Q:You have recruited healthy control subjects and patients with Alzheimer disease for an fMRI study monitoring disease progression. But when you analyze the brain fMRI studies, you find unexpected imaging abnormalities in some of your controls. It's not clear if these abnormalities are clinically significant because the scans contain only selected images necessary for the research protocol. Do you tell the research control subjects about these incidentally discovered findings? If so, what exactly should you tell them and how should you proceed?
Dr. Bernat is professor of neurology and medicine at Dartmouth Medical School. He is the former chair of the AAN Ethics Law & Humanities Committee and is the author of Ethical Issues in Neurology, 3rd ed. (Lippincott Williams & Wilkins, 2008).
DR. JAMES L. BERNAT: The issue of how researchers should respond to possibly abnormal neuroimaging findings discovered incidentally in the course of volunteering for a study falls into the realm of the emerging field of neuroethics. As the term is generally used, neuroethics refers to ethical issues generated by neuroscience research. For scholars in this nascent field, the approach to handling incidental neuroimaging research findings in normal volunteer research subjects was one of the first topics studied. There is now a standard approach to optimally manage such cases that I review here.
These issues arise commonly so neurologists practicing where neuroimaging studies are conducted are likely to be consulted in such cases. I have performed neurological consultations in a number of cases, including a young woman found to have arrested hydrocephalus, a middle-aged woman found with an unruputured middle cerebral artery aneurysm, an elderly man found to have asymptomatic bilateral subdural hematomas, a woman found to have diffuse demyelinating-appearing cerebral hemispheric white matter lesions, and a man found to have extensive age-related white matter hemispheric lesions. In each case, the research subject was anxious about the meaning of the abnormality.
In a study of the frequency and severity of incidental research fMRI findings, Judy Illes, PhD, and colleagues found them in 6.6 percent of people participating as normal subjects. Older subjects had more findings than younger subjects but the findings in the younger subjects were more serious. That these findings occur frequently in normal people is not surprising because the incidence of asymptomatic brain MRI abnormalities is high. In the Rotterdam Study of neurologically asymptomatic “normal” people aged 45–97, 7.2 percent had asymptomatic brain infarctions, 1.8 percent, cerebral aneurysms, and 1.6 percent benign tumors, mostly meningiomas. Cerebral hemispheric white matter lesions were commonly seen with advancing age.
Research subjects nearly always want to be notified when imaging abnormalities are discovered. In a 2006 paper in the Journal of Magnetic Resonance Imaging, Matthew Kirschen, MD, and colleagues reported that 97 percent of surveyed research subjects wished to be told if abnormal findings were detected and most subjects preferred that the communication come from a member of the research team.
The issue of notification is complicated by the fact that the research MRI or fMRI images selected for the goal of the study often are not as complete as in a clinical MRI study. Further, it is common in some protocols for the images to be viewed only by psychologists, cognitive neuroscientists, or others who are not trained in neuroradiology. These limitations affect the determination of whether an observed finding is abnormal, exactly what it represents, and whether it is significant. Some protocols have overcome this limitation by requiring clinical neuroradiological interpretation, but this feature adds greatly to the protocol's expense.
In a Neurology editorial in 2004, Robert Grossman, MD, and I outlined the elements of consent and disclosure for research subject participation that follow established ethical guidelines for clinical research. The informed consent process and form should itemize all the risks of participating including increased anxiety resulting from learning an abnormality may be present, financial costs resulting from additional testing that may become necessary to better define the purported abnormality, and health risks resulting from additional tests to further clarify the purported abnormality. The consent form also should clarify if the scan sequences are adequate for clinical purposes, whether the scan will be interpreted by a trained clinician, who will notify the patient in the event of an abnormality, and who will be responsible for paying for further medical care that becomes required because of the finding.
In 2005, the Working Group on Incidental Findings in Brain Imaging Research was impaneled and met to study this problem and issue guidelines, with support of three divisions of NIH: the NINDS, the National Institute of Mental Health, and the National Institute of Biomedical Imaging and Bioengineering. Their preliminary report was published in Science in 2006 and their final report was published in Neurology in 2007. The Working Group's recommendations provide the current standard for managing this problem.
