Incidental Findings in CT Colonography: Literature Review and Survey of Current Research Practice

CT colonography (sometimes called "virtual colonoscopy") uses CT to image the abdomen and pelvis as an alternative to traditional colonoscopy. However, because CT colonography images most of the torso (from the base of the lungs to the pubis) it regularly yields incidental findings of potential medical significance beyond the colon (extracolonic findings). This is a major problem in CT colonography research, as the majority of research subjects have an extracolonic finding and the proper handling of incidental findings in human subjects research generally is highly controversial. CT colonography research offers a model for managing incidental findings in human subjects research, especially in the context of research imaging. This article was produced by a team largely based at Mayo Clinic as part of an NIH-funded grant project on "Managing Incidental Findings in Human Subjects Research" (Susan M. Wolf, Principal Investigator). The article combines an analysis of the literature with an empirical survey of CT colonography researchers and normative analysis. By combining reports in the subset of the literature meeting certain criteria, the authors derive the number of human subjects with incidental findings, the number of subjects with incidental findings requiring medical or surgical attention, the number who actually receive a follow-up intervention, and the mean cost per incidental finding. The survey of researchers then sheds light on self-reported practice in handling incidental findings. While the Virtual Colonoscopy Working Group has proposed a reporting system (CRADS) with five categories (E0 through E4) in increasing order of clinical importance, practice varies. Fifty-eight percent of research programs surveyed reported all extracolonic findings, whereas 42% reported only potentially significant extracolonic findings. The mechanism of conveying extracolonic findings varied widely. Eighty-three percent of research programs generated a clinical report to detail extracolonic findings, with 67% contacting the subject’s physician directly and 42% contacting the subject directly. Ninety-two percent of programs discussed the potential for discovering incidental (extracolonic) findings of medical significance verbally during the informed consent process, but only 42% reported that their written consent form told subjects if and when extracolonic findings would be reviewed. CT colonography thus constitutes a research domain in which incidental findings are common. Indeed, the incidence of incidental (extracolonic) findings exceeds that of colonic findings. CT colonography researchers have accepted a responsibility to indentify, analyze, and report incidental findings, unlike researchers in a number of other domains. The approach to incidental findings in CT colonography research in many ways offers a useful paradigm for handling incidental findings in imaging research more generally. When research images are acquired that are anatomically and clinically meaningful and incidental findings are anticipated, researchers should develop a plan for detecting and communicating them to research subjects and their physicians. Despite inconsistency in approach to informed consent among CT colonography research programs, the benefits and burdens of incidental findings should be discussed with subjects during the informed consent process, in both verbal discussions and written consent documents.