In a 2023 study published in the Annals of Emergency Medicine, European researchers fed the AI system ChatGPT information on 30 ER patients. Details included physician notes on the patients’ symptoms, physical exams, and lab results. ChatGPT made the correct diagnosis in 97% of patients compared to 87% for human doctors.

AI 1, Physicians 0

JAMA Cardiology reported in 2021 that an AI trained on nearly a million ECGs performed comparably to or exceeded cardiologist clinical diagnoses and the MUSE (GE Healthcare) system›s automated ECG analysis for most diagnostic classes.

AI 2, Physicians 0

Google’s medically focused AI model (Med-PaLM2) scored 85%+ when answering US Medical Licensing Examination–style questions. That›s an «expert» physician level and far beyond the accuracy threshold needed to pass the actual exam.

AI 3, Physicians 0

A new AI tool that uses an online finger-tapping test outperformed primary care physicians when assessing the severity of Parkinson’s disease.

AI 4, Physicians 0

JAMA Ophthalmology reported in 2024 that a chatbot outperformed glaucoma specialists and matched retina specialists in diagnostic and treatment accuracy.

AI 5, Physicians 0

Should we stop? Because we could go on. In the last few years, these AI vs Physician studies have proliferated, and guess who’s winning?

65% of Doctors are Concerned

Now, the standard answer with anything AI-and-Medicine goes something like this: AI is coming, and it will be a transformative tool for physicians and improve patient care.

But the underlying unanswered question is: Physicians spend many years and a lot of money to become really good at what they do. How, exactly, should a doctor feel about a machine that can suddenly do the job better and faster?

The Medscape 2023 Physician and AI Report surveyed 1043 US physicians about their views on AI. In total, 65% are concerned about AI making diagnosis and treatment decisions, but 56% are enthusiastic about having it as an adjunct.

Cardiologists, anesthesiologists, and radiologists are most enthusiastic about AI, whereas family physicians and pediatricians are the least enthusiastic.

To get a more personal view of how physicians and other healthcare professionals are feeling about this transformative tech, I spoke with a variety of practicing doctors, a psychotherapist, and a third-year Harvard Medical School student.

‘Abysmally Poor Understanding’

Alfredo A. Sadun, MD, PhD, has been a neuro-ophthalmologist for nearly 50 years. A graduate of MIT and vice-chair of ophthalmology at UCLA, he’s long been fascinated by AI’s march into medicine. He’s watched it accomplish things that no ophthalmologist can do, such as identify gender, age, and risk for heart attack and stroke from retinal scans. But he doesn›t see the same level of interest and comprehension among the medical community.

“There’s still an abysmally poor understanding of AI among physicians in general,” he said. “It’s striking because these are intelligent, well-educated people. But we tend to draw conclusions based on what we’re familiar with, and most doctors’ experience with computers involves EHRs [electronic health records] and administrative garbage. It’s the reason they’re burning out.”

Easing the Burden

Anthony Philippakis, MD, PhD, left his cardiology practice in 2015 to become the chief data officer at the Broad Institute of MIT and Harvard. While there, he helped develop an AI-based method for identifying patients at risk for atrial fibrillation. Now, he’s a general partner at Google Ventures with the goal of bridging the gap between data sciences and medicine. His perspective on AI is unique, given that he’s seen the issue from both sides.

“I am not a bitter physician, but to be honest, when I was practicing, way too much of my time was spent staring at screens and not enough laying hands on patients,” he said. “Can you imagine what it would be like to speak to the EHR naturally and say, ‘Please order the following labs for this patient and notify me when the results come in.’ Boy, would that improve healthcare and physician satisfaction. Every physician I know is excited and optimistic about that. Almost everyone I’ve talked to feels like AI could take a lot of the stuff they don’t like doing off their plates.”

Indeed, the dividing line between physician support for AI and physician suspicion or skepticism of AI is just that. In our survey, more than three quarters of physicians said they would consider using AI for office administrative tasks, scheduling, EHRs, researching medical conditions, and even summarizing a patient’s record before a visit. But far fewer are supportive of it delivering diagnoses and treatments. This, despite an estimated 800,000 Americans dying or becoming permanently disabled each year because of diagnostic error.

Could AI Have Diagnosed This?

John D. Nuschke, MD, has been a primary care physician in Allentown, Pennsylvania, for 40 years. He’s a jovial general physician who insists his patients call him Jack. He’s recently started using an AI medical scribe called Freed. With the patient’s permission, it listens in on the visit and generates notes, saving Dr. Nuschke time and helping him focus on the person. He likes that type of assistance, but when it comes to AI replacing him, he’s skeptical.

“I had this patient I diagnosed with prostate cancer,” he explained. “He got treated and was fine for 5 years. Then, he started losing weight and feeling awful — got weak as a kitten. He went back to his urologist and oncologist who thought he had metastatic prostate cancer. He went through PET scans and blood work, but there was no sign his cancer had returned. So the specialists sent him back to me, and the second he walked in, I saw he was floridly hyperthyroid. I could tell across the room just by looking at him. Would AI have been able to make that diagnosis? Does AI do physical exams?”

Dr. Nuschke said he’s also had several instances where patients received their cancer diagnosis from the lab through an automated patient-portal system rather than from him. “That’s an AI of sorts, and I found it distressing,” he said.

Empathy From a Robot

All the doctors I spoke to were hopeful that by freeing them from the burden of administrative work, they would be able to return to the reason they got into this business in the first place — to spend more time with patients in need and support them with grace and compassion.

But suppose AI could do that too?

In a 2023 study conducted at the University of California San Diego and published in JAMA Internal Medicine, three licensed healthcare professionals compared the responses of ChatGPT and physicians to real-world health questions. The panel rated the AI’s answers nearly four times higher in quality and almost 10 times more empathetic than physicians’ replies.

A similar 2024 study in Nature found that Google’s large-language model AI matched or surpassed physician diagnostic accuracy in all six of the medical specialties considered. Plus, it outperformed doctors in 24 of 26 criteria for conversation quality, including politeness, explanation, honesty, and expressing care and commitment.

Nathaniel Chin, MD, is a gerontologist at the University of Wisconsin and advisory board member for the Alzheimer’s Foundation of America. Although he admits that studies like these “sadden me,” he’s also a realist. “There was hesitation among physicians at the beginning of the pandemic to virtual care because we missed the human connection,” he explained, “but we worked our way around that. We need to remember that what makes a chatbot strong is that it’s nothuman. It doesn’t burn out, it doesn’t get tired, it can look at data very quickly, and it doesn’t have to go home to a family and try to balance work with other aspects of life. A human being is very complex, whereas a chatbot has one single purpose.”

“Even if you don’t have AI in your space now or don’t like the idea of it, that doesn’t matter,” he added. “It’s coming. But it needs to be done right. If AI is implemented by clinicians for clinicians, it has great potential. But if it’s implemented by businesspeople for business reasons, perhaps not.”

‘The Ones Who Use the Tools the Best Will Be the Best’

One branch of medicine that stands to be dramatically affected by AI is mental health. Because bots are natural data-crunchers, they are becoming adept at analyzing the many subtle clues (phrasing in social media posts and text messages, smartwatch biometrics, therapy session videos…) that could indicate depression or other psychological disorders. In fact, its availability via smartphone apps could help democratize and destigmatize the practice.

“There is a day ahead — probably within 5 years — when a patient won’t be able to tell the difference between a real therapist and an AI therapist,” said Ken Mallon, MS, LMFT, a clinical psychotherapist and data scientist in San Jose, California. “That doesn’t worry me, though. It’s hard on therapists’ egos, but new technologies get developed. Things change. People who embrace these tools will benefit from them. The ones who use the tools the best will be the best.”

Time to Restructure Med School

Aditya Jain is in his third year at Harvard Medical School. At age 24, he’s heading into this brave new medical world with excitement and anxiety. Excitement because he sees AI revolutionizing healthcare on every level. Although the current generations of physicians and patients may grumble about its onset, he believes younger ones will feel comfortable with “DocGPT.” He’s excited that his generation of physicians will be the “translators and managers of this transition” and redefine “what it means to be a doctor.”

His anxiety, however, stems from the fact that AI has come on so fast that “it has not yet crossed the threshold of medical education,” he said. “Medical schools still largely prepare students to work as solo clinical decision makers. Most of my first 2 years were spent on pattern recognition and rote memorization, skills that AI can and will master.”

Indeed, Mr. Jain said AI was not a part of his first- or second-year curriculum. “I talk to students who are a year older than me, graduating, heading to residency, and they tell me they wish they had gotten a better grasp of how to use these technologies in medicine and in their practice. They were surprised to hear that people in my year hadn’t started using ChatGPT. We need to expend a lot more effort within the field, within academia, within practicing physicians, to figure out what our role will be in a world where AI is matching or even exceeding human intelligence. And then we need to restructure the medical education to better accomplish these goals.”

So Are You Ready for AI to Be a Better Doctor Than You?

“Yes, I am,” said Dr. Philippakis without hesitation. “When I was going through my medical training, I was continually confronted with the reality that I personally was not smart enough to keep all the information in my head that could be used to make a good decision for a patient. We have now reached a point where the amount of information that is important and useful in the practice of medicine outstrips what a human being can know. The opportunity to enable physicians with AI to remedy that situation is a good thing for doctors and, most importantly, a good thing for patients. I believe the future of medicine belongs not so much to the AI practitioner but to the AI-enabled practitioner.”

“Quick story,” added Dr. Chin. “I asked ChatGPT two questions. The first was ‘Explain the difference between Alzheimer’s and dementia’ because that’s the most common misconception in my field. And it gave me a pretty darn good answer — one I would use in a presentation with some tweaking. Then I asked it, ‘Are you a better doctor than me?’ And it replied, ‘My purpose is not to replace you, my purpose is to be supportive of you and enhance your ability.’ ”

A version of this article appeared on Medscape.com.

In a 2023 study published in the Annals of Emergency Medicine, European researchers fed the AI system ChatGPT information on 30 ER patients. Details included physician notes on the patients’ symptoms, physical exams, and lab results. ChatGPT made the correct diagnosis in 97% of patients compared to 87% for human doctors.

AI 1, Physicians 0

JAMA Cardiology reported in 2021 that an AI trained on nearly a million ECGs performed comparably to or exceeded cardiologist clinical diagnoses and the MUSE (GE Healthcare) system›s automated ECG analysis for most diagnostic classes.

AI 2, Physicians 0

Google’s medically focused AI model (Med-PaLM2) scored 85%+ when answering US Medical Licensing Examination–style questions. That›s an «expert» physician level and far beyond the accuracy threshold needed to pass the actual exam.

AI 3, Physicians 0

A new AI tool that uses an online finger-tapping test outperformed primary care physicians when assessing the severity of Parkinson’s disease.

AI 4, Physicians 0

JAMA Ophthalmology reported in 2024 that a chatbot outperformed glaucoma specialists and matched retina specialists in diagnostic and treatment accuracy.

AI 5, Physicians 0

Should we stop? Because we could go on. In the last few years, these AI vs Physician studies have proliferated, and guess who’s winning?

65% of Doctors are Concerned

Now, the standard answer with anything AI-and-Medicine goes something like this: AI is coming, and it will be a transformative tool for physicians and improve patient care.

But the underlying unanswered question is: Physicians spend many years and a lot of money to become really good at what they do. How, exactly, should a doctor feel about a machine that can suddenly do the job better and faster?

The Medscape 2023 Physician and AI Report surveyed 1043 US physicians about their views on AI. In total, 65% are concerned about AI making diagnosis and treatment decisions, but 56% are enthusiastic about having it as an adjunct.

Cardiologists, anesthesiologists, and radiologists are most enthusiastic about AI, whereas family physicians and pediatricians are the least enthusiastic.

To get a more personal view of how physicians and other healthcare professionals are feeling about this transformative tech, I spoke with a variety of practicing doctors, a psychotherapist, and a third-year Harvard Medical School student.

‘Abysmally Poor Understanding’

Alfredo A. Sadun, MD, PhD, has been a neuro-ophthalmologist for nearly 50 years. A graduate of MIT and vice-chair of ophthalmology at UCLA, he’s long been fascinated by AI’s march into medicine. He’s watched it accomplish things that no ophthalmologist can do, such as identify gender, age, and risk for heart attack and stroke from retinal scans. But he doesn›t see the same level of interest and comprehension among the medical community.

“There’s still an abysmally poor understanding of AI among physicians in general,” he said. “It’s striking because these are intelligent, well-educated people. But we tend to draw conclusions based on what we’re familiar with, and most doctors’ experience with computers involves EHRs [electronic health records] and administrative garbage. It’s the reason they’re burning out.”

Easing the Burden

Anthony Philippakis, MD, PhD, left his cardiology practice in 2015 to become the chief data officer at the Broad Institute of MIT and Harvard. While there, he helped develop an AI-based method for identifying patients at risk for atrial fibrillation. Now, he’s a general partner at Google Ventures with the goal of bridging the gap between data sciences and medicine. His perspective on AI is unique, given that he’s seen the issue from both sides.

“I am not a bitter physician, but to be honest, when I was practicing, way too much of my time was spent staring at screens and not enough laying hands on patients,” he said. “Can you imagine what it would be like to speak to the EHR naturally and say, ‘Please order the following labs for this patient and notify me when the results come in.’ Boy, would that improve healthcare and physician satisfaction. Every physician I know is excited and optimistic about that. Almost everyone I’ve talked to feels like AI could take a lot of the stuff they don’t like doing off their plates.”

Indeed, the dividing line between physician support for AI and physician suspicion or skepticism of AI is just that. In our survey, more than three quarters of physicians said they would consider using AI for office administrative tasks, scheduling, EHRs, researching medical conditions, and even summarizing a patient’s record before a visit. But far fewer are supportive of it delivering diagnoses and treatments. This, despite an estimated 800,000 Americans dying or becoming permanently disabled each year because of diagnostic error.

Could AI Have Diagnosed This?

John D. Nuschke, MD, has been a primary care physician in Allentown, Pennsylvania, for 40 years. He’s a jovial general physician who insists his patients call him Jack. He’s recently started using an AI medical scribe called Freed. With the patient’s permission, it listens in on the visit and generates notes, saving Dr. Nuschke time and helping him focus on the person. He likes that type of assistance, but when it comes to AI replacing him, he’s skeptical.

“I had this patient I diagnosed with prostate cancer,” he explained. “He got treated and was fine for 5 years. Then, he started losing weight and feeling awful — got weak as a kitten. He went back to his urologist and oncologist who thought he had metastatic prostate cancer. He went through PET scans and blood work, but there was no sign his cancer had returned. So the specialists sent him back to me, and the second he walked in, I saw he was floridly hyperthyroid. I could tell across the room just by looking at him. Would AI have been able to make that diagnosis? Does AI do physical exams?”

Dr. Nuschke said he’s also had several instances where patients received their cancer diagnosis from the lab through an automated patient-portal system rather than from him. “That’s an AI of sorts, and I found it distressing,” he said.

Empathy From a Robot

All the doctors I spoke to were hopeful that by freeing them from the burden of administrative work, they would be able to return to the reason they got into this business in the first place — to spend more time with patients in need and support them with grace and compassion.

But suppose AI could do that too?

In a 2023 study conducted at the University of California San Diego and published in JAMA Internal Medicine, three licensed healthcare professionals compared the responses of ChatGPT and physicians to real-world health questions. The panel rated the AI’s answers nearly four times higher in quality and almost 10 times more empathetic than physicians’ replies.

A similar 2024 study in Nature found that Google’s large-language model AI matched or surpassed physician diagnostic accuracy in all six of the medical specialties considered. Plus, it outperformed doctors in 24 of 26 criteria for conversation quality, including politeness, explanation, honesty, and expressing care and commitment.

Nathaniel Chin, MD, is a gerontologist at the University of Wisconsin and advisory board member for the Alzheimer’s Foundation of America. Although he admits that studies like these “sadden me,” he’s also a realist. “There was hesitation among physicians at the beginning of the pandemic to virtual care because we missed the human connection,” he explained, “but we worked our way around that. We need to remember that what makes a chatbot strong is that it’s nothuman. It doesn’t burn out, it doesn’t get tired, it can look at data very quickly, and it doesn’t have to go home to a family and try to balance work with other aspects of life. A human being is very complex, whereas a chatbot has one single purpose.”

“Even if you don’t have AI in your space now or don’t like the idea of it, that doesn’t matter,” he added. “It’s coming. But it needs to be done right. If AI is implemented by clinicians for clinicians, it has great potential. But if it’s implemented by businesspeople for business reasons, perhaps not.”

‘The Ones Who Use the Tools the Best Will Be the Best’

One branch of medicine that stands to be dramatically affected by AI is mental health. Because bots are natural data-crunchers, they are becoming adept at analyzing the many subtle clues (phrasing in social media posts and text messages, smartwatch biometrics, therapy session videos…) that could indicate depression or other psychological disorders. In fact, its availability via smartphone apps could help democratize and destigmatize the practice.

