A Conversation with Dr. Bert Vogelstein about Whole-Genome Sequencing to Predict Cancer Risk
Johns Hopkins researchers have used mathematical models based on clinical data from identical twins to test the ability of genomic testing to predict risk for 24 common diseases, including cancer. Dr. Bert Vogelstein, director of the Ludwig Center for Cancer Genetics and Therapeutics at Johns Hopkins University and an investigator with the Howard Hughes Medical Institute, recently spoke with the NCI Cancer Bulletin about the research, which was published April 2 in Science Translational Medicine.
What was the goal of your study, and what did you learn?
Our goal was to provide a realistic interpretation of what the public can expect to obtain from whole-genome sequencing at some point in the future, in the best circumstances, assuming that the technology will be developed that enables us to fully interpret all of the sequencing information. One question we asked was: In this best possible case, what would whole-genome sequencing achieve for the average individual?
Our results suggest that more than half of people who undergo whole-genome sequencing could, in this best-case scenario, obtain results indicating that they have an increased genetic risk for one or more diseases.
However, most individuals will receive negative test results for most diseases, including most cancers, meaning that they have less than a 10 percent risk of developing those diseases. Therefore, whole-genome sequencing will generally not predict exactly what diseases they will get or what diseases they will die from.
What do your findings mean specifically for estimating cancer risk?
As one example, we estimate that as many as 2 percent of women who have a whole-genome sequencing test could get a positive result for ovarian cancer, which would mean that they have a more than 10 percent risk of developing ovarian cancer, several times higher than the risk of the average woman.
Whole-genome sequencing will not alleviate the need for effective cancer prevention and early detection measures.
One of the implications of this study is that whole-genome sequencing will not alleviate the need for effective cancer prevention and early detection measures, which we already know can reduce cancer incidence and death. Equally important, we need more research to find better ways to prevent cancer or to detect it early.
What could people who have a positive result on a test, such as the one for ovarian cancer, do with that information?
[Patients] could get the best available surveillance, including ultrasound and gynecological exams to detect cancer as early as possible. And, hopefully, future research will result in more sophisticated and sensitive tests for detecting cancers early.
There are also environmental factors at play. Obesity is a risk factor for ovarian cancer, for example, and women at increased risk for this disease would certainly want to avoid that.
It’s been suggested that people would be better off spending their money on a gym membership than on genome sequencing. What do your results say about that line of thinking?
Our position is that the more knowledge individuals have, the better informed they will be, and the better the choices they will make. They may look at our data and decide that, for most diseases, whole-genome sequencing is not really going to be much help to them, but they know for a fact that getting more exercise will reduce their risk for several diseases, so they will choose to spend their money on the gym membership.
Another person may say, “My mother died of ovarian cancer, and if I’m at high risk, I want to know about it.” That person might want to buy both the gym membership and the genome-sequencing test. Someone else may say, “Knowing I have a slightly increased risk for some diseases is going to wreak havoc on my psyche, and it’s not worth it.”
I would not make that decision for anyone; like all health care decisions, it is extremely personal. I simply hope that our data will help medical professionals and others involved in health care inform their patients about what this technology can and can’t do.
What was your rationale for doing this study now?
This study evolved from our previous research. One of our goals is to try to understand the basis for familial cancers. In 2009, Alison Klein, Ralph Hruban, Sian Jones, and other members of our group reported on a family with familial pancreatic cancer not caused by a mutation in any of the genes known at that time to be involved in this disease.
Through genome-wide sequencing, we identified a heritable mutation in a gene called PALB2 in this family. This was the first time genome-wide sequencing had been used to identify the basis for a hereditary disease. Since then, a number of studies have identified the value of genome-wide sequencing for identifying the genetic basis of diseases in families.
So we began to wonder: If we could do this in families with a clear predisposition for cancer, could it be done for everybody? As whole-genome sequencing becomes less expensive, this becomes feasible. A few years ago it cost about $50,000 to sequence all of the approximately 20,000 genes in a human cell. Now it can be done for less than $5,000, and a few years from now the cost will be less than $1,000. We began to wonder about the public health implications and the value this might have for consumers.
What are the biggest caveats to note about this study?
The main caveat is that we’re not talking about what’s possible now but about what might be possible at some time in the future. Any perceived value of this approach [to learning about genetic risk] should be tempered by the realization that we’re not there yet.
Much of your cancer data was obtained from Scandinavian twin registries. Would the results have been different if you had looked at people of other ethnicities?
From other studies, and given what we know about cancer, I don’t think the generalities would be different. You can’t change the genes people are born with. You can change their environment, but with cancer, random influences play a large role. In different populations, you would find slightly different results, but we believe that the general conclusions would stand.
What are the implications for public health?
The best possible finding for a study like ours would be that a large part of the screened population [that test positive] might be at high risk for cancer and, conversely, that those who didn’t test positive would have a very low risk. In that scenario, we might have been able to recommend that prevention and surveillance measures be concentrated on individuals at increased risk.
But this study and earlier studies show that individuals at high genetic risk account for only a minority of those who will die of cancer, so the public health implications of whole-genome sequencing are not as strong as we might have wished for.
—Interviewed by Eleanor Mayfield