In its preliminary report, the Working Group concluded that: 1) it is ethically desirable for suspicious incidental findings to be disclosed to the research subject because of the subject's right to know; 2) the potentially harmful consequences of false-positive reports on normal volunteers has not been adequately studied; 3) wide variability exists about when and how incidental findings are reported to human subjects; 4) vulnerable subjects may require special assistance in arranging follow-up medical consultations; 5) it is desirable, when possible, to have a physician validate the presence of a suspected incidental finding; 6) it is desirable, when possible, to have a physician communicate the incidental finding to the subject; 7) institutional review boards (IRBs) should require clarification of the complete process for handling the incidental findings; 8) there is no ethical requirement for the researcher to obtain clinical scans on the subject; and 9) more research is necessary to study the costs and benefits of identifying incidental findings and for referring subjects for appropriate medical follow-up.
The Working Group final report guidelines provide a range of appropriate options for studies conducted in different settings, ranging from those in a psychology department of a liberal arts college to those in a neuroradiology department of an academic medical center. For example, in a fMRI study conducted in a college psychology department in which images are obtained that lack both the resolution and completeness of a clinical scan, and for which no neuroradiological interpretation is conducted, they recommend that no action need be taken other than to define the plan for handling the incidental findings as part of the consent process.
The Working Group final report identified the key elements of informed consent for research subjects: 1) the project is designed for a research purpose and is neither intended nor designed for clinical diagnosis; 2) the investigators are not trained in radiological diagnosis and the scans are not adequate for diagnostic purposes; 3) it is possible that the imaging study may discover a possible abnormality; 4) being told of a possibly abnormal findings may produce anxiety and create the need and expense for further assessment; 5) if a possible abnormality is discovered, the investigator may consult with a clinician (for example, radiologist or neurologist) to determine whether the finding merits further attention, and the subject will be notified; 6) the decision to proceed with further evaluation will be made by the research subject who will bear the costs for such evaluation; and 7) the research team is available for further information and discussion. The final report also issued guidelines for investigators who plan to conduct this type of research, admonishing them to: 1) anticipate the possibility of discovering incidental findings; 2) establish a comprehensive process for managing incidental findings that is clearly spelled out to the IRB and in the consent form; 3) permit but do not mandate the involvement of a medical professional to assist interpretation of the images; 4) have the principal investigator or clinical designate disclose incidental findings to the subject or surrogate in a timely manner in proportion to the seriousness of the finding; and 5) pay attention to potential burdens when choosing special subject groups.
In the present case, the process for managing disclosure to research subjects and action on incidental findings should have been included in the research consent form, and the steps followed. If they were not, the investigators now should notify the research subjects and arrange timely neurological consultations. They then should draft a comprehensive plan for future clinical studies to manage incidental findings containing the elements discussed here that is approved by the IRB.•
YOUR QUESTIONS ANSWERED
Do you have an ethical question involving your practice that you'd like answered? If so, submit your question or case challenge for consideration for the Ask the Neuroethicist column: neurotoday@LWWNY.com.
• Illes J, Kirschen MP, Edwards E, et al. Incidental findings in brain imaging research. Science
• Illes J, Kirschen MP, Edwards E, et al. Practical approaches to incidental findings in brain imaging research. Neurology 2007;70:384–390.
• Vernooij MW, Ikram MA, Tanghe HL, et al. Incidental findings on brain MRI in the general population. N Engl J Med
• Kirschen MP, Jaworska A, Illes J. Subjects' expectations in neuroimaging research. J Magn Reson Imaging 2006;23:205–209.
• Grossman RI, Bernat JL. Incidental research imaging findings: Pandora's costly box. Neurology
• Wolf SM, Paradise J, Nelson AC, et al (eds). Symposium: Incidental Findings in Human Subjects Research: From Imaging to Genomics. J Law Med Ethics