乌鸦传媒

WASHINGTON -- The FDA should wait for more data before approving pembrolizumab (Keytruda) as neoadjuvant therapy for high-risk triple-negative breast cancer (TNBC), an advisory committee concluded.

By a unanimous 10-0 vote, the Oncologic Drugs Advisory Committee (ODAC) essentially sided with FDA staff members who expressed doubt about the existing strength of evidence from an ongoing phase III clinical trial. In a report submitted to ODAC, FDA staff concluded that the magnitude of improvement in the co-primary endpoint of pathologic complete response (pCR) with a pembrolizumab-containing regimen, combined with evidence from the metastatic setting, "are not supportive of clinical benefit in the neoadjuvant setting."

Data for the co-primary outcome of event-free survival (EFS) remain immature, as do data for the secondary outcome of overall survival (OS).

Possibly summing up the sentiments of the entire committee, Deborah Armstrong, MD, of Johns Hopkins Medicine in Baltimore, said, "I hope that this is a positive study, but at this point in time, it's premature to start treating patients with this therapy."

ODAC chair Philip Hoffman, MD, of the University of Chicago, added, "I think that many, if not most, of us believe that this trial may well turn out to be positive, as time goes on with further analyses and as we see what the effect of the pathologic CR rate does in fact lead to. But I think there certainly is consensus among the committee that hasn't yet been demonstrated sufficiently to warrant going ahead with full FDA approval right now."

The meeting set a precedent for Hoffman's 2-year tenure as ODAC chair. Prior meetings involved consideration of a first approval for a cancer therapy. In contrast, the committee was convened this time to consider a new indication for an approved drug, and to base a decision on data from interim analyses of an ongoing trial.

No drug in the checkpoint inhibitor class has received a breast cancer indication outside the metastatic setting. Atezolizumab (Tecentriq) has an indication for metastatic TNBC, based on results of the IMpassion130 trial, and pembrolizumab recently received accelerated approval for the same indication. The KEYNOTE-522 trial sought to push the checkpoint inhibitor envelope into the neoadjuvant setting.

The trial included 1,174 patients with untreated, high-risk, early-stage TNBC. They were randomized 2:1 to receive neoadjuvant therapy with carboplatin-paclitaxel chemotherapy plus pembrolizumab or placebo, followed by pembrolizumab or placebo as adjuvant therapy. The trial had co-primary endpoints of pCR rate and EFS.

Results for pCR were unblinded after an interim analysis showed a statistically significant absolute difference of 13.6% in favor of the pembrolizumab arm. Follow-up in the trial continues for EFS and OS. In the meantime, the pCR difference has shrunk to 7.5%.

In a submission to ODAC, Merck representatives expressed confidence that the difference in pCR would eventually translate into a survival advantage for patients who received the pembrolizumab regimen. During the meeting, considerable discussion revolved around an for assessing endpoints such as EFS on the basis of pCR results. Although the meta-analysis was limited to neoadjuvant chemotherapy, the findings from KEYNOTE-522 appear to fall within ranges established by the meta-analysis, according to Sunita Zalani, PhD, of Merck.

"We don't know what magnitude of pCR benefit will lead to a benefit in a clinical endpoint such as EFS, particularly for immunotherapy," Zalani said. "However, what we're seeing here is based on early data for a benefit which appears to exceed what we would expect from pCR, based on the FDA meta-analysis, which is consistent with the mechanisms of action of immunotherapy, where frequently the effects on long-term outcomes are not completely captured in response data."

The most recent results showed about a 1% incidence of fatal adverse events (AEs) in the pembrolizumab arm, and immune-related AEs did not resolve in many instances, even after discontinuation of pembrolizumab. Several ODAC panelists referenced the data in expressing their support for putting off the decision until data on EFS and OS are clearer.

The uncertainty surrounding the data was compounded by the absence of a reliable biomarker for patients with TNBC, said ODAC panelist Matthew Ellis, MD, of Baylor College of Medicine in Houston.

"This is not a disease that we understand well, and I think we need to go back to the drawing board, to some extent, and work hard on molecular profiling techniques that might get us to a better place," said Ellis, himself a cancer survivor.

"I too hope that the next analysis produces convincing results," he added. "I don't think we should approve drugs based on projections. We need solid evidence to prescribe drugs."

Comment