Over 500 clinical trials combining radiation (RT) and immune checkpoint blockade (ICB) have been initiated based on preclinical evidence that RT can augment local immunity and improve the efficacy of ICB. However, many recent clinical trials have not found a benefit of combining RT and ICB, raising questions about whether a synergy exists. We examined whether RT and ICB interact to beneficially stimulate the immune response in patients and identified biomarkers of response to RT and ICB. We performed a molecular analysis of 1,740 patients from 3 cohorts. The COSINR dataset is a randomized clinical trial of 22 non-small cell lung cancer (NSCLC) patients treated with concurrent or sequential SBRT and ipilimumab/nivolumab. Paired pre- and on-treatment biopsies of an irradiated metastasis underwent whole exome sequencing and RNA-seq. On-treatment biopsies were obtained after SBRT and prior to ICB (sequential) or after SBRT and one cycle of ICB (concurrent). The UC cohort consisted of targeted DNA sequencing of 58 NSCLC patients treated with ICB alone, sequential RT+ICB, or concurrent RT+ICB. The MSKCC dataset is a pan-cancer cohort of targeted DNA sequencing of 1,660 patients treated with ICB. Aneuploidy score (AS) was defined as the fraction of chromosome arms with arm-level copy number alterations. Survival analyses utilized the Kaplan-Meier method and multivariable Cox proportional hazards models. In the COSINR trial, SBRT+ICB increased, whereas SBRT alone decreased, expression of effector T cell IFN-gamma and adaptive immune signatures (P<0.05). Established biomarkers of ICB response, including IFN-gamma signature, tumor mutational burden (TMB), PD-L1 expression, and neoantigen burden were not associated with survival (P>0.05). However, patients whose tumors exhibited high (≥median) but not low, AS had improved survival when treated with concurrent vs. sequential SBRT+ICB (1-year overall survival [OS] 100% vs. 17%, P = 0.025). Our findings were corroborated in the UC cohort: high AS tumors treated with RT + ICB had superior 1-year OS compared to those treated with ICB alone (59% vs. 31%, P = 0.021). Among those who received RT + ICB, concurrent treatment improved OS relative to sequential (1-year OS 76% vs. 38%). RT did not improve OS in patients with low (<median) AS. In the MSKCC cohort, high AS was associated with poor OS following ICB after controlling for TMB and ICB paradigm (hazard ratio [95% CI]: 1.44 [1.07-1.94], P = 0.014). Among tumors with low TMB (<80th percentile), but not high TMB, high AS (≥median) was associated with poor OS (P = 0.003). Our findings distinguish the genomic and transcriptomic effects of RT versus RT+ICB and challenge the prevailing paradigm that local ablative RT positively stimulates the immune response. We propose the use of tumor aneuploidy as a biomarker in personalizing treatment approaches for patients with various cancers.