FDA Panel Votes for Limits on Some GI Cancer Immunotherapy

Late last week, a US Food and Drug Administration (FDA) panel met to discuss whether to limit the use of nivolumab (Opdivo) and pembrolizumab (Keytruda) in patients with unresectable or metastatic esophageal squamous cell carcinoma and human epidermal growth factor receptor 2 (HER2)–negative, microsatellite stable gastric/gastroesophageal adenocarcinoma.

During the meeting, the FDA’s Oncologic Drugs Advisory Committee (ODAC) voted in favor of restricting the use of these immunotherapy agents to patients with programmed death-ligand 1 (PD-L1) expression of 1% or higher. 

The agency usually follows the ODAC’s advice.

The FDA had originally approved the two immune checkpoint inhibitors for both indications in combination with chemotherapy, regardless of patients’ PD-L1 status. The FDA had also approved nivolumab in combination with ipilimumab for esophageal cancer, regardless of PD-L1 expression. The approvals were based on overall benefit in intent-to-treat populations, not on specific PD-L1 expression subgroups.

Since then, additional studies — including the agency’s own pooled analyses of the approval trials — have found that overall survival benefits are limited to patients with PD-L1 expression of 1% or higher.

These findings have raised concerns that patients with no or low PD-L1 expression face the risks associated with immunotherapy, which include death, but minimal to no benefits.

In response, the FDA has considered changing the labelling for these indications to require a PD-L1 cutoff point of 1% or higher. The move would mirror guidelines from the American Society of Clinical Oncology and the National Comprehensive Cancer Network that already recommend use with chemotherapy only at certain PD-L1 cutoffs.

Before the agency acts, however, the FDA wanted the advice of the ODAC. It asked the 14-member panel whether the risk-benefit assessment is “favorable for the use of PD-1 inhibitors in first line” for the two indications among patients with PD-L1 expression below 1%.

In two nearly unanimous decisions for each indication, the panel voted that risk-benefit assessment was not favorable. In other words, the risks do outweigh the benefits for this patient population with no or low PD-L1 expression.

The determination also applies to tislelizumab (Tevimbra), an immune checkpoint inhibitor under review by the FDA for the same indications.

Voting came after hours of testimony from FDA scientists and the three drug companies involved in the decisions.

Merck, maker of pembrolizumab, was against any labelling change. Nivolumab’s maker, Bristol Myers Squibb (BMS), also wanted to stick with the current PD-L1 agnostic indications but said that any indication change should be class-wide to avoid confusion. BeiGene USA, maker of tislelizumab, had no problem with a PD-L1 cutoff of 1% because its approval trial showed benefit only in patients at that level or higher.

In general, Merck and BMS said the drug benefits correspond with higher PD-L1 expression but noted that patients with low or no PD-L1 expression can sometimes benefit from treatment. The companies had several patients testify to the benefits of the agents and noted patients like this would likely lose access. But an ODAC panelist noted that patients who died from immunotherapy complications weren’t there to respond.

The companies also expressed concern about taking treatment decisions out of the hands of oncologists as well as the need for additional biopsies to determine if tumors cross the proposed PD-L1 threshold at some point during treatment. With this potential new restriction, the companies were worried that insurance companies would be even less likely to pay for their checkpoint inhibitors in low or no PD-L1 expressors.

ODAC wasn’t moved. With consistent findings across multiple trials, the strength of the FDA’s data carried the day.

In a pooled analysis of the three companies’ unresectable or metastatic HER2–negative, microsatellite-stable gastric/gastroesophageal adenocarcinoma approval trials across almost 4000 patients, those with PD-L1 levels below 1% did not demonstrate a significant overall survival benefit (hazard ratio [HR], 0.91; 95% CI, 0.75-1.09). The median overall survival with immunotherapy plus chemotherapy was only 1 month more — 13.4 months vs 12.4 months with chemotherapy alone.

FDA’s pooled analysis for unresectable or metastatic esophageal squamous cell carcinoma also showed no overall survival benefit (HR, 1.1; 95% CI, 0.76-1.58), with a trend suggesting harm. Median overall survival with immunotherapy plus chemotherapy was 14.6 months vs 9.8 months with chemotherapy alone.

Despite the vote on esophageal squamous cell carcinoma, panelists had reservations about making decisions based on just over 160 patients with PD-L1 levels below 1% in the three esophageal squamous cell carcinoma trials.

Still, one panelist said, it’s likely “the best dataset we will get.”

The companies all used different methods to test PD-L1 levels, and attendees called for a single standardized PD-L1 test. Richard Pazdur, MD, head of the FDA’s Oncology Center of Excellence, said the agency has been working with companies for years to get them to agree to such a test, with no luck.

If the FDA ultimately decides to restrict immunotherapy use in this patient population based on PD-L1 levels, insurance company coverage may become more limited. Pazdur asked the companies if they would be willing to expand their patient assistance programs to provide free coverage of immune checkpoint blockers to patients with low or no PD-L1 expression.

BeiGene and BMS seemed open to the idea. Merck said, “We’ll have to…think about it.”

M. Alexander Otto is a physician assistant with a master’s degree in medical science and a journalism degree from Newhouse. He is an award-winning medical journalist who worked for several major news outlets before joining Medscape Medical News. Email: aotto@mdedge.com.