“There is a day ahead — probably within 5 years — when a patient won’t be able to tell the difference between a real therapist and an AI therapist,” said Ken Mallon, MS, LMFT, a clinical psychotherapist and data scientist in San Jose, California. “That doesn’t worry me, though. It’s hard on therapists’ egos, but new technologies get developed. Things change. People who embrace these tools will benefit from them. The ones who use the tools the best will be the best.”

Time to Restructure Med School

Aditya Jain is in his third year at Harvard Medical School. At age 24, he’s heading into this brave new medical world with excitement and anxiety. Excitement because he sees AI revolutionizing healthcare on every level. Although the current generations of physicians and patients may grumble about its onset, he believes younger ones will feel comfortable with “DocGPT.” He’s excited that his generation of physicians will be the “translators and managers of this transition” and redefine “what it means to be a doctor.”

His anxiety, however, stems from the fact that AI has come on so fast that “it has not yet crossed the threshold of medical education,” he said. “Medical schools still largely prepare students to work as solo clinical decision makers. Most of my first 2 years were spent on pattern recognition and rote memorization, skills that AI can and will master.”

Indeed, Mr. Jain said AI was not a part of his first- or second-year curriculum. “I talk to students who are a year older than me, graduating, heading to residency, and they tell me they wish they had gotten a better grasp of how to use these technologies in medicine and in their practice. They were surprised to hear that people in my year hadn’t started using ChatGPT. We need to expend a lot more effort within the field, within academia, within practicing physicians, to figure out what our role will be in a world where AI is matching or even exceeding human intelligence. And then we need to restructure the medical education to better accomplish these goals.”

So Are You Ready for AI to Be a Better Doctor Than You?

“Yes, I am,” said Dr. Philippakis without hesitation. “When I was going through my medical training, I was continually confronted with the reality that I personally was not smart enough to keep all the information in my head that could be used to make a good decision for a patient. We have now reached a point where the amount of information that is important and useful in the practice of medicine outstrips what a human being can know. The opportunity to enable physicians with AI to remedy that situation is a good thing for doctors and, most importantly, a good thing for patients. I believe the future of medicine belongs not so much to the AI practitioner but to the AI-enabled practitioner.”

“Quick story,” added Dr. Chin. “I asked ChatGPT two questions. The first was ‘Explain the difference between Alzheimer’s and dementia’ because that’s the most common misconception in my field. And it gave me a pretty darn good answer — one I would use in a presentation with some tweaking. Then I asked it, ‘Are you a better doctor than me?’ And it replied, ‘My purpose is not to replace you, my purpose is to be supportive of you and enhance your ability.’ ”

A version of this article appeared on Medscape.com.

In a 2023 study published in the Annals of Emergency Medicine, European researchers fed the AI system ChatGPT information on 30 ER patients. Details included physician notes on the patients’ symptoms, physical exams, and lab results. ChatGPT made the correct diagnosis in 97% of patients compared to 87% for human doctors.

AI 1, Physicians 0

JAMA Cardiology reported in 2021 that an AI trained on nearly a million ECGs performed comparably to or exceeded cardiologist clinical diagnoses and the MUSE (GE Healthcare) system›s automated ECG analysis for most diagnostic classes.

AI 2, Physicians 0

Google’s medically focused AI model (Med-PaLM2) scored 85%+ when answering US Medical Licensing Examination–style questions. That›s an «expert» physician level and far beyond the accuracy threshold needed to pass the actual exam.

AI 3, Physicians 0

A new AI tool that uses an online finger-tapping test outperformed primary care physicians when assessing the severity of Parkinson’s disease.

AI 4, Physicians 0

JAMA Ophthalmology reported in 2024 that a chatbot outperformed glaucoma specialists and matched retina specialists in diagnostic and treatment accuracy.

AI 5, Physicians 0

Should we stop? Because we could go on. In the last few years, these AI vs Physician studies have proliferated, and guess who’s winning?

65% of Doctors are Concerned

Now, the standard answer with anything AI-and-Medicine goes something like this: AI is coming, and it will be a transformative tool for physicians and improve patient care.

But the underlying unanswered question is: Physicians spend many years and a lot of money to become really good at what they do. How, exactly, should a doctor feel about a machine that can suddenly do the job better and faster?

The Medscape 2023 Physician and AI Report surveyed 1043 US physicians about their views on AI. In total, 65% are concerned about AI making diagnosis and treatment decisions, but 56% are enthusiastic about having it as an adjunct.

Cardiologists, anesthesiologists, and radiologists are most enthusiastic about AI, whereas family physicians and pediatricians are the least enthusiastic.

To get a more personal view of how physicians and other healthcare professionals are feeling about this transformative tech, I spoke with a variety of practicing doctors, a psychotherapist, and a third-year Harvard Medical School student.

‘Abysmally Poor Understanding’

Alfredo A. Sadun, MD, PhD, has been a neuro-ophthalmologist for nearly 50 years. A graduate of MIT and vice-chair of ophthalmology at UCLA, he’s long been fascinated by AI’s march into medicine. He’s watched it accomplish things that no ophthalmologist can do, such as identify gender, age, and risk for heart attack and stroke from retinal scans. But he doesn›t see the same level of interest and comprehension among the medical community.

“There’s still an abysmally poor understanding of AI among physicians in general,” he said. “It’s striking because these are intelligent, well-educated people. But we tend to draw conclusions based on what we’re familiar with, and most doctors’ experience with computers involves EHRs [electronic health records] and administrative garbage. It’s the reason they’re burning out.”

Easing the Burden

Anthony Philippakis, MD, PhD, left his cardiology practice in 2015 to become the chief data officer at the Broad Institute of MIT and Harvard. While there, he helped develop an AI-based method for identifying patients at risk for atrial fibrillation. Now, he’s a general partner at Google Ventures with the goal of bridging the gap between data sciences and medicine. His perspective on AI is unique, given that he’s seen the issue from both sides.

“I am not a bitter physician, but to be honest, when I was practicing, way too much of my time was spent staring at screens and not enough laying hands on patients,” he said. “Can you imagine what it would be like to speak to the EHR naturally and say, ‘Please order the following labs for this patient and notify me when the results come in.’ Boy, would that improve healthcare and physician satisfaction. Every physician I know is excited and optimistic about that. Almost everyone I’ve talked to feels like AI could take a lot of the stuff they don’t like doing off their plates.”

Indeed, the dividing line between physician support for AI and physician suspicion or skepticism of AI is just that. In our survey, more than three quarters of physicians said they would consider using AI for office administrative tasks, scheduling, EHRs, researching medical conditions, and even summarizing a patient’s record before a visit. But far fewer are supportive of it delivering diagnoses and treatments. This, despite an estimated 800,000 Americans dying or becoming permanently disabled each year because of diagnostic error.

Could AI Have Diagnosed This?

John D. Nuschke, MD, has been a primary care physician in Allentown, Pennsylvania, for 40 years. He’s a jovial general physician who insists his patients call him Jack. He’s recently started using an AI medical scribe called Freed. With the patient’s permission, it listens in on the visit and generates notes, saving Dr. Nuschke time and helping him focus on the person. He likes that type of assistance, but when it comes to AI replacing him, he’s skeptical.

“I had this patient I diagnosed with prostate cancer,” he explained. “He got treated and was fine for 5 years. Then, he started losing weight and feeling awful — got weak as a kitten. He went back to his urologist and oncologist who thought he had metastatic prostate cancer. He went through PET scans and blood work, but there was no sign his cancer had returned. So the specialists sent him back to me, and the second he walked in, I saw he was floridly hyperthyroid. I could tell across the room just by looking at him. Would AI have been able to make that diagnosis? Does AI do physical exams?”

Dr. Nuschke said he’s also had several instances where patients received their cancer diagnosis from the lab through an automated patient-portal system rather than from him. “That’s an AI of sorts, and I found it distressing,” he said.

Empathy From a Robot

All the doctors I spoke to were hopeful that by freeing them from the burden of administrative work, they would be able to return to the reason they got into this business in the first place — to spend more time with patients in need and support them with grace and compassion.

But suppose AI could do that too?

In a 2023 study conducted at the University of California San Diego and published in JAMA Internal Medicine, three licensed healthcare professionals compared the responses of ChatGPT and physicians to real-world health questions. The panel rated the AI’s answers nearly four times higher in quality and almost 10 times more empathetic than physicians’ replies.

A similar 2024 study in Nature found that Google’s large-language model AI matched or surpassed physician diagnostic accuracy in all six of the medical specialties considered. Plus, it outperformed doctors in 24 of 26 criteria for conversation quality, including politeness, explanation, honesty, and expressing care and commitment.

Nathaniel Chin, MD, is a gerontologist at the University of Wisconsin and advisory board member for the Alzheimer’s Foundation of America. Although he admits that studies like these “sadden me,” he’s also a realist. “There was hesitation among physicians at the beginning of the pandemic to virtual care because we missed the human connection,” he explained, “but we worked our way around that. We need to remember that what makes a chatbot strong is that it’s nothuman. It doesn’t burn out, it doesn’t get tired, it can look at data very quickly, and it doesn’t have to go home to a family and try to balance work with other aspects of life. A human being is very complex, whereas a chatbot has one single purpose.”

“Even if you don’t have AI in your space now or don’t like the idea of it, that doesn’t matter,” he added. “It’s coming. But it needs to be done right. If AI is implemented by clinicians for clinicians, it has great potential. But if it’s implemented by businesspeople for business reasons, perhaps not.”

‘The Ones Who Use the Tools the Best Will Be the Best’

One branch of medicine that stands to be dramatically affected by AI is mental health. Because bots are natural data-crunchers, they are becoming adept at analyzing the many subtle clues (phrasing in social media posts and text messages, smartwatch biometrics, therapy session videos…) that could indicate depression or other psychological disorders. In fact, its availability via smartphone apps could help democratize and destigmatize the practice.

“There is a day ahead — probably within 5 years — when a patient won’t be able to tell the difference between a real therapist and an AI therapist,” said Ken Mallon, MS, LMFT, a clinical psychotherapist and data scientist in San Jose, California. “That doesn’t worry me, though. It’s hard on therapists’ egos, but new technologies get developed. Things change. People who embrace these tools will benefit from them. The ones who use the tools the best will be the best.”

Time to Restructure Med School

Aditya Jain is in his third year at Harvard Medical School. At age 24, he’s heading into this brave new medical world with excitement and anxiety. Excitement because he sees AI revolutionizing healthcare on every level. Although the current generations of physicians and patients may grumble about its onset, he believes younger ones will feel comfortable with “DocGPT.” He’s excited that his generation of physicians will be the “translators and managers of this transition” and redefine “what it means to be a doctor.”

His anxiety, however, stems from the fact that AI has come on so fast that “it has not yet crossed the threshold of medical education,” he said. “Medical schools still largely prepare students to work as solo clinical decision makers. Most of my first 2 years were spent on pattern recognition and rote memorization, skills that AI can and will master.”

Indeed, Mr. Jain said AI was not a part of his first- or second-year curriculum. “I talk to students who are a year older than me, graduating, heading to residency, and they tell me they wish they had gotten a better grasp of how to use these technologies in medicine and in their practice. They were surprised to hear that people in my year hadn’t started using ChatGPT. We need to expend a lot more effort within the field, within academia, within practicing physicians, to figure out what our role will be in a world where AI is matching or even exceeding human intelligence. And then we need to restructure the medical education to better accomplish these goals.”

So Are You Ready for AI to Be a Better Doctor Than You?

“Yes, I am,” said Dr. Philippakis without hesitation. “When I was going through my medical training, I was continually confronted with the reality that I personally was not smart enough to keep all the information in my head that could be used to make a good decision for a patient. We have now reached a point where the amount of information that is important and useful in the practice of medicine outstrips what a human being can know. The opportunity to enable physicians with AI to remedy that situation is a good thing for doctors and, most importantly, a good thing for patients. I believe the future of medicine belongs not so much to the AI practitioner but to the AI-enabled practitioner.”

“Quick story,” added Dr. Chin. “I asked ChatGPT two questions. The first was ‘Explain the difference between Alzheimer’s and dementia’ because that’s the most common misconception in my field. And it gave me a pretty darn good answer — one I would use in a presentation with some tweaking. Then I asked it, ‘Are you a better doctor than me?’ And it replied, ‘My purpose is not to replace you, my purpose is to be supportive of you and enhance your ability.’ ”

A version of this article appeared on Medscape.com.

E-cigarettes entered the market as consumer products without comprehensive toxicological testing,based on the assessment that they were 95% less harmful than traditional cigarettes. Further, consumer dvertising suggests that e-cigarettes are a good alternative to conventional combustible cigarettes and can serve as a gateway to quitting smoking.

However, hen considering damage to the endothelium and toxicity, e-cigarettes have a negative impact like that of conventional cigarettes. Moreover, switching to e-cigarettes often leads to dual use, said Stefan Andreas, MD, director of the Lungenfachklinik in Immenhausen, Germany, at the Congress of the German Respiratory Society and Intensive Care Medicine. 
 

Subclinical Atherosclerosis

Because e-cigarettes have emerged relatively recently, long-term studies on their cardiac consequences are not yet available. Dr. Andreas explained that the impact on endothelial function is relevant for risk assessment. Endothelial function is a biomarker for early, subclinical atherosclerosis. “If endothelial function is impaired, the risk for heart attack and stroke is significantly increased 5-10 years later,” said Dr. Andreas.

The results of a crossover study showed reduced vascular elasticity after consuming both tobacco cigarettes and e-cigarettes. The study included 20 smokers, and endothelial function was measured using flow-mediated vasodilation.

Significant effects on the vessels were also found in a study of 31 participants who had never smoked. The study participants inhaled a nicotine-free aerosol from e-cigarettes. Before and after, parameters of endothelial function were examined using a 3.0-T MRI. After aerosol inhalation, the resistance index was 2.3% higher (P < .05), and flow-mediated vascular dilation was reduced by 34% (P < .001).

A recent review involving 372 participants from China showed that e-cigarettes lead to an increase in pulse wave velocity, with a difference of 3.08 (P < .001). “Pulse wave velocity is also a marker of endothelial function: The stiffer the vessels, the higher the pulse wave velocity,” said Dr. Andreas. The authors of the review concluded that “e-cigarettes should not be promoted as a healthier alternative to tobacco smoking.”
 

No Harmless Alternative

A recent review compared the effects of tobacco smoking and e-cigarettes. The results showed that vaping e-cigarettes causes oxidative stress, inflammation, endothelial dysfunction, and related cardiovascular consequences. The authors attributed the findings to overlapping toxic compounds in vapor and tobacco smoke and similar pathomechanical features of vaping and smoking. Although the toxic mixture in smoke is more complex, both e-cigarettes and tobacco cigarettes “impaired endothelial function to a similar extent,” they wrote. The authors attributed this finding to oxidative stress as the central mechanism.

“There is increasing evidence that e-cigarettes are not a harmless alternative to tobacco cigarettes,” wrote Thomas Münzel, MD, professor of cardiology at the University of Mainz and his team in their 2020 review, which examined studies in humans and animals. They provided an overview of the effects of tobacco/hookah smoking and e-cigarette vaping on endothelial function. They also pointed to emerging adverse effects on the proteome, transcriptome, epigenome, microbiome, and circadian clock.

Finally, a toxicological review of e-cigarettes also found alarmingly high levels of carcinogens and toxins that could have long-term effects on other organs, including the development of neurological symptoms, lung cancer, cardiovascular diseases, and cavities.

Dr. Andreas observed that even small amounts, such as those obtained through secondhand smoking, can be harmful. In 2007, Dr. Andreas and his colleagues showed that even low exposure to tobacco smoke can lead to a significant increase in cardiovascular events.
 

Conflicts of Interest 

Dr. Andreas recommended closely examining the studies that suggest that e-cigarettes are less risky. “It is noticeable that there is a significant difference depending on whether publications were supported by the tobacco industry or not,” he emphasized.

Danish scientists found that a conflict of interest (COI) has a strong influence on study results. “In studies without a COI, e-cigarettes are found to cause damage 95% of the time. In contrast, when there is a strong conflict of interest, the result is often ‘no harm,’” said Dr. Andreas.

This effect is quite relevant for the discussion of e-cigarettes. “If scientists make a critical statement in a position paper, there will always be someone who says, ‘No, it’s different, there are these and those publications.’ The true nature of interest-driven publications on e-cigarettes is not always easy to discern,” said Dr. Andreas.
 

No Gateway to Quitting 

E-cigarettes are used in clinical studies for tobacco cessation. The results of a randomized study showed that significantly more smokers who were switched to e-cigarettes quit smoking, compared with controls. But there was no significant difference in complete smoking cessation between groups. Moreover, 45% of smokers who switched to e-cigarettes became dual users, compared with 11% of controls.

