NEW YORK (GenomeWeb) – Over the past decade, as genomics research has transformed understanding of the origins and persistence of cancer, some experts predicted that one day drugs would be approved based on the molecular changes that drive abnormal cells, regardless of whether these cells are in the lung, colon, or breast.
With the approval of a new, tissue-agnostic, biomarker-guided indication for Merck's Keytruda (pembrolizumab) the US Food and Drug Administration this week proved those prognosticators correct. Drug, diagnostic, and genomics experts agreed that FDA's action, though restricted to advanced patients who have failed other therapies, represents the start of a major shift in how the field thinks about cancer care, and propels precision medicine into a new phase. And while the new indication didn't come with an FDA-approved companion test, it places a spotlight on the need for well-validated diagnostics, and on next-generation sequencing cancer tests from companies such as Foundation Medicine, Personal Genome Diagnostics, and Guardant Health.
"This whole notion of a tissue-agnostic, biology-driven indication has become a reality," said Roy Baynes, senior VP of global clinical development at Merck Research Laboratories. "One hopes this is a useful advance for the whole area of precision medicine."
"I always thought that genomics would drive how cancers are treated," said Luis Diaz, who led the research that provided the first prospective clinical data suggesting that a pan-tumor indication may be a realistic ambition for Keytruda. "But in this case, it really melts the borders between cancer types and it bases treatment purely on a genomic signature."
The agency granted accelerated approval to Keytruda for adult and pediatric patients with unresectable or metastatic solid tumors characterized by high microsatellite instability or mismatch repair deficiency, after they've progressed on alternative drugs. Mismatch repair deficiency is a condition that makes cancer cells unable to repair the DNA damage that occurs during replication so that they become inundated by genetic mutations.
Researchers at Johns Hopkins University in 1993 described how this hobbled DNA repair mechanism resulted in the accumulation of mutations in repetitive DNA regions called microsatellites and caused a hereditary condition called Lynch Syndrome, which increases the risk for colorectal and other cancers. Colorectal cancers that are mismatch repair deficient have 10 to 100 times more genomic mutations than tumors with this corrective mechanism intact. The FDA estimated that these hypermutated tumors were common in 5 percent of colorectal patients, and also occurred with varying frequency in endometrial, gastrointestinal, breast, prostate, bladder, thyroid, and other cancers.
Several years ago, with the increasing availability of genomic testing and advances in immunotherapy, Diaz, while at JHU, began noticing that patients with mismatch repair deficiencies appeared to be responding better to immunotherapies like Keytruda than patients with normal mismatch repair functions. He and a team of researchers tested this observation prospectively in a study involving 41 cancer patients and showed that those with hundreds to thousands of genetic mutations in their tumors were more likely to benefit from Keytruda than those without hypermutated tumors.
Keytruda works by blocking PD-1 receptors on T-cells from binding to PD-L1 receptors on tumor cells, hindering a mechanism cancer cells use to evade recognition and attack from the immune system. Diaz's team hypothesized that highly mutated tumor cells, for example, colorectal cancers with mismatch repair deficiency, would be more likely to produce neoantigens that are recognized by the immune system and might respond particularly well to Keytruda,
In 2015, they published in the New England Journal of Medicine that 40 percent of colorectal cancer patients with mismatch repair deficiency responded to Keytruda, but there was no response among colorectal cancer patients without this genomic characteristic. Intriguingly, five out of seven patients with other types of cancers also responded to the drug. Additionally, 78 percent of colorectal cancer patients and 67 percent of patients with other types of cancers and mismatch repair deficiency had immune-related progression free survival at 20 weeks, compared to only 11 percent of colorectal cancer patients with mismatch repair intact tumors. Diaz and colleagues concluded that the results "show an approach for the treatment of a specific class of tumors that is based solely on genetic status — that is, without regard to the underlying tumor type."
"It's pretty incredible. This was a concept that we came up with, that we executed on, and we paid for … through philanthropy," said Diaz, now head of Memorial Sloan Kettering's solid tumor oncology division. "And it resulted in the first pan-tumor indication in the history of the FDA that is genetically derived. And it encompasses adults and kids. It's almost too much to think about, you know?"
The NEJM study, done without pharma support, cost under $500,000, and is included in the drug's label. While Merck wasn't involved in that investigation, the findings certainly got the company's attention. "When we started to see the results of very high response rates in colorectal cancer patients, as well as high response rates in other tumor types, we mounted a fairly substantial internal program," Baynes said.
Merck worked with Diaz and his colleagues to conduct the studies that resulted in the latest approval. The company submitted data to the FDA from five single-armed studies involving 149 patients with 15 types of solid tumors characterized by high microsatellite instability or mismatch repair deficiency. After treatment with Keytruda, 39.6 percent had a complete or partial response, and in 78 percent the response lasted for six months or more.
According to Keytruda's label, microsatellite instability and mismatch repair status for 135 out of 149 patients in the five studies was determined prospectively using local, laboratory-developed PCR testing or immunohistochemistry, or both. The rest of the patients had their microsatellite instability-high status determined retrospectively using PCR within a central lab.
IHC tests detect loss of protein expression in one of four mismatch repair genes (MLH1, PMS2, MSH2, and MSH6) and PCR testing measures variations in the lengths of microsatellite DNA segments. Although in order for advanced cancer patients to receive Keytruda they'll have to be tested for these tumor characteristics, the agency did not wait to approve the new indication with an FDA-approved companion diagnostic. The agency "didn't want to restrict the field," Diaz suggested, "especially with a therapy that could potentially lead to a transformative result in patients who otherwise failed standard therapies."
"In the case of Keytruda for this indication, based on the magnitude of the results that were demonstrated in patients with different life-threatening cancers that were identified as microsatellite-high or mismatch repair deficient, the FDA believed that it was important to approve Keytruda without an approved companion diagnostic test," an FDA spokesperson said.
Moreover, IHC and PCR assays for mismatch repair deficiency and microsatellite instability have been available as lab-developed tests for years, and guidelines bodies recommend testing for all colorectal cancer patients to evaluate prognosis and screen for Lynch syndrome. However, the latest indication for Keytruda means that this kind of testing will now be done much more frequently in metastatic cancer patients who have failed on prior therapies. Around 4 percent of all cancer patients, or 25,000 to 30,000 patients per year, have microsatellite instability.
There are also newer, next-generation sequencing tests for determining microsatellite instability, as well as NGS tests for assessing tumor mutational burden, which is another way of identifying tumors with genomically aberrant tumors. "Mutational burden [testing] will be part of the next iteration of genomically guided immunotherapy," but it needs to be evaluated prospectively like microsatellite instability has been, Diaz said.
Foundation Medicine, which markets an NGS panel that analyzes more than 300 cancer-linked genes, last year presented data from more than 60,000 patients' samples and showed that tumor mutational burden was particularly high in samples with mismatch repair deficiency. Baynes cited that work as suggesting "a high degree of correlation between tumor mutational burden and responsiveness" to immunotherapy.
In Merck's efforts to identify patients most likely to respond to Keytruda, the firm looked at both PD-L1 expression and high mutational burden within tumors, and screened 30 cancer types for screening in Phase II studies. "We have already declared fairly striking activity in 25 of them," Baynes said. "So, it worked as quite a good screening approach to identify potential utility."
Based on studies showing that advanced non-small cell lung cancer patients with tumors expressing PD-L1 protein are more likely to respond to Keytruda, Merck has launched the drug with a companion diagnostic in this setting. In addition to exploring PD-L1 expression as a predictive marker for Keytruda, Merck is also investigating gene signatures that can determine the degree of immune suppressiveness in the tumor microenviroment, as well as aspects of mutational burden.
"There will be additional work done on refining the diagnostics," Baynes assured. FDA's accelerated approval of Keytruda's new indication comes with a post-marketing commitment for Merck to submit analytical and clinical validation data on an IHC-based diagnostic and a genetic test for identifying mismatch repair deficient tumors by June 2019.
The latest approval is a boost for molecular diagnostics companies that have historically played a supportive role in drug development programs. One study recently showed that although more drugs are being developed with companion tests, their availability and quality are not well managed once a drug is launched, resulting in many patients missing out on treatment they'd benefit from. One example is PD-L1 testing, where the availability of multiple assays and nuanced indications have created confusion about which tests to use and whether to test patients at all.
But pan-tumor, genomically guided indications will put a spotlight on diagnostics and may create more urgency to better integrate testing in patient care and address these issues.
Anticipating a future where cancers will be grouped by biomarker characteristics and more drugs will target these features, some diagnostic companies are developing so-called universal companion tests. A Foundation Medicine spokesperson said its FoundationOne NGS platform currently gauges microsatellite instability, tumor mutational burden, and major mismatch repair genes. The company has validated its MSI method with established PCR and IHC methods and claims 97 percent concordance.
Foundation will submit these and other data to the FDA to garner regulatory approval for FoundationOne as a pan-cancer, universal companion diagnostic assay. "Further down the line with clinical validation, we believe these markers could be associated directly with immunotherapy drugs as a CDx," the spokesperson said. "This could be a place where an effort to standardize as much as possible across therapeutic agents and possibly across tissue would be of high benefit to prescribers and patients."
Personal Genome Diagnostics, a JHU spinout that Diaz co-founded, has been offering microsatellite instability analysis as part of a 125-gene tissue-based NGS test and more recently, as part of a circulating tumor DNA test that analyzes 64 genes in plasma samples. John Simmons, director of translational science and diagnostics, added that PGDx is planning to take a tissue-based NGS microsatellite instability assay through FDA later this year. The company also recently received nearly $223,000 in grant funding from the National Cancer Institute that it will use to advance a liquid biopsy tests for determining tumor mutational burden.
Guardant Health is another firm developing a 500-gene liquid biopsy test that can be used in drug development and clinical trials, and in particular will be able to gauge tumor mutational burden for personalizing immunotherapies. The company has partnerships with a number of drug companies, including Merck, and has said it will launch the panel in mid-2017.
The latest tissue-agnostic indication for Keytruda is just the start, Diaz predicted. Microsatellite instability and mismatch repair deficiency might predict response to other PD-1 or PD-L1 inhibitors, though he cautioned that this would need to be proven in clinical studies. "Like a domino effect, we'll start seeing more and more of these," he said.