Comparison of tumor response assessment methods in patients with metastatic colorectal cancer after locoregional therapy.

Abstract

We evaluated the performance of the Response Evaluation Criteria in