“Translating these results means that for one person who quits smoking by using e-cigarettes, they gain five people who use both traditional cigarettes and e-cigarettes,” explained Dr. Andreas.

In their recent review, Münzel and colleagues pointed out that the assessment that e-cigarettes could help with quitting might be wrong. Rather, it seems that “e-cigarettes have the opposite effect.” They also note that the age of initiation for e-cigarettes is generally lower than for tobacco cigarettes: Consumption often starts at age 13 or 14 years. And the consumption of e-cigarettes among children and adolescents increased by 7% from 2016 to 2023.

A meta-analysis published at the end of February also shows that e-cigarettes are about as dangerous as tobacco cigarettes. They are more dangerous than not smoking, and dual use is more dangerous than tobacco cigarettes alone. “There is a need to reassess the assumption that e-cigarette use provides substantial harm reduction across all cigarette-caused diseases, particularly accounting for dual use,” wrote the authors.

“One must always consider that e-cigarettes have only been available for a relatively short time. We can only see the cumulative toxicity in 10, 20 years when we have patients who have smoked e-cigarettes only for 20 years,” said Dr. Andreas. Ultimately, however, e-cigarettes promote dual use and, consequently, additive toxicity.
 

Nicotine Replacement Therapies 

Quitting smoking reduces the risk of cardiovascular events and premature death by 40%, even among patients with cardiovascular disease, according to a Cochrane meta-analysis. Smoking cessation reduces the risk for cardiovascular death by 39%, the risk for major adverse cardiovascular events by 43%, the risk for heart attack by 36%, the risk for stroke by 30%, and overall mortality by 40%.

Quitting smoking is the most effective measure for risk reduction, as a meta-analysis of 20 studies in patients with coronary heart disease found. Smoking cessation was associated with a 36% risk reduction compared with 29% risk reduction for statin therapy, 23% risk reduction with beta-blockers and ACE inhibitors and 15% risk reduction with aspirin.

Dr. Andreas emphasized that nicotine replacement therapies are well-researched and safe even in cardiovascular disease, as shown by a US study that included patients who had sustained a heart attack. A group of the participants was treated with nicotine patches for 10 weeks, while the other group received a placebo. After 14 weeks, 21% of the nicotine patch group achieved abstinence vs 9% of the placebo group (P = .001). Transdermal nicotine application does not lead to a significant increase in cardiovascular events in high-risk patients.

The German “Nonsmoker Heroes” app has proven to be an effective means of behavioral therapeutic coaching. A recent study of it included 17 study centers with 661 participants. About 21% of the subjects had chronic obstructive pulmonary disease, 19% had asthma. Smoking onset occurred at age 16 years. The subjects were highly dependent: > 72% had at least moderate dependence, > 58% had high to very high dependence, and the population had an average of 3.6 quit attempts. The odds ratio for self-reported abstinence was 2.2 after 6 months. “The app is not only effective, but also can be prescribed on an extrabudgetary basis,” said Dr. Andreas.

This story was translated from the Medscape German edition using several editorial tools, including AI, as part of the process. Human editors reviewed this content before publication. A version of this article appeared on Medscape.com.

E-cigarettes entered the market as consumer products without comprehensive toxicological testing,based on the assessment that they were 95% less harmful than traditional cigarettes. Further, consumer dvertising suggests that e-cigarettes are a good alternative to conventional combustible cigarettes and can serve as a gateway to quitting smoking.

However, hen considering damage to the endothelium and toxicity, e-cigarettes have a negative impact like that of conventional cigarettes. Moreover, switching to e-cigarettes often leads to dual use, said Stefan Andreas, MD, director of the Lungenfachklinik in Immenhausen, Germany, at the Congress of the German Respiratory Society and Intensive Care Medicine. 
 

Subclinical Atherosclerosis

Because e-cigarettes have emerged relatively recently, long-term studies on their cardiac consequences are not yet available. Dr. Andreas explained that the impact on endothelial function is relevant for risk assessment. Endothelial function is a biomarker for early, subclinical atherosclerosis. “If endothelial function is impaired, the risk for heart attack and stroke is significantly increased 5-10 years later,” said Dr. Andreas.

The results of a crossover study showed reduced vascular elasticity after consuming both tobacco cigarettes and e-cigarettes. The study included 20 smokers, and endothelial function was measured using flow-mediated vasodilation.

Significant effects on the vessels were also found in a study of 31 participants who had never smoked. The study participants inhaled a nicotine-free aerosol from e-cigarettes. Before and after, parameters of endothelial function were examined using a 3.0-T MRI. After aerosol inhalation, the resistance index was 2.3% higher (P < .05), and flow-mediated vascular dilation was reduced by 34% (P < .001).

A recent review involving 372 participants from China showed that e-cigarettes lead to an increase in pulse wave velocity, with a difference of 3.08 (P < .001). “Pulse wave velocity is also a marker of endothelial function: The stiffer the vessels, the higher the pulse wave velocity,” said Dr. Andreas. The authors of the review concluded that “e-cigarettes should not be promoted as a healthier alternative to tobacco smoking.”
 

No Harmless Alternative

A recent review compared the effects of tobacco smoking and e-cigarettes. The results showed that vaping e-cigarettes causes oxidative stress, inflammation, endothelial dysfunction, and related cardiovascular consequences. The authors attributed the findings to overlapping toxic compounds in vapor and tobacco smoke and similar pathomechanical features of vaping and smoking. Although the toxic mixture in smoke is more complex, both e-cigarettes and tobacco cigarettes “impaired endothelial function to a similar extent,” they wrote. The authors attributed this finding to oxidative stress as the central mechanism.

“There is increasing evidence that e-cigarettes are not a harmless alternative to tobacco cigarettes,” wrote Thomas Münzel, MD, professor of cardiology at the University of Mainz and his team in their 2020 review, which examined studies in humans and animals. They provided an overview of the effects of tobacco/hookah smoking and e-cigarette vaping on endothelial function. They also pointed to emerging adverse effects on the proteome, transcriptome, epigenome, microbiome, and circadian clock.

Finally, a toxicological review of e-cigarettes also found alarmingly high levels of carcinogens and toxins that could have long-term effects on other organs, including the development of neurological symptoms, lung cancer, cardiovascular diseases, and cavities.

Dr. Andreas observed that even small amounts, such as those obtained through secondhand smoking, can be harmful. In 2007, Dr. Andreas and his colleagues showed that even low exposure to tobacco smoke can lead to a significant increase in cardiovascular events.
 

Conflicts of Interest 

Dr. Andreas recommended closely examining the studies that suggest that e-cigarettes are less risky. “It is noticeable that there is a significant difference depending on whether publications were supported by the tobacco industry or not,” he emphasized.

Danish scientists found that a conflict of interest (COI) has a strong influence on study results. “In studies without a COI, e-cigarettes are found to cause damage 95% of the time. In contrast, when there is a strong conflict of interest, the result is often ‘no harm,’” said Dr. Andreas.

This effect is quite relevant for the discussion of e-cigarettes. “If scientists make a critical statement in a position paper, there will always be someone who says, ‘No, it’s different, there are these and those publications.’ The true nature of interest-driven publications on e-cigarettes is not always easy to discern,” said Dr. Andreas.
 

No Gateway to Quitting 

E-cigarettes are used in clinical studies for tobacco cessation. The results of a randomized study showed that significantly more smokers who were switched to e-cigarettes quit smoking, compared with controls. But there was no significant difference in complete smoking cessation between groups. Moreover, 45% of smokers who switched to e-cigarettes became dual users, compared with 11% of controls.

“Translating these results means that for one person who quits smoking by using e-cigarettes, they gain five people who use both traditional cigarettes and e-cigarettes,” explained Dr. Andreas.

In their recent review, Münzel and colleagues pointed out that the assessment that e-cigarettes could help with quitting might be wrong. Rather, it seems that “e-cigarettes have the opposite effect.” They also note that the age of initiation for e-cigarettes is generally lower than for tobacco cigarettes: Consumption often starts at age 13 or 14 years. And the consumption of e-cigarettes among children and adolescents increased by 7% from 2016 to 2023.

A meta-analysis published at the end of February also shows that e-cigarettes are about as dangerous as tobacco cigarettes. They are more dangerous than not smoking, and dual use is more dangerous than tobacco cigarettes alone. “There is a need to reassess the assumption that e-cigarette use provides substantial harm reduction across all cigarette-caused diseases, particularly accounting for dual use,” wrote the authors.

“One must always consider that e-cigarettes have only been available for a relatively short time. We can only see the cumulative toxicity in 10, 20 years when we have patients who have smoked e-cigarettes only for 20 years,” said Dr. Andreas. Ultimately, however, e-cigarettes promote dual use and, consequently, additive toxicity.
 

Nicotine Replacement Therapies 

Quitting smoking reduces the risk of cardiovascular events and premature death by 40%, even among patients with cardiovascular disease, according to a Cochrane meta-analysis. Smoking cessation reduces the risk for cardiovascular death by 39%, the risk for major adverse cardiovascular events by 43%, the risk for heart attack by 36%, the risk for stroke by 30%, and overall mortality by 40%.

Quitting smoking is the most effective measure for risk reduction, as a meta-analysis of 20 studies in patients with coronary heart disease found. Smoking cessation was associated with a 36% risk reduction compared with 29% risk reduction for statin therapy, 23% risk reduction with beta-blockers and ACE inhibitors and 15% risk reduction with aspirin.

Dr. Andreas emphasized that nicotine replacement therapies are well-researched and safe even in cardiovascular disease, as shown by a US study that included patients who had sustained a heart attack. A group of the participants was treated with nicotine patches for 10 weeks, while the other group received a placebo. After 14 weeks, 21% of the nicotine patch group achieved abstinence vs 9% of the placebo group (P = .001). Transdermal nicotine application does not lead to a significant increase in cardiovascular events in high-risk patients.

The German “Nonsmoker Heroes” app has proven to be an effective means of behavioral therapeutic coaching. A recent study of it included 17 study centers with 661 participants. About 21% of the subjects had chronic obstructive pulmonary disease, 19% had asthma. Smoking onset occurred at age 16 years. The subjects were highly dependent: > 72% had at least moderate dependence, > 58% had high to very high dependence, and the population had an average of 3.6 quit attempts. The odds ratio for self-reported abstinence was 2.2 after 6 months. “The app is not only effective, but also can be prescribed on an extrabudgetary basis,” said Dr. Andreas.

This story was translated from the Medscape German edition using several editorial tools, including AI, as part of the process. Human editors reviewed this content before publication. A version of this article appeared on Medscape.com.

E-cigarettes entered the market as consumer products without comprehensive toxicological testing,based on the assessment that they were 95% less harmful than traditional cigarettes. Further, consumer dvertising suggests that e-cigarettes are a good alternative to conventional combustible cigarettes and can serve as a gateway to quitting smoking.

However, hen considering damage to the endothelium and toxicity, e-cigarettes have a negative impact like that of conventional cigarettes. Moreover, switching to e-cigarettes often leads to dual use, said Stefan Andreas, MD, director of the Lungenfachklinik in Immenhausen, Germany, at the Congress of the German Respiratory Society and Intensive Care Medicine. 
 

Subclinical Atherosclerosis

Because e-cigarettes have emerged relatively recently, long-term studies on their cardiac consequences are not yet available. Dr. Andreas explained that the impact on endothelial function is relevant for risk assessment. Endothelial function is a biomarker for early, subclinical atherosclerosis. “If endothelial function is impaired, the risk for heart attack and stroke is significantly increased 5-10 years later,” said Dr. Andreas.

The results of a crossover study showed reduced vascular elasticity after consuming both tobacco cigarettes and e-cigarettes. The study included 20 smokers, and endothelial function was measured using flow-mediated vasodilation.

Significant effects on the vessels were also found in a study of 31 participants who had never smoked. The study participants inhaled a nicotine-free aerosol from e-cigarettes. Before and after, parameters of endothelial function were examined using a 3.0-T MRI. After aerosol inhalation, the resistance index was 2.3% higher (P < .05), and flow-mediated vascular dilation was reduced by 34% (P < .001).

A recent review involving 372 participants from China showed that e-cigarettes lead to an increase in pulse wave velocity, with a difference of 3.08 (P < .001). “Pulse wave velocity is also a marker of endothelial function: The stiffer the vessels, the higher the pulse wave velocity,” said Dr. Andreas. The authors of the review concluded that “e-cigarettes should not be promoted as a healthier alternative to tobacco smoking.”
 

No Harmless Alternative

A recent review compared the effects of tobacco smoking and e-cigarettes. The results showed that vaping e-cigarettes causes oxidative stress, inflammation, endothelial dysfunction, and related cardiovascular consequences. The authors attributed the findings to overlapping toxic compounds in vapor and tobacco smoke and similar pathomechanical features of vaping and smoking. Although the toxic mixture in smoke is more complex, both e-cigarettes and tobacco cigarettes “impaired endothelial function to a similar extent,” they wrote. The authors attributed this finding to oxidative stress as the central mechanism.

“There is increasing evidence that e-cigarettes are not a harmless alternative to tobacco cigarettes,” wrote Thomas Münzel, MD, professor of cardiology at the University of Mainz and his team in their 2020 review, which examined studies in humans and animals. They provided an overview of the effects of tobacco/hookah smoking and e-cigarette vaping on endothelial function. They also pointed to emerging adverse effects on the proteome, transcriptome, epigenome, microbiome, and circadian clock.

Finally, a toxicological review of e-cigarettes also found alarmingly high levels of carcinogens and toxins that could have long-term effects on other organs, including the development of neurological symptoms, lung cancer, cardiovascular diseases, and cavities.

Dr. Andreas observed that even small amounts, such as those obtained through secondhand smoking, can be harmful. In 2007, Dr. Andreas and his colleagues showed that even low exposure to tobacco smoke can lead to a significant increase in cardiovascular events.
 

Conflicts of Interest 

Dr. Andreas recommended closely examining the studies that suggest that e-cigarettes are less risky. “It is noticeable that there is a significant difference depending on whether publications were supported by the tobacco industry or not,” he emphasized.

Danish scientists found that a conflict of interest (COI) has a strong influence on study results. “In studies without a COI, e-cigarettes are found to cause damage 95% of the time. In contrast, when there is a strong conflict of interest, the result is often ‘no harm,’” said Dr. Andreas.

This effect is quite relevant for the discussion of e-cigarettes. “If scientists make a critical statement in a position paper, there will always be someone who says, ‘No, it’s different, there are these and those publications.’ The true nature of interest-driven publications on e-cigarettes is not always easy to discern,” said Dr. Andreas.
 

No Gateway to Quitting 

E-cigarettes are used in clinical studies for tobacco cessation. The results of a randomized study showed that significantly more smokers who were switched to e-cigarettes quit smoking, compared with controls. But there was no significant difference in complete smoking cessation between groups. Moreover, 45% of smokers who switched to e-cigarettes became dual users, compared with 11% of controls.

“Translating these results means that for one person who quits smoking by using e-cigarettes, they gain five people who use both traditional cigarettes and e-cigarettes,” explained Dr. Andreas.

In their recent review, Münzel and colleagues pointed out that the assessment that e-cigarettes could help with quitting might be wrong. Rather, it seems that “e-cigarettes have the opposite effect.” They also note that the age of initiation for e-cigarettes is generally lower than for tobacco cigarettes: Consumption often starts at age 13 or 14 years. And the consumption of e-cigarettes among children and adolescents increased by 7% from 2016 to 2023.

A meta-analysis published at the end of February also shows that e-cigarettes are about as dangerous as tobacco cigarettes. They are more dangerous than not smoking, and dual use is more dangerous than tobacco cigarettes alone. “There is a need to reassess the assumption that e-cigarette use provides substantial harm reduction across all cigarette-caused diseases, particularly accounting for dual use,” wrote the authors.

“One must always consider that e-cigarettes have only been available for a relatively short time. We can only see the cumulative toxicity in 10, 20 years when we have patients who have smoked e-cigarettes only for 20 years,” said Dr. Andreas. Ultimately, however, e-cigarettes promote dual use and, consequently, additive toxicity.
 

Nicotine Replacement Therapies 

Quitting smoking reduces the risk of cardiovascular events and premature death by 40%, even among patients with cardiovascular disease, according to a Cochrane meta-analysis. Smoking cessation reduces the risk for cardiovascular death by 39%, the risk for major adverse cardiovascular events by 43%, the risk for heart attack by 36%, the risk for stroke by 30%, and overall mortality by 40%.

Quitting smoking is the most effective measure for risk reduction, as a meta-analysis of 20 studies in patients with coronary heart disease found. Smoking cessation was associated with a 36% risk reduction compared with 29% risk reduction for statin therapy, 23% risk reduction with beta-blockers and ACE inhibitors and 15% risk reduction with aspirin.

