FDA Restricts Pembrolizumab and Nivolumab for PD-L1-Low Gastric Cancer

The U.S. Food and Drug Administration (FDA) has imposed new restrictions on the use of pembrolizumab (Keytruda) and nivolumab (Opdivo) for patients suffering from gastric and gastroesophageal cancers with low PD-L1 expression levels, specifically those with combined positive scores (CPS) of less than 1. This decision, announced on June 22, 2025, comes after analyses from pivotal phase 3 clinical trials—CheckMate 649 and KEYNOTE-859—indicated no significant difference in overall survival (OS) between patients receiving these immunotherapies and those undergoing standard chemotherapy treatments.

The restrictions were communicated to the drug manufacturers in official letters from the FDA, which also mandated additional studies to validate diagnostic testing for PD-L1 expression to further clarify treatment eligibility. The FDA’s Oncologic Advisory Drug Committee (ODAC) previously voted against the use of PD-L1 expression as a predictive biomarker for gastric cancers in a 10-to-2 decision, citing the lack of an observable OS benefit in patients with low PD-L1 scores.

Dr. Hanna K. Sanoff, a professor of Gastrointestinal Oncology at UNC Lineberger Comprehensive Cancer Center and an ODAC member, stated, "I hope we can see how this evolves and how we can get immunotherapies to be effective in this PD-L1-negative population. Until we do that, I did not see enough evidence that we are helping people and not harming them." This sentiment underscores the ongoing debate within the oncology community regarding the efficacy and safety of immunotherapy in subgroups of patients with low PD-L1 expression.

Data from the CheckMate 649 trial, involving 1581 patients, illustrated that median OS for those treated with nivolumab combined with chemotherapy was 13.1 months, compared to 12.5 months for chemotherapy alone in the PD-L1–low group. Furthermore, among those with a PD-L1 CPS of 5 or greater, the median OS was significantly better at 14.4 months with nivolumab versus 11.1 months with a placebo (hazard ratio [HR] of 0.71, 95% confidence interval [CI] of 0.61-0.83; p < 0.0001).

In a similar vein, the KEYNOTE-859 trial, which enrolled 1579 patients, reported a median OS of 12.7 months for those receiving chemotherapy with pembrolizumab, while those on chemotherapy alone had a median OS of 12.2 months. This data highlights that the combination therapy may be beneficial in patients with higher PD-L1 scores.

The FDA’s decision reflects a growing emphasis on the need for more tailored treatment protocols in oncology, particularly as clinical evidence continues to evolve. As research progresses, the industry anticipates that clearer guidelines will emerge regarding the appropriate use of PD-L1 inhibitors in gastric cancer treatment.

In conclusion, while the introduction of pembrolizumab and nivolumab represented significant advancements in cancer treatment, the current limitations underscore the complexities involved in immunotherapy efficacy, particularly in subpopulations defined by PD-L1 expression levels. The FDA’s decision may prompt further research and discussions on the role of biomarkers in guiding therapeutic strategies for gastric and gastroesophageal cancers.