The study findings can inform physician/patient discussions, but does not appear to represent strong selection criteria.

“Even combining the 3 markers,it’s not yet an accurate predictor,” he said. “Not all patients with all 3 markers responded and likewise, there were responders with low inflammation and low TMB. It’s not 100% accurate.”

The KEYNOTE-028 findings will help researchers design better studies that will “point patients toward mechanism-based immunotherapies,” but the goal of matching immunotherapies to patients is still on the horizon, Dr Ott said.

In KEYNOTE-028, 475 patients withPD-L1–positive solid tumors received pembrolizumab (10 mg/kg every 2 weeks) for 2 years or until progression or discontinuation for toxicity.Objective response rates (ORRs) to treatment ranged from 0 for pancreatic cancer to 33% for small-cell lung cancer, the authors reported.1

Continue Reading

“Higher response rates and longer progression-free survival were demonstrated in tumors with higher T-cell-inflamed GEP [gene expression profile], PD-L1 expression, and/or TMB,”the authors reported.1

KEYNOTE-028 does offer clinicians preliminary, nonrandomized trial evidence for pembrolizumab efficacy against some rare cancer types, like salivary gland andt hyroid cancers, Dr Ott noted.

“For the firsttime, with this manuscript, there is evidence for some tumor types for which pembrolizumab efficacy hasn’t previously been reported,” he said.

Dr Ott and colleagues hope to identify additional tumor-biology biomarkers, such as immunohistochemical CD8+assays in melanoma, to improve predictive power, Dr Ott said.

“Ideally, this will be done in a more prospective way,” he said. 


  1. Ott PA, Bang YJ, Piha-Paul SA, et al. T-cell-inflamedgene-expression profile, programmed death ligand 1 expression, and tumormutational burden predict efficacy in patients treated with pembrolizumabacross 20 cancers: KEYNOTE-028. JClin Oncol. 2019;37(4):318-327.