The FDA is once again gathering external experts to scrutinize PD-1 inhibitors as a group. This time, the agency is considering limiting these immunotherapies’ use in stomach cancer and esophageal cancer.
The FDA will hold an Oncologic Drugs Advisory Committee (ODAC) meeting Sept. 26 to consider whether approvals for checkpoint inhibitors in advanced gastric and gastroesophageal junction (GEJ) adenocarcinoma—as well as in esophageal squamous cell carcinoma (ESCC)—should be restricted based on tumors’ PD-L1 expression, the agency said in a public filing (PDF).
Existing approvals for Bristol Myers Squibb’s Opdivo and anti-CTLA4 agent Yervoy, plus Merck & Co.’s Keytruda, will be discussed. BeiGene’s two open applications for Tevimbra will also face scrutiny.
“Cumulative data have shown that PD-L1 expression appears to be a predictive biomarker of treatment efficacy in this patient population,” the FDA said in its filing. The agency noted that “clinical trials have used different approaches to assess PD-L1 expression and different thresholds to define PD-L1 positivity.”
PD-1 inhibitors are known to work better in tumors that express PD-L1, and their efficacy in stomach cancer has been less than impressive. In gastric cancer, the drugs have generally been shown to reduce patients’ risk of death by about 20% versus chemo-based regimens.
While the upcoming ODAC meeting will focus on HER2-negative stomach cancer, Merck last year voluntarily asked the FDA to narrow Keytruda’s previous accelerated approval in first-line HER2-positive gastric and GEJ cancer to PD-L1-positive tumors only. The FDA followed suit and revised the drug’s indication in November.
At an interim analysis of the phase 3 KEYNOTE-811 trial, Keytruda’s combination with Roche’s Herceptin and chemotherapy pared down the risk of progression or death by about 30% compared with Herceptin-chemo alone in HER2-positive PD-L1-positive cases, as defined by a combined positive score (CPS) of at least 1.
In a small subset of patients with PD-L1-negative disease, the Keytruda regimen showed no tumor progression advantage but a potential detriment to patients’ lives. In that group, investigators noticed a 41% increased risk of death at a later data cut.
The trial eventually met statistical significance on overall survival in the entire trial population—regardless of PD-L1 expression—but the poor performance in PD-L1-negative disease was impossible to ignore.
Just a week after the FDA restricted Keytruda’s label to PD-L1-positive disease in HER2-positive gastric and GEJ cancer, the agency in November approved Keytruda’s use alongside chemo for HER2-negative disease without constraints related to PD-L1 expression based on results from the KEYNOTE-859 trial.
In that study, Keytruda and chemo pared down the risk of death by 22% and improved median overall survival by just 1.4 months versus chemo alone in HER2-negative disease. A subgroup analysis showed that in patients with PD-L1-negative tumors, the death risk reduction was low at 8%.
Because the overall trial result met the statistical significance bar, and the PD-L1-negative group’s performance trended in the right direction, the FDA handed Keytruda a broad indication. Now, the agency apparently is having second thoughts.
In a statement to Fierce Pharma, a Merck spokesperson said Keytruda plays “an important role” in helping patients with certain advanced gastric and esophageal cancers.
“We look forward to a productive conversation with the FDA’s ODAC regarding the role of PD-L1 expression as a predictive biomarker of treatment efficacy in patients with certain gastric and esophageal cancers,” the spokesperson added.
A similar pattern of smaller benefit in PD-L1-negative patients has been seen in studies of BMS’ Opdivo and BeiGene’s Tevimbra.
In the phase 3 RATIONALE-305 trial in first-line gastric and GEJ cancer, the addition of Tevimbra to chemo reduced the risk of death by 29% at final analysis in PD-L1-positive patients, as defined by a tumor area positivity (TAP) score of 5%. In PD-L1-negative patients, the size of improvement fell to just 8% versus chemo alone. The death-risk reduction in the all-comers population, at 20%, met statistical significance.
As for the CheckMate-649 trial for Opdivo in previously untreated gastric, GEJ and esophageal adenocarcinoma, a three-year analysis found the drug only reduced the risk of death by a mere 5% in PD-L1-negative patients, as defined by a CPS score below 1. In PD-L1-positive patients, the survival improvement was 25%. Similarly, the overall survival data in all randomized patients was statistically significant.
Merck, BMS and BeiGene conducted separate trials for their respective meds in ESCC. Earlier this month, BeiGene said an FDA decision for its first-line ESCC application was postponed because of a delay in scheduling clinical site inspections. Now, it looks like the FDA wants expert input on whether the drug deserves a broad label.
In a statement to Fierce Pharma, a BeiGene spokesperson confirmed that the company is seeking FDA approvals in all-comers populations for both the gastric/GEJ cancers and ESCC applications and that it is “committed to working with the agency to bring Tevimbra to patients in need.”
As to why BeiGene picked the TAP 5% score for the PD-L1 stratification, the spokesperson said the cutoff was derived from an analysis of an early-phase, single-arm trial of Tevimbra monotherapy in second- or later-line treatment of gastric or esophageal cancer.
Opdivo in combination with either chemo or Yervoy scored full FDA approval in first-line ESCC in 2022 based on all-comers overall survival results from the phase 3 CheckMate-648 trial.
“With proven survival benefits, Opdivo-based regimens have changed the outlook for patients with GC, GEJC, EAC and ESCC regardless of PD-L1 status,” Ian Waxman, M.D., BMS’ senior global program lead of oncology late development, said in a statement. “We look forward to the opportunity to discuss the importance of Opdivo-based regimens as treatment options for appropriate gastric and esophageal cancer patients more in depth with the Committee.”
This isn’t the first time the FDA has convened an ODAC to zero in on checkpoint inhibitors. At a meeting in 2021, the agency scrutinized several PD-1/L1 accelerated approvals that hadn’t met their confirmatory trial commitments. Some of the indications were later withdrawn or narrowed.