Experts Support FDA’s Plan for Restrictions on PD-1 Drugs in Gastric, Esophageal Cancers

On September 26, 2024, the Oncologic Drugs Advisory Committee (ODAC) convened to discuss the use of immune checkpoint inhibitors, specifically PD-1 inhibitors like pembrolizumab (Keytruda) and nivolumab (Opdivo), in the treatment of gastric and esophageal cancers. The discussion centered around the efficacy of these drugs in patients with varying levels of PD-L1 expression and whether continued use in the broader population was justified, given emerging data. 

“Very commonly we’ll use biomarkers to try to tell us if we can add on additional targeted therapies,” Dr. Nicholas Hornstein, medical oncologist at Northwell Health, told SurvivorNet during an in-depth discussion on gastric cancer treatment. “Some of these targeted therapies [include] trastuzumab or anti-HER2 or anti-PD-L1 therapy such as pembrolizumab.

“There’s been some controversy recently for which patients qualify for these treatments, especially pembrolizumab, which was previously given to any patient receiving anti-HER2 therapy and now is only given to patients where their tumor cells express PD-L1. On some level, this is a somewhat complicated topic,” Dr. Hornstein added.

PD-1 inhibitors have been approved for the treatment of several cancers, including metastatic or unresectable gastric and gastroesophageal junction (GEJ) cancers. As Dr. Hornstein mentioned, initially, these drugs were approved for patients regardless of PD-L1 expression, a marker that indicates how likely a patient is to benefit from immune checkpoint inhibition.

Trial Data Leading to FDA Review

Key studies were central to the FDA’s review: the CheckMate-649 (CM-649), KEYNOTE-859 (KN-859), and RATIONALE-305 (RN-305) trials. These randomized trials examined the efficacy of PD-1 inhibitors combined with chemotherapy in patients with advanced gastric and esophageal adenocarcinomas. They stratified patients by PD-L1 expression levels, using different assays and cutoffs for PD-L1 positivity, such as CPS ≥1 or CPS ≥10 (combined positive score). Results consistently showed that higher PD-L1 expression was correlated with better outcomes, while lower or no PD-L1 expression led to marginal or no benefit.

CheckMate-649: Nivolumab in Gastric Cancer

The CheckMate-649 trial evaluated nivolumab with chemotherapy versus chemotherapy alone in patients with PD-L1 CPS ≥1 and ≥5. While patients with higher PD-L1 expression (CPS ≥10) experienced significant improvements in overall survival (OS), patients with lower expression (CPS <5) saw little benefit. This raised questions about the benefit-risk profile for nivolumab in the broader gastric cancer population.

KEYNOTE-859: Pembrolizumab in Gastric Cancer

The KEYNOTE-859 trial similarly found that pembrolizumab, when added to chemotherapy, was most effective in patients with high PD-L1 expression (CPS ≥10). Subgroup analyses suggested marginal efficacy in patients with lower PD-L1 levels (CPS <1), leading to uncertainty about its use in this population. Notably, the European Medicines Agency (EMA) already limits pembrolizumab to patients with CPS ≥1, a more restrictive guideline than in the U.S.

RATIONALE-305: Tislelizumab’s Efficacy

The RATIONALE-305 trial, which studied tislelizumab, a newer PD-1 inhibitor, used a different assay (Tumor Area Positivity) to determine PD-L1 expression. Again, results demonstrated that patients with higher PD-L1 expression (TAP ≥10) derived the most significant survival benefit, while those with lower expression (TAP <1) showed minimal improvement. These findings were consistent with those seen in the other trials.

What Was Decided?

The FDA presented analyses showing that patients with low or no PD-L1 expression derive little benefit from these drugs, while still being exposed to significant risks, including immune-related adverse events such as pneumonitis, colitis, and endocrinopathies.

An important theme of the meeting was the role of PD-L1 as a predictive biomarker. While some patients with low PD-L1 expression responded to treatment, these cases were rare, and the overall data suggest that PD-L1 expression should guide the use of PD-1 inhibitors. The FDA’s exploratory analyses indicated that the greatest magnitude of benefit was seen in patients with CPS ≥10 or TAP ≥10, across all studies.

If the FDA decides to restrict the use of PD-1 inhibitors based on PD-L1 expression, it could significantly alter the treatment landscape for gastric and esophageal cancers. Physicians would need to ensure that PD-L1 testing is performed routinely, and patients with low expression may no longer be candidates for these therapies. This shift would align the U.S. with European guidelines and help avoid exposing patients to potentially ineffective treatments.

A key concern raised by the FDA was the potential for harm in patients with low PD-L1 expression. These patients are less likely to benefit from PD-1 inhibitors but still face the risk of severe adverse events. The agency also emphasized that the continued use of these drugs in the broader population could lead to unnecessary healthcare costs and patient morbidity without corresponding survival benefits.

One challenge highlighted in the meeting was the variability in PD-L1 testing methods. Different trials used different assays (e.g., Dako 28-8 for nivolumab, 22C3 for pembrolizumab, and SP263 for tislelizumab), which complicates direct comparisons of results. The FDA called for standardization of testing to ensure consistency in patient selection for PD-1 therapies.

Future Directions and Trials

The committee discussed the need for future clinical trials to focus on biomarker-selected populations, especially those with high PD-L1 expression. Trials that combine PD-1 inhibitors with other therapies, such as anti-CTLA-4 antibodies or targeted therapies, may also be necessary to improve outcomes for patients with low PD-L1 expression.

For patients with gastric or esophageal cancer, these changes could mean that access to PD-1 inhibitors would be limited to those with higher PD-L1 expression. While this could reduce the use of these drugs in some patients, it would also prevent unnecessary side effects in those unlikely to benefit, ultimately improving overall treatment outcomes.