimal Pathologist Involvement

Only 3 steps in the process (request review, tissue review, and completion of an interpretive report) require a pathologist, which are necessary for optimal care and to address barriers to precision oncology.3 While the laboratory may consume only 5% of a health system budget, optimal laboratory use could prevent as much as 30% of avoidable costs.4 These estimates are widely recognized and addressed by campaigns such as Choosing Wisely, as well as programming of alerts and hard stops in EHR systems to reduce duplicate or otherwise inappropriate orders. The tests associated with precision oncology, such as CGP assays, require more nuanced consideration that is best achieved through pathology consultation. In vetting requests for such tests, the pathologist needs information that ordering HCPs do not routinely provide when ordering other tests. A consult asking for such information allows an ordering HCP to efficiently convey this information without having to call the laboratory to circumvent a hard stop.

Regardless of whether a formal electronic consult is used, pathologists must be involved in the review of requests. Creation of an original in-house report also provides an opportunity for pathologists to offer their expertise and maximize the contribution of pathology to patient care. If outside (other VHA facility or non-VHA facility/institution) reports are simply scanned into the EHR without review and issuance of an interpretive report by an in-house pathologist, then an interpretation by a pathologist with access to the patient’s complete chart is never provided. Testing companies are not provided with every patient diagnosis, so in patients with multiple neoplastic conditions, a report may seem to indicate that a detected mutation is from 1 tumor when it is actually from another. Even when all known diagnoses are considered, a variant may be detected that the medical record could reveal to indicate a new diagnosis.

Variation in reporting between companies necessitates pathologist review to standardize care. Some companies indicate which variants may represent clonal hematopoiesis, while others will simply list the pathogenic variants. An oncologist who sees a high volume of hematolymphoid neoplasia may recognize which variants may represent clonal hematopoiesis, but others may not. Reports from the same company may vary, and their interpretation often requires a pathologist's expertise. For example, even if a sample meets the technical requirements for analysis, the report may indicate that the quality or quantity of DNA has reduced the sensitivity for genomic alteration detection. A pathologist would know how to use this information in deciding how to proceed. In a situation where quantity was the issue, the pathologist may know there is additional tissue that could be sent for testing. If quality is the issue, the pathologist may know that additional blocks from the same case likely have the same quality of DNA and would also be unsuitable for testing.

Pathologist input is necessary for precision oncology testing. Some tasks that would ideally be completed by a molecular pathologist (eg, creation of reports to indicate which variants may represent clonal hematopoiesis of indeterminate potential) may be sufficiently completed by a pathologist without fellowship training in molecular pathology.

There are about 15,000 full-time pathologists in the US.⁴ In the 20 years since molecular genetic pathology was formally recognized as a specialty, there have been < 500 pathologists who have pursued fellowship training in this specialty.5 With the inundation of molecular variants uncovered by routine next-generation sequencing (NGS), there are too few fellowship-trained molecular pathologists to provide all such aforementioned input; it is incumbent on surgical pathologists in general to take on such responsibilities.

Consult Implementation Data

These results support the feasibility and effectiveness of the consult process. Prior to consult implementation, many requests were not compliant with VHA National Precision Oncology Program (NPOP) testing guidelines. Since enactment of the consult, > 90% of requests have been in compliance. In the year preceding the consult (January 2020 to December 2021), 55 of 211 (26.1%) metastatic lung and prostate cancers samples eligible for NGS were tested and 126 (59.7%) NGS vendor reports were scanned into the EHR. The mean time from metastasis to NGS result was 151 days. In the year following enactment of the consult (January 2021 to December 2022), 168 of 224 (75.0%) of metastatic lung and prostate cancers eligible for NGS were tested and all 224 NGS vendor reports were scanned into the EHR. The mean time from metastasis to NGS result was 83 days. These data indicate that the practices recommended increase test use, appropriateness of orders, standardization of reporting, and efficiency of care.

CONCLUSIONS

Processing precision oncology testing requires substantial work for pathology departments. Laboratory workforce shortages and ever-expanding indications necessitate additional study of pathology processes to manage increasing workload and maintain the highest quality of cancer care through maximal efficiency and the development of appropriate staffing models. The use of a consult for anatomic pathology molecular testing is one process that can increase test use, appropriateness of orders, standardization of reporting, and efficiency of care. This report provides a comprehensive description and mapping of the process, highlights best practices, identifies inefficiencies, and provides a description and mapping of a target state.