In KEYNOTE-061 randomised study, PD-1 blockade with immune checkpoint inhibitor (ICI) pembrolizumab did not significantly improve overall survival (OS) over paclitaxel in second-line treatment for advanced gastric/gastro-oesophageal junction (GEJ) adenocarcinoma with PD-L1 combined positive score (CPS) ≥1. The exploratory study showed an association between tissue tumour mutational burden (tTMB) and efficacy of pembrolizumab. Data further suggest that tTMB is a significant and independent predictor beyond PD-L1 status according to Dr. Kohei Shitara of the Department of Gastroenterology and Gastrointestinal Oncology, National Cancer Center Hospital East in Chiba, Japan and colleagues who published the findings in the August 2021 issue of the Annals of Oncology.

Additional post hoc analyses in patients with PD-L1 CPS ≥10 demonstrated prolonged OS, higher response rates, and durable responses in 53 patients treated with pembrolizumab versus 55 patients treated with paclitaxel. The authors commented that pembrolizumab is not approved in the second-line setting for advanced gastric/GEJ adenocarcinoma. In the exploratory analysis from KEYNOTE-061, they evaluated the association of tTMB with PD-L1 and MSI-H and their relationship with clinical outcomes.

In patients with whole exome sequencing (WES) data, the KEYNOTE-061 investigators evaluated association of tTMB with objective response rate (ORR), progression-free survival (PFS), and OS. tTMB was also evaluated by using FoundationOne®CDx. Prespecified equivalent cut-offs of 175 mut/exome for WES and 10 mut/Mb for FoundationOne®CDx were used.

The WES data were available in 420 of 592 patients (71%), of whom 218 were treated with pembrolizumab and 202 with paclitaxel. According univariate analysis, WES-tTMB was significantly associated with ORR, PFS, and OS in patients treated with pembrolizumab (all p < 0.001), but not in patients treated with paclitaxel.

In univariate analysis the area under the receiver operating characteristics curve for WES-tTMB and response was 0.68 (95% confidence interval [CI] 0.56-0.81) for pembrolizumab and 0.51 (95% CI 0.39-0.63) for paclitaxel.

There was low correlation between WES-tTMB and CPS in both treatment groups (r ≤ 0.16). The WES-tTMB remained significantly associated with all clinical endpoints with pembrolizumab after adjusting for CPS and with PFS and OS after excluding known MSI-H tumours (n = 26).

FoundationOne®CDx-tTMB demonstrated a positive association with ORR, PFS, and OS in pembrolizumab-treated patients (all p ≤ 0.003), but not PFS or OS in paclitaxel-treated patients.

The authors concluded that tTMB is a significant and independent predictor beyond PD-L1 status.

However, in a letter to editors published in the Annals of Oncology, a group of physicians from the Department of Medicine, Memorial Sloan Kettering Cancer Center in New York, NY, US questioned the conclusions from this post hoc KEYNOTE-061 analysis by Shitara et al. They wrote that the study rather confuses than clarifies the predictive value of TMB in defining clinical ICI benefit.


Shitara K, Özgüroğlu M, Bang Y-J, et al. Molecular determinants of clinical outcomes with pembrolizumab versus paclitaxel in a randomized, open-label, phase III trial in patients with gastroesophageal adenocarcinoma. Annals of Oncology 2021;32(9):1127-1136. DOI:

Foote MB, Maron SB, Cercek A, et al. TMB cut-offs fail to predict benefit of PD-1 blockade in gastroesophageal adenocarcinoma in KEYNOTE-061. Annals of Oncology 2021;32(9):1188-1189. DOI: