Targeted CT follow-up proposed for incidental lung nodules

CHICAGO – Limiting CT follow-up of incidental pulmonary nodules to a small area around the suspicious nodule may help resolve a common clinical dilemma and an Achilles heel of lung cancer screening, researchers suggest.

It could also reduce the radiation dose by more than 80%.

"Why should we radiate the entire thorax on follow-up?" Dr. Gregory D. Pearson asked rhetorically at the annual meeting of the Radiological Society of North America.

In the pivotal National Lung Cancer Screening Trial, annual low-dose CT reduced lung cancer deaths by 20% and all-cause mortality by 6.7% among heavy smokers, compared with traditional X-rays. Approximately 40% of the CT group, however, had findings of small, indeterminate pulmonary nodules considered suspicious for lung cancer on at least one scan, with 1.4% experiencing a complication as a result of additional testing (N. Engl. J. Med. 2011;365:395-409).

Several societies are working to devise the best protocol for further workup of detected nodules, but none have come up with the solution. The Fleischner Society, an international society for thoracic radiology, previously proposed serial follow-up imaging at intervals of less than 1 year for solid nodules smaller than 8 mm detected incidentally at nonscreening CT (Radiology 2005;237:395-400). For all semisolid nodules and ground-glass nodules at least 5 mm in size, researchers recommended a 3-month follow-up (Radiology 2005;237:395-400).

Follow-up CT scans, however, add costs and contribute to a patient’s cumulative radiation burden, said Dr. Pearson, a thoracic radiologist at New York-Presbyterian Hospital/Columbia University in New York City.

In an effort to develop a more practical solution, Dr. Pearson and his colleagues analyzed the variability in nodule location among 50 patients with subcentimeter nodules identified on CT screening for lung cancer and emphysema, and then devised a protocol for targeted CT follow-up that was validated in 50 additional patients.

Two experts and one novice reader, a medical student, independently measured the distances of the nodules from two anatomical landmarks, the lung apex and carina, on baseline and follow-up scans by cross-referencing the axial images with the CT scanogram on a PACS (picture archiving and communication system) workstation.

The interobserver variability was quite low, with mean differences of just 2-3 mm when the two experts were paired together or individually paired with the novice reader, Dr. Pearson said. Results were significantly better measuring from the apex than from the carina for two of the three reader pairings (P = .005 and .03), although the absolute difference was just 1 mm.

Interscan variability between the baseline and follow-up scans was slightly larger, averaging 5.3 mm from the apex and 6.2 mm from the carina. The maximal difference in nodule location between the two scans was 24 mm from the apex and 29 mm from the carina. Both outcomes significantly favored the measurement from the apex (P = .01).

"The measurements were highly reproducible, [there was] no benefit in measuring from the carina, and we figured that if the nodules were normally distributed throughout the lungs, that 99% of nodules should range between the mean plus 3 standard deviations, which would be about a 38-mm range," Dr. Pearson said.

To allow for a greater margin of error, measurements in the validation phase of the study were only from the ipsilateral lung apex and covered a 60-mm range. At baseline, the expert identified the nodule and placed a cut line on the scan to measure the distance from the apex. Medical students then measured the distance of the nodule from the apex on the follow-up scan, and reviewed a region of about 6 cm to determine whether the nodule would have been included in its entirety if the range of the CT scan had been narrowed to just 60 mm.

"The results here are pretty simple; 100% of the nodules in follow-up were scanned in their entirety," Dr. Pearson said.

The average craniocaudal coverage on follow-up CT was 363 mm.

By using a targeted 60-mm range, the average craniocaudal coverage and radiation dose would be reduced by 83%, although the actual dose reduction will depend on the location of the nodules, he added.

In addition, use of targeted CTs would cut interpretation and report times due to the decreased anatomical coverage, would not disrupt work flow, and would require minimal technologist training.

Still, the technique needs to be validated prospectively in larger groups, and would require buy-in from radiologists, medical societies, and Medicare and private insurers, Dr. Pearson said.

Session comoderator Dr. H. Page McAdams, division chief of cardiac and thoracic imaging at Duke University in Durham, N.C., said in an interview that the concept is intriguing, but raises a number of potential societal, reimbursement, and medical-legal issues. For example, what happens if a physician does a limited follow-up CT and misses a new nodule? The other question is, how many patients in these screening populations really have a simple nodule? If most of them have multiple nodules, it would obviate the benefits of targeted CT.

"I think it’s a great idea, but there’s a lot of work that still needs to be done in terms of answering those particular questions," Dr. McAdams said.

Roughly 75% of the 100 consecutive patients in the study had a single nodule, Dr. Pearson told this publication.

Dr. McAdams reported a research grant from General Electric, serving as a consultant for American College of Radiology Image Metrix, and working as an author for Reed Elsevier and UpToDate Inc.

