Skip to main content
Premium Trial:

Request an Annual Quote

Foundation Medicine Study Finds Potentially Treatable Mutations in Vast Majority of CUP Patients

Premium

NEW YORK (GenomeWeb) – A study by researchers from Foundation Medicine of 200 patients with cancer of unknown primary origin, has found that the company's comprehensive sequencing approach was able to identify a potentially targetable mutation in a vast majority of this cohort.

In the study, published this week in the newly launched JAMA Oncology, sequencing revealed a potentially targetable alteration — defined as a mutation either linked to a cancer drug currently on the market or required for entry into a mechanism-driven, registered clinical trial — in 169 patients, or 85 percent of the full cohort.

Although the researchers have only collected a small amount of follow-up data on the sequenced CUP patients, they also reported that at least two subjects with an identified targetable alteration went on to receive a mutation-targeted cancer therapy with dramatic positive results.

One, a woman with a MET amplification who was treated with Pfizer's Xalkori (crizotinib), had a sustained complete clinical response for more than three years. Another female was found to have an ALK fusion, and also has maintained a complete response to Xalkori, the group reported.

Jeffrey Ross, the study's first author and medical director at Foundation Medicine, told GenomeWeb that both subjects are still in remission.

Several methods to help guide treatment for patients with cancers of unknown primary origin have been developed in recent years, but most that have been commercialized have been focused on linking a CUP to a particular origin location in the body.

BioTheranostics markets a 92-gene CUP test called CancerType ID, Rosetta Genomics offers the microRNA-based Cancer Origin Test, and Response Genetics sells the only FDA-approved CUP assay, the Tumor of Origin test originally developed by Pathwork Diagnostics.

By identifying widespread metastases specifically as "breast cancer" or "colon cancer," rather than an unknown malignancy, such tests allow physicians to treat patients according to their particular cancer type, hopefully improving outcomes.

But according to the Foundation Medicine group, these methods can still fail to identify an origin site for 25 percent or more of CUP patients. Moreover, studies of how such testing actually impacts patient outcomes have shown relatively limited benefit over standard chemotherapeutic strategies.

For example, a BioTheranostics study of its gene expression-based CancerType ID in 252 patients found that those who received site-specific treatment survived an average of 12.5 months.

Although patients in the study who were predicted to have more treatment-responsive tumor types, like breast and ovarian cancer, had even longer survival — an average of 13.4 months and higher than 20 months for some of the most responsive types — the average improvement in outcome across the whole cohort was modest — only a few months more than the average survival seen in previous large studies of CUP using standard empiric chemotherapy strategies.

Meanwhile, growing evidence from sequencing studies has raised questions across the board about whether cancers should best be defined and treated according to their location and other clinical features or by their molecular features, the authors wrote.

In their study, the Foundation Medicine team set out not to predict where a CUP patient's cancer originated in the body, but simply to identify whether and how often these cancers were characterized by genetic alterations that indicate sensitivity to molecularly targeted drugs.

"We saw that these patients with CUP often have a very expensive workup to try to find the primary tumor, but then when we looked at outcome data in the published literature, their response to therapy didn't seem to improve all that much whether they thought they found [the origin] or not. Either way, [the outcome] is dismal, just very poor," Ross said.

"We were very much hoping that if we used the Foundation One test, this comprehensive profiling approach, we could discover targets that could put patients on completely different treatments than the ones they classically would go on," he added.

In the study, the researchers performed Foundation Medicine's Foundation One, a broad sequencing assay covering more than 300 cancer-associated genes, on 200 CUP patients.

According to Ross, most of these had had extensive imaging testing, which failed to identify a primary origin for their cancer. A small percentage also had been tested using either the Biotheranostics CancerType ID or Pathwork (now Response Genetics) Tissue of Origin test, with negative results.

Among these 200 patients, 192, or 96 percent, had at least one genetic alteration, and 169, or 85 percent, had an alteration that the company researchers classified as potentially targetable.

"Now having done more than 35,000 tumors with Foundation One overall and looking at breast, lung, colorectal, [and] prostate cancers all separately, we have never seen an alteration rate as high as in this 200-patient set," Ross said.

Narrowing further to only the most immediately targetable alterations, the team reported that 26 of the alterations identified were associated directly with currently approved targeted drugs, and another 14 cases had an "off-target" association with a current drug.

The remaining actionable mutations, while not linked to approved therapies, could make patients eligible for hundreds of registered clinical trials, the authors wrote.

The study also divided the total cohort into two subsets, those with features of adenocarcinoma, and those without, observing disparities in the distribution of several genes between the two types. For example, actionable alterations in the RTK/Ras singling pathway were present in 72 percent of the adenocarcinoma-featured CUPs, but only 39 percent of the non-ACUPs.

According to Ross, while the Foundation Medicine group doesn't have the evidence yet to know whether sequencing does or does not significantly improve CUP patient outcomes by directing them to targeted therapies, they hope to be able to demonstrate this in further retrospective analysis of this initial study group, as well as in an upcoming prospective study.

The company is planning to partner with a major academic oncology practice that Foundation Medicine has not yet named for that prospective study, and aims to accrue at least 100 patients, hopefully starting accrual by the end of this quarter.

All the subjects will be tested using Foundation One, and the group will compare outcomes in the treatment cohort, which will receive targeted treatment based on the test results, to a control group, which will be treated using standard of care chemotherapy.

The Scan

Less Than Half

An effort to reproduce key findings from high-profile preclinical cancer studies finds less than half could be replicated, according to the Scientist.

Still Some SARS-CoV-2 Sequencing Holes

The Wall Street Journal reports that viral genomic surveillance has improved in the US, though says there are still gaps.

Avoiding Passing Them On

People with known disease-linked genetic variants are turning to in vitro fertilization so as to not pass those variants to their children, the Washington Post says.

PNAS Papers on Long Cell-Free DNA in Maternal Plasma, Genetic Propensity for Voting

In PNAS this week: long, cell-free DNA of maternal and fetal origins identified in maternal plasma, and more.