Skip to main content
Premium Trial:

Request an Annual Quote

Basket Study Supported Keytruda Pan-Cancer Indication But Challenges Remain


PARIS – As part of the evidence supporting Keytruda's approval in a tumor-agnostic, biomarker-informed drug indication, Merck submitted data from a basket trial.

Although cancer researchers, drugmakers, and even the US Food and Drug Administration is enthusiastic about the use of basket trials ― an adaptive approach that allows the investigation of treatments by genomic signature instead of tumor site ― conducting them is challenging since the research enterprise in the US isn't set up to efficiently enable such investigations, said Roy Baynes, senior VP of global clinical development at Merck Research Laboratories, on Monday at an annual symposium hosted in Paris by Worldwide Innovative Networking (WIN) in Personalized Cancer Medicine. WIN is a nonprofit, nongovernmental entity involving 41 member organizations with the common goal of bringing precision oncology to patients globally.

In May, the FDA granted accelerated approval for Keytruda as a treatment 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. When the DNA repair mechanism is hobbled in this way, it results in the accumulation of mutations in repetitive DNA regions called microsatellites.

Merck decided to seek approval in this biomarker indication regardless of tumor type based on a study published two years ago by Johns Hopkins researchers led by Luis Diaz. They showed in a 41-patient study that colorectal cancer patients with mismatch repair deficiency responded remarkably well to Keytruda, while those lacking this deficiency didn't. Intriguingly, patients with mismatch repair deficient cancers of other types also responded well to the drug.

Diaz and others have highlighted the impact of that study as a win for academic research, particularly since no drug company wanted to fund the study, and investigators had to raise the funds, around $500,000, through philanthropy.

While the prospective data from Diaz's group were impressive, it wasn't all that was required for FDA approval, said Baynes, according to whom a lot more data had to be submitted and many more millions had to be spent before the agency gave its nod to Keytruda's pan-cancer indication. "It was actually a tour de force of effort," he said. Merck had to verify each data point, produce a dataset in a format that the FDA could look at, adjudicate an imaging endpoint, and initiate a randomized confirmatory study.

Among the evidence submitted to the agency from five single-arm studies involving around 150 patients, Merck had data from a basket trial, KEYNOTE-158, investigating a number of rare tumor types, and it included a cohort that was enrolling any non-colorectal cancer solid tumors characterized by high microsatellite instability.

"Now there is a challenge with conducting these types of trials," Baynes said. "If the prevalence of the [biomarker] is low, there will be an extraordinarily high treatment screen failure rate, and that's clearly frustrating to the patients, physicians, and the study apparatus." Between 2 percent and 16 percent of cancers display high microsatellite instability, and heavily treated tumors have a higher likelihood of displaying this hypermutated characteristic.

Even so, rare biomarkers will be especially difficult to find in rare tumor types, and the research enterprise in this country isn't set up to easily identify patients in this way. "For the most part cancer centers aren't organized by molecular biology," Baynes observed. "They are organized by tumor types. So, actually finding a champion to lead such an effort is not without its challenges."

Precision oncology approaches reveal rare, targetable tumor biomarkers, splintering traditional groupings based on the site where the cancer occurred. In reaction, drugmakers and cancer researchers have been employing clinical trial approaches, such as basket trials, which allow them to explore biomarker-drug hypotheses in a few patients with a molecular marker of interest, quickly retire arms with limited activity, and continue to study promising biomarker-driven indications in larger cohorts.

Though the FDA has been willing to accept data from basket studies as part of the evidence supporting premarket review of a therapy, drugmakers still have to submit data from traditional clinical trial approaches, especially if the treatment was approved on an accelerated timeline.

Baynes highlighted that another trial supporting Keytruda's pan-cancer approval was KEYNOTE-164, a single-arm study that looked at the efficacy of the drug in colorectal cancer patients with high microsatellite instability or mismatch repair deficiency who had failed prior treatments, and this analysis replicated the findings of the first study by Diaz and colleagues. 

Meanwhile, in order to convert from accelerated or conditional approval to full approval, Merck also had to conduct a confirmatory, randomized study. Baynes noted that KEYNOTE-177, a study randomizing microsatellite instability-high colorectal cancer patients to Keytruda or standard of care in the frontline setting, is currently under way.

This study may also be critical for the approval of this pan-cancer indication in Europe, where according to Baynes, "there has been less appetite for the acceptance of single-arm data." While this reluctance is changing among European regulators, payment bodies still want to see data from randomized studies.

The Joint Federal Committee of Physicians, Dentists, Hospitals, and Health Insurance Funds (GBA) is one such body that weighs in on whether a drug approved by the European Medicines Agency should be reimbursed. "The GBA, in the absence of control data will invariably have a finding of no benefit, and therefore will not reimburse," Baynes said. However, because many European countries reference the GBA, "this produces an enormous hurdle to getting important medications to patients in a timely fashion, and patients will probably have to await the results of the randomized trial," he added.

There is some data suggesting that patients with mismatch repair deficient or microsatellite instability-high tumors that have not be previously treated may not do as well on Keytruda as heavily treated tumors. KEYNOTE-177 will provide definitive evidence in first-line colorectal cancer, but due to extensive off-label use of the drug, Baynes noted it might be particularly challenging to enroll a randomized study. However, in KEYNOTE-177 patients in the standard therapy arm can cross over to Keytruda once they progress.

"I'm wondering if it's even ethical to randomize those patients, based on the data that has been published," said Jose Baselga, chief medical officer at Memorial Sloan Kettering Cancer Center. "The survival hasn't been reached as of follow up."

At MSKCC, researchers are using a next-generation sequencing panel of more than 400 genes to identify personalized treatment options for cancer patients, and the approach has also helped advance accrual in numerous basket studies.

One attendee at the WIN Symposium questioned whether FDA's willingness to approve the drug so quickly and in a novel indication had to do with the election of President Donald Trump, who has repeatedly denigrated the agency as slow and outdated, and has promised to dismantle regulations to expedite drug approvals.

On the contrary, the agency's statistics show it has greatly improved its review timelines over the years. In 2016, the agency's median review time was 10.1 months for standard applications and eight months for priority applications. In comparison, in 1993 review times were 20.8 months and 16.3 months, respectively. The FDA said it granted Keytruda priority review, which places the agency on a six-month clock to arrive at a decision.

Moreover, the agency has been particularly willing to accept data from single-arm studies for molecularly-guided cancer drugs, where randomized trials are difficult to enroll in rare patient subpopulations. "I'm not a politician, but the FDA has shown remarkable flexibility in working with the oncology community," replied Baynes, highlighting that the breakthrough therapy designation spearheaded by Richard Pazdur, director of the FDA's Oncology Center of Excellence, has helped to speed a lot of cancer drugs to patients.