CHICAGO – Limiting CT follow-up of incidental pulmonary nodules to a small area around the suspicious nodule may help resolve a common clinical dilemma and an Achilles heel of lung cancer screening, researchers suggest.

It could also reduce the radiation dose by more than 80%.

"Why should we radiate the entire thorax on follow-up?" Dr. Gregory D. Pearson asked rhetorically at the annual meeting of the Radiological Society of North America.

In the pivotal National Lung Cancer Screening Trial, annual low-dose CT reduced lung cancer deaths by 20% and all-cause mortality by 6.7% among heavy smokers, compared with traditional X-rays. Approximately 40% of the CT group, however, had findings of small, indeterminate pulmonary nodules considered suspicious for lung cancer on at least one scan, with 1.4% experiencing a complication as a result of additional testing (N. Engl. J. Med. 2011;365:395-409).

Several societies are working to devise the best protocol for further workup of detected nodules, but none have come up with the solution. The Fleischner Society, an international society for thoracic radiology, previously proposed serial follow-up imaging at intervals of less than 1 year for solid nodules smaller than 8 mm detected incidentally at nonscreening CT (Radiology 2005;237:395-400). For all semisolid nodules and ground-glass nodules at least 5 mm in size, researchers recommended a 3-month follow-up (Radiology 2005;237:395-400).

Follow-up CT scans, however, add costs and contribute to a patient’s cumulative radiation burden, said Dr. Pearson, a thoracic radiologist at New York-Presbyterian Hospital/Columbia University in New York City.

In an effort to develop a more practical solution, Dr. Pearson and his colleagues analyzed the variability in nodule location among 50 patients with subcentimeter nodules identified on CT screening for lung cancer and emphysema, and then devised a protocol for targeted CT follow-up that was validated in 50 additional patients.

Two experts and one novice reader, a medical student, independently measured the distances of the nodules from two anatomical landmarks, the lung apex and carina, on baseline and follow-up scans by cross-referencing the axial images with the CT scanogram on a PACS (picture archiving and communication system) workstation.

The interobserver variability was quite low, with mean differences of just 2-3 mm when the two experts were paired together or individually paired with the novice reader, Dr. Pearson said. Results were significantly better measuring from the apex than from the carina for two of the three reader pairings (P = .005 and .03), although the absolute difference was just 1 mm.

Interscan variability between the baseline and follow-up scans was slightly larger, averaging 5.3 mm from the apex and 6.2 mm from the carina. The maximal difference in nodule location between the two scans was 24 mm from the apex and 29 mm from the carina. Both outcomes significantly favored the measurement from the apex (P = .01).

"The measurements were highly reproducible, [there was] no benefit in measuring from the carina, and we figured that if the nodules were normally distributed throughout the lungs, that 99% of nodules should range between the mean plus 3 standard deviations, which would be about a 38-mm range," Dr. Pearson said.

To allow for a greater margin of error, measurements in the validation phase of the study were only from the ipsilateral lung apex and covered a 60-mm range. At baseline, the expert identified the nodule and placed a cut line on the scan to measure the distance from the apex. Medical students then measured the distance of the nodule from the apex on the follow-up scan, and reviewed a region of about 6 cm to determine whether the nodule would have been included in its entirety if the range of the CT scan had been narrowed to just 60 mm.

"The results here are pretty simple; 100% of the nodules in follow-up were scanned in their entirety," Dr. Pearson said.

The average craniocaudal coverage on follow-up CT was 363 mm.

By using a targeted 60-mm range, the average craniocaudal coverage and radiation dose would be reduced by 83%, although the actual dose reduction will depend on the location of the nodules, he added.

In addition, use of targeted CTs would cut interpretation and report times due to the decreased anatomical coverage, would not disrupt work flow, and would require minimal technologist training.

Still, the technique needs to be validated prospectively in larger groups, and would require buy-in from radiologists, medical societies, and Medicare and private insurers, Dr. Pearson said.

Session comoderator Dr. H. Page McAdams, division chief of cardiac and thoracic imaging at Duke University in Durham, N.C., said in an interview that the concept is intriguing, but raises a number of potential societal, reimbursement, and medical-legal issues. For example, what happens if a physician does a limited follow-up CT and misses a new nodule? The other question is, how many patients in these screening populations really have a simple nodule? If most of them have multiple nodules, it would obviate the benefits of targeted CT.

"I think it’s a great idea, but there’s a lot of work that still needs to be done in terms of answering those particular questions," Dr. McAdams said.

Roughly 75% of the 100 consecutive patients in the study had a single nodule, Dr. Pearson told this publication.

Dr. McAdams reported a research grant from General Electric, serving as a consultant for American College of Radiology Image Metrix, and working as an author for Reed Elsevier and UpToDate Inc.

