Predictive Biomarkers for Checkpoint Blockade in Urothelial Cancer: A Systematic Review.

Author information

Abstract

BACKGROUND:

Immune checkpoint inhibitors have had a major impact on the management of advanced urothelial cancer. Despite their impact, only a minority of patients derive benefit. In this context, predictive biomarkers that can assist in patient selection are needed. This systematic review aims to survey the current biomarkers that have been investigated in clinical studies, and their potential for patient selection.

METHODS:

We conducted a search of MEDLINE and EMBASE, as well as a manual review of major meeting abstracts, for studies in humans of immune checkpoint inhibitors given for urothelial cancer that included both biomarkers and clinical outcomes. Studies had to report the correlation between biomarkers and outcomes to be included. Results published in abstract form only were included since several important biomarker studies have yet to be published.

RESULTS:

Our systematic review retrieved 1236 studies, of which 921 were unique and screened; 144 met criteria for full review, and 25 were included in the analysis. The manual search yielded 1 additional entry not included in our systematic review, for a total of 26 entries. The checkpoint inhibitors used in these studies included atezolizumab, avelumab, durvalumab, ipilimumab, nivolumab and pembrolizumab. The biomarkers tested included PD-L1 immunohistochemistry, molecular subtyping and immune gene expression analysis by RNA sequencing, targeted gene panels for mutations in DNA damage repair genes and estimation of tumor mutational burden, genomic alterations and total mutational burden by exome sequencing, analysis of tumor immune infiltrate by immunohistochemistry and T-cell receptor sequencing, and analysis of circulating immune cells and cytokines.

DISCUSSION:

No single biomarker has been able to predict response to immune checkpoint inhibitors accurately. Most studies included only a treatment arm, and without a comparator arm it is not possible to ascertain if biomarkers are predictive or merely prognostic. While PD-L1 immunohistochemistry has been largely unsuccessful, other biomarkers that reflect the immunogenicity of the underlying tumor, the characteristics of the immune infiltrate and the properties of the patient's immune system have shown promising data, but are all in need of validation.