1506P - Superiority of Choi vs Recist Criteria in Evaluating Outcome of Advanced Soft Tissue Sarcoma (STS) Patients Treated with Sorafenib

Abstract

ABSTRACT Background Objective tumor response may be underestimated by RECIST criteria, since biological agents can cause metabolic response (necrosis) not related to tumor size. For this reason, CHOI criteria were designed to evaluate both tumor density and size. We compared CHOI and RECIST criteria in evaluating response to sorafenib in patients with advanced STS, treated within a phase II trial. The objective was to compare CHOI and RECIST criteria and to relate response to progression-free (PFS) and overall survival (OS). Patients and methods Sixty one advanced STS patients received sorafenib 400 mg twice daily for progressive or relapsed disease after anthracycline-based chemotherapy. Treatment was continued until progression or major toxicity. Contrast-enhanced CT was performed every 3 months. Thirty patients were evaluable for response, according to both RECIST and CHOI criteria. Results A total of 30 patients were included in the analysis. According to RECIST, we observed 1 (3%) complete response (CR), 1 (3%) partial remission (PR), 18 (60%) stable diseases (SD) and 10 (34%) progressive diseases (PD) at 3 months. According to CHOI, we observed 1 (3%) CR, 10 (34%) PR, 5 (16%) SD and 14 (47%) PD. The agreement between the two techniques was low (cohen Kappa: 0.36; CI 95%, 0.17-0.55). Of the 18 pts with SD according to RECIST, a total of 13 (72%) pts were reallocated to responder (8 PR, 44%) and non-responder (5 PD, 28%) groups, while only in 5 cases SD was confirmed using CHOI criteria. Median PFS and OS for SD RECIST pts were 8.3 and 22.7 months, respectively. In these 18 pts, we compared median PFS and OS of responders (CR + PR, n= 8) vs non-responders (SD + PD, n= 10) according to CHOI criteria. PFS was 11.2 vs 7.9 (p= 0.05), and OS 23.3 vs 15 months (p= 0.21). Conclusions Our analysis suggests that CHOI criteria may be helpful to define response to a biological treatment in STS. According to CHOI criteria, SD pts as defined by RECIST could be better divided in two subgroups (responders and non-responders) with a different outcome. Therefore, CHOI criteria may have an early significant predictive value for PFS in STS pts treated with biological agents. Disclosure All authors have declared no conflicts of interest.