CHICAGO – Limiting CT follow-up of incidental pulmonary nodules to a small area around the suspicious nodule may help resolve a common clinical dilemma and an Achilles heel of lung cancer screening, researchers suggest.

It could also reduce the radiation dose by more than 80%.

"Why should we radiate the entire thorax on follow-up?" Dr. Gregory D. Pearson asked rhetorically at the annual meeting of the Radiological Society of North America.

In the pivotal National Lung Cancer Screening Trial, annual low-dose CT reduced lung cancer deaths by 20% and all-cause mortality by 6.7% among heavy smokers, compared with traditional X-rays. Approximately 40% of the CT group, however, had findings of small, indeterminate pulmonary nodules considered suspicious for lung cancer on at least one scan, with 1.4% experiencing a complication as a result of additional testing (N. Engl. J. Med. 2011;365:395-409).

Several societies are working to devise the best protocol for further workup of detected nodules, but none have come up with the solution. The Fleischner Society, an international society for thoracic radiology, previously proposed serial follow-up imaging at intervals of less than 1 year for solid nodules smaller than 8 mm detected incidentally at nonscreening CT (Radiology 2005;237:395-400). For all semisolid nodules and ground-glass nodules at least 5 mm in size, researchers recommended a 3-month follow-up (Radiology 2005;237:395-400).

Follow-up CT scans, however, add costs and contribute to a patient’s cumulative radiation burden, said Dr. Pearson, a thoracic radiologist at New York-Presbyterian Hospital/Columbia University in New York City.

In an effort to develop a more practical solution, Dr. Pearson and his colleagues analyzed the variability in nodule location among 50 patients with subcentimeter nodules identified on CT screening for lung cancer and emphysema, and then devised a protocol for targeted CT follow-up that was validated in 50 additional patients.

Two experts and one novice reader, a medical student, independently measured the distances of the nodules from two anatomical landmarks, the lung apex and carina, on baseline and follow-up scans by cross-referencing the axial images with the CT scanogram on a PACS (picture archiving and communication system) workstation.

The interobserver variability was quite low, with mean differences of just 2-3 mm when the two experts were paired together or individually paired with the novice reader, Dr. Pearson said. Results were significantly better measuring from the apex than from the carina for two of the three reader pairings (P = .005 and .03), although the absolute difference was just 1 mm.

Interscan variability between the baseline and follow-up scans was slightly larger, averaging 5.3 mm from the apex and 6.2 mm from the carina. The maximal difference in nodule location between the two scans was 24 mm from the apex and 29 mm from the carina. Both outcomes significantly favored the measurement from the apex (P = .01).

"The measurements were highly reproducible, [there was] no benefit in measuring from the carina, and we figured that if the nodules were normally distributed throughout the lungs, that 99% of nodules should range between the mean plus 3 standard deviations, which would be about a 38-mm range," Dr. Pearson said.

To allow for a greater margin of error, measurements in the validation phase of the study were only from the ipsilateral lung apex and covered a 60-mm range. At baseline, the expert identified the nodule and placed a cut line on the scan to measure the distance from the apex. Medical students then measured the distance of the nodule from the apex on the follow-up scan, and reviewed a region of about 6 cm to determine whether the nodule would have been included in its entirety if the range of the CT scan had been narrowed to just 60 mm.

"The results here are pretty simple; 100% of the nodules in follow-up were scanned in their entirety," Dr. Pearson said.

The average craniocaudal coverage on follow-up CT was 363 mm.

By using a targeted 60-mm range, the average craniocaudal coverage and radiation dose would be reduced by 83%, although the actual dose reduction will depend on the location of the nodules, he added.

In addition, use of targeted CTs would cut interpretation and report times due to the decreased anatomical coverage, would not disrupt work flow, and would require minimal technologist training.

Still, the technique needs to be validated prospectively in larger groups, and would require buy-in from radiologists, medical societies, and Medicare and private insurers, Dr. Pearson said.

Session comoderator Dr. H. Page McAdams, division chief of cardiac and thoracic imaging at Duke University in Durham, N.C., said in an interview that the concept is intriguing, but raises a number of potential societal, reimbursement, and medical-legal issues. For example, what happens if a physician does a limited follow-up CT and misses a new nodule? The other question is, how many patients in these screening populations really have a simple nodule? If most of them have multiple nodules, it would obviate the benefits of targeted CT.

"I think it’s a great idea, but there’s a lot of work that still needs to be done in terms of answering those particular questions," Dr. McAdams said.

Roughly 75% of the 100 consecutive patients in the study had a single nodule, Dr. Pearson told this publication.

Dr. McAdams reported a research grant from General Electric, serving as a consultant for American College of Radiology Image Metrix, and working as an author for Reed Elsevier and UpToDate Inc.