<h3>Objectives:</h3> Genomically unstable tumors, characterized by <i>BRCA1/2</i> alterations and homologous recombination deficiency (HRD), are hypothesized to be more mutated and potentially more sensitive to immune checkpoint inhibitors. To explore this hypothesis we analyzed outcomes in the IMagyn050 trial according to <i>BRCA1/2,</i> HRD, and PD-L1 status [Moore, ESMO 2020]. <h3>Methods:</h3> IMagyn050 (NCT03038100) is a double-blind randomized phase III trial evaluating the efficacy and safety of adding atezolizumab/placebo to carboplatin, paclitaxel, and bevacizumab (CPB) followed by maintenance bevacizumab plus atezolizumab/placebo. PD-L1 status was determined centrally using VENTANA SP142 (PD-L1+ defined as ≥1% tumor-infiltrating immune cells expressing PD-L1). Deleterious tumor germline/somatic <i>BRCA1/2</i> alterations (<i>BRCA1/2</i>m), genomic loss of heterozygosity (gLOH), tumor mutation burden (TMB), and microsatellite instability (MSI) were evaluated using the FoundationOne assay (Foundation Medicine, Inc., Cambridge, MA). HRD and homologous recombination proficiency (HRP) were defined as gLOH ≥16% and <16%, respectively, regardless of <i>BRCA1/2</i>m status. Potential associations between clinical outcome (RECIST-defined progression-free survival [PFS]) and <i>BRCA1/2,</i> HRD, and PD-L1 status were evaluated using standard correlation analyses and Kaplan-Meier estimates. <h3>Results:</h3> Among 1301 randomized patients, samples from 1050 were evaluable for <i>BRCA1/2</i> (22% were <i>BRCA1/2</i>m, 78% <i>BRCA1/2</i> nonmutant) and 980 were evaluable for gLOH (46% were HRD, 54% HRP). Median TMB was similarly low in <i>BRCA1/2</i>m and <i>BRCA1/2</i> nonmutant (3.78 vs 2.52 Mut/Mb, respectively), and HRD and HRP (3.78 vs 2.52 Mut/Mb, respectively) tumors. Only 3% (29/1024) of evaluable tumors had TMB ≥10 Mut/Mb and 0.3% (3/1022) were MSI-high (1 mixed, 1 undifferentiated, 1 other). All high-grade serous cases were MS-stable. PFS prognosis (assessed in the placebo + CPB arm) was improved in patients with <i>BRCA1/2</i>m (hazard ratio [HR] 0.62, 95% CI 0.46–0.84) and HRD (HR 0.63, 95% CI 0.49–0.80) tumors. There was a suggested association between PD-L1+ and HRD (HRD prevalence: 40% vs 25% in PD-L1+ vs PD-L1– subgroups, respectively; exploratory Fisher's exact test p=0.0001) but not with <i>BRCA1/2</i>m (<i>BRCA1/2</i>m prevalence: 21% vs 14%, respectively; exploratory Fisher exact test p=0.064). Adding atezolizumab to CPB did not improve PFS, irrespective of <i>BRCA1/2</i> or HRD status (table). In the PD-L1+ population, HRs were similar in <i>BRCA1/2</i>m and <i>BRCA1/2</i> nonmutant subgroups, and in HRD and HRP subgroups. In the PD-L1– population, atezolizumab did not improve PFS. <h3>Conclusions:</h3> We showed that the majority of ovarian tumors have low TMB scores regardless of <i>BRCA1/2</i> or HRD. Neither <i>BRCA1/2</i>m nor HRD was associated with greater clinical benefit from adding atezolizumab to CPB. PD-L1 status was more reliably associated with numerically longer PFS with atezolizumab + CPB. This is the first randomized double-blind trial in ovarian cancer to demonstrate that genomic instability triggered by <i>BRCA1/2</i>m or HRD does not improve sensitivity to immune checkpoint inhibitors.