Dr. Andreas emphasized that nicotine replacement therapies are well-researched and safe even in cardiovascular disease, as shown by a US study that included patients who had sustained a heart attack. A group of the participants was treated with nicotine patches for 10 weeks, while the other group received a placebo. After 14 weeks, 21% of the nicotine patch group achieved abstinence vs 9% of the placebo group (P = .001). Transdermal nicotine application does not lead to a significant increase in cardiovascular events in high-risk patients.

The German “Nonsmoker Heroes” app has proven to be an effective means of behavioral therapeutic coaching. A recent study of it included 17 study centers with 661 participants. About 21% of the subjects had chronic obstructive pulmonary disease, 19% had asthma. Smoking onset occurred at age 16 years. The subjects were highly dependent: > 72% had at least moderate dependence, > 58% had high to very high dependence, and the population had an average of 3.6 quit attempts. The odds ratio for self-reported abstinence was 2.2 after 6 months. “The app is not only effective, but also can be prescribed on an extrabudgetary basis,” said Dr. Andreas.

This story was translated from the Medscape German edition using several editorial tools, including AI, as part of the process. Human editors reviewed this content before publication. A version of this article appeared on Medscape.com.

TOPLINE:

Endoscopic sleeve gastroplasty (ESG) is more cost-effective, and achieves and sustains greater weight loss, than semaglutide over a 5-year period in patients with class II obesity.

METHODOLOGY:

• Researchers used a Markov cohort model to assess the cost-effectiveness of semaglutide vs ESG over 5 years in people with class II obesity (body mass index [BMI], 35-39.9), with the model costs based on the US healthcare system.
• A 45-year-old patient with a BMI of 37 was included as the base case in this study.
• The model simulated hypothetical patients with class II obesity who received ESG, semaglutide, or no treatment (reference group with zero treatment costs).
• The model derived clinical data for the first year from two randomized clinical trials, STEP 1 (semaglutide) and MERIT (ESG); for the following years, data were derived from published studies and publicly available data sources.
• Study outcomes were total costs, quality-adjusted life years (QALYs), and incremental cost-effectiveness ratio (ICER).

TAKEAWAY:

• ESG led to better weight loss outcomes (BMI, 31.7 vs 33.0) and added 0.06 more QALYs relative to semaglutide in the modelled patients over the 5-year time horizon; about 20% of the patients receiving semaglutide dropped out owing to medication intolerance or other reasons.
• The semaglutide treatment was $33,583 more expensive than the ESG treatment over the 5-year period.
• ESG became more cost-effective than semaglutide at 2 years and remained so over a 5-year time horizon, with an ICER of — $595,532 per QALY for the base case.
• The annual price of semaglutide would need to be reduced from $13,618 to $3591 to achieve nondominance compared with ESG.

IN PRACTICE:

“The strategic choice of cost saving yet effective treatment such as ESG compared with semaglutide for specific patient groups could help alleviate the potential budget strain expected from the use of semaglutide,” the authors wrote.

SOURCE:

Muhammad Haseeb, MD, MSc, Division of Gastroenterology, Hepatology and Endoscopy, Brigham and Women’s Hospital, Boston, led this study, which was published online on April 12, 2024, in JAMA Network Open.

LIMITATIONS:

The study did not look at benefits associated with improvements in comorbidities from either treatment strategy, and the model used did not account for any microlevel follow-up costs such as routine clinic visits. The authors acknowledged that semaglutide’s prices may fall in the future when more anti-obesity drugs get approved.

DISCLOSURES:

This study was supported in part by the National Institutes of Health. Some authors declared receiving personal fees, royalty payments, and/or grants and having other ties with several sources.

A version of this article appeared on Medscape.com.

TOPLINE:

Endoscopic sleeve gastroplasty (ESG) is more cost-effective, and achieves and sustains greater weight loss, than semaglutide over a 5-year period in patients with class II obesity.

METHODOLOGY:

• Researchers used a Markov cohort model to assess the cost-effectiveness of semaglutide vs ESG over 5 years in people with class II obesity (body mass index [BMI], 35-39.9), with the model costs based on the US healthcare system.
• A 45-year-old patient with a BMI of 37 was included as the base case in this study.
• The model simulated hypothetical patients with class II obesity who received ESG, semaglutide, or no treatment (reference group with zero treatment costs).
• The model derived clinical data for the first year from two randomized clinical trials, STEP 1 (semaglutide) and MERIT (ESG); for the following years, data were derived from published studies and publicly available data sources.
• Study outcomes were total costs, quality-adjusted life years (QALYs), and incremental cost-effectiveness ratio (ICER).

TAKEAWAY:

• ESG led to better weight loss outcomes (BMI, 31.7 vs 33.0) and added 0.06 more QALYs relative to semaglutide in the modelled patients over the 5-year time horizon; about 20% of the patients receiving semaglutide dropped out owing to medication intolerance or other reasons.
• The semaglutide treatment was $33,583 more expensive than the ESG treatment over the 5-year period.
• ESG became more cost-effective than semaglutide at 2 years and remained so over a 5-year time horizon, with an ICER of — $595,532 per QALY for the base case.
• The annual price of semaglutide would need to be reduced from $13,618 to $3591 to achieve nondominance compared with ESG.

IN PRACTICE:

“The strategic choice of cost saving yet effective treatment such as ESG compared with semaglutide for specific patient groups could help alleviate the potential budget strain expected from the use of semaglutide,” the authors wrote.

SOURCE:

Muhammad Haseeb, MD, MSc, Division of Gastroenterology, Hepatology and Endoscopy, Brigham and Women’s Hospital, Boston, led this study, which was published online on April 12, 2024, in JAMA Network Open.

LIMITATIONS:

The study did not look at benefits associated with improvements in comorbidities from either treatment strategy, and the model used did not account for any microlevel follow-up costs such as routine clinic visits. The authors acknowledged that semaglutide’s prices may fall in the future when more anti-obesity drugs get approved.

DISCLOSURES:

This study was supported in part by the National Institutes of Health. Some authors declared receiving personal fees, royalty payments, and/or grants and having other ties with several sources.

A version of this article appeared on Medscape.com.

TOPLINE:

Endoscopic sleeve gastroplasty (ESG) is more cost-effective, and achieves and sustains greater weight loss, than semaglutide over a 5-year period in patients with class II obesity.

METHODOLOGY:

• Researchers used a Markov cohort model to assess the cost-effectiveness of semaglutide vs ESG over 5 years in people with class II obesity (body mass index [BMI], 35-39.9), with the model costs based on the US healthcare system.
• A 45-year-old patient with a BMI of 37 was included as the base case in this study.
• The model simulated hypothetical patients with class II obesity who received ESG, semaglutide, or no treatment (reference group with zero treatment costs).
• The model derived clinical data for the first year from two randomized clinical trials, STEP 1 (semaglutide) and MERIT (ESG); for the following years, data were derived from published studies and publicly available data sources.
• Study outcomes were total costs, quality-adjusted life years (QALYs), and incremental cost-effectiveness ratio (ICER).

TAKEAWAY:

• ESG led to better weight loss outcomes (BMI, 31.7 vs 33.0) and added 0.06 more QALYs relative to semaglutide in the modelled patients over the 5-year time horizon; about 20% of the patients receiving semaglutide dropped out owing to medication intolerance or other reasons.
• The semaglutide treatment was $33,583 more expensive than the ESG treatment over the 5-year period.
• ESG became more cost-effective than semaglutide at 2 years and remained so over a 5-year time horizon, with an ICER of — $595,532 per QALY for the base case.
• The annual price of semaglutide would need to be reduced from $13,618 to $3591 to achieve nondominance compared with ESG.

IN PRACTICE:

“The strategic choice of cost saving yet effective treatment such as ESG compared with semaglutide for specific patient groups could help alleviate the potential budget strain expected from the use of semaglutide,” the authors wrote.

SOURCE:

Muhammad Haseeb, MD, MSc, Division of Gastroenterology, Hepatology and Endoscopy, Brigham and Women’s Hospital, Boston, led this study, which was published online on April 12, 2024, in JAMA Network Open.

LIMITATIONS:

The study did not look at benefits associated with improvements in comorbidities from either treatment strategy, and the model used did not account for any microlevel follow-up costs such as routine clinic visits. The authors acknowledged that semaglutide’s prices may fall in the future when more anti-obesity drugs get approved.

DISCLOSURES:

This study was supported in part by the National Institutes of Health. Some authors declared receiving personal fees, royalty payments, and/or grants and having other ties with several sources.

A version of this article appeared on Medscape.com.

A smartphone-based tool that tracks mental health status by detecting changes in voice may complement traditional psychiatric assessments and improve an individual’s ability to self-monitor depressive and other mental health symptoms, new research suggested.

Investigators used the Mental Fitness Vocal Biomarker (MFVB) scoring algorithm, which is incorporated into a smartphone voice journaling application, to detect increased or decreased risk for elevated mental health symptom severity by analyzing 30-second free speech voice recordings for specific vocal patterns previously linked to mental health.

A comparison between MFVB scores and commonly used clinical mental health assessments revealed a statistically significant correlation, researchers noted.

“While the MFVB tool is not intended to diagnose or treat mental health conditions, these findings provide a robust initial foundation upon which to further explore its potential in personalized wellness tracking, which has so far not yet been able to extend measurement of physical health to mental wellbeing,” reported the researchers, led by Erik Larsen, PhD, with Boston-based Sonde Health, which developed the tool.

The study was published online in Frontiers in Psychiatry.
 

Eight Vocal Features

The potential value of vocal biomarkers for mental health assessment has gained increasing attention.

“Somebody that is depressed often sounds more monotone; they may have less inflection in their voice and speak slower with less energy, which can be recognized in voice recordings,” Dr. Larsen told this news organization.

“This is an area which has received quite a bit of research in the last few decades to find out what specific aspects of acoustics and rhythm of speech could point to conditions like depression,” Dr. Larsen said.

In the current study, the researchers set out to validate the ability of the MFVB platform to detect mental health symptoms.

With the tool, users record their thoughts and feelings as a 30-second voice journal. The tool analyzes the recordings for eight vocal features previously shown to be relevant to mental health. These include jitter, shimmer, pitch variability, energy variability, vowel space, phonation duration, speech rate, and pause duration.

The tool calculates a real-time MFVB score ranging from 0 to 100. A score of 80-100 is defined as “excellent” and 70-79 as “good,” while a score of 0-69 is categorized as “pay attention.” It was trained on more than 1 million voice samples to optimize performance across a diverse range of cultures, languages, and socioeconomic groups.

The current study included 104 outpatient psychiatric patients (73% women) with anxiety-related diagnoses, trauma, and stress-related disorders or depressive disorders. The cohort was mostly made up of White, non-Hispanic young adults. Patients with a history of substance abuse or who were taking psychiatric medications that may affect voice and speech were excluded.

During the 4-week study period, participants conducted 1336 app sessions with voice recordings, resulting in an average of 12.8 sessions per participant, or 3.2 per week.

MFVB scores were cross-referenced against the results of participants’ M3 Checklist, a clinically validated mental health assessment tool.

Over a period of 2 weeks, participants were twice as likely to report elevated mental health symptoms if their MFVB scores remained in the “pay attention” range vs in the “excellent” range, the researchers found.

The effect was more pronounced in those who used the app more frequently, with frequent users 8.5 times more likely to show elevated symptoms.

The correlation between MFVB scores and established mental health assessments was “not only statistically significant but also meaningful for participants,” researchers wrote. Subgroup analyses suggest the app works best for depression and stress- and trauma-related disorders.

The tool provides psychiatric outpatients with “immediate quantitative feedback on their mental health symptom severity,” the researchers noted.

In their paper, the investigators caution that the results highlight the “general ability” of MFVB score categories to differentiate mental health symptom severity levels but do not distinguish what type of symptoms these may be, such as depression, anxiety, or posttraumatic stress disorder.

In a statement, study investigator Lindsey Venesky, PhD, psychologist and clinical director at the Cognitive Behavior Institute in Pittsburgh, noted that the ability to collect mental health data from patients between clinic visits “could transform how we monitor symptoms and optimize treatment plans.”

“Voice-based health tracking technology can provide accurate insights into a client’s mental health status over time and can do so seamlessly and unobtrusively, with little added effort for clients,” Dr. Venesky said.

Need for Replication, Validation

Commenting on the findings, John Torous, MD, director of the division of digital psychiatry at Beth Israel Deaconess Medical Center, Boston, noted that “over the last 20 years, there has been a lot of interest in voice biomarkers, yet somehow that research has never been translated into mainstream clinical care.”

Voice biomarkers are “relevant and have potential” in mental health, he noted. The findings in this study are “interesting, but they need to be thoroughly externally replicated and validated to show that these biomarkers are valid and reliable,” Dr. Torous added.

Changes in voice are part of the mental status exam, Dr. Torous said, “but it’s only one piece of information that we collect in a clinical assessment of many pieces of information.”

Dr. Torous also cautioned that “as a practicing psychiatrist, it can be tricky to be given new data if you don’t know how to interpret it or what it means. An important step would be education, outreach, and resources for physicians to learn about potential voice biomarkers.”

The authors received internal financial support for the research, authorship, and/or publication of this article. The pilot phase of the study at St. Joseph’s Healthcare Hamilton was partially supported through Mitacs Accelerate International, Canada. Dr. Larsen and three coauthors are employed by Sonde Health. Dr. Torous had no relevant conflicts of interest.

A version of this article appeared on Medscape.com.

A smartphone-based tool that tracks mental health status by detecting changes in voice may complement traditional psychiatric assessments and improve an individual’s ability to self-monitor depressive and other mental health symptoms, new research suggested.

Investigators used the Mental Fitness Vocal Biomarker (MFVB) scoring algorithm, which is incorporated into a smartphone voice journaling application, to detect increased or decreased risk for elevated mental health symptom severity by analyzing 30-second free speech voice recordings for specific vocal patterns previously linked to mental health.

A comparison between MFVB scores and commonly used clinical mental health assessments revealed a statistically significant correlation, researchers noted.

“While the MFVB tool is not intended to diagnose or treat mental health conditions, these findings provide a robust initial foundation upon which to further explore its potential in personalized wellness tracking, which has so far not yet been able to extend measurement of physical health to mental wellbeing,” reported the researchers, led by Erik Larsen, PhD, with Boston-based Sonde Health, which developed the tool.

The study was published online in Frontiers in Psychiatry.
 

Eight Vocal Features

The potential value of vocal biomarkers for mental health assessment has gained increasing attention.

“Somebody that is depressed often sounds more monotone; they may have less inflection in their voice and speak slower with less energy, which can be recognized in voice recordings,” Dr. Larsen told this news organization.

“This is an area which has received quite a bit of research in the last few decades to find out what specific aspects of acoustics and rhythm of speech could point to conditions like depression,” Dr. Larsen said.

In the current study, the researchers set out to validate the ability of the MFVB platform to detect mental health symptoms.

With the tool, users record their thoughts and feelings as a 30-second voice journal. The tool analyzes the recordings for eight vocal features previously shown to be relevant to mental health. These include jitter, shimmer, pitch variability, energy variability, vowel space, phonation duration, speech rate, and pause duration.

The tool calculates a real-time MFVB score ranging from 0 to 100. A score of 80-100 is defined as “excellent” and 70-79 as “good,” while a score of 0-69 is categorized as “pay attention.” It was trained on more than 1 million voice samples to optimize performance across a diverse range of cultures, languages, and socioeconomic groups.

The current study included 104 outpatient psychiatric patients (73% women) with anxiety-related diagnoses, trauma, and stress-related disorders or depressive disorders. The cohort was mostly made up of White, non-Hispanic young adults. Patients with a history of substance abuse or who were taking psychiatric medications that may affect voice and speech were excluded.

During the 4-week study period, participants conducted 1336 app sessions with voice recordings, resulting in an average of 12.8 sessions per participant, or 3.2 per week.

MFVB scores were cross-referenced against the results of participants’ M3 Checklist, a clinically validated mental health assessment tool.

Over a period of 2 weeks, participants were twice as likely to report elevated mental health symptoms if their MFVB scores remained in the “pay attention” range vs in the “excellent” range, the researchers found.

The effect was more pronounced in those who used the app more frequently, with frequent users 8.5 times more likely to show elevated symptoms.

The correlation between MFVB scores and established mental health assessments was “not only statistically significant but also meaningful for participants,” researchers wrote. Subgroup analyses suggest the app works best for depression and stress- and trauma-related disorders.

The tool provides psychiatric outpatients with “immediate quantitative feedback on their mental health symptom severity,” the researchers noted.

In their paper, the investigators caution that the results highlight the “general ability” of MFVB score categories to differentiate mental health symptom severity levels but do not distinguish what type of symptoms these may be, such as depression, anxiety, or posttraumatic stress disorder.

