NGS, maybe, sort of, sometimes clinic ready

Detailed genomic profiling is a sine qua non of precision medicine, and at tertiary cancer centers, the comprehensive profiling of tumors and matched germline DNA is becoming ever more routine.

The technological advances in tumor analysis are transforming the way cancer is conceptualized. The most recent example is the FDA’s first approval of a drug – the checkpoint blocker Keytruda (pembrolizumab, Merck & Co. Inc.) – to treat patients with a specific genomic alteration regardless of a tumor’s anatomical origin. (See BioWorld Today, May 25, 2017.)

Clinical studies, though, “have been less compelling,” Eli Van Allen, assistant professor of medicine at Harvard Medical School, told the audience at a session during the American Society of Clinical Oncology meeting in Chicago this month on “Translating Tumor Sequencing Into Clinical Decision-Making: The Record to Date.”

In fact, the first proof-of-concept trial testing the idea that assigning patients to treatments based on molecular profiling failed to prove its concept. In 2015, the phase II SHIVA trial showed no difference in progression-free survival between patients who were assigned to one of eight targeted therapies based on molecular tumor profiling and those who received chemotherapy. Overall survival, which was presented at this year’s meeting, was also not significantly longer for patients who received targeted therapies based on genomic analysis of their tumors.

Still, at the session and elsewhere in the ASCO meeting, researchers presented work showing that it is technically feasible to make genomic profiling part of community-based cancer care, and engaged in lively discussions about how to enable genomic profiling to have a clinical impact.

Several presentations at ASCO demonstrated that, technically, the routine extensive profiling of tumors is within reach.

In one session, Olivier Tredan, chair of the Department of Medical Oncology at the French Centre Léon Bérard, presented data from the ProfiLER study, which used either next-generation sequencing (NGS) of 69 genes or whole-genome hybridization, depending on the size and purity of the tissue biopsy, to suggest targeted therapies for patients with advanced refractory tumors. At the meeting, Tredan presented data from 2,676 patients.

The good news was, logistically speaking, the method is ready for prime time.

Theoretically, we could test ‘every patient in France’

“Routine genomic testing is feasible in a local and regional setting,” Tredan said. “Theoretically, we could do this testing for every patient in France.”

Unfortunately, for now, doing the testing for every patient in France would not equate to doing much good.

Of the roughly 2,700 patients enrolled to date, 143, or 7 percent, were treated with a targeted therapy that was based on their tumor profiling results, despite the fact that roughly half of them had actionable mutations.

Tredan and his team were able to make a molecularly targeted recommendation for nearly 700 people. But the majority of patients did not get the treatment that was recommended.

Partly, this is because getting those treatments depended in part on the availability of clinical trials testing them. And because patients on the trial were extremely ill, if such a trial was not immediately available, a sizable fraction of them progressed or died before treatment recommendations could be implemented.

The problem is not limited to studies outside of the U.S., though the situation is improving as more basket trials are being conducted.

Erin Cobain, a medical oncologist at the University of Michigan, described a genomic profiling study of 500 patients with metastatic cancer, in which about 20 percent of patients “had subsequent clinical management that was informed by their sequencing results.” Whether that number is high or low is in the eye of the beholder, but Cobain noted that the number of patients who were able to enroll in appropriate clinical trials had sharply increased over the course of the study as more basket studies were initiated.

Overall, though, Cobain’s conclusion was that NGS is “of limited utility performed late in metastatic disease course.”

Success rates will increase as more trials become available and patients are screened earlier.

Collect lots of plasma

As sequencing methods evolve, it will also be possible to screen for ever-increasing numbers of potential targets – though here, bigger is not necessarily better.

Harvard’s Van Allen warned that as the amount of information available to clinicians has increased exponentially, their brain sizes and working memories have not, resulting in “clinical data fatigue.”

Good training and decision support is one way to prevent data fatigue. Another is to make sure that what is collected is the optimal amount of data rather than the most data possible.

Sometimes, more data allow for better decisions. But other times, more data add to the noise, not the signal.

Geoffrey Oxnard, a medical oncologist at the Dana-Farber Cancer Institute, told his audience that in looking for actionable mutations in colorectal cancer, liquid biopsies agreed with tissue biopsies about 80 percent of the time. The agreement, however, was 100 percent for seven key genes that are known to be important drivers in colorectal cancer.

Oxnard’s conclusion was that “if there are seven genes you are interested in colorectal cancer, you might as well just run seven genes.”

At the meeting, Oxnard gave an overview of current state as well as future potential of liquid biopsies.

So far, the routine clinical use of liquid biopsies is limited to the search for specific targeted mutations.

But in experimental clinical settings, they are being used for capturing both the overall mutational landscape of heterogeneous tumors, and how that landscape changes over time, in a way that is not feasible with tissue biopsies. (See BioWorld Today, June 7, 2016.)

Oxnard was clear-eyed about the utility of liquid biopsies. In discussing one study that reported discovering “hypothetically targetable” mutations in 99 percent of about 440 liquid biopsies of tumors of unknown primary origin, his dry assessment was that “targetability is a tricky concept” that depends not just on the number of available clinical trials, but also on a hard to quantify variable – “the optimism of the clinician making the call.”

Overall, though, he was bullish on the utility of liquid biopsies for both clinical and research advances. His advice to his fellow oncologists? “Collect plasma – I mean, as much plasma as you can.”

Read more ASCO meeting coverage from BioWorld Today: “HIM2? Zytiga prostate win just the start as test proceeds into development.”

And don’t miss “The Evolving Utilization of Specific Biomarker Roles within Trial Design,” a free webinar on Thursday, June 15, for an exploration of trends in the application of specific biomarker roles (therapeutic effect marker, toxic effect marker, disease marker) within active clinical trials.

More information and register