NEW YORK (GenomeWeb) – Cancer Research UK's stratified medicine program has evolved from testing with a variety of methods, including Sanger sequencing, to a standardized next-generation sequencing panel over the five years since it first launched.
The goal of the first phase was to figure out how to implement genetic testing for cancer patients within hospitals. Eventually, such tumor profiling might become standard of care, according to Emily Shaw, clinical lead of the stratified medicine unit at CRUK.
Researchers from CRUK have now assessed over 9,000 patients' tumors under the Stratified Medicine Program and published the results of the first 7,850 samples that were analyzed before they introduced the 28-gene NGS panel.
Shaw told GenomeWeb that for the second phase of the program, aside from screening patients with the NGS panel, the group is working with pharmaceutical companies to set up multi-arm clinical trials for drugs targeted to patients' alterations. Last year, the group launched the National Lung Matrix Trial, which sought to identify genetic markers in patients with non-small cell lung cancer that would stratify them to one of 14 drugs developed by AstraZeneca or Pfizer.
In a publication in the journal Drug Discovery Today, the researchers reported on the analysis of tumors from 1,874 breast, 1,634 colorectal, 1,889 lung, 535 malignant melanoma, 559 ovarian, and 1,359 prostate samples.
For each cancer type, between three and five genes were analyzed. Shaw said one challenge of this first phase was the lack of one validated method that all laboratories used to analyze the genes. Labs could use whatever method they had already validated or were most comfortable with. That sometimes resulted in lengthy turnaround times and more work to analyze and compare results, she said.
For instance, one goal of the program was to keep turnaround times to 15 days, but that was often "challenging to achieve, because the genes most likely to fail and require repeat testing were individual exons of the numerous exons examined within the tumor suppressor genes," the authors wrote. The PTEN and TP53 genes had failure rates of up to 18 percent compared to failure rates of 5 percent or less for hotspot mutations in the BRAF and KRAS genes.
Shaw said the group moved to the NGS panel to overcome some of these challenges. Although there was initially a significant amount of upfront work to validate the NGS panel and to get the labs comfortable with the bioinformatics and data, it should help in the longer term to decrease turnaround time and provide more consistent results, Shaw said.
In addition, while the cost and turnaround time for NGS panels might be greater than for individual hot-spot tests, "as the required number of markers per tumor increases, economies of scale mean that this will ultimately be a more cost-efficient and durable approach," the authors wrote.
From the first 7,850 samples, researchers found that the mutation rate varied significantly between tumor types. For instance, 65 percent of all the lung cancer samples were wild type for all genes analyzed, while only 19 percent of colorectal cancer samples had no mutations in the analyzed genes. Prostate cancer samples were dominated by the TMPRSS2-ERG gene fusion, found in 40 percent of samples. And only 2.1 percent of prostate cancer samples had mutations in more than one gene. In malignant melanoma, BRAF was the most commonly mutated gene, altered in 48 percent of samples.
Unsurprisingly, TP53 was the most commonly mutated gene in all cancer types, mutated in 23 percent of breast cancer samples, 67 percent of colorectal cancer samples, and 48 percent of ovarian cancer samples.
Shaw said the researchers have not yet finished analyzing data from the NGS panel, so could not say whether the results from the NGS panel have shown to be consistent with the prior technology. But, she said, NGS is more sensitive and can detect mutations at a lower allele frequency than the other methods, so samples may have more mutations. The NGS panel is also analyzing more genes.
Shaw added that one important lesson learned from the study was how to do patient consent. Genetic testing is still "new and there can be concerns around genetic data and the Pandora's box of what you might find if you start looking at your DNA or your tumor's DNA," she said. But, during a pilot phase, she said, researchers were able to "talk to patients about testing and their concerns," and once the program was launched consent rates exceeded 95 percent.
The primary intention of the first phase of the Stratified Medicine Program was not to use the results to guide treatment, Shaw said, but was "mainly to get the infrastructure going and prime the system." The laboratories performing the testing were all approved for clinical testing for the National Health Service, but the results were marked as research, Shaw said.
In addition, at the time, there were very few approved drugs for the targets being tested, she said. Nevertheless, there were a handful of targets tested for which drugs are approved, including EGFR and ALK mutations in lung cancer, BRAF mutations in melanoma, and KRAS and NRAS mutations in colorectal cancer. Oncologists were able to access patients' tests for these results, she said.
In addition, during the later stages of the program, the team partnered with a couple of clinical trials for which CRUK provided the pre-screening in the context of the trial, Shaw said. For instance, the researchers partnered with Roche to provide pre-screening for a clinical trial of a drug that targeted BRAF.
The initial partnership with Roche at the end of the first phase of the program led to the launch of the National Lung Matrix trial, Shaw said. Eventually, she anticipates that the trial will expand to other tumor types and other drug/genetic target combinations.
Ultimately, she said that she thinks tumor profiling with NGS panels will become part of routine cancer care. Currently, one challenge is that there are so few approved drugs linked to genomic targets. But, looking at the pipelines of targeted drugs in both the UK and the US, "there will clearly be a need to characterize the genetic profile of tumors," she said. "I think it is just a matter of how and what scope of analysis do we need," she said, adding that aside from targeted panels, transcriptomic, proteomic, and even data on the tumor's immune environment could all be important.
"We do see the [Stratified Medicine Program] continuing to move into other tumor types and, eventually, hopefully becoming standard of care within our National Health Service," she said.