SAN FRANCISCO – The immune checkpoint inhibitor pembrolizumab (Keytruda) may soon unseat chemotherapy as standard second-line therapy for certain advanced cancers of the esophagus or gastroesophageal junction, according to data from the global phase 3 KEYNOTE-181 trial.
“Patients with advanced esophageal cancer after first-line therapy have a poor prognosis and limited treatment options,” said lead investigator Takashi Kojima, MD, of the National Cancer Center Hospital East, Kashiwa, Japan. “Taxanes and irinotecan are commonly used after first-line chemotherapy; however, no overall survival benefit has been demonstrated for chemotherapy in a phase 3 study.”
The 628 patients in KEYNOTE-181 were randomly assigned to chemotherapy (paclitaxel, docetaxel, or irinotecan, left to investigator’s choice) or pembrolizumab, an antibody to programmed death 1 (PD-1). Currently, pembrolizumab is approved in the United States for use as third- or later-line therapy for gastric or gastroesophageal junction cancer that is positive for programmed death ligand 1 (PD-L1) as defined by a combined positive score (CPS) of 1 or greater, among many other indications.
Main trial results reported at the 2019 GI Cancers Symposium showed that among patients with high PD-L1 expression, defined by a CPS of 10 or higher, pembrolizumab reduced risk of death by about one-third, prolonging survival by 2.6 months. The difference met the predefined threshold for statistical significance.
There was a more modest, nonsignificant benefit among patients with tumors having squamous cell carcinoma histology and among the entire intention-to-treat population.
The rate of treatment-related adverse events of grade 3-5 was roughly half as high with pembrolizumab versus chemotherapy (18.2% vs. 40.9%).
“These data suggest that pembrolizumab should be considered a new standard of care for patients with PD-L1 CPS of 10 or greater metastatic esophageal cancer in the second-line setting,” Dr. Kojima concluded.
Implications for practice
“In the intention-to-treat population, the KEYNOTE-181 study failed to meet its primary endpoint of overall survival, so pembrolizumab is not indicated in unselected esophageal cancer patients,” said invited discussant Harry H. Yoon, MD, cochair of the Esophageal/Gastric Cancer Disease Group at the Mayo Clinic Cancer Center, Rochester, Minn.
For the patients with squamous histology, the negative findings are unlikely to be due to underpowering and may instead be related to the trial’s use of multiple primary endpoints, in his opinion. “Some may advocate using pembrolizumab off protocol [in this population], particularly for patients who cannot tolerate chemotherapy, because it is after all better tolerated than chemo. This can be a discussion point for guideline committees,” he said.
The results for the group with PD-L1 CPS scores of 10 or higher are statistically significant and clinically meaningful, as well as internally valid – with the caveat that patients were not stratified by PD-L1 status and some favorable risk factors were more common in the pembrolizumab group, according to Dr. Yoon. The 43% survival rate at 12 months translates to a number needed to treat of just four patients for one patient to be alive at that time point.
“A multivariate analysis could help clarify whether the positive results in the PD-L1 CPS 10-or-higher subgroup are explained by a higher frequency of favorable patient characteristics in the pembrolizumab arm,” he noted. “The strength of those results could influence guideline recommendations and implementation in clinical practice.”
Subgroup analyses suggested benefit was mainly seen in Asian patients, who tend to have higher prevalence of squamous tumors, Dr. Yoon said. Potential molecular differences at play here may be elucidated by ongoing research.
Ultimately, the findings in the PD-L1 CPS 10-or-higher group have potential implications for biomarker testing. “For the patient in front of me, I currently order PD-L1 and HER2 at first metastatic diagnosis in gastroesophageal adenocarcinomas,” he elaborated. “It’s reasonable to consider a practice change. This means ordering PD-L1 at first metastatic diagnosis in patients with squamous carcinoma of the esophagus. This would also mean some pathology labs may need to report a more detailed PD-L1 CPS score, if they don’t already.”
These findings also have potential implications for treatment. “For the second-line setting, for squamous carcinoma of the esophagus, esophageal adenocarcinomas, and Siewert 1 adenocarcinomas with a PD-L1 CPS of 10 or more, it’s reasonable to consider pembrolizumab,” Dr. Yoon noted. “This should be discussed within guideline committees, and the results will be submitted to regulatory authorities, who will have access to more detailed data. It’s possible that these recommendations could be modified in the near future.”