A recent study by Banchereau et al. published in Nature Communications investigated the molecular determinants of response to PD-L1 blockade across different cancer types. The investigators performed molecular analysis of archived tissue from three atezolizumab monotherapy studies, including patients with advanced urothelial cancer (UC, n=208), non-small cell lung cancer (NSCLC, n= 81), and renal cell carcinoma (RCC, n= 77). PD-L1 expression by immunohistochemistry and tumor mutational burden (TMB) (categorized based on the median, 16.3 mutations/megabase) had poor predictive capacity.
The investigators performed bulk RNA sequencing of pretreatment tumor samples. A machine learning-based method was applied to identify a transcriptional signature predictive to atezolizumab and complement PD-L1 expression and TMB. A 58-gene signature was identified in this cohort with a receiver operating characteristic (ROC) curve of AUC=0.99. Applying this signature in another independent cohort of advanced UC, NSCLC, and RCC tumors treated with atezolizumab in a phase I basket clinical trial showed low performance in predicting response (AUC<0.65).
The investigators identified the genes differentially expressed between anti-PD-L1 responders and nonresponses. Upregulation of CDKN2A was enriched in the responding subgroup. On the other hand, there was a trend toward poor response rates in tumors with CDKN2A loss. These observations support the rationale for combining CDK4/6 inhibitors with ICB. However, the challenge of establishing predictive signatures reflects the significant heterogeneity in the mechanisms of response to ICB and requires an individualized approach.
Written by: Bishoy M. Faltas, MD, Director of Bladder Cancer Research, Englander Institute for Precision Medicine, Weill Cornell Medicine, New York City, New York
- Banchereau R, Leng N, Zill O, Sokol E, Liu G, Pavlick D, et al. Molecular determinants of response to PD-L1 blockade across tumor types. Nat Commun. 2021;12(1):1–11. PMID: 34172722
Read the Abstract