This story was originally published Nov. 4.
NEW YORK (GenomeWeb) — Genome sequencing can contribute more to personalized breast cancer prevention strategies than previously projected, according to a recent estimate by researchers at Stanford University School of Medicine's Department of Health Research and Policy.
In a paper published online in Cancer Epidemiology, Biomarkers and Prevention last month, the scientists, led by Alice Whittemore, a professor of epidemiology and biostatistics at Stanford, reported that estimating breast cancer risk distribution based on 86 known breast cancer variants and targeting those in the top 25 percent for screening or prevention measures would include about half of future breast cancer cases, leading to a greater gain in disease prevention than previously assumed. Those gains could increase even further in the future, they wrote, as our understanding of the genetics of breast cancer improves.
Whittemore told Clinical Sequencing News that the study, which used data from the literature about breast cancer risk variants but is theoretical in nature, is a response to a paper published two years ago in Science Translational Medicine by a group at Johns Hopkins University.
That study, also theoretical, estimated the ability of whole-genome sequencing of healthy individuals to predict those at risk for 24 common diseases, including breast cancer. It concluded that most individuals would receive negative results for most diseases, which would "not be very informative" because their risk would still be close to that of the general population. On the other hand, more than 90 percent of individuals would find out about a clinically significant predisposition to at least one disease, the study found.
The Hopkins study was based on an analysis of 53,000 identical twin pairs, who share genetic risk factors, but it did not take into account specific genes and risk variants that have been identified for certain conditions, and was criticized for that omission. "It occurred to me that we could do that for breast cancer," said Whittemore, pulling together all known breast cancer susceptibility loci, their frequency and associated relative risk.
She and her colleagues looked at 86 breast cancer risk loci in total and assigned risk scores based on their presence or combination in individuals. They then calculated the distribution of risk scores across a population of European ancestry to see if they could pick out those at highest genetic risk for breast cancer. "If it was not spread out at all, if everybody had the same risk … then it would be impossible to pick out anybody," Whittemore explained. "But it was much more spread out than that."
For example, those 10 percent with the highest risk scores contain 32 percent of breast cancer cases, and the top 25 percent contain about 50 percent of breast cancer cases.
According to Whittemore, the results are "not tremendously more optimistic, but they are more optimistic" than those of the Hopkins team, "and we're hopeful that with the identification of new genes, the story can get even better."
Bert Vogelstein, director of the Ludwig Center at Johns Hopkins and a senior author of the 2012 study, agreed. "The predictions are slightly more optimistic but roughly consistent with those made in the [2012] paper, particularly given the completely different models used and the confidence limits in the estimates," he told CSN in an e-mail.
"In our view, the Whittemore and colleagues modeling provides valuable new information that will surely be helpful for assessing the benefits of genome-based predictions of breast cancer," he said.
But the fact that genome sequencing can identify those at risk for breast cancer better than anticipated does not necessarily mean every woman should undergo genome sequencing yet, Whittemore cautioned.
"We're not there yet, we still have ways to go," she said. One the one hand, not all risk genes have been discovered yet, and on the other hand, there are not enough highly effective preventive measures available yet to make sure those determined to be at highest risk do not get the disease, or have it discovered at an early stage.
One encouraging point, she said, is that those at high genetic risk can help lower their overall risk to a greater degree than those at low risk by keeping an eye on modifiable risk factors, such as alcohol consumption, weight, age of motherhood, and number of children. For example, someone with a high genetic risk might be able to drop their risk by a third through those measures, for instance from 54 percent to 36 percent, whereas the same modifications would only lower the risk from 6 percent to 4 percent — also a third — in someone with a low genetic risk.