CHICAGO (GenomeWeb) – Molecular tumor boards are continuing to gain treatment clues from individuals' cancer sequences, at the same time as growing cancer data collections reveal broader tumor features across patients.
At the American Association for Cancer Research annual meeting here yesterday, Arul Chinnaiyan reported on some of the results coming out of the University of Michigan's "Michigan Oncology Sequencing Center" (MiOncoSeq) research program, which has sequenced matched tumor and normal samples and generated tumor transcriptomes for roughly 3,000 cancer patients so far.
In addition to identifying clinically actionable mutations, Chinnaiyan said, a few broader clues have come up along the way. In particular, he highlighted an analysis of metastatic castration-resistant prostate cancer cases with tumors that had biallelic mutations in the CDK12 gene.
When the team took a closer look at those tumors, it saw signs that copy number or other changes leading to CDK12 loss may demarcate a distinct tumor subgroup with a particularly high gene fusion-induced neoantigen load, hinting that advanced or metastatic CDK12-mutated castration-resistant prostate cancers may respond to immune checkpoint immunotherapy drugs or targeted treatments.
The neoantigen load of the CDK12-mutated castration-resistant prostate cancers comes in second to that of prostate tumors with mismatch repair deficiencies, Chinnaiyan said, and exceeds that of cancers with mutations affecting homologous recombination mechanisms for repairing double-strand DNA breaks.
Based on those findings, the team has undertaken a small pilot study to investigate anti-PD1 immune checkpoint treatment outcomes in individuals with CDK12-altered tumors who already failed conventional late stage treatments.
The MiOncoSeq pipeline and results from the first 100 patients were described in a Science Translational Medicine study in 2012, a year into the cancer sequencing program, and somatic and germline profiles for 500 individuals sequenced through the MiOncoSeq program were reported in Nature in 2015.
There, Chinnaiyan and colleagues described an abundance of mutations in metastatic tumors, beyond those found in corresponding primary tumor samples. They also presented a germline analysis suggesting roughly 12 percent of individuals profiled carried pathogenic variants implicated in hereditary cancer risk.
Today, roughly 20 to 30 cancer patients per week move through the MiOncoSeq pipeline, Chinnaiyan said, making it difficult to discuss every case in a molecular tumor review board. He estimated that the team has held more than 100 tumor boards so far, to discuss and deliberate over potentially clinically actionable mutations in more than one third of cases profiled.
During the same AACR meeting session, Michael Korn, a gastroenterology and medical oncology researcher at the University of California, San Francisco, shared insights from the first few years of the UCSF Molecular Tumor Board — a multi-disciplinary group ranging from oncologists to clinical trial experts and genetic counselors that is tasked with interpreting high-throughput sequence data for matched tumor and normal samples assessed using the UCSF500 panel.
Across clinical cases profiled from April 2015 to November 2017, the researchers identified germline cancer-related variants in around 15 percent of individuals. They also tracked down mutations deemed clinically actionable in a significant subset of cases.
While survival did appear to increase in patients with molecular tumor board recommendations, Korn cautioned that many of the individuals who received targeted therapy recommendations based on this molecular data did not receive the recommended treatment.
In some cases, patients were already responding quite well to conventional therapies, he explained. But in other instances, patients either opted against the targeted treatment or declined significantly between the time of the molecular tumor board's recommendation and the start of treatment, which typically takes more than 120 days, in part due to drug access issues.
A molecular oncology knowledgebase founded on the molecular tumor board experiences so far is also helping to integrate the data and provide real-time feedback to clinicians. The availability of treatments will likely continue evolving with new decisions from the US Food and Drug Administration and the Centers for Medicare and Medicaid Services, Korn added.
At the Memorial Sloan Kettering Cancer Center's Kravis Center for Molecular Oncology, meanwhile, investigators are using a somewhat similar strategy to search for molecular drivers in paired tumor-normal sequence data generated with the MSK-IMPACT panel, MSKCC researcher David Solit said during a presentation.
Given the volume of cases moving through that program — 150 to 200 per week, on average — it is not feasible to bring together a molecular tumor board for most of the cases, he explained. Instead, that team has worked out methods for automating the annotation and analysis process, while establishing physician support tools that provide information on the levels of evidence supporting potentially actionable mutations.
Among other things, the molecular profiling effort has facilitated basket studies on targeted therapies for specific mutations presented at relatively low frequencies across multiple cancer types, Solit explained. The data can also serve as a starting point for delving into the function and significance of certain genetic changes in the tumors.
The team plans to continue expanding and generating evidence on MSK-IMPACT, with an eye to profiling additional genes, adding viral probes, and interrogating microsatellite instability patterns in various tumor types.