Incidental findings at chest CT: a needs assessment survey of radiologists' knowledge.

Rationale And Objectives: To assess practice patterns in evaluating incidental findings at chest computed tomography (CT) to determine the need for further education.

Materials And Methods: A survey was given to 1600 radiologists, presenting four clinical case questions regarding the evaluation/significance of the following incidental findings at chest CT: thyroid lesion; enlarged mediastinal lymph nodes; asymptomatic, small pulmonary embolus; and small lung nodule. The respondents' answers were compared with "truth," as defined by the best evidence available in the medical literature. Additional questions elicited the respondents' demographics and comfort levels in addressing the findings. Analysis of variance models with a Tukey correction for post hoc comparisons and chi-square tests were used to determine if any demographic factors or comfort levels were predictive of higher correct response rates.

Results: The overall survey response rate was 28% (445/1600). Correct case response rates ranged from 26% (115/442) to 79% (343/445). Only 6% (28/438) of respondents chose the correct answers for all cases. Up to 80% (353/440) of respondents felt comfortable in addressing findings, and only 57% (252/443) of respondents felt that they needed more training in this area. Fellowship training in cardiothoracic radiology, working in a teaching practice, and subspecialization in abdominal or cardiothoracic radiology were predictive of higher correct response rates. Except for one case question, the comfort level was not predictive of correct response rate.

Conclusions: There was considerable variability among radiologists and substantial deviation from best medical practice with regard to the interpretation/evaluation of incidental findings at chest CT, signifying a significant need for further education.