We couldn’t resist comparing central with local RECIST1.1 and with Choi assessment: Exploratory analysis of tumor imaging in EORTC STBSG Phase 2 trial 1317 “CaboGIST.

e23563 Background: Gastrointestinal stromal tumor (GIST) is commonly driven by activating mutations in KIT or PDGFRA. Advanced GIST is treated with tyrosine kinase inhibitors (TKIs) but develops resistance over time. EORTC 1317 assessed the safety and activity of cabozantinib, a multi-TKI targeting KIT, MET, AXL and proangiogenic pathways, in GIST patients who had progressed on imatinib and sunitinib. The efficacy analysis of the trial, which met its primary endpoint, was based on local assessment of RECIST1.1 response. RECIST neglects myxoid degeneration, necrosis and vascular remodeling induced by TKIs without major volumetric changes. Density changes on CT scans can predict clinical benefit and can be assessed by Choi criteria. Methods: We describe results of a post hoc exploratory analysis of CT scans performed centrally using RECIST 1.1 and Choi criteria. Results: Week 12 scans were available and evaluable by central review in 43 pts, Choi in 42 cases. Comparisons between local and central RECIST1.1 outcome revealed discrepancies in 17/43 evaluable cases (39.5%). When comparing Choi with local and central RECIST1.1 at week 12, discrepancies were observed in 27/42 (64.3%) and 21/42 (50%) evaluable cases, respectively. In summary, 70% of evaluable patients were progression-free and alive at week 12 based on local assessment, 86% and 83% according to central RECIST1.1 and Choi criteria, respectively. The main difference was the rate of objective response with cabozantinib in week 12: 5 PR (12%) with local RECIST1.1, 3 PR (7%) with central assessment, and 21 PR (50%) with Choi criteria. Conclusions: RECIST1.1 remains an unsatisfactory tool for response assessment in GIST, illustrated by the high inter-rater variability of response outcome comparing local versus central analysis. RECIST1.1 clearly underestimates the anti-tumor activity of TKIs in GIST. Cabozantinib did not only meet the primary endpoint of this trial when applying RECIST1.1 per protocol, but achieved objective responses in 50% of evaluable patients in week 12 when using Choi criteria. [Table: see text]