NEW YORK (GenomeWeb) — Researchers from the Vanderbilt University Medical Center have released a retrospective study showing that Foundation Medicine's FoundationOne test identified a potentially actionable mutation in the majority of the first 103 cancer patients to receive the test at the center.
The study appeared this month in The Oncologist. Overall, the researchers found that 83 percent of 103 patients tested between April 2012 and August 2013 had a "potentially actionable" genetic alteration. The authors reported that 21 percent of the patients then received genotype-directed therapy with an appropriate targeted drug, and most in the context of a clinical trial.
In a handful of cases, the group reported, this genotype-driven treatment also led to "significant benefit."
Douglas Johnson, the study's lead author and fellow at Vanderbilt, said that he and his colleagues had two main goals in performing their retrospective analysis. "Previous studies have validated the [FoundationOne] platform and described the testing characteristics, and also the fact that most tumors have potentially actionable mutations or alterations," Johnson said. "But what hadn’t been described as well was actually how that played out in clinical practice … This was something we wanted to start to address, though obviously this study doesn’t definitively address it."
Johnson noted that the group was specifically interested in two things: whether these mutations predict benefit to approved treatments and currently marketed drugs, or to more experimental treatments; and whether patients were actually getting treatment based on the testing.
In the study, Johnson and his colleagues retrospectively reviewed records from patients seen at the Vanderbilt Ingram Cancer Center who were tested with FoundationOne between April 2012 and August 2013.
The researchers first analyzed the percentage of patients for whom the testing uncovered novel therapy options based on the presence of a potentially actionable mutation, defining actionable alterations as those associated with susceptibility to an approved drug or an experimental therapy being tested anywhere in the US.
The team then assessed the percentage of these patients who actually received genotype-directed therapy.
According to the authors FoundationOne identified at least one genetic alteration in 97 of the tested patients, and 86 had an alteration that was considered "potentially actionable" according to the study criteria, with an average of two actionable mutations per patient.
About a quarter of the patients with an actionable mutation had an alteration linked to a drug already approved for the tumor type in question, and another 17 percent had alterations that could be targeted by a drug approved for another tumor type. Half the patients were potential candidates for genotype-directed therapy in the context of a clinical trial of a drug demonstrating "at least early activity in the clinical setting." A final 6 percent were eligible for clinical trials of drugs with only "preclinical rationale for their use," the authors wrote.
The researchers then looked to see how many of these patients with the potential to be treated based on their genotype actually were. They found 18 of the tested patients — 21 percent of those with actionable alterations — received genotype-directed therapy.
Seven of these, the authors wrote, received a clinically approved agent, while 11 were enrolled in clinical trials. The researchers did not consider therapy given in response to testing other than FoundationOne to be "genotype-directed," for example trastuzumab given in HER2-amplified breast cancer that had been classified as HER2/neu by other testing methods.
"I think maybe going in I would have thought [the number receiving genotype-directed therapy] would be a little higher than what we saw, but looking at the data it makes a lot of sense," Johnson said. "Some of the tumors, for example, had maybe a p53 mutation in isolation. Certainly there are trials around the country enrolling those patients, but they are not necessarily widely available."
But, he added, this is also a moving target. "For example, we saw a lot of mutations in the CDK pathways so as CDK 4/6 inhibitors become more widespread and there are more trials in different tumor types we should see an evolution of those numbers over time."
The researchers also collected data on patients' outcomes, including one patient with T-cell prolymphocytic leukemia found to harbor a JAK mutation who improved on a JAK 1/2 inhibitor, and a patient with melanoma found to have a BRAFV600E mutation not identified by previous testing, who had "dramatic durable improvement" on dabrafenib.
The group also noted a patient initially diagnosed with mucoepidermoid lung cancer who had an early response to crizotinib after discovery of an EML4-ALK fusion. Based on the presence of the fusion, the authors wrote, the tumor was probably actually a misdiagnosed adenocarcinoma.
However, the authors wrote, while several patients did respond to genotype-driven therapy, the converse was also true. This, they said, highlights the "still underdeveloped understanding of the pathophysiologic implications of many genetic alterations."
Vanderbilt began using Foundation Medicine's test as an adjunct and in some cases a replacement for a previous hotspot panel called SNaPshot.
According to Johnson, the center currently does not have distinct criteria for which patients are eligible for FoundationOne or not, leaving the decision ultimately up to treating physicians. The center also still uses SNaPshot routinely in some tumor types as a pre-screening tool to rapidly identify certain alterations with strong validated therapeutic links.
"The general approach is to use [SNaPshot] as screening, then for the negative patients or for tumor types that don’t have a SNaPshot panel we then use FoundationONe," he said.
Most of the time, he added, the center is using Foundation Medicine's test for patients who have already failed first line therapies although, as it is at the discretion of treating physicians whether to order the test, it could also potentially be run for newly-diagnosed patients.
Since the study, the number of patients who have received the test at Vanderbilt is up to 280, Johnson said. The researchers have continued to collect data on sequencing results, and plans to take another look at the prevalence of actionable mutations and how that translates into therapeutic decision-making once they have amassed a larger cohort.
According to Johnson, the center is also working on developing an in-house sequencing strategy that could potentially replace the use of FoundationOne some time in the future.
"I think our ultimate goal is to have a targeted NGS platform in house but that is a complicated undertaking and we want to make sure that all our processes are aligned before moving forward with that … But having this tool [in the mean time] has been very helpful to be able to provide this for patients," he said.