Background

The immune checkpoint inhibitor (ICI) pembrolizumab is FDA-approved for treatment of solid tumors with high tumor mutational burden (TMB-high, ≥10 variants/Mb). However, the extent to which TMB generalizes as a predictive biomarker to diverse patient populations and platforms is largely unknown. Herein, we investigate the interplay between genetic ancestry, TMB, and clinical outcomes of ICIs in solid tumors.

Methods

In this retrospective study, we analyzed genomic and clinical data from patients treated with ICIs at a single institution and sequenced on a tumor-only panel, Oncopanel. We inferred continental genetic ancestry (West African, East Asian, and European), developed a method to calibrate tumor-only estimated TMB accounting for ancestry-specific biases, and evaluated its performance as a biomarker across genetic ancestry groups. Primary outcomes were overall survival (OS) and time to ICI failure (TTF).

Results

1614 patients treated with ICIs were diagnosed with one of seven cancer types including non-small cell lung cancer (NSCLC). Relative to self-reported race (Black/Asian), 44% and 29% more patients had detectable African and Asian ancestry, respectively. TMB-high misclassification due to tumor-only sequencing was significantly higher in patients with African (57%) and Asian (49%) ancestry compared to European (32%). Misclassified TMB-high was not associated with OS and was corrected through ancestry-specific recalibration. For NSCLC, TMB-high was only significantly associated with improved OS and TTF in European patients.

Conclusions

We demonstrate that tumor-only TMB generalizes poorly to non-European ancestries, with implications for other emerging biomarkers, and propose an ancestry-informed recalibration to improve patient stratification.

VIEW THE PUBLICATION