In a statement, study investigator Lindsey Venesky, PhD, psychologist and clinical director at the Cognitive Behavior Institute in Pittsburgh, noted that the ability to collect mental health data from patients between clinic visits “could transform how we monitor symptoms and optimize treatment plans.”

“Voice-based health tracking technology can provide accurate insights into a client’s mental health status over time and can do so seamlessly and unobtrusively, with little added effort for clients,” Dr. Venesky said.

Need for Replication, Validation

Commenting on the findings, John Torous, MD, director of the division of digital psychiatry at Beth Israel Deaconess Medical Center, Boston, noted that “over the last 20 years, there has been a lot of interest in voice biomarkers, yet somehow that research has never been translated into mainstream clinical care.”

Voice biomarkers are “relevant and have potential” in mental health, he noted. The findings in this study are “interesting, but they need to be thoroughly externally replicated and validated to show that these biomarkers are valid and reliable,” Dr. Torous added.

Changes in voice are part of the mental status exam, Dr. Torous said, “but it’s only one piece of information that we collect in a clinical assessment of many pieces of information.”

Dr. Torous also cautioned that “as a practicing psychiatrist, it can be tricky to be given new data if you don’t know how to interpret it or what it means. An important step would be education, outreach, and resources for physicians to learn about potential voice biomarkers.”

The authors received internal financial support for the research, authorship, and/or publication of this article. The pilot phase of the study at St. Joseph’s Healthcare Hamilton was partially supported through Mitacs Accelerate International, Canada. Dr. Larsen and three coauthors are employed by Sonde Health. Dr. Torous had no relevant conflicts of interest.

A version of this article appeared on Medscape.com.

A smartphone-based tool that tracks mental health status by detecting changes in voice may complement traditional psychiatric assessments and improve an individual’s ability to self-monitor depressive and other mental health symptoms, new research suggested.

Investigators used the Mental Fitness Vocal Biomarker (MFVB) scoring algorithm, which is incorporated into a smartphone voice journaling application, to detect increased or decreased risk for elevated mental health symptom severity by analyzing 30-second free speech voice recordings for specific vocal patterns previously linked to mental health.

A comparison between MFVB scores and commonly used clinical mental health assessments revealed a statistically significant correlation, researchers noted.

“While the MFVB tool is not intended to diagnose or treat mental health conditions, these findings provide a robust initial foundation upon which to further explore its potential in personalized wellness tracking, which has so far not yet been able to extend measurement of physical health to mental wellbeing,” reported the researchers, led by Erik Larsen, PhD, with Boston-based Sonde Health, which developed the tool.

The study was published online in Frontiers in Psychiatry.
 

Eight Vocal Features

The potential value of vocal biomarkers for mental health assessment has gained increasing attention.

“Somebody that is depressed often sounds more monotone; they may have less inflection in their voice and speak slower with less energy, which can be recognized in voice recordings,” Dr. Larsen told this news organization.

“This is an area which has received quite a bit of research in the last few decades to find out what specific aspects of acoustics and rhythm of speech could point to conditions like depression,” Dr. Larsen said.

In the current study, the researchers set out to validate the ability of the MFVB platform to detect mental health symptoms.

With the tool, users record their thoughts and feelings as a 30-second voice journal. The tool analyzes the recordings for eight vocal features previously shown to be relevant to mental health. These include jitter, shimmer, pitch variability, energy variability, vowel space, phonation duration, speech rate, and pause duration.

The tool calculates a real-time MFVB score ranging from 0 to 100. A score of 80-100 is defined as “excellent” and 70-79 as “good,” while a score of 0-69 is categorized as “pay attention.” It was trained on more than 1 million voice samples to optimize performance across a diverse range of cultures, languages, and socioeconomic groups.

The current study included 104 outpatient psychiatric patients (73% women) with anxiety-related diagnoses, trauma, and stress-related disorders or depressive disorders. The cohort was mostly made up of White, non-Hispanic young adults. Patients with a history of substance abuse or who were taking psychiatric medications that may affect voice and speech were excluded.

During the 4-week study period, participants conducted 1336 app sessions with voice recordings, resulting in an average of 12.8 sessions per participant, or 3.2 per week.

MFVB scores were cross-referenced against the results of participants’ M3 Checklist, a clinically validated mental health assessment tool.

Over a period of 2 weeks, participants were twice as likely to report elevated mental health symptoms if their MFVB scores remained in the “pay attention” range vs in the “excellent” range, the researchers found.

The effect was more pronounced in those who used the app more frequently, with frequent users 8.5 times more likely to show elevated symptoms.

The correlation between MFVB scores and established mental health assessments was “not only statistically significant but also meaningful for participants,” researchers wrote. Subgroup analyses suggest the app works best for depression and stress- and trauma-related disorders.

The tool provides psychiatric outpatients with “immediate quantitative feedback on their mental health symptom severity,” the researchers noted.

In their paper, the investigators caution that the results highlight the “general ability” of MFVB score categories to differentiate mental health symptom severity levels but do not distinguish what type of symptoms these may be, such as depression, anxiety, or posttraumatic stress disorder.

In a statement, study investigator Lindsey Venesky, PhD, psychologist and clinical director at the Cognitive Behavior Institute in Pittsburgh, noted that the ability to collect mental health data from patients between clinic visits “could transform how we monitor symptoms and optimize treatment plans.”

“Voice-based health tracking technology can provide accurate insights into a client’s mental health status over time and can do so seamlessly and unobtrusively, with little added effort for clients,” Dr. Venesky said.

Need for Replication, Validation

Commenting on the findings, John Torous, MD, director of the division of digital psychiatry at Beth Israel Deaconess Medical Center, Boston, noted that “over the last 20 years, there has been a lot of interest in voice biomarkers, yet somehow that research has never been translated into mainstream clinical care.”

Voice biomarkers are “relevant and have potential” in mental health, he noted. The findings in this study are “interesting, but they need to be thoroughly externally replicated and validated to show that these biomarkers are valid and reliable,” Dr. Torous added.

Changes in voice are part of the mental status exam, Dr. Torous said, “but it’s only one piece of information that we collect in a clinical assessment of many pieces of information.”

Dr. Torous also cautioned that “as a practicing psychiatrist, it can be tricky to be given new data if you don’t know how to interpret it or what it means. An important step would be education, outreach, and resources for physicians to learn about potential voice biomarkers.”

The authors received internal financial support for the research, authorship, and/or publication of this article. The pilot phase of the study at St. Joseph’s Healthcare Hamilton was partially supported through Mitacs Accelerate International, Canada. Dr. Larsen and three coauthors are employed by Sonde Health. Dr. Torous had no relevant conflicts of interest.

A version of this article appeared on Medscape.com.

SAN DIEGO — A new blood test that combines cell-free DNA fragmentomes and protein biomarkers to screen for ovarian cancer shows promising results, according to an initial analysis. 

The test, under development by Delfi Diagnostics, “looks very sensitive for detecting ovarian cancer early,” said company founder and board member Victor E. Velculescu, MD, PhD, codirector of Cancer Genetics and Epigenetics at Johns Hopkins University, Baltimore. 

The assay uses machine learning to integrate cell-free DNA fragment patterns with concentrations of two ovarian cancer biomarkers — CA125 and HE4 — to detect tumors. 

While fragmentation patterns are organized in healthy people, they are chaotic in cancer and reveal both its presence and location, said Velculescu who presented the findings at the American Association for Cancer Research annual meeting.

The researchers tested the assay in 134 women with ovarian cancer, 204 women without cancer, and 203 women with benign adnexal masses. The approach identified 69% of stage 1 cancers, 76% of stage 2, 85% of stage 3, and 100% of stage 4 at a specificity of over 99% and an area under the curve (AUC) of 0.97.

The test identified 91% of high-grade serous ovarian cancers — the most common type of ovarian cancer.

The AUC for distinguishing benign masses from cancer was 0.87, with 60% of ovarian cancers detected at a specificity of 95%. 

“In the preoperative setting where lower specificity is acceptable, this approach may improve management of adnexal masses,” the investigators said in their abstract.

Dr. Velculescu cautioned that the report “is an initial analysis” and that his team is working on validating the finding on a larger scale in both average and high-risk women.

If validated, the test “could enable population-wide ovarian cancer screening,” he added.

Delfi recently launched a lung cancer screening blood test — FirstLook Lung— that also uses a “fragmentomics” approach to detect tumors. The company is hopeful it will reach the market with a similar test for ovarian cancer, but it’s not a certainty. 

With lung cancer, we know screening helps. For ovarian cancer, however, it’s unclear whether this will help or not, said Dr. Velculescu. But based on the study findings, but “we are now optimistic that this could make an impact. We have more work to do.” 

This presentation was one of many at the meeting about liquid biopsies using DNA, RNA, and proteins to detect cancer, including a new assay for pancreatic cancer, another cancer that like ovarian cancer is difficult to detect in the early stages. 

“This is the future,” said study moderator Roy S. Herbst, MD, PhD, chief of medical oncology at Yale University in New Haven, Connecticut. 

He called liquid biopsy “a great advance” in many oncology settings, including cancer screening because finding tumors early offers the best chance at cure. 

However, one of the main concerns about rolling out liquid biopsies for wide-scale cancer screening is the possibility that a test will come back positive, but no tumor will be seen on diagnostic imaging, said Herbst. It won’t be clear if the test was a false positive or if the patient has a brewing tumor that can’t be located and treated, a difficult situation for both patients and doctors. 

What to do in that situation is “a policy question that the entire country is asking now as liquid biopsies are moving forward,” he said. We are going to have to come together to figure it out and learn how to use these tests.

The work was funded by Delfi Diagnostics, the National Institutes of Health, and others. Dr. Velculescu, in addition to founding Delfi, holds patents on the technology. Dr. Herbst is a consultant, researcher, and/or holds stock in many companies, including AstraZeneca, Pfizer, and Checkpoint Therapeutics. 
 

A version of this article appeared on Medscape.com .

SAN DIEGO — A new blood test that combines cell-free DNA fragmentomes and protein biomarkers to screen for ovarian cancer shows promising results, according to an initial analysis. 

The test, under development by Delfi Diagnostics, “looks very sensitive for detecting ovarian cancer early,” said company founder and board member Victor E. Velculescu, MD, PhD, codirector of Cancer Genetics and Epigenetics at Johns Hopkins University, Baltimore. 

The assay uses machine learning to integrate cell-free DNA fragment patterns with concentrations of two ovarian cancer biomarkers — CA125 and HE4 — to detect tumors. 

While fragmentation patterns are organized in healthy people, they are chaotic in cancer and reveal both its presence and location, said Velculescu who presented the findings at the American Association for Cancer Research annual meeting.

The researchers tested the assay in 134 women with ovarian cancer, 204 women without cancer, and 203 women with benign adnexal masses. The approach identified 69% of stage 1 cancers, 76% of stage 2, 85% of stage 3, and 100% of stage 4 at a specificity of over 99% and an area under the curve (AUC) of 0.97.

The test identified 91% of high-grade serous ovarian cancers — the most common type of ovarian cancer.

The AUC for distinguishing benign masses from cancer was 0.87, with 60% of ovarian cancers detected at a specificity of 95%. 

“In the preoperative setting where lower specificity is acceptable, this approach may improve management of adnexal masses,” the investigators said in their abstract.

Dr. Velculescu cautioned that the report “is an initial analysis” and that his team is working on validating the finding on a larger scale in both average and high-risk women.

If validated, the test “could enable population-wide ovarian cancer screening,” he added.

Delfi recently launched a lung cancer screening blood test — FirstLook Lung— that also uses a “fragmentomics” approach to detect tumors. The company is hopeful it will reach the market with a similar test for ovarian cancer, but it’s not a certainty. 

With lung cancer, we know screening helps. For ovarian cancer, however, it’s unclear whether this will help or not, said Dr. Velculescu. But based on the study findings, but “we are now optimistic that this could make an impact. We have more work to do.” 

This presentation was one of many at the meeting about liquid biopsies using DNA, RNA, and proteins to detect cancer, including a new assay for pancreatic cancer, another cancer that like ovarian cancer is difficult to detect in the early stages. 

“This is the future,” said study moderator Roy S. Herbst, MD, PhD, chief of medical oncology at Yale University in New Haven, Connecticut. 

He called liquid biopsy “a great advance” in many oncology settings, including cancer screening because finding tumors early offers the best chance at cure. 

However, one of the main concerns about rolling out liquid biopsies for wide-scale cancer screening is the possibility that a test will come back positive, but no tumor will be seen on diagnostic imaging, said Herbst. It won’t be clear if the test was a false positive or if the patient has a brewing tumor that can’t be located and treated, a difficult situation for both patients and doctors. 

What to do in that situation is “a policy question that the entire country is asking now as liquid biopsies are moving forward,” he said. We are going to have to come together to figure it out and learn how to use these tests.

The work was funded by Delfi Diagnostics, the National Institutes of Health, and others. Dr. Velculescu, in addition to founding Delfi, holds patents on the technology. Dr. Herbst is a consultant, researcher, and/or holds stock in many companies, including AstraZeneca, Pfizer, and Checkpoint Therapeutics. 
 

A version of this article appeared on Medscape.com .

SAN DIEGO — A new blood test that combines cell-free DNA fragmentomes and protein biomarkers to screen for ovarian cancer shows promising results, according to an initial analysis. 

The test, under development by Delfi Diagnostics, “looks very sensitive for detecting ovarian cancer early,” said company founder and board member Victor E. Velculescu, MD, PhD, codirector of Cancer Genetics and Epigenetics at Johns Hopkins University, Baltimore. 

The assay uses machine learning to integrate cell-free DNA fragment patterns with concentrations of two ovarian cancer biomarkers — CA125 and HE4 — to detect tumors. 

While fragmentation patterns are organized in healthy people, they are chaotic in cancer and reveal both its presence and location, said Velculescu who presented the findings at the American Association for Cancer Research annual meeting.

The researchers tested the assay in 134 women with ovarian cancer, 204 women without cancer, and 203 women with benign adnexal masses. The approach identified 69% of stage 1 cancers, 76% of stage 2, 85% of stage 3, and 100% of stage 4 at a specificity of over 99% and an area under the curve (AUC) of 0.97.

The test identified 91% of high-grade serous ovarian cancers — the most common type of ovarian cancer.

The AUC for distinguishing benign masses from cancer was 0.87, with 60% of ovarian cancers detected at a specificity of 95%. 

“In the preoperative setting where lower specificity is acceptable, this approach may improve management of adnexal masses,” the investigators said in their abstract.

Dr. Velculescu cautioned that the report “is an initial analysis” and that his team is working on validating the finding on a larger scale in both average and high-risk women.

If validated, the test “could enable population-wide ovarian cancer screening,” he added.

Delfi recently launched a lung cancer screening blood test — FirstLook Lung— that also uses a “fragmentomics” approach to detect tumors. The company is hopeful it will reach the market with a similar test for ovarian cancer, but it’s not a certainty. 

With lung cancer, we know screening helps. For ovarian cancer, however, it’s unclear whether this will help or not, said Dr. Velculescu. But based on the study findings, but “we are now optimistic that this could make an impact. We have more work to do.” 

This presentation was one of many at the meeting about liquid biopsies using DNA, RNA, and proteins to detect cancer, including a new assay for pancreatic cancer, another cancer that like ovarian cancer is difficult to detect in the early stages. 

“This is the future,” said study moderator Roy S. Herbst, MD, PhD, chief of medical oncology at Yale University in New Haven, Connecticut. 

He called liquid biopsy “a great advance” in many oncology settings, including cancer screening because finding tumors early offers the best chance at cure. 

However, one of the main concerns about rolling out liquid biopsies for wide-scale cancer screening is the possibility that a test will come back positive, but no tumor will be seen on diagnostic imaging, said Herbst. It won’t be clear if the test was a false positive or if the patient has a brewing tumor that can’t be located and treated, a difficult situation for both patients and doctors. 

What to do in that situation is “a policy question that the entire country is asking now as liquid biopsies are moving forward,” he said. We are going to have to come together to figure it out and learn how to use these tests.

The work was funded by Delfi Diagnostics, the National Institutes of Health, and others. Dr. Velculescu, in addition to founding Delfi, holds patents on the technology. Dr. Herbst is a consultant, researcher, and/or holds stock in many companies, including AstraZeneca, Pfizer, and Checkpoint Therapeutics. 
 

A version of this article appeared on Medscape.com .

TOPLINE:

Black patients with schizophrenia are less likely to receive a clozapine prescription compared with White patients, a new study shows. The findings held even after the researchers controlled for demographic variables, social determinants of health, and care access patterns.

