PD-L1 Test Debated in Gastroesophageal Cancer Immunotherapy

Laird Harrison

April 20, 2022

Clinicians are struggling to make sense of conflicting regulations from the Food and Drug Administration and the European Medicines Agency on whether measuring programmed death–ligand 1 (PD-L1) is essential before prescribing checkpoint inhibitors for gastroesophageal cancer.

"In the last couple of years, the incorporation of PD-1 antibodies is really changing our standard of care and international guidelines in this disease," said Ian Chau, MD, a consultant medical oncologist at the Royal Marsden Hospital, London.

He moderated a debate at the 2022 Gastrointestinal Cancers Symposium on the importance of measuring PD-L1 expression levels before administering immune checkpoint inhibitor therapy.

Tumor cells can use PD-1 signaling to deactivate the response of T cells that would otherwise destroy them, and several new drugs are designed to block that signaling.

Multiple randomized controlled trials have shown the benefit of adding such checkpoint inhibitors to chemotherapy for gastroesophageal cancer and currently, chemotherapy plus a checkpoint inhibitor is standard care, Chau said.

PD-1–blocking antibodies include pembrolizumab (Keytruda, Merck) for colorectal cancer, gastric cancer, esophageal cancer, hepatocellular carcinoma, and renal cell carcinoma, among other cancers. And, nivolumab (Opdivo, Bristol-Myers Squibb) approved for renal cell carcinoma, colorectal cancer, hepatocellular carcinoma, and esophageal squamous cell carcinoma, among other cancers.

Regulators differ on whether these treatments should be limited to patients whose expression level of PD-L1 reaches a defined threshold. The FDA has not required the measurement of any biomarker before starting therapy with either of these drugs. However, the EMA requires a PD-L1 combined positive score (CPS) of at least 10 for pembrolizumab and at least 5 for nivolumab.

In a poll conducted before the start of the debate, 83% of physician attendees said they favor of the EMA position, while 17% disagreed.

Florian Lordick, MD, PhD, director of the University of Leipzig (Germany) Cancer Center, argued that tests for PD-L1 expression are accurate. About half of patients have CPS of at least 1. And, pathologists are in agreement in interpreting the tests about 97% of the time.

Pivotal Decision-Making Clinical Trials

The EMA requires a PD-L1 assay primarily based on its interpretation of the data from KEYNOTE-590 and CheckMate-648.

KEYNOTE-590 included 749 patients with esophageal carcinoma who were randomized to receive either pembrolizumab with standard of care chemotherapy, or placebo and standard of care chemotherapy. Patients receiving pembrolizumab who had PD-L1 CPS scores of 10 or more survived a few months longer on average. But in a post hoc analysis, the investigators found that, in patients with PD-L1 CPS scores less than 10, the difference between the treatment and placebo groups was not statistically significant.

In CheckMate-648, 970 patients with esophageal carcinoma were randomized to receive nivolumab with ipilimumab, nivolumab with chemotherapy, or chemotherapy alone. Investigators used a slightly different measurement of PD-L1 tumor proportion score (TPS), in comparing chemotherapy plus nivolumab to chemotherapy alone for advanced esophageal squamous cell carcinoma.

The median survival time of patients with TPS of at least 1% was 15.4 months in the nivolumab group and 9.2 months in the control group. But among patients with a TPS of less than 1%, the median overall survival was 12.0 months in the nivolumab group and 12.2 in the control group. CPS thresholds of 5 or 1 resulted in similar effects.

Lordick cited a systematic review of four studies in press at ESMO Open. Hazard ratios for three of the studies favored immunotherapy only in patients with CPS of at least 10.

Deciding which patients to treat matters because these drugs are expensive, he said. A single dose of 240 mg nivolumab costs $7,228.70, and treatment for 1 year can cost $173,488.80 in addition to costs for hospitalization because of immune-related adverse events, labs, imaging, colonoscopy, and other related costs.

"This is a lot of money, isn't it? It's the same price the hospital pay for two registered nurses in the U.S., at least when we are talking about the average price. I'm not sure I want to spend this money for a drug that does not work," Lordick said.

Aaron Scott, MD, an oncologist with the University of Arizona Cancer Center, Tucson, said that "patients and clinicians want and need options" because there may be other factors that should be considered.

"PD-L1 has shown inconsistent results. And while I agree it is the best that we have for predictive biomarker in the space, it is far from perfect. PD-L1 has not been predictive for response in a variety of settings and trials. In first-line, second-line, third-line trials we have examples where it does not predict response," he said.

JUPITER06 compared toripalimab or placebo with paclitaxel and cisplatin for patients with esophageal squamous cell carcinoma. Patients who received toripalimab lived longer than patients who received the placebo, but within the toripalimab group, there was no difference in median overall survival between those patients above and those below the threshold of CPS 1.

ORIENT-15 compared sintilimab or placebo with chemotherapy as first-line therapy for patients with esophageal squamous cell carcinoma. Although the treatment group fared better, survival rates were the same whether patients were above or below the CPS 10 threshold.

Scott cited three other trials in which PD-L1 was not predictive of response to checkpoint inhibitors.

The differences among studies could be attributed to different assays, he said. "Where you biopsy and when you biopsy seems to matter."

In a second opinion poll conducted after the presentations, the proportion of physician attendees saying PD-L1 was essential before initiating checkpoint inhibitors, dropped to about two-thirds.

Chau reported financial relationships with Bristol-Myers Squibb, Merck Serono, and other pharmaceutical companies. Lordick reported financial relationships Bristol-Myers Squibb, Merck Sharp & Dohme, Merck, and other pharmaceutical companies.

This article originally appeared on MDedge.com, part of the Medscape Professional Network.


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.