CHICAGO (GenomeWeb) – Research presented at the annual meeting of the American Society of Clinical Oncology this week included a wealth of data from programs using next-generation sequencing to guide treatment for cancer patients.
The results indicated, as previous reports have also found, that NGS panels and broader sequencing strategies identify cancer-associated alterations in a large minority, and in some cases a majority, of patients, a significant proportion of which are potentially actionable.
Several studies that were presented also shared data on the percentage of patients who have an actionable mutation discovered by NGS that actually go on to receive treatment targeting that alteration, and of those, how many appear to benefit from targeted treatment.
These data are critical to establishing the clinical utility of NGS-driven therapy and prospective trials comparing genotype-driven versus non-genotype-driven therapy. In the meantime, data from the pioneering programs that have adopted the approach offer hints at its impact on patient care and outcomes.
One study presented at the meeting this week by researchers from the Sarah Cannon Research Institute reported on the results of targeted NGS performed as part of SCRI's initiative to match cancer patients to clinical trials based on their molecular profile.
From October 2012 through December 2013, the institute consented 1,040 patients, of which 936 had sufficient tumor tissue for sequencing using a targeted hotspot panel covering 35 cancer-related genes. The effort uncovered at least one mutation in 420, or about half, of those patients and what the team determined to be an actionable mutation in 100 patients.
Overall, the group reported that they were able to return results to treating physicians within 12 days. In addition, 206 patients, or 22 percent of those sequenced, were enrolled into clinical trials, but more than half had no mutations detected by sequencing, and another six percent had what the team deemed "non-actionable" mutations.
Fifty patients with actionable mutations were enrolled in trials specifically targeting an alteration discovered by the team's NGS analysis, while 53 went onto a trial not specifically targeting their cancer driver mutation.
Todd Bauer, the first author of the study, told Clinical Sequencing News that if the important questions to ask about NGS-driven cancer therapy are whether it can identify mutations, whether those mutations are actionable, whether patients can be treated based on those findings, and whether that treatment benefits them, the answers from the SCRI study have been: "Yes, yes, yes, and it's hard to say."
"If you break it down, for example, for PIK3 mutations, we had 90 [patients], but only about one-third went on trials," Bauer said. "Why did that happen? Well, some progressed too quickly and never got any more treatment. Some are on surveillance either having finished adjuvant treatment or are stable on treatment."
The researchers shared some data in their poster at the meeting tracking patient outcomes comparing those who received treatment targeted at their genetic alteration and those who did not.
"Looking at PIK3 again," Bauer said, "the poster shows two bars, one purple and one green. The purple bar is those [patients] put on a matched trial and the green bar is an unmatched trial, and the length of those bars is the average weeks on study."
For those with PIK3 mutations, the group has recorded only a small difference between how long patients remain in a clinical trial of a genomically targeted agent versus how long they last in other non-targeted trials: 12 weeks, on average, compared to 11.5 weeks.
For other mutations, like KRAS, the difference SCRI measured has been a little larger — 12 weeks for genotype-directed trials versus eight for non-directed trials. In patients with BRAF mutations, those on genotype-driven trials have remained, on average, 29.5 weeks compared to 10 weeks for those on non-targeted drugs.
But it gets more complicated, Bauer said. "The caveat is, it's late-stage patients on multiple phase I trials, which aren't designed to look at activity. So if you take another example, FGFR mutations, it's flip-flopped."
In this subset, the group's data shows that the two patients in the cohort with FGFR mutations on genotype-directed trials have progressed, while the single patient with an FGFR mutation on a non-directed trial is "doing great," Bauer said.
Moving forward, Bauer said his team plans to continue to track outcomes. They are also including in their program and their analyses alternative sequencing results that patients may have received elsewhere, not just the 35-gene panel offered by SCRI.
Other studies
In another poster presented at the meeting, a team from Princess Margaret Cancer Center in Toronto described results from molecular profiling of colorectal cancer patients using both a Sequenom panel targeting 279 mutations in 23 genes and the Illumina TruSeq cancer panel covering 48 genes.
The group initiated their molecular profiling strategy with the Sequenom panel, but have since switched exclusively to the running the TruSeq panel on the Illumina MiSeq, researchers told CSN at the conference.
In their poster, the researchers reported on 190 patients enrolled between March 2012 and October 2013. Among 153 patients analyzed with the Sequenom panel, 55 percent were found to have at least one cancer mutation, while the MiSeq strategy found mutations in 89 percent of 37 patients analyzed.
KRAS mutations were the most frequent, present in 35 percent of the cohort. Other mutations included BRAF and NRAS mutations, PIK3CA mutations, CTNNB1, ERBB2, and EGFR mutations, as well as a number of other alterations detected only by the MiSeq panel.
The team has collected limited data so far on the impact of sequencing results on patients' treatment and outcomes. At the meeting, lead author Joanne Wing-Yan Chiu told CSN that of about 240 patients who have undergone molecular analysis to date at Princess Margaret, only around 10 have gone on to a trial of a targeted therapy.
In another conference abstract, researchers from the University of Texas MD Anderson Cancer Center shared data from the first 500 patients with advanced cancer prospectively enrolled in an IRB-approved phase I clinical trial through the center's Department of Investigational Cancer Therapeutics.
Under this effort, MD Anderson tested archival tumor specimens using the 46-gene Ion Torrent AmpliSeq panel. Of 500 patients analyzed, 293 had at least one detectable mutation, the group reported. The most commonly mutated genes were TP53 (38 percent), KRAS (11 percent), and PIK3CA (10 percent). According to the team's report, "matching patients to molecularly targeted therapies is underway."
Another study provided data from a long list of oncology practices enrolled in a study using Foundation Medicine's FoundationOne cancer sequencing panel to analyze their patients. The study examined the rate at which physicians changed their treatment decisions based on FoundationOne results, and it is also looking at the impact of such changes on patient survival.
According to the presentation, the FoundationOne test caused physicians to change their proposed therapy in 26 of 128 patients while 69 patients were treated according to physicians' plans prior to sequencing. Another 33 patients were not treated based on physicians' prescriptions pre- or post-testing.
According to the authors, "data collection for [survival] analysis and correlation of results to tumor genomic profiles are ongoing."
In a retrospective study of NGS-driven therapy, researchers from a large group of community oncology practices reported that of 632 patients who received NGS testing, 360 harbored actionable mutations. Reviewing a subset of about 300 of these cases, the group found that mutational testing guided treatment in 34 percent.
A report from Canada's British Columbia Cancer Agency presented at the meeting tracked results of Ion Torrent AmpliSeq testing, as well as additional DNA and RNA sequencing, on 65 patients with a variety of advanced cancers recruited between 2012 and 2014, 56 of which were successfully sequenced.
According to the group, the AmpliSeq panel only yielded actionable targets in 40 percent of cases, while full genomic and transcriptomic data offered more comprehensive information about driver pathways and was "informative" in 70 percent of the subjects.
The sequencing results were actionable in 21 of 30 patients for whom clinical data was available, the researchers reported. Eight patients were treated based on their genotype, with six deriving clinical benefit from genomic-directed treatment, according to the authors.
Though the group wrote that "further studies are needed to determine the utility of this technology," they concluded that whole-genome sequencing offers more comprehensive information and additional clinically actionable data over targeted panels.