METHODOLOGY:

• The study drew on structured electronic health record data on 3160 adult patients with schizophrenia.
• The mean age at first recorded diagnosis was 39.5 years; 70% of participants were male, 53% Black, and 91% resided in an urban setting.
• The researchers used the social vulnerability index (SVI) to quantify social determinants of health.
• Descriptive data analysis, logistic regression, and sensitivity analysis were used to identify differences between those who received a clozapine prescription and those who were prescribed antipsychotic medications other than clozapine.

TAKEAWAY:

• Overall, 401 patients received a clozapine prescription, 51% of whom were White and 40% were Black.
• Moreover, 19% of all White patients in the study received clozapine vs 10% of Black patients.
• After the researchers controlled for demographic variables, SVI scores, and care patterns, White patients were significantly more likely to receive a clozapine prescription than Black patients (adjusted odds ratio [aOR], 1.71; P < .001).
• Factors that had a statistically significant influence on the likelihood of receiving a clozapine prescription were minority status and language (OR, 2.97; P < .007), treatment duration (OR, 1.36; P < .001), and socioeconomic status (OR, 0.27; P = .001).

IN PRACTICE:

“The reasons for the underprescription of clozapine among Black patients with schizophrenia are multifactorial and may include concerns about benign ethnic neutropenia, prescriber bias, prescribers’ anticipation of patients’ nonadherence to the treatment, and the notion that the medication is less effective for Black patients,” the authors wrote.

SOURCE:

Xiaoming Zeng, MD, PhD, professor of psychiatry, University of North Carolina, Chapel Hill, North Carolina, was the senior and corresponding on the study. It was published online on March 19 in Psychiatric Services.

LIMITATIONS:

Due to the study’s cross-sectional and single-site design, the findings may not be generalizable to other geographic areas or institutions. The study lacked information on substance use disorders, common health conditions, or other patient-level data. A question remains whether all patients who received clozapine actually had treatment-resistant schizophrenia because other research has shown that there is an overdiagnosis of schizophrenia among Black patients.

DISCLOSURES:

The study was supported by a grant from the Foundation of Hope. Dr. Zeng reported no relevant financial relationships. The other authors’ disclosures are listed on the original paper.

A version of this article appeared on Medscape.com.

TOPLINE:

Black patients with schizophrenia are less likely to receive a clozapine prescription compared with White patients, a new study shows. The findings held even after the researchers controlled for demographic variables, social determinants of health, and care access patterns.

METHODOLOGY:

• The study drew on structured electronic health record data on 3160 adult patients with schizophrenia.
• The mean age at first recorded diagnosis was 39.5 years; 70% of participants were male, 53% Black, and 91% resided in an urban setting.
• The researchers used the social vulnerability index (SVI) to quantify social determinants of health.
• Descriptive data analysis, logistic regression, and sensitivity analysis were used to identify differences between those who received a clozapine prescription and those who were prescribed antipsychotic medications other than clozapine.

TAKEAWAY:

• Overall, 401 patients received a clozapine prescription, 51% of whom were White and 40% were Black.
• Moreover, 19% of all White patients in the study received clozapine vs 10% of Black patients.
• After the researchers controlled for demographic variables, SVI scores, and care patterns, White patients were significantly more likely to receive a clozapine prescription than Black patients (adjusted odds ratio [aOR], 1.71; P < .001).
• Factors that had a statistically significant influence on the likelihood of receiving a clozapine prescription were minority status and language (OR, 2.97; P < .007), treatment duration (OR, 1.36; P < .001), and socioeconomic status (OR, 0.27; P = .001).

IN PRACTICE:

“The reasons for the underprescription of clozapine among Black patients with schizophrenia are multifactorial and may include concerns about benign ethnic neutropenia, prescriber bias, prescribers’ anticipation of patients’ nonadherence to the treatment, and the notion that the medication is less effective for Black patients,” the authors wrote.

SOURCE:

Xiaoming Zeng, MD, PhD, professor of psychiatry, University of North Carolina, Chapel Hill, North Carolina, was the senior and corresponding on the study. It was published online on March 19 in Psychiatric Services.

LIMITATIONS:

Due to the study’s cross-sectional and single-site design, the findings may not be generalizable to other geographic areas or institutions. The study lacked information on substance use disorders, common health conditions, or other patient-level data. A question remains whether all patients who received clozapine actually had treatment-resistant schizophrenia because other research has shown that there is an overdiagnosis of schizophrenia among Black patients.

DISCLOSURES:

The study was supported by a grant from the Foundation of Hope. Dr. Zeng reported no relevant financial relationships. The other authors’ disclosures are listed on the original paper.

A version of this article appeared on Medscape.com.

TOPLINE:

Black patients with schizophrenia are less likely to receive a clozapine prescription compared with White patients, a new study shows. The findings held even after the researchers controlled for demographic variables, social determinants of health, and care access patterns.

METHODOLOGY:

• The study drew on structured electronic health record data on 3160 adult patients with schizophrenia.
• The mean age at first recorded diagnosis was 39.5 years; 70% of participants were male, 53% Black, and 91% resided in an urban setting.
• The researchers used the social vulnerability index (SVI) to quantify social determinants of health.
• Descriptive data analysis, logistic regression, and sensitivity analysis were used to identify differences between those who received a clozapine prescription and those who were prescribed antipsychotic medications other than clozapine.

TAKEAWAY:

• Overall, 401 patients received a clozapine prescription, 51% of whom were White and 40% were Black.
• Moreover, 19% of all White patients in the study received clozapine vs 10% of Black patients.
• After the researchers controlled for demographic variables, SVI scores, and care patterns, White patients were significantly more likely to receive a clozapine prescription than Black patients (adjusted odds ratio [aOR], 1.71; P < .001).
• Factors that had a statistically significant influence on the likelihood of receiving a clozapine prescription were minority status and language (OR, 2.97; P < .007), treatment duration (OR, 1.36; P < .001), and socioeconomic status (OR, 0.27; P = .001).

IN PRACTICE:

“The reasons for the underprescription of clozapine among Black patients with schizophrenia are multifactorial and may include concerns about benign ethnic neutropenia, prescriber bias, prescribers’ anticipation of patients’ nonadherence to the treatment, and the notion that the medication is less effective for Black patients,” the authors wrote.

SOURCE:

Xiaoming Zeng, MD, PhD, professor of psychiatry, University of North Carolina, Chapel Hill, North Carolina, was the senior and corresponding on the study. It was published online on March 19 in Psychiatric Services.

LIMITATIONS:

Due to the study’s cross-sectional and single-site design, the findings may not be generalizable to other geographic areas or institutions. The study lacked information on substance use disorders, common health conditions, or other patient-level data. A question remains whether all patients who received clozapine actually had treatment-resistant schizophrenia because other research has shown that there is an overdiagnosis of schizophrenia among Black patients.

DISCLOSURES:

The study was supported by a grant from the Foundation of Hope. Dr. Zeng reported no relevant financial relationships. The other authors’ disclosures are listed on the original paper.

A version of this article appeared on Medscape.com.

TOPLINE:

Diet has only a marginal impact on microbiome development in infancy, although metabolite profiles differ between breast- and formula-fed infants; circadian rhythm of the gut microbiome is detectable as early as 2 weeks after birth.

METHODOLOGY:

• A randomized, controlled interventional trial compared microbiota development in 210 newborns who were exclusively breastfed or received one of four formulas: Un-supplemented formula, Bifidobacterium-supplemented formula, galacto-oligosaccharide (GOS)-supplemented, or formula containing GOSs and bifidobacteria. Exclusively breastfed infants served as a reference group to evaluate the impact of infant formula feeding.
• Researchers tracked the infants’ microbiota and metabolite profiles in response to the different feeding modes via stool samples collected periodically during the first 1-2 years of life.
• They also made note of the time of day that the stool sample was collected to assess 24-hour oscillations of the microbiome in relation to dietary exposure.

TAKEAWAY:

• Global microbiota assembly of infants is primarily affected by age and less so by diet. All infants showed a gradual increase in gut microbe diversity, and at 24 months, there was no observable difference between the groups.
• However, gut metabolite profiles differed significantly between exclusively formula-fed and exclusively breastfed infants. None of the supplemented formulas were able to fully recreate the breast milk-related microbial environment.
• GOS-supplemented formula was more effective at promoting sustained levels of bifidobacteria than formula containing bifidobacteria.
• Metabolic and bacterial profiling revealed 24-hour fluctuations and circadian networks as early as 2 weeks after birth. Infant microbes maintained circadian rhythms when grown in continuous culture, even in the absence of external light or host cues, suggesting an intrinsic clock mechanism in bacteria.

IN PRACTICE:

“Our findings warrant the need for further analysis of circadian fluctuations of both bacteria and metabolites and their functional role in contributing to the benefits of infant nutrition,” the study authors wrote.

SOURCE:

The study was published online April 2 in Cell Host & Microbe.

LIMITATIONS:

The group size for exclusively formula-fed infants was limited, and the explicit contribution of breast milk, relative to infant formula, to bacterial rhythms remains unclear. A possible limitation of the circadian analysis is that the number of fecal samples collected during the night was lower than during the daytime and decreased with age.

DISCLOSURES:

This research was supported by Töpfer GmbH, the German Research Foundation, the Joint Programming Initiative of the European Union, and the German Ministry of Education and Research. The authors had disclosed no relevant conflicts of interest.

A version of this article appeared on Medscape.com.

TOPLINE:

Diet has only a marginal impact on microbiome development in infancy, although metabolite profiles differ between breast- and formula-fed infants; circadian rhythm of the gut microbiome is detectable as early as 2 weeks after birth.

METHODOLOGY:

• A randomized, controlled interventional trial compared microbiota development in 210 newborns who were exclusively breastfed or received one of four formulas: Un-supplemented formula, Bifidobacterium-supplemented formula, galacto-oligosaccharide (GOS)-supplemented, or formula containing GOSs and bifidobacteria. Exclusively breastfed infants served as a reference group to evaluate the impact of infant formula feeding.
• Researchers tracked the infants’ microbiota and metabolite profiles in response to the different feeding modes via stool samples collected periodically during the first 1-2 years of life.
• They also made note of the time of day that the stool sample was collected to assess 24-hour oscillations of the microbiome in relation to dietary exposure.

TAKEAWAY:

• Global microbiota assembly of infants is primarily affected by age and less so by diet. All infants showed a gradual increase in gut microbe diversity, and at 24 months, there was no observable difference between the groups.
• However, gut metabolite profiles differed significantly between exclusively formula-fed and exclusively breastfed infants. None of the supplemented formulas were able to fully recreate the breast milk-related microbial environment.
• GOS-supplemented formula was more effective at promoting sustained levels of bifidobacteria than formula containing bifidobacteria.
• Metabolic and bacterial profiling revealed 24-hour fluctuations and circadian networks as early as 2 weeks after birth. Infant microbes maintained circadian rhythms when grown in continuous culture, even in the absence of external light or host cues, suggesting an intrinsic clock mechanism in bacteria.

IN PRACTICE:

“Our findings warrant the need for further analysis of circadian fluctuations of both bacteria and metabolites and their functional role in contributing to the benefits of infant nutrition,” the study authors wrote.

SOURCE:

The study was published online April 2 in Cell Host & Microbe.

LIMITATIONS:

The group size for exclusively formula-fed infants was limited, and the explicit contribution of breast milk, relative to infant formula, to bacterial rhythms remains unclear. A possible limitation of the circadian analysis is that the number of fecal samples collected during the night was lower than during the daytime and decreased with age.

DISCLOSURES:

This research was supported by Töpfer GmbH, the German Research Foundation, the Joint Programming Initiative of the European Union, and the German Ministry of Education and Research. The authors had disclosed no relevant conflicts of interest.

A version of this article appeared on Medscape.com.

TOPLINE:

Diet has only a marginal impact on microbiome development in infancy, although metabolite profiles differ between breast- and formula-fed infants; circadian rhythm of the gut microbiome is detectable as early as 2 weeks after birth.

METHODOLOGY:

• A randomized, controlled interventional trial compared microbiota development in 210 newborns who were exclusively breastfed or received one of four formulas: Un-supplemented formula, Bifidobacterium-supplemented formula, galacto-oligosaccharide (GOS)-supplemented, or formula containing GOSs and bifidobacteria. Exclusively breastfed infants served as a reference group to evaluate the impact of infant formula feeding.
• Researchers tracked the infants’ microbiota and metabolite profiles in response to the different feeding modes via stool samples collected periodically during the first 1-2 years of life.
• They also made note of the time of day that the stool sample was collected to assess 24-hour oscillations of the microbiome in relation to dietary exposure.

TAKEAWAY:

• Global microbiota assembly of infants is primarily affected by age and less so by diet. All infants showed a gradual increase in gut microbe diversity, and at 24 months, there was no observable difference between the groups.
• However, gut metabolite profiles differed significantly between exclusively formula-fed and exclusively breastfed infants. None of the supplemented formulas were able to fully recreate the breast milk-related microbial environment.
• GOS-supplemented formula was more effective at promoting sustained levels of bifidobacteria than formula containing bifidobacteria.
• Metabolic and bacterial profiling revealed 24-hour fluctuations and circadian networks as early as 2 weeks after birth. Infant microbes maintained circadian rhythms when grown in continuous culture, even in the absence of external light or host cues, suggesting an intrinsic clock mechanism in bacteria.

IN PRACTICE:

“Our findings warrant the need for further analysis of circadian fluctuations of both bacteria and metabolites and their functional role in contributing to the benefits of infant nutrition,” the study authors wrote.

SOURCE:

The study was published online April 2 in Cell Host & Microbe.

LIMITATIONS:

The group size for exclusively formula-fed infants was limited, and the explicit contribution of breast milk, relative to infant formula, to bacterial rhythms remains unclear. A possible limitation of the circadian analysis is that the number of fecal samples collected during the night was lower than during the daytime and decreased with age.

DISCLOSURES:

This research was supported by Töpfer GmbH, the German Research Foundation, the Joint Programming Initiative of the European Union, and the German Ministry of Education and Research. The authors had disclosed no relevant conflicts of interest.

A version of this article appeared on Medscape.com.

Ultrasound for assessing lymph nodal involvement may be substituted for sentinel lymph node biopsy with no change in outcomes in patients with early breast cancer, a new study finds.

This was the conclusion of research on the agenda at the American Society of Breast Surgeons annual meeting.

Sentinel lymph node biopsy (SLNB) is the standard of care for individuals with early-stage HR+HER2- breast cancer to assess nodal involvement, but SLNB can bring complications including postoperative arm problems and lasting lymphedema, according to Andreas Giannakou, MD, of Brigham and Women’s Hospital and the Dana-Farber Cancer Institute, Boston, the presenter of this new research.

The SOUND (Sentinel Node vs. Observation After Axillary Ultra-Sound) trial, published in JAMA Oncology in 2023, showed that ultrasound nodal imaging was a safe and effective alternative to SLNB in certain patients with early-stage breast cancers, but real-world validation was needed, Dr. Giannakou said during a press briefing in advance of the meeting.

Why Was the SOUND Trial Important?

The SOUND trial randomized 1,463 individuals with early stage (cT1NO) breast cancer (tumors less than 2 cm) and negative findings on axillary ultrasound to either SLNB or no axillary surgical staging.

The 5-year rate of distant disease-free survival was 97.7% in the SLNB group vs. 98% in the no axillary surgery group, suggesting that omission of staging was noninferior to SLNB in these patients and a safe and effective option.

In current practice, nodal status remains a key factor in decision-making for adjuvant systemic therapy in premenopausal patients and in patients with HER2+ and triple-negative breast cancer, Dr. Giannakou said during the press briefing.

“The SOUND trial is a potentially practice-changing study that can spare a specific patient population from axillary surgical staging,” Dr. Giannakou said in an interview. “Before broadly applying clinical trial results to practice, it is important to ensure that the trial population is representative of the population being treated in real world practice,” he said.

What Did the New Study Show? 

In the new study, the researchers identified 312 patients meeting the SOUND trial eligibility criteria in a large database from a single center, and compared disease characteristics and outcomes with the 708 patients in the SLNB arm of the SOUND trial.

The researchers found a similarly high rate of negative SLNB results and very low recurrence in the study population. Notably, only 11.3% of the patients in the current study and 13.1% of patients in the SOUND trial had 1-3 positive lymph nodes, and less than 1% of patients in both cohorts had 4 or more positive nodes, Dr. Giannakou said.

The population of the current study was similar to that of the SOUND trial population with respect to treatment characteristics and nodal disease burden,” Dr. Giannakou said during the interview. These findings suggest that omission of sentinel lymph node in the new study cohort would have also likely been oncologically safe.

“These results are confirmatory but not surprising,” he said. Previous studies have shown that the sensitivity and accuracy of axillary ultrasound is comparable to the sentinel lymph node biopsy in patients with early breast cancer and only one abnormal lymph node on the ultrasound. 
 

What Are the Clinical Implications?

