In November 2000, sports photographer Mickey Pfleger was standing on the sidelines of an NFL football game when one of the players knocked him down during an errant play. Pfleger was unconscious for several minutes and was taken to hospital, where a brain scan revealed that, although he had suffered no ill effects from the accident, he did have a large brain tumour. The chance discovery of his tumour, an ‘incidental finding’ in medical terms, allowed Pfleger to undergo the appropriate treatment earlier and may have improved his chances of long‐term survival.
Although the circumstances are rarely as dramatic as being knocked out by a 250‐pound football player, incidental findings are fairly common during medical and scientific studies on human volunteers, and they pose a dilemma for both researchers and their subjects. When Pfleger received his diagnosis he was already a patient, and doctors guided him to the appropriate therapy. In the research setting, however, scientists, who are often not experienced clinicians, face a complex mixture of risks and responsibilities when deciding when and how to hand over a subject with a potential medical problem to the healthcare system.
Although virtually any study involving a human subject could reveal an incidental finding, researchers in neuroimaging and genetics in particular have turned their attention to this issue in recent years. Because of technological advances that make it possible to examine the human brain and the human genome in far more detail than ever before, these research areas are rapidly expanding. With that expansion comes the likelihood that more incidental findings can and will be detected.
So far, assessments of the frequency and types of incidental findings in neuroimaging have focused on magnetic resonance imaging (MRI) studies (Illes et al, 2004b; Kim et al, 2002; Katzman et al, 1999). The rate of abnormalities was in the range of 2–8% and comprised mostly benign findings, such as chronic sinusitis, that did not require any clinical follow‐up. More serious conditions included blood vessel malformations (Fig 1), cysts and tumours.
… incidental findings are fairly common during medical and scientific studies on human volunteers, and they pose a dilemma for both researchers and their subjects
Genetics researchers have to grapple not only with true incidental findings, such as the discovery of a discrepancy in paternity during a family study, said Flavia Facio, a genetic counsellor at the US National Human Genome Research Institute (NHGRI; Bethesda, MD, USA), but also with sequence variations, called variants of unknown clinical significance, whose effects are simply not known. Today, most people who participate in genetic studies have symptoms or a family history of the disease under consideration, so unlike participants in neuroimaging studies, they already suspect that they have a medical problem. Facio predicts, however, that as genetics research advances, researchers will begin to recruit more reportedly healthy volunteers. Thus, geneticists face some of the same ethical questions as neuroimagers in the arena of incidental findings.
Incidental findings can be a particular problem in brain research because the scientists running the protocols are usually not medical doctors and are not qualified to draw clinical conclusions from their data, said Judy Illes of Stanford University's Center for Biomedical Ethics (CA, USA), who studies the ethics of incidental findings in neuroimaging. In addition, the data are frequently not of clinically interpretable quality. Thus, a suspicious finding is just that—a suspicion—unless the researchers consult with medical experts. Some neuroimaging groups therefore include neurologists and radiologists on their research teams, just in case they unexpectedly need a medical opinion on one of their scans. When Stanford student Matthew Kirschen discovered a golf‐ball‐sized structure in the brain of one of his volunteers, he called in his principal investigator, the director of the imaging center, the manager of the MRI magnet and a Stanford neurologist. The group was able to assess the problem quickly, and within hours had helped the subject obtain appropriate medical care. All of the people Kirschen consulted were officially part of the research team. “We do that on purpose,” said Kirschen, “so that if something like this happens, we have the ability to share the information with [the team] or get advice from them without having to go outside of the confidentiality or privacy domains.” Lining up a team of medical experts may be relatively straightforward when the research imaging facility is affiliated with a large hospital, but the task can be more daunting for researchers who work in a purely academic setting.
Before genetics researchers discuss a finding of potential clinical significance with a subject, they will generally verify it with an independent laboratory carrying Clinical Laboratory Improvement Amendment (CLIA) certification. If the CLIA laboratory confirms the researchers' results, they can feel confident that they are giving their subjects accurate information, according to Vivian Ota Wang, a Programme Director at the NHGRI. However, CLIA laboratories do not always provide a clinical interpretation of the result, Facio stressed. That is still left to the judgement of the researchers, who, given the increasingly widespread use of genetic techniques by scientists from diverse backgrounds, may not have formal training in genetics.
Genetics researchers have to grapple not only with true incidental findings … but also with sequence variations … whose effects are simply not known
Researchers also face the difficult problem of communicating a suspected problem to a study participant. The responsibility of informing the subject may, by default, fall to the principal investigator as leader of the group; however, he or she may not always be the most appropriate person. In a study on the expectations of participants in neuroimaging, Kirschen, Illes and colleagues found that a majority of subjects would prefer to learn about an incidental finding from a physician on the research team. “Part or most of the motivation for that is that they want somebody who can answer their questions,” commented Kirschen. Consequently, consent forms for neuroimaging studies at Stanford University ask for contact information for participants' primary care physicians, so that they can be consulted if necessary (Illes et al, 2004b). The speed with which suspicious results are reviewed and shared with subjects is also important, especially in a field such as neuroimaging in which some conditions require urgent medical attention (Illes et al, 2004b). Another important aspect of communication is assessing the subject's understanding of and reaction to the finding. Even though she discusses the possibility of variants of unknown clinical significance during pre‐test counselling, Facio finds that “it's not unusual for people who receive these results to be disappointed or even surprised.”
Incidental findings can be a particular problem in brain research because the scientists running the protocols are usually not medical doctors and are not qualified to draw clinical conclusions from their data…
The question of how carefully to screen subjects for abnormalities is still open in the neuroimaging community. At one end of the spectrum, the US National Institutes of Health (NIH; Bethesda, MD, USA) gives a full clinical brain scan to everyone who volunteers for its neuroimaging studies (Check, 2005). This practice offers participants maximal reassurance that any problems will be discovered, and minimizes the chances that researchers will miss an abnormality. However, it is also extremely expensive in both time and money and some researchers feel very strongly that the increased costs are not worth it. “In the present financial climate [giving every participant a clinical scan] means that half as much research gets done,” said Alan Evans, a neuroimaging researcher from McGill University (Montreal, QC, Canada). “And that has, in my mind, a much greater impact on society and the health of society than the very, very low incidence of incidental findings which are actually correct and an even lower incidence where there is something you could have done.” In addition, doubling the scanning time may exceed the patience of some volunteers, particularly in paediatric studies, Evans added. Other researchers have adopted a less comprehensive, but still expensive, policy in which a radiologist reads every research scan, according to a recent survey (Illes et al, 2004a). Evans' group, and the majority of those surveyed, take a moderate approach, which Evans described as a ‘Good Samaritan’ policy. They inform participants upfront that their scans are not of clinical quality and are not being interpreted by medically qualified personnel. However, if they do notice something suspicious, they will consult with a radiologist for confirmation and inform the subject of the results. At the other end of the spectrum, about 18% of neuroimaging researchers who were not required by their institutions to discuss volunteers' brain scans with a radiologist said they never did.
A final issue that troubles researchers is legal liability—the risk that they will be sued if they misdiagnose or miss a serious medical condition in one of their subjects. Beyond the impact on the individual researchers, Evans worries that a lawsuit might prompt the affected institution to disband its clinical research entirely, and hopes that the NIH will formulate some guidelines to help researchers protect themselves from lawsuits. So far, the problem is hypothetical. Neither Wang nor Illes knows of any legal action arising from an incidental finding in genetics or neuroimaging. However, “We don't need a legal event to be proactive. Our goal is to offer some ethical guidelines that protect our subjects, the investigators, and the research,” Illes said.
For the research subject, an incidental finding may at first appear to be a blessing; in reality however, it can impose a substantial psychological and financial burden. Clinically evaluating the finding often requires extensive follow‐up tests. Moreover, said Illes, in the USA “where one has to purchase healthcare and isn't guaranteed health insurance by any means”, a finding may create additional costs for the patient and even affect a subject's future insurability. In an anonymous letter to Nature, one person described how he was denied disability insurance because of a brain tumour that was discovered during a research scan (Anonymous, 2005). The anxiety is particularly onerous when the medical implications of the finding are not clear. For example, a woman who learns that she has a sequence variation of unknown clinical significance in her BRCA1 gene, which is implicated in hereditary breast and ovarian cancer, may end up undergoing more frequent clinical breast exams or other screening, even though her particular variation may not increase her risk of breast cancer at all. In cases where the woman is already deeply concerned about breast cancer, perhaps because of a family history of the disease, such a result can “insert even more ambiguity and more anxiety, and…feeling of loss of control,” said Facio. In fact, some consider the potential negative impact of an incidental finding to be a risk that needs to be explicitly discussed during the informed consent process before participating in a research study (Grossman & Bernat, 2004). Facio has known people who declined to participate in a study because they do not wish to deal with the consequences of an ambiguous result.
To address some of these issues, Illes and others organized a workshop in January 2005 on the ‘Detection and Disclosure of Incidental Findings in Neuroimaging Research’. About 50 neuroimaging and genetics researchers and representatives of the ethical and legal communities met for two days for debate and discussion in Bethesda, MD, USA. Despite some areas of disagreement, the group ultimately adopted several recommendations. Among other things, they concluded that researchers should anticipate the possibility of incidental findings and decide beforehand how to handle them; that someone capable of rendering a medical opinion on a scan should be available to the research team; that a qualified member of the research team should promptly inform the subject of any clinically significant findings; that the possibility of incidental findings should be discussed with subjects during the informed consent process; and that guidelines on incidental findings should outline the responsibilities of researchers while not impeding the progress of research. Even though the focus was on neuroimaging, geneticist Wang felt the discussion was broadly applicable because it allowed people from different backgrounds “to cross‐fertilize a little bit better and really learn from the lessons of other professions”. Kirschen called it “absolutely astounding” to watch the diverse assembly debate the issues and said that it did an “excellent job” of settling on broadly applicable consensus points.
…a suspicious finding is just that—a suspicion—unless the researchers consult with medical experts
Evans was more critical of the meeting. He is concerned that the debate will lead to unnecessary bureaucratic regulations that will have a “chilling effect” on clinical research. “It's already a tough enough business to do clinical research…[If you] raise it to a further level of paperwork complexity and bureaucratic obstructionism, people will just say ‘forget it’.” The meeting's organizers are acutely aware of the need to develop guidelines that will protect researchers as well as their subjects. Co‐organizer Emmeline Edwards, Deputy Director for Extramural Research at the US National Institute of Neurological Disorders and Stroke (Bethesda, MD, USA), cites the desire to strike the right balance as the main reason why some aspects of the issue are still under discussion. “We certainly do not want to hamper neuroimaging research,” she said, “but we also want to raise awareness of the responsibility of the researcher…to manage the finding.”
On the basis of the conclusions of the workshop, Illes and her colleagues at Stanford have developed a pathway for managing incidental findings as well as some suggested language for informed consent forms. The group is also continuing its efforts to inform and learn from the research community. Kirschen hopes that their work will resonate beyond the neuroimaging community, noting that the aspects of incidental findings that they are debating “will also be relevant in other aspects of research and medicine as we push technology forward”.
- Copyright © 2005 European Molecular Biology Organization