CHICAGO – A few months ahead of the commercial launch of its genomic profiling test for clinical oncology, Foundation Medicine is receiving around 25 patient samples a week from physicians seeking to use its next-gen sequencing-based approach to inform cancer treatment, a company official said this week at the annual meeting of the American Association for Cancer Research.
Vincent Miller, senior vice president of clinical development at Foundation Medicine, said during a presentation at the conference that that the 25 clinical samples it sees each week are in addition to samples it is processing as part of formal collaborations with pharmaceutical companies and academic groups.
The samples it has seen so far have come from "a broad range of tumor types including common, unusual, and rare tumors," he said.
During his talk, Miller also shared some data from clinical reports the company has provided for treating physicians over the last two weeks.
In one example, for a patient who had an adenocarcinoma of the pancreas, the company detected three key genomic alterations: a BRAF V600E mutation and loss of CDK2A and CDK2B, which would suggest the use of a BRAF inhibitor such as vemurafenib (Roche's Zelboraf) or a CDK5/6 inhibitor. The BRAF mutation highlights the opportunity for the company's approach, Miller said, because BRAF inhibitors are currently not indicated for pancreatic cancer.
He acknowledged that there are challenges for this approach, as well, however, since most patients have multiple alterations and it is not clear whether clinicians should act on one or all of them, and in what sequence.
For example, a patient with salivary gland adenocarcinoma harbored a PI3K mutation indicating that a PI3K inhibitor might be effective; an MDM2 amplification, which would point toward an MDM2 inhibitor; and CDKN2A and CDKN2B loss, which would suggest a CDK4/6 inhibitor, but these findings in themselves do not offer insight into how and when to prescribe existing therapies or enroll the patient in clinical trials investigating new drugs.
"There could be the potential for sequential clinical trials or rational combination therapeutics," Miller said.
In a follow-up interview, Miller told PGx Reporter via e-mail that "the decision-making process and treatment strategy based on this information is completely up to the discretion of the physician."
And while the company's reports include information and search terms that can help physicians identify trials for investigational treatments that may be appropriate for their patients, "we do not actively help place patients on clinical trials," Miller said.
Nevertheless, the company is "actively speaking with our biopharma partners and with patient advocacy groups to explore way to use the information we provide to support clinical trial enrollment," he added.
Foundation is not actively following up with physicians to gather information on which treatments are prescribed and how effective they are, though he said that some physicians have provided that information to the company.
In addition, he said that clinical information on the patients under study has not been provided in many cases, so the recommendations it makes to physicians are based on sequence alone.
Foundation plans to launch its test in June. It has a turnaround time of 14 days and a list price of $5,800.
Miller said this week that it’s "still too early" to comment on reimbursement, but said that as with any new test, "it will take time and appropriate clinical studies to secure broad third-party payor coverage and we are currently working with major academic institutions on studies that will generate the data we require."
New Feasibility Data
Foundation's test sequences 189 cancer-related genes to up to 1000-fold coverage.
In February, researchers from the company and the Dana-Farber Cancer Institute published a paper in Nature Medicine that demonstrated the test's ability to identify mutations and gene fusions for which known and experimental therapies exist (Clinical Sequencing News, 2/22/2012).
In that study, the researchers tested the assay on 40 colorectal cancer and 24 non-small cell lung cancer formalin-fixed, paraffin-embedded biopsy samples.
At AACR this week, company officials and collaborators provided additional details about the technical feasibility of the platform with additional tumor types.
In a presentation, Roman Yelensky, director of clinical genomic analysis at Foundation, discussed an internal validation study that used the platform to sequence 58 archived FFPE tumor samples, including 16 breast, 16 colon, 14 lung, and 12 renal cell cancers.
The average sequencing coverage was 850x, Yelensky said.
Of the 58 cases, 39 — or 67 percent — had an alteration that was "potentially actionable," in the sense that it would likely confer sensitivity or resistance to approved or experimental targeted therapies, he said. Furthermore, 40 percent of these genetic variants were not at the "hotspots" that are typically included in cancer genotyping panels.
Of the potentially clinically actionable mutations, 57 percent occurred at known sites of oncogene activation, 27 percent involved gene amplification or deletion, 14 percent resulted in tumor suppressor loss, and one resulted in the EML4-ALK fusion gene. In total, Foundation identified 121 mutations and 30 copy number alterations across the samples.
Additionally, mutation frequencies were consistent with previous studies for the tissue types studied. For example, 56 of the breast cancer samples carried TP53 mutations, 38 percent had PIK3CA mutations, and 13 percent had HER2 amplifications. In colon cancer samples, 75 percent harbored APC mutations, 63 percent carried TP53 mutations, and 38 percent, 31 percent, and 25 percent had KRAS, BRAF, and PIK3CA mutations, respectively. In non-small cell lung cancer, 57 percent of the samples had TP53 mutations, and 42 percent harbored mutations in KRAS. Among renal cell carcinoma samples, 17 percent had mutations in VHL.
In another presentation, Michael Berger, a collaborator at Memorial Sloan-Kettering Cancer Institute, discussed a comparison of Foundation's approach to an in-house genotyping panel that MSKCC currently runs on the Sequenom MassArray platform.
Berger said that the MassArray panel, which MSKCC has been running since 2009, covers 91 mutations in eight genes. As of January, the group had used it to profile more than 6,000 tumors.
In the comparison study, the MSKCC team selected 71 surgically resected FFPE tumors — 42 lung, 26 colorectal, 3 melanoma — that had previously been genotyped on the Sequenom platform.
Berger said that there was high concordance between the Sequenom and Foundation assays, with 60 mutation calls in common, two unique to Sequenom, and five unique to Foundation.
In addition, Foundation's assay identified 73 variants at sites of known recurrent somatic mutations and 30 loss-of-function variants in tumor suppressor genes not tested by the Sequenom panel.
These additional variants are "not traditionally interrogated in the clinical setting," Berger said, and many of the represent plausibly actionable mutations that could influence treatment decisions.
Berger said that his lab did not evaluate the cost of the Foundation test compared to the Sequenom assay and has not yet decided whether to adopt the next-gen sequencing approach.
Have topics you'd like to see covered in Pharmacogenomics Reporter? Contact the editor at btoner [at] genomeweb [.] com.