The current study findings make an important contribution to the effort to de-escalate axillary surgery in early breast cancer, Dr. Giannakou said during the interview. Although SLNB is less morbid than axillary lymph node dissection, the lymphedema risk still exists, and identifying which patients actually benefit from SLNB is critical, he said.

“In our multidisciplinary team, we are working to define selection criteria for postmenopausal patients with HR+HER2- breast cancer who would have met eligibility criteria for the SOUND trial and for whom omission of SLNB would not change adjuvant treatment considerations,” he said.

“Breast surgeons have been moving towards less aggressive axillary surgery based on evidence showing its safety in specific patient cohorts, particularly those with low-risk factors such as older age (70 years and above) and early-stage hormone receptor-positive breast cancer,” Sarah Blair, MD, professor and vice chair in the department of surgery at UC San Diego Health, said in an interview.

“The Choosing Wisely recommendations, issued by the Society of Surgical Oncology, advise against routine use of sentinel lymph node biopsy in women aged 70 and older with early-stage hormone receptor–positive breast cancer; these recommendations are based on clinical trials demonstrating oncologic safety in this population,” said Dr. Blair, who was not involved in the SOUND trial or the current study.

The data from the new study are encouraging and highlight the generalizability of the SOUND results, Mediget Teshome, MD, chief of breast surgery at UCLA Health, said in an interview. The results help to define a low-risk group of patients for which sentinel node staging may be omitted, after multidisciplinary discussion to ensure that nodal staging will not impact adjuvant systemic therapy or radiation decision-making, said Dr. Teshome, who was not involved in the SOUND trial or the current study.
 

What Are the Limitations of the SOUND trial and the New Study?

The current study limitations included its design having been a retrospective review of a prospective database with selection bias, lack of standard criteria for preoperative axillary ultrasound, and the lack of SLNB for many patients older than 70 years based on the Choosing Wisely criteria, Dr. Giannakou said in the press briefing.

“Despite the evidence supporting axillary surgery de-escalation, it can be challenging for surgeons to change their practice based on a single study,” Dr. Blair said an interview. However, the SOUND trial findings support current evidence, giving surgeons more confidence to discuss multidisciplinary treatment options, she said.
 

What Additional Research is Needed?

“Longer follow-up is needed to make definitive conclusions about the oncologic outcomes of axillary surgery de-escalation in this patient population,” said Dr. Blair. “Given that slow-growing tumors are involved, the time to recurrence may extend beyond the typical follow-up period of three years.

“Ongoing research and collaboration among multidisciplinary teams are essential to ensure optimal treatment decisions and patient outcomes,” she emphasized.

Dr. Giannakou, Dr. Blair, and Dr. Teshome had no financial conflicts to disclose.

Ultrasound for assessing lymph nodal involvement may be substituted for sentinel lymph node biopsy with no change in outcomes in patients with early breast cancer, a new study finds.

This was the conclusion of research on the agenda at the American Society of Breast Surgeons annual meeting.

Sentinel lymph node biopsy (SLNB) is the standard of care for individuals with early-stage HR+HER2- breast cancer to assess nodal involvement, but SLNB can bring complications including postoperative arm problems and lasting lymphedema, according to Andreas Giannakou, MD, of Brigham and Women’s Hospital and the Dana-Farber Cancer Institute, Boston, the presenter of this new research.

The SOUND (Sentinel Node vs. Observation After Axillary Ultra-Sound) trial, published in JAMA Oncology in 2023, showed that ultrasound nodal imaging was a safe and effective alternative to SLNB in certain patients with early-stage breast cancers, but real-world validation was needed, Dr. Giannakou said during a press briefing in advance of the meeting.

Why Was the SOUND Trial Important?

The SOUND trial randomized 1,463 individuals with early stage (cT1NO) breast cancer (tumors less than 2 cm) and negative findings on axillary ultrasound to either SLNB or no axillary surgical staging.

The 5-year rate of distant disease-free survival was 97.7% in the SLNB group vs. 98% in the no axillary surgery group, suggesting that omission of staging was noninferior to SLNB in these patients and a safe and effective option.

In current practice, nodal status remains a key factor in decision-making for adjuvant systemic therapy in premenopausal patients and in patients with HER2+ and triple-negative breast cancer, Dr. Giannakou said during the press briefing.

“The SOUND trial is a potentially practice-changing study that can spare a specific patient population from axillary surgical staging,” Dr. Giannakou said in an interview. “Before broadly applying clinical trial results to practice, it is important to ensure that the trial population is representative of the population being treated in real world practice,” he said.

What Did the New Study Show? 

In the new study, the researchers identified 312 patients meeting the SOUND trial eligibility criteria in a large database from a single center, and compared disease characteristics and outcomes with the 708 patients in the SLNB arm of the SOUND trial.

The researchers found a similarly high rate of negative SLNB results and very low recurrence in the study population. Notably, only 11.3% of the patients in the current study and 13.1% of patients in the SOUND trial had 1-3 positive lymph nodes, and less than 1% of patients in both cohorts had 4 or more positive nodes, Dr. Giannakou said.

The population of the current study was similar to that of the SOUND trial population with respect to treatment characteristics and nodal disease burden,” Dr. Giannakou said during the interview. These findings suggest that omission of sentinel lymph node in the new study cohort would have also likely been oncologically safe.

“These results are confirmatory but not surprising,” he said. Previous studies have shown that the sensitivity and accuracy of axillary ultrasound is comparable to the sentinel lymph node biopsy in patients with early breast cancer and only one abnormal lymph node on the ultrasound. 
 

What Are the Clinical Implications?

The current study findings make an important contribution to the effort to de-escalate axillary surgery in early breast cancer, Dr. Giannakou said during the interview. Although SLNB is less morbid than axillary lymph node dissection, the lymphedema risk still exists, and identifying which patients actually benefit from SLNB is critical, he said.

“In our multidisciplinary team, we are working to define selection criteria for postmenopausal patients with HR+HER2- breast cancer who would have met eligibility criteria for the SOUND trial and for whom omission of SLNB would not change adjuvant treatment considerations,” he said.

“Breast surgeons have been moving towards less aggressive axillary surgery based on evidence showing its safety in specific patient cohorts, particularly those with low-risk factors such as older age (70 years and above) and early-stage hormone receptor-positive breast cancer,” Sarah Blair, MD, professor and vice chair in the department of surgery at UC San Diego Health, said in an interview.

“The Choosing Wisely recommendations, issued by the Society of Surgical Oncology, advise against routine use of sentinel lymph node biopsy in women aged 70 and older with early-stage hormone receptor–positive breast cancer; these recommendations are based on clinical trials demonstrating oncologic safety in this population,” said Dr. Blair, who was not involved in the SOUND trial or the current study.

The data from the new study are encouraging and highlight the generalizability of the SOUND results, Mediget Teshome, MD, chief of breast surgery at UCLA Health, said in an interview. The results help to define a low-risk group of patients for which sentinel node staging may be omitted, after multidisciplinary discussion to ensure that nodal staging will not impact adjuvant systemic therapy or radiation decision-making, said Dr. Teshome, who was not involved in the SOUND trial or the current study.
 

What Are the Limitations of the SOUND trial and the New Study?

The current study limitations included its design having been a retrospective review of a prospective database with selection bias, lack of standard criteria for preoperative axillary ultrasound, and the lack of SLNB for many patients older than 70 years based on the Choosing Wisely criteria, Dr. Giannakou said in the press briefing.

“Despite the evidence supporting axillary surgery de-escalation, it can be challenging for surgeons to change their practice based on a single study,” Dr. Blair said an interview. However, the SOUND trial findings support current evidence, giving surgeons more confidence to discuss multidisciplinary treatment options, she said.
 

What Additional Research is Needed?

“Longer follow-up is needed to make definitive conclusions about the oncologic outcomes of axillary surgery de-escalation in this patient population,” said Dr. Blair. “Given that slow-growing tumors are involved, the time to recurrence may extend beyond the typical follow-up period of three years.

“Ongoing research and collaboration among multidisciplinary teams are essential to ensure optimal treatment decisions and patient outcomes,” she emphasized.

Dr. Giannakou, Dr. Blair, and Dr. Teshome had no financial conflicts to disclose.

Ultrasound for assessing lymph nodal involvement may be substituted for sentinel lymph node biopsy with no change in outcomes in patients with early breast cancer, a new study finds.

This was the conclusion of research on the agenda at the American Society of Breast Surgeons annual meeting.

Sentinel lymph node biopsy (SLNB) is the standard of care for individuals with early-stage HR+HER2- breast cancer to assess nodal involvement, but SLNB can bring complications including postoperative arm problems and lasting lymphedema, according to Andreas Giannakou, MD, of Brigham and Women’s Hospital and the Dana-Farber Cancer Institute, Boston, the presenter of this new research.

The SOUND (Sentinel Node vs. Observation After Axillary Ultra-Sound) trial, published in JAMA Oncology in 2023, showed that ultrasound nodal imaging was a safe and effective alternative to SLNB in certain patients with early-stage breast cancers, but real-world validation was needed, Dr. Giannakou said during a press briefing in advance of the meeting.

Why Was the SOUND Trial Important?

The SOUND trial randomized 1,463 individuals with early stage (cT1NO) breast cancer (tumors less than 2 cm) and negative findings on axillary ultrasound to either SLNB or no axillary surgical staging.

The 5-year rate of distant disease-free survival was 97.7% in the SLNB group vs. 98% in the no axillary surgery group, suggesting that omission of staging was noninferior to SLNB in these patients and a safe and effective option.

In current practice, nodal status remains a key factor in decision-making for adjuvant systemic therapy in premenopausal patients and in patients with HER2+ and triple-negative breast cancer, Dr. Giannakou said during the press briefing.

“The SOUND trial is a potentially practice-changing study that can spare a specific patient population from axillary surgical staging,” Dr. Giannakou said in an interview. “Before broadly applying clinical trial results to practice, it is important to ensure that the trial population is representative of the population being treated in real world practice,” he said.

What Did the New Study Show? 

In the new study, the researchers identified 312 patients meeting the SOUND trial eligibility criteria in a large database from a single center, and compared disease characteristics and outcomes with the 708 patients in the SLNB arm of the SOUND trial.

The researchers found a similarly high rate of negative SLNB results and very low recurrence in the study population. Notably, only 11.3% of the patients in the current study and 13.1% of patients in the SOUND trial had 1-3 positive lymph nodes, and less than 1% of patients in both cohorts had 4 or more positive nodes, Dr. Giannakou said.

The population of the current study was similar to that of the SOUND trial population with respect to treatment characteristics and nodal disease burden,” Dr. Giannakou said during the interview. These findings suggest that omission of sentinel lymph node in the new study cohort would have also likely been oncologically safe.

“These results are confirmatory but not surprising,” he said. Previous studies have shown that the sensitivity and accuracy of axillary ultrasound is comparable to the sentinel lymph node biopsy in patients with early breast cancer and only one abnormal lymph node on the ultrasound. 
 

What Are the Clinical Implications?

The current study findings make an important contribution to the effort to de-escalate axillary surgery in early breast cancer, Dr. Giannakou said during the interview. Although SLNB is less morbid than axillary lymph node dissection, the lymphedema risk still exists, and identifying which patients actually benefit from SLNB is critical, he said.

“In our multidisciplinary team, we are working to define selection criteria for postmenopausal patients with HR+HER2- breast cancer who would have met eligibility criteria for the SOUND trial and for whom omission of SLNB would not change adjuvant treatment considerations,” he said.

“Breast surgeons have been moving towards less aggressive axillary surgery based on evidence showing its safety in specific patient cohorts, particularly those with low-risk factors such as older age (70 years and above) and early-stage hormone receptor-positive breast cancer,” Sarah Blair, MD, professor and vice chair in the department of surgery at UC San Diego Health, said in an interview.

“The Choosing Wisely recommendations, issued by the Society of Surgical Oncology, advise against routine use of sentinel lymph node biopsy in women aged 70 and older with early-stage hormone receptor–positive breast cancer; these recommendations are based on clinical trials demonstrating oncologic safety in this population,” said Dr. Blair, who was not involved in the SOUND trial or the current study.

The data from the new study are encouraging and highlight the generalizability of the SOUND results, Mediget Teshome, MD, chief of breast surgery at UCLA Health, said in an interview. The results help to define a low-risk group of patients for which sentinel node staging may be omitted, after multidisciplinary discussion to ensure that nodal staging will not impact adjuvant systemic therapy or radiation decision-making, said Dr. Teshome, who was not involved in the SOUND trial or the current study.
 

What Are the Limitations of the SOUND trial and the New Study?

The current study limitations included its design having been a retrospective review of a prospective database with selection bias, lack of standard criteria for preoperative axillary ultrasound, and the lack of SLNB for many patients older than 70 years based on the Choosing Wisely criteria, Dr. Giannakou said in the press briefing.

“Despite the evidence supporting axillary surgery de-escalation, it can be challenging for surgeons to change their practice based on a single study,” Dr. Blair said an interview. However, the SOUND trial findings support current evidence, giving surgeons more confidence to discuss multidisciplinary treatment options, she said.
 

What Additional Research is Needed?

“Longer follow-up is needed to make definitive conclusions about the oncologic outcomes of axillary surgery de-escalation in this patient population,” said Dr. Blair. “Given that slow-growing tumors are involved, the time to recurrence may extend beyond the typical follow-up period of three years.

“Ongoing research and collaboration among multidisciplinary teams are essential to ensure optimal treatment decisions and patient outcomes,” she emphasized.

Dr. Giannakou, Dr. Blair, and Dr. Teshome had no financial conflicts to disclose.

Safety concerns persist for out-of-hospital births in the United States with multiple potential risk factors and few safety requirements, according to a paper published in the American Journal of Obstetrics and Gynecology.

In 2022, the Centers for Disease Control and Prevention (CDC) reported the highest number of planned home births in 30 years. The numbers rose 12% from 2020 to 2021, the latest period for which complete data are available. Home births rose from 45,646 (1.26% of births) in 2020 to 51,642 (1.41% of births).

Amos Grünebaum, MD, and Frank A. Chervenak, MD, with Northwell Health, and the Department of Obstetrics and Gynecology, Lenox Hill Hospital, Zucker School of Medicine in New Hyde Park, New York, reviewed the latest safety data surrounding community births in the United States along with well-known perinatal risks and safety requirements for safe out-of-hospital births.

“Most planned home births continue to have one or more risk factors that are associated with an increase in adverse pregnancy outcomes,” they wrote.
 

Birth Certificate Data Analyzed

The researchers used the CDC birth certificate database and analyzed deliveries between 2016 and 2022 regarding the incidence of perinatal risks in community births. The risks included were prior cesarean, first baby, mother older than 35 years, twins, breech presentation, gestational age of less than 37 weeks or more than 41 weeks, newborn weight over 4,000 grams, adequacy of prenatal care, grand multiparity (5 or more prior pregnancies), and a prepregnancy body mass index of at least 35.

The incidence of perinatal risks for out-of-hospital births ranged individually from 0.2% to 28.54% among birthing center births and 0.32% to 24.4% for planned home births.

“The ACOG committee opinion on home births states that for every 1000 home births, 3.9 babies will die,” the authors noted, or about twice the risk of hospital births. The deaths are “potentially avoidable with easy access to an operating room,” they wrote.

Among the safety concerns for perinatal morbidity and mortality in community births, the authors cited the lack of:

• Appropriate patient selection for out-of-hospital births through standardized guidelines.
• Availability of a Certified Nurse Midwife, a Certified Midwife, or midwife whose education and licensure meet International Confederation of Midwives’ (ICM) Global Standards for Midwifery Education.
• Providers practicing obstetrics within an integrated and regulated health system with ready access and availability of board-certified obstetricians to provide consultation for qualified midwives.
• Standardized guidelines on when transport to a hospital is necessary.

“While prerequisites for a safe out-of-hospital delivery may be in place in other high-income countries, these prerequisites have not been actualized in the United States,” the authors wrote.

Incorporating Patient Preferences Into Delivery Models

Yalda Afshar, MD, PhD, maternal-fetal medicine subspecialist and a physician-scientist at UCLA Health in California, said obstetricians are responsible for offering the most evidence-based care to pregnant people.

“What this birth certificate data demonstrates,” she said, “is a tendency among birthing people to opt for out-of-hospital births, despite documented risks to both the pregnant person and the neonate. This underscores the need to persist in educating on risk stratification, risk reduction, and safe birthing practices, while also fostering innovation. Innovation should stem from our commitment to incorporate the preferences of pregnant people into our healthcare delivery model.”

Dr. Afshar, who was not part of the study, said clinicians should develop innovative ways to effectively meet the needs of pregnant patients while ensuring their safety and well-being.

“Ideally, we would establish safe environments within hospital systems and centers that emulate home-like birthing experiences, thereby mitigating risks for these families,” she said.

