By mining epidemiological data from several long-term health studies and combining it with genomic data, a team led by the Dana-Farber Cancer Institute and Harvard Medical School has shown that colorectal cancer patients with PIK3CA mutations may benefit from treatment with aspirin, and that PIK3CA mutation status could serve as biomarker to predict response to aspirin treatment.
The study, published last month in the New England Journal of Medicine, evaluated data from 964 patients with colon or rectal cancer from the Nurses' Health Study and the Health Professionals Follow-up Study. It found that patients with PIK3CA-mutated cancers who regularly took aspirin after their diagnosis had significantly longer survival, while those with wild-type cancers showed no benefit from aspirin treatment.
According to the researchers, led by Dana Farber's Shuji Ogino, the results suggest that aspirin might be worth testing as an adjuvant treatment for the approximately 20 percent of colorectal cancer patients with PIK3CA mutations.
"What we conclude is that this PK3CA mutation can be a predictive biomarker and based on molecular testing, doctors could strongly or weakly recommend aspirin," Ogino told PGx Reporter.
According to the group, numerous observational and other studies have suggested that aspirin might play a protective role in colorectal cancer. Aspirin is currently prescribed to some colorectal cancer patients, Ogino said, but so far there has been no way to predict which patients are likely to actually benefit from it.
Ogino said his team's previous research found that levels of the enzyme PTGS2 could predict response to aspirin treatment, but the association didn't reach statistical significance. And because of a lack of good standards for measuring PTGS2 using immunohistochemistry, the group wanted to search for a better, more objective marker.
According to the group, other experiments have suggested that as aspirin inhibits PTGS2 it also down-regulates PI3K signaling, which hinted that PIK3CA mutations could be a potential marker as well.
"Based on previous studies, we hypothesized that PIK3CA mutation may be a good marker for aspirin response," Ogino said. Testing this hypothesis prospectively, he said, would have taken decades, but by using epidemiological data coupled with molecular data the group was able to find an answer much more quickly.
In the recent NEJM study, Ogino and his colleagues compared the survival of colorectal patients who reported that they regularly used aspirin after their diagnosis with those who didn't, and further subdivided the group into those with PIK3CA mutations and those without.
The team studied samples from a subset of 964 patients from the two large longitudinal health studies for which the relevant aspirin use data was available, collecting specimens from the registries and using pyrosequencing to establish PIK3CA mutation status for each patient's tumor. The group also recorded whether samples had BRAF or KRAS mutations.
The researchers found that patients with PIK3CA mutations who reported regular aspirin use had a significantly improved five-year survival rate — 97 percent — over those who didn't take aspirin — 74 percent.
In contrast, patients without the mutation showed no difference in survival whether they took aspirin regularly or not.
Because the group had previously found that PTGS2 levels were also predictive of response to aspirin use, the researchers evaluated whether a combination of both markers could serve an even greater predictor. According to the study authors, the strongest effect of aspirin use was indeed in patients with both markers, though this finding did not have high statistical significance.
Because the study sampled patients treated before 2006, the group assumed that chemotherapy treatment was similar for the PIK3CA-mutated cases and the wild-type cases. According to the researchers, information on patients' mutation status was not available to treating physicians at the time of the studies.
The team also distinguished between aspirin use before and after diagnosis, finding that pre-diagnosis use did not seem to influence the relationship between PIK3CA and post-diagnosis aspirin.
Ogino said that the group is pursuing avenues to validate the findings. Unfortunately, relatively few trials of aspirin treatment in colorectal cancer have been conducted.
One option, he said, would be to analyze data from a trial of celecoxib (Pfizer's Celebrex), a similar drug to aspirin, instead. But it's not an ideal solution. If the results reflect what the group found in its aspirin study it would shore up the aspirin finding. However, if the results do not match up it would be unclear what that might mean about the group's original findings.
Potentially, the researchers could also use mouse models or cell lines, but this route has several downsides. Most important, Ogino said, is the fact that aspirin likely affects inflammation more than cancer cells themselves. "Cancer is not just the cancer cell, it’s a much more complicated system so you can't assess it in the test tube, basically," he said.