Comparison of iRECIST and choi criteria for evaluation of response to ipilimumab plus nivolumab in advanced non-translocated sarcomas.

e23524 Background: Patients (pts) with locally advanced or metastatic sarcoma have limited treatment options and a poor prognosis, with PFS of < 6mos and OS of < 15mos. A phase 2 trial showed encouraging response with combination ipilimumab and nivolumab immunotherapy (IO). RECIST and iRECIST both determine response rate. Recent data within GIST shows the Choi criteria better predicts clinical responses (assessing size and density of target lesions) than RECIST. We compared response to IO using iRECIST and Choi. Methods: In this single-institution retrospective analysis, pts with sarcoma were treated with at least one cycle of ipilimumab and nivolumab until disease progression (PD) or unacceptable toxicity between 2018-2021. Routine imaging was performed with CT or MRI to monitor for response. Responses were assessed using iRECIST, and Choi criteria for stable disease (SD), partial response (PR) and PD were correlated with clinical outcomes such as best response and progression free survival (PFS). The Kaplan-Meier method was used for survival analysis. Results: For 74 pts demographics were as follows: median age 55, 47% male; 11% African American, 88% Caucasian, 1% Asian; 78% metastatic disease, and 22% locally advanced disease. 85% of pts had high grade, 3% had intermediate grade, and 12% had low grade sarcoma. At the start of treatment, median tumor mutation burden was low (< 10mut/Mb) in 39%, high (> 10mut/Mb) in 3% and unknown in 58%. The PDL1 was negative in 4%, < 20 in 5% and > 20 in 1% with unknown PDL1 status in majority of patients, 89%. Overall response rate (ORR) was 45.3% (N = 59). The best response was SD in 34%, PR in 11% and PD in 55% based on iRESIST and Choi criteria. Median overall survival (OS) was 10.9m (95% CI 5.0-17.7) from the start of treatment. The mPFS was 4.1m (95% CI 2.8-5.7) based on iRESIST compared to mPFS was 2.9m (95% CI 2.1-5.7) based on Choi criteria. 55% of patients experienced any grade immune mediated toxicity with 54% requiring systemic steroids, and 19% resulting in treatment delay. The mOS for patients with toxicity was 14.7mo (95% CI 10.9-NR) versus 3.4m (95% CI 2.1-10.3) for those without IO toxicity. Conclusions: This real-world IO experience in advanced sarcomas shows remarkable ORR even with majority of patients having low TMB and unknown PD-L1 status. mPFS differed greatly between iRECIST and Choi criteria. iRECIST requires two scans to confirm PD, whereas Choi identifies PD with one scan, thus not delaying declaring PD. Another observation is that the Choi criteria was created for CT scans in GIST but has limitations within MRI. The development of immune mediated toxicity appears to be an independent predictor for prolonged OS in advanced sarcoma pts receiving IO therapy. A larger study is needed for confirmation of these trends.