Though not explicitly stated in the data, she added, it is crucial to emphasize the need for continuous risk assessment throughout pregnancy and childbirth, “with a paramount focus on the safety of the pregnant individual.”

The authors and Dr. Afshar have no relevant financial disclosures.

Safety concerns persist for out-of-hospital births in the United States with multiple potential risk factors and few safety requirements, according to a paper published in the American Journal of Obstetrics and Gynecology.

In 2022, the Centers for Disease Control and Prevention (CDC) reported the highest number of planned home births in 30 years. The numbers rose 12% from 2020 to 2021, the latest period for which complete data are available. Home births rose from 45,646 (1.26% of births) in 2020 to 51,642 (1.41% of births).

Amos Grünebaum, MD, and Frank A. Chervenak, MD, with Northwell Health, and the Department of Obstetrics and Gynecology, Lenox Hill Hospital, Zucker School of Medicine in New Hyde Park, New York, reviewed the latest safety data surrounding community births in the United States along with well-known perinatal risks and safety requirements for safe out-of-hospital births.

“Most planned home births continue to have one or more risk factors that are associated with an increase in adverse pregnancy outcomes,” they wrote.
 

Birth Certificate Data Analyzed

The researchers used the CDC birth certificate database and analyzed deliveries between 2016 and 2022 regarding the incidence of perinatal risks in community births. The risks included were prior cesarean, first baby, mother older than 35 years, twins, breech presentation, gestational age of less than 37 weeks or more than 41 weeks, newborn weight over 4,000 grams, adequacy of prenatal care, grand multiparity (5 or more prior pregnancies), and a prepregnancy body mass index of at least 35.

The incidence of perinatal risks for out-of-hospital births ranged individually from 0.2% to 28.54% among birthing center births and 0.32% to 24.4% for planned home births.

“The ACOG committee opinion on home births states that for every 1000 home births, 3.9 babies will die,” the authors noted, or about twice the risk of hospital births. The deaths are “potentially avoidable with easy access to an operating room,” they wrote.

Among the safety concerns for perinatal morbidity and mortality in community births, the authors cited the lack of:

• Appropriate patient selection for out-of-hospital births through standardized guidelines.
• Availability of a Certified Nurse Midwife, a Certified Midwife, or midwife whose education and licensure meet International Confederation of Midwives’ (ICM) Global Standards for Midwifery Education.
• Providers practicing obstetrics within an integrated and regulated health system with ready access and availability of board-certified obstetricians to provide consultation for qualified midwives.
• Standardized guidelines on when transport to a hospital is necessary.

“While prerequisites for a safe out-of-hospital delivery may be in place in other high-income countries, these prerequisites have not been actualized in the United States,” the authors wrote.

Incorporating Patient Preferences Into Delivery Models

Yalda Afshar, MD, PhD, maternal-fetal medicine subspecialist and a physician-scientist at UCLA Health in California, said obstetricians are responsible for offering the most evidence-based care to pregnant people.

“What this birth certificate data demonstrates,” she said, “is a tendency among birthing people to opt for out-of-hospital births, despite documented risks to both the pregnant person and the neonate. This underscores the need to persist in educating on risk stratification, risk reduction, and safe birthing practices, while also fostering innovation. Innovation should stem from our commitment to incorporate the preferences of pregnant people into our healthcare delivery model.”

Dr. Afshar, who was not part of the study, said clinicians should develop innovative ways to effectively meet the needs of pregnant patients while ensuring their safety and well-being.

“Ideally, we would establish safe environments within hospital systems and centers that emulate home-like birthing experiences, thereby mitigating risks for these families,” she said.

Though not explicitly stated in the data, she added, it is crucial to emphasize the need for continuous risk assessment throughout pregnancy and childbirth, “with a paramount focus on the safety of the pregnant individual.”

The authors and Dr. Afshar have no relevant financial disclosures.

Safety concerns persist for out-of-hospital births in the United States with multiple potential risk factors and few safety requirements, according to a paper published in the American Journal of Obstetrics and Gynecology.

In 2022, the Centers for Disease Control and Prevention (CDC) reported the highest number of planned home births in 30 years. The numbers rose 12% from 2020 to 2021, the latest period for which complete data are available. Home births rose from 45,646 (1.26% of births) in 2020 to 51,642 (1.41% of births).

Amos Grünebaum, MD, and Frank A. Chervenak, MD, with Northwell Health, and the Department of Obstetrics and Gynecology, Lenox Hill Hospital, Zucker School of Medicine in New Hyde Park, New York, reviewed the latest safety data surrounding community births in the United States along with well-known perinatal risks and safety requirements for safe out-of-hospital births.

“Most planned home births continue to have one or more risk factors that are associated with an increase in adverse pregnancy outcomes,” they wrote.
 

Birth Certificate Data Analyzed

The researchers used the CDC birth certificate database and analyzed deliveries between 2016 and 2022 regarding the incidence of perinatal risks in community births. The risks included were prior cesarean, first baby, mother older than 35 years, twins, breech presentation, gestational age of less than 37 weeks or more than 41 weeks, newborn weight over 4,000 grams, adequacy of prenatal care, grand multiparity (5 or more prior pregnancies), and a prepregnancy body mass index of at least 35.

The incidence of perinatal risks for out-of-hospital births ranged individually from 0.2% to 28.54% among birthing center births and 0.32% to 24.4% for planned home births.

“The ACOG committee opinion on home births states that for every 1000 home births, 3.9 babies will die,” the authors noted, or about twice the risk of hospital births. The deaths are “potentially avoidable with easy access to an operating room,” they wrote.

Among the safety concerns for perinatal morbidity and mortality in community births, the authors cited the lack of:

• Appropriate patient selection for out-of-hospital births through standardized guidelines.
• Availability of a Certified Nurse Midwife, a Certified Midwife, or midwife whose education and licensure meet International Confederation of Midwives’ (ICM) Global Standards for Midwifery Education.
• Providers practicing obstetrics within an integrated and regulated health system with ready access and availability of board-certified obstetricians to provide consultation for qualified midwives.
• Standardized guidelines on when transport to a hospital is necessary.

“While prerequisites for a safe out-of-hospital delivery may be in place in other high-income countries, these prerequisites have not been actualized in the United States,” the authors wrote.

Incorporating Patient Preferences Into Delivery Models

Yalda Afshar, MD, PhD, maternal-fetal medicine subspecialist and a physician-scientist at UCLA Health in California, said obstetricians are responsible for offering the most evidence-based care to pregnant people.

“What this birth certificate data demonstrates,” she said, “is a tendency among birthing people to opt for out-of-hospital births, despite documented risks to both the pregnant person and the neonate. This underscores the need to persist in educating on risk stratification, risk reduction, and safe birthing practices, while also fostering innovation. Innovation should stem from our commitment to incorporate the preferences of pregnant people into our healthcare delivery model.”

Dr. Afshar, who was not part of the study, said clinicians should develop innovative ways to effectively meet the needs of pregnant patients while ensuring their safety and well-being.

“Ideally, we would establish safe environments within hospital systems and centers that emulate home-like birthing experiences, thereby mitigating risks for these families,” she said.

Though not explicitly stated in the data, she added, it is crucial to emphasize the need for continuous risk assessment throughout pregnancy and childbirth, “with a paramount focus on the safety of the pregnant individual.”

The authors and Dr. Afshar have no relevant financial disclosures.

TOPLINE:

A recent survey highlighted ethical concerns US oncologists have about using artificial intelligence (AI) to help make cancer treatment decisions and revealed some contradictory views about how best to integrate these tools into practice. Most respondents, for instance, said patients should not be expected to understand how AI tools work, but many also felt patients could make treatment decisions based on AI-generated recommendations. Most oncologists also felt responsible for protecting patients from biased AI, but few were confident that they could do so.

METHODOLOGY:

• The US Food and Drug Administration (FDA) has  for use in various medical specialties over the past few decades, and increasingly, AI tools are being integrated into cancer care.
• However, the uptake of these tools in oncology has raised ethical questions and concerns, including challenges with AI bias, error, or misuse, as well as issues explaining how an AI model reached a result.
• In the current study, researchers asked 204 oncologists from 37 states for their views on the ethical implications of using AI for cancer care.
• Among the survey respondents, 64% were men and 63% were non-Hispanic White; 29% were from academic practices, 47% had received some education on AI use in healthcare, and 45% were familiar with clinical decision models.
• The researchers assessed respondents’ answers to various questions, including whether to provide informed consent for AI use and how oncologists would approach a scenario where the AI model and the oncologist recommended a different treatment regimen.

TAKEAWAY:

• Overall, 81% of oncologists supported having patient consent to use an AI model during treatment decisions, and 85% felt that oncologists needed to be able to explain an AI-based clinical decision model to use it in the clinic; however, only 23% felt that patients also needed to be able to explain an AI model.
• When an AI decision model recommended a different treatment regimen than the treating oncologist, the most common response (36.8%) was to present both options to the patient and let the patient decide. Oncologists from academic settings were about 2.5 times more likely than those from other settings to let the patient decide. About 34% of respondents said they would present both options but recommend the oncologist’s regimen, whereas about 22% said they would present both but recommend the AI’s regimen. A small percentage would only present the oncologist’s regimen (5%) or the AI’s regimen (about 2.5%).
• About three of four respondents (76.5%) agreed that oncologists should protect patients from biased AI tools; however, only about one of four (27.9%) felt confident they could identify biased AI models.
• Most oncologists (91%) felt that AI developers were responsible for the medico-legal problems associated with AI use; less than half (47%) said oncologists or hospitals (43%) shared this responsibility.

IN PRACTICE:

“Together, these data characterize barriers that may impede the ethical adoption of AI into cancer care. The findings suggest that the implementation of AI in oncology must include rigorous assessments of its effect on care decisions, as well as decisional responsibility when problems related to AI use arise,” the authors concluded.

SOURCE:

The study, with first author Andrew Hantel, MD, from Dana-Farber Cancer Institute, Boston, was published last month in JAMA Network Open.

LIMITATIONS:

The study had a moderate sample size and response rate, although demographics of participating oncologists appear to be nationally representative. The cross-sectional study design limited the generalizability of the findings over time as AI is integrated into cancer care.

DISCLOSURES:

The study was funded by the National Cancer Institute, the Dana-Farber McGraw/Patterson Research Fund, and the Mark Foundation Emerging Leader Award. Dr. Hantel reported receiving personal fees from AbbVie, AstraZeneca, the American Journal of Managed Care, Genentech, and GSK.

A version of this article appeared on Medscape.com.

TOPLINE:

A recent survey highlighted ethical concerns US oncologists have about using artificial intelligence (AI) to help make cancer treatment decisions and revealed some contradictory views about how best to integrate these tools into practice. Most respondents, for instance, said patients should not be expected to understand how AI tools work, but many also felt patients could make treatment decisions based on AI-generated recommendations. Most oncologists also felt responsible for protecting patients from biased AI, but few were confident that they could do so.

METHODOLOGY:

• The US Food and Drug Administration (FDA) has  for use in various medical specialties over the past few decades, and increasingly, AI tools are being integrated into cancer care.
• However, the uptake of these tools in oncology has raised ethical questions and concerns, including challenges with AI bias, error, or misuse, as well as issues explaining how an AI model reached a result.
• In the current study, researchers asked 204 oncologists from 37 states for their views on the ethical implications of using AI for cancer care.
• Among the survey respondents, 64% were men and 63% were non-Hispanic White; 29% were from academic practices, 47% had received some education on AI use in healthcare, and 45% were familiar with clinical decision models.
• The researchers assessed respondents’ answers to various questions, including whether to provide informed consent for AI use and how oncologists would approach a scenario where the AI model and the oncologist recommended a different treatment regimen.

TAKEAWAY:

• Overall, 81% of oncologists supported having patient consent to use an AI model during treatment decisions, and 85% felt that oncologists needed to be able to explain an AI-based clinical decision model to use it in the clinic; however, only 23% felt that patients also needed to be able to explain an AI model.
• When an AI decision model recommended a different treatment regimen than the treating oncologist, the most common response (36.8%) was to present both options to the patient and let the patient decide. Oncologists from academic settings were about 2.5 times more likely than those from other settings to let the patient decide. About 34% of respondents said they would present both options but recommend the oncologist’s regimen, whereas about 22% said they would present both but recommend the AI’s regimen. A small percentage would only present the oncologist’s regimen (5%) or the AI’s regimen (about 2.5%).
• About three of four respondents (76.5%) agreed that oncologists should protect patients from biased AI tools; however, only about one of four (27.9%) felt confident they could identify biased AI models.
• Most oncologists (91%) felt that AI developers were responsible for the medico-legal problems associated with AI use; less than half (47%) said oncologists or hospitals (43%) shared this responsibility.

IN PRACTICE:

“Together, these data characterize barriers that may impede the ethical adoption of AI into cancer care. The findings suggest that the implementation of AI in oncology must include rigorous assessments of its effect on care decisions, as well as decisional responsibility when problems related to AI use arise,” the authors concluded.

SOURCE:

The study, with first author Andrew Hantel, MD, from Dana-Farber Cancer Institute, Boston, was published last month in JAMA Network Open.

LIMITATIONS:

The study had a moderate sample size and response rate, although demographics of participating oncologists appear to be nationally representative. The cross-sectional study design limited the generalizability of the findings over time as AI is integrated into cancer care.

DISCLOSURES:

The study was funded by the National Cancer Institute, the Dana-Farber McGraw/Patterson Research Fund, and the Mark Foundation Emerging Leader Award. Dr. Hantel reported receiving personal fees from AbbVie, AstraZeneca, the American Journal of Managed Care, Genentech, and GSK.

A version of this article appeared on Medscape.com.

TOPLINE:

A recent survey highlighted ethical concerns US oncologists have about using artificial intelligence (AI) to help make cancer treatment decisions and revealed some contradictory views about how best to integrate these tools into practice. Most respondents, for instance, said patients should not be expected to understand how AI tools work, but many also felt patients could make treatment decisions based on AI-generated recommendations. Most oncologists also felt responsible for protecting patients from biased AI, but few were confident that they could do so.

METHODOLOGY:

• The US Food and Drug Administration (FDA) has  for use in various medical specialties over the past few decades, and increasingly, AI tools are being integrated into cancer care.
• However, the uptake of these tools in oncology has raised ethical questions and concerns, including challenges with AI bias, error, or misuse, as well as issues explaining how an AI model reached a result.
• In the current study, researchers asked 204 oncologists from 37 states for their views on the ethical implications of using AI for cancer care.
• Among the survey respondents, 64% were men and 63% were non-Hispanic White; 29% were from academic practices, 47% had received some education on AI use in healthcare, and 45% were familiar with clinical decision models.
• The researchers assessed respondents’ answers to various questions, including whether to provide informed consent for AI use and how oncologists would approach a scenario where the AI model and the oncologist recommended a different treatment regimen.

TAKEAWAY:

• Overall, 81% of oncologists supported having patient consent to use an AI model during treatment decisions, and 85% felt that oncologists needed to be able to explain an AI-based clinical decision model to use it in the clinic; however, only 23% felt that patients also needed to be able to explain an AI model.
• When an AI decision model recommended a different treatment regimen than the treating oncologist, the most common response (36.8%) was to present both options to the patient and let the patient decide. Oncologists from academic settings were about 2.5 times more likely than those from other settings to let the patient decide. About 34% of respondents said they would present both options but recommend the oncologist’s regimen, whereas about 22% said they would present both but recommend the AI’s regimen. A small percentage would only present the oncologist’s regimen (5%) or the AI’s regimen (about 2.5%).
• About three of four respondents (76.5%) agreed that oncologists should protect patients from biased AI tools; however, only about one of four (27.9%) felt confident they could identify biased AI models.
• Most oncologists (91%) felt that AI developers were responsible for the medico-legal problems associated with AI use; less than half (47%) said oncologists or hospitals (43%) shared this responsibility.

IN PRACTICE:

“Together, these data characterize barriers that may impede the ethical adoption of AI into cancer care. The findings suggest that the implementation of AI in oncology must include rigorous assessments of its effect on care decisions, as well as decisional responsibility when problems related to AI use arise,” the authors concluded.

SOURCE:

The study, with first author Andrew Hantel, MD, from Dana-Farber Cancer Institute, Boston, was published last month in JAMA Network Open.

LIMITATIONS:

The study had a moderate sample size and response rate, although demographics of participating oncologists appear to be nationally representative. The cross-sectional study design limited the generalizability of the findings over time as AI is integrated into cancer care.

DISCLOSURES:

The study was funded by the National Cancer Institute, the Dana-Farber McGraw/Patterson Research Fund, and the Mark Foundation Emerging Leader Award. Dr. Hantel reported receiving personal fees from AbbVie, AstraZeneca, the American Journal of Managed Care, Genentech, and GSK.

A version of this article appeared on Medscape.com.