Whole-Blood Gene Expression Profiles Correlate with Response to Immune Checkpoint Inhibitors in Patients with Metastatic Renal Cell Carcinoma

SIMPLE SUMMARY: Some metastatic renal cell carcinoma (mRCC) patients do not respond to immune checkpoint inhibitor (ICI) therapy. However, since predicting who will be a poor responder is difficult, a non-invasive, high-resolution genomic biomarker to accurately predict clinical responses to ICIs is urgently required. We found a minimum set of 14 genes that change in response to treatment and this panel can be used to accurately classify responder patients. Our results suggest that the gene signatures identified from a whole-blood transcriptome approach are clinically useful biomarkers for predicting ICI responses in patients with mRCC. ABSTRACT: In metastatic renal cell carcinoma (mRCC), the clinical response to immune checkpoint inhibitors (ICIs) is limited in a subset of patients and the need exists to identify non-invasive, blood-based, predictive biomarkers for responses. We performed RNA sequencing using whole-blood samples prospectively collected from 49 patients with mRCC prior to the administration of ipilimumab (IPI) and/or nivolumab (NIVO) to determine whether gene expression profiles were associated with responses. An analysis from 33 mRCC patients with complete responses (n = 5), partial responses (n = 14), and progressive disease (n = 14) showed 460 differentially expressed genes (DEGs) related to immune responses between the responder and non-responder groups with significant differences. A set of 14 genes generated from the initial 460 DEGs accurately classified responders (sensitivity 94.7% and specificity 50.0%) while consensus clustering defined clusters with significantly differing response rates (92.3% and 35.0%). These clustering results were replicated in a cohort featuring 16 additional SD patients (49 total patients): response rates were 95.8% and 48.0%. Collectively, whole-blood gene expression profiles derived from mRCC patients treated with ICIs clearly differed by response and hierarchical clustering using immune response DEGs accurately classified responder patients. These results suggest that such screening may serve as a predictor for ICI responses in mRCC patients.