Genetic testing of tumors is becoming increasingly common in cancer care. The molecular alterations found in a tumor can provide critical information for making an accurate diagnosis and determining the best treatment.
Although current clinical testing usually focuses on a panel of specific mutations, cancer centers are developing programs to analyze entire cancer genomes routinely — an approach made possible by less-expensive sequencing costs — in order to individualize care. This raises a thorny question: What happens when a genome analysis of a person’s tumor reveals that he or she is at risk for developing a different type of cancer or another disease?
In August 2013, Memorial Sloan Kettering Clinical Genetics Service Chief Kenneth Offit, Clinical Genetics Service Clinic Director Mark E. Robson, and researcher Yvonne Bombard published a viewpoint in the Journal of the American Medical Association regarding this question of incidental genetic findings, which cancer researchers have dubbed the “incidentalome.”
We asked Dr. Robson to discuss some of the issues surrounding incidental genetic findings and what Memorial Sloan Kettering is doing to address them.
What is an example of a genetic variation that might be discovered incidentally while sequencing the genome of a patient’s tumor?
Let’s say you’re sequencing a lung cancer tumor in search of an EGFR mutation to target with an anticancer drug, and you find a mutation in BRCA1, which is associated with increased risk for breast and ovarian cancer. Since most of a tumor’s DNA sequence is identical to the sequence of a normal cell from that same patient, and BRCA1 mutations are not associated with lung cancer, this additional variation is probably inherited — it’s what is called a germline mutation.
One can come across mutations another way. A sample of normal DNA is often sequenced to serve as a comparison for the tumor DNA sequence — a baseline, if you like. If this nontumor DNA were analyzed separately, the BRCA1 mutation may be identified directly.
In these situations, what is the clinician’s responsibility to inform the patient? It’s a very complex question. There are many variables to consider, such as whether the testing was for clinical or research purposes, the individual preference of the patient involved, whether anything can be done to control risk, and whether other people — such as close relatives — may be affected.
Has this actually become a problem for doctors and researchers, or is it still a hypothetical situation at the moment?
Most clinical testing of tumors has been for a relatively limited number of specific mutations, not the full genome. But soon we’re going to be testing for a much broader panel of genes, increasing the chances of incidental findings.
On the research side, it’s quickly becoming an issue. Many tumor samples that have been stored in tissue banks for years or decades are now being fully sequenced. If incidental discoveries are made during that process, is there an obligation to try to find those patients and inform them? This has not been established, and there are obvious practical barriers. We need to lay the ethical groundwork for how we’re going to respond to these questions.
What steps have been taken at Memorial Sloan Kettering to address the issue?
Last summer, our Institutional Review Board (IRB), which oversees all of our patient-related research, updated part of our patient consent policy for biospecimen collection studies. When patients agree to have a tissue sample taken, they are asked whether they are open to being re-contacted if an investigator finds something that might affect their health.
Under the new procedure, if a researcher finds something that might be important to communicate to the patient, the specific finding will be put before the IRB and carefully considered. If there is agreement the information should be conveyed, and the patient has indicated that he or she wants to be contacted, we’ll reach out to that person. We think this protects the people participating in our studies without restricting important research.
With all the genetic research taking place at Memorial Sloan Kettering, are the IRB and the Clinical Genetics Service facing a deluge of these cases?
So far, no. The way the analyses are being conducted is that the computer looks for mutations in specific spots and subtracts all other information about the inherited genetic sequence before the investigator sees it. In other words, if you have genetic variants present in the tumor that are also in the normal cells, they are being filtered out by the software. The investigator ends up seeing only variants that are in the tumor, which was the purpose of the research test, and not the inherited changes.
But it is still possible for researchers to accidentally discover or be able to infer from a tumor sample that an inherited genetic change is present. In other cases, the researchers may want to know about inherited as well as tumor changes. If results are to be known, and potentially shared, then this will require a greatly increased effort to provide counseling to our patients. The Clinical Genetics Service is exploring ways to gear up our efforts to meet the expected demand for “genomic” counseling based on tumor-normal sequencing.
Have you gotten a sense from patients about what their preference usually is regarding being informed of these incidental genetic discoveries?
Genomic tests can reveal information about many disease risks. Commonly, people say, “I want to know everything,” but the devil’s in the details when you start considering the risk for diseases that can’t be prevented or treated. The Clinical Genetics Service is setting up focus groups of patients and unaffected people to try to understand how people think when they are confronted with these situations and how they prioritize different types of genetic information. We also have an active IRB protocol in which we are giving people who had their sequence determined as part of research studies the opportunity to learn their results. Jada Hamilton [of Psychiatry and Behavioral Science] and I are also beginning a study where we will interview patients who have had tumor genomic profiling to get a sense of their thoughts about these issues.
Right now, it’s not clear what the dividing lines are. We want to reach a point where mutations are sorted into different categories in which certain incidental findings are nearly always appropriate to communicate to patients, others almost never so, and some that require more context to determine.
We’re moving from the traditional model of asking patients if they would like to hear the results of a specific test before that test is performed to this brave new world where we’re trying to help people make decisions about genetic information revealed incidentally that is not possible to fully anticipate. It’s a very complicated issue, but it also offers a tremendous opportunity to benefit patients.
If you are interested in participating in the focus group, call 646-888-4867. Everyone is welcome, including patients, relatives, Memorial Sloan Kettering employees, and the general public. No sequencing is provided.