After an embarrassing admonition from the FDA, Merck & Co. has returned with updated Keytruda data it hopes will change the regulators’ mind on approval in early triple-negative breast cancer (TNBC).
The use of Keytruda around surgery pared down the risk of disease worsening, cancer returning or death by 37% in patients with early-stage, nonmetastatic TNBC, Merck unveiled at a European Society for Medical Oncology virtual plenary session.
Keytruda hit this so-called event-free survival (EFS) goal when coupled with chemotherapy before surgery and used alone after surgery in the phase 3 KEYNOTE-522 trial. The improvement was seen over the control group of patients, who only got chemo before surgical removal of tumors and placebo post-surgery.
FDA previously rejected Merck’s application for Keytruda’s use in TNBC patients before and after surgery, or in the neoadjuvant and adjuvant settings, partly because the event-free survival data were immature when Merck filed for approval with data from its third interim analysis of the KEYNOTE-522 study. But now, after a median follow-up of 39 months, the trial has passed the statistical significance bar.
Among the 1,174 patients enrolled in the study, 15.7% of patients in the Keytruda arm and 23.8% of the control group experienced disease progression, cancer recurrence, a second primary cancer or death. Median even-free survival wasn’t reached in either group.
Merck has filed the data with the FDA, and the company believes they address the previous complete response letter, Roy Baynes, chief medical officer of Merck Research Laboratories, said in an interview ahead of the data presentation.
The FDA took issue with the initial application, criticizing Merck for jumping the gun not just because it lacked a clear signal for EFS benefits. Its data on a surrogate marker for clinical benefit known as pathologic complete response (pCR)—defined as the absence of residual cancer cells from tissue samples after treatment—were weak in the eyes of the FDA. PCR remains a controversial marker with limited evidence that it can predict patient survival.
Merck had shown that adding Keytruda to chemotherapy before surgery could help more patients achieve no invasive residual cancer in breast or lymph nodes at time of surgery. At the first and prespecified primary analysis for that pCR marker, the absolute difference between the two arms was 13.6 percentage points in favor of Keytruda.
Baynes noted that Keytruda’s pCR rate, at 64.8%, was the highest reported for a treatment in neoadjuvant TNBC, and that its advantage over control was statistically significant. But the FDA had told the company before the original filing that it would like to see a larger pCR benefit to consider an accelerated approval. What’s more, the agency noted that Keytruda’s pCR margin over placebo had dropped to 7.5 points at the third interim analysis when all 1,174 enrolled patients were counted.
Still, as Baynes pointed out, pCR was just a surrogate trial marker. “The important outcome here is ultimately, does the patient progress or not? And we have shown very clearly that this [Keytruda] regimen reduces the risk of progression,” he said in the interview.
But Keytruda’s declining pCR advantage raises the possibility of a similar pattern for EFS as more patients experience cancer progression or recurrence. At least KEYNOTE-522’s EFS data have remained stable thus far, with the risk reduction at 37% and 35%, respectively, at the second and third interim analyses.
Keytruda appeared to have benefited patients regardless of whether they achieved a pCR, and Merck saw “really no difference” in efficacy between patients with or without PD-L1 expression in their tumors, Baynes said.
A conditional nod doled out in November allows Keytruda to be used alongside chemo in metastatic TNBC but only in patients with PD-L1 levels at a combined positive score of 10 or higher.
Another concern the FDA has simply can’t be resolved with more KEYNOTE-522 data: The agency doesn’t like the trial’s continuous neoadjuvant and adjuvant design, because it doesn’t tease out the contribution of each treatment phase to EFS. Past trials and drug applications have separated filings for neoadjuvant and adjuvant uses. Based on the discussion during an advisory committee meeting in February around additional side effects Keytruda causes, the FDA appears to lean toward limiting the PD-1 inhibitor’s use.
Nevertheless, Merck “may submit” another application when it has further EFS data based on the same KEYNOTE-522 trial, an official with the FDA said at the meeting.
The ball is now back in the FDA’s court. If the agency approves Keytruda for the entire neoadjuvant-adjuvant regimen without requirement for further trials, it could be a relief for other immuno-oncology agents running similarly styled studies, at least from a regulatory point of view.