FDA Narrows Use of PD-1 Drugs in GI Cancers

Highlights

• The FDA has restricted the use of Merck's Keytruda and BMS's Opdivo to PD-L1-positive patients for gastric, GEJ, and esophageal cancers.
• Combination therapy of Opdivo and Yervoy for esophageal squamous cell carcinoma is also now limited to PD-L1-positive disease.
• The decision follows analysis showing minimal benefit for PD-L1-negative subgroups in phase 3 trials.
• BeOne's Tevimbra was the first PD-1 therapy in this category to receive a PD-L1-specific approval.
• The move continues a pattern of the FDA reevaluating broad checkpoint inhibitor approvals in light of emerging efficacy data.

Regulatory Shift in Immunotherapy Indications

In a significant policy shift, the U.S. Food and Drug Administration (FDA) has officially restricted the use of two leading PD-1 inhibitors—Merck & Co.'s Keytruda and Bristol Myers Squibb's (BMS) Opdivo—for certain gastrointestinal cancers. This move follows an intensive regulatory review that cast doubt on the benefits of PD-1 inhibitors for patients lacking PD-L1 expression.

According to the FDA, BMS's Opdivo, when used in combination with chemotherapy, is now only approved for advanced or metastatic gastric, gastroesophageal junction (GEJ), and esophageal cancers in patients whose tumors express PD-L1. In addition, the label for the Opdivo-Yervoy combination in first-line esophageal squamous cell carcinoma has been similarly narrowed to PD-L1-positive disease.

Merck's Keytruda has undergone comparable restrictions. Its use in gastric, GEJ, and esophageal cancers is now limited to patients with PD-L1-positive tumors. The label change specifically affects HER2-negative gastric or GEJ adenocarcinomas, while HER2-positive indications had already been revised in March to require PD-L1 expression.

Scientific Basis for Label Updates

The FDA's actions are grounded in recent trial data and risk-benefit analyses. In the phase 3 Keynote-859 study, Keytruda combined with chemotherapy demonstrated statistically significant overall survival benefits in a broad patient group with HER2-negative gastric or GEJ adenocarcinoma. However, an exploratory subset analysis revealed only an 8% reduction in death risk for PD-L1-negative patients—prompting the agency to conclude that the overall benefit stemmed largely from PD-L1-positive cases.

The FDA noted in correspondence with the companies that, “the revised indications reflect the patient populations for whom the drugs have favorable risk-benefit assessments.” These findings echo concerns raised by the agency in 2023 regarding the use of checkpoint inhibitors in PD-L1-negative gastrointestinal cancers.

Industry Response and Regulatory Precedent

Despite initial resistance from Merck and BMS, both companies submitted revised indications in February 2025. “Then Merck and BMS, despite having argued for continued broad labels at the advisory committee meeting, filed revisions to their existing indications late February,” the FDA said.

This decision follows a similar outcome for BeOne's Tevimbra, which became the first PD-1 inhibitor restricted to PD-L1-positive patients in this setting when it was approved in December 2024 for use with chemotherapy in first-line HER2-negative gastric or GEJ cancers.

The FDA's updated view had already been backed by an external advisory committee in September 2024, which voted overwhelmingly in favor of limiting PD-1 inhibitor use to PD-L1-positive patients based on risk-benefit concerns.

Historical Context of PD-1 Label Restrictions

This is not the first instance of the FDA revisiting broad approvals for PD-1/L1 inhibitors. In 2018, the agency restricted the use of Keytruda and Roche's Tecentriq for certain bladder cancer patients not eligible for cisplatin-based chemotherapy after observing potential survival risks. In 2021, Keytruda's monotherapy indication in newly diagnosed bladder cancer was further limited to patients ineligible for any platinum-based chemotherapy.