Peripheral blood correlates of outcomes with adjuvant tremelimumab plus durvalumab for patients with renal cell carcinoma at high or intermediate risk of relapse: Initial results from the TransRAMPART study.

527 Background: Despite optimal surgery, approximately 1/3 of patients with renal cell carcinoma (RCC) develop incurable relapse. Adjuvant pembrolizumab reduces the risk of relapse in patients with clear cell RCC and is considered standard care. Post nephrectomy plasma KIM-1 levels have been shown to be prognostic in RCC in the ASSURE & IMmotion010 trials. RAMPART (NCT03288532) is an investigator-led, international randomized phase III study, that enrolled patients with RCC at intermediate and high risk of relapse. Patients were randomized to placebo (Arm A), durvalumab (Arm B), durvalumab + tremelimumab (Arm C). A DFS benefit has recently been reported for Arm C compared to the control Arm A (DFS at 3 years 81% vs. 73%, HR 0.65, 95% CI 0.45 to 0.93, p = 0.009). Methods: TransRAMPART collected longitudinal blood, tissue and urine samples in parallel to the main study. The aims included understanding prognostic or predictive markers for adjuvant treatment. A total of 187 patients were recruited. Samples were collected at baseline, 16 weeks, 32 weeks, and at later timepoints up to 5 years. Blood samples were profiled by MSD cytokine array (plasma), and by high resolution flow cytometry (PBMCs) by IMU Biosciences. Here we report the initial results of these exploratory blood-based analyses for Arm A and Arm C. Results: High baseline KIM-1 levels ( > 86 pg/mL, as used in analysis of the IMmotion010 trial) were associated with worse DFS outcomes in the combined A+C population (p = 0.0075, HR = 3.4, 95% CI 1.3-8.9, n = 77). Analyzing the patients by arm, KIM-1 remained negatively prognostic in the control Arm A (p = 0.0015, HR = 8.0, 95% CI 1.7-37, n = 45), however the effect was no longer seen with adjuvant combination immunotherapy in arm C (p = 0.72, HR = 1.3, 95% CI 0.31-5.1, n = 32). Peripheral immune profiling revealed an immune profile associated with Arm C treatment which returned toward baseline following the treatment period (64 patients in Arm A, 50 patients in Arm C). KIM-1 levels did not correlate with immune changes; instead, various immune cell lineages showed differential phenotypic enrichment according to baseline KIM-1. Tracking these phenotypes through the treatment period revealed divergence between Arm A and Arm C patients. Conclusions: We confirm that baseline KIM-1 levels are prognostic in RCC in an independent population. Although this exploratory analysis only assessed a subset of the whole trial population, this prognostic effect seems to be negated by adjuvant immunotherapy. There is an immune signature associated with KIM-1 levels distinct from the impact of treatment, suggesting an interplay between KIM-1 and immune activity in the postoperative period. Deeper characterization of these immune changes is ongoing. Clinical trial information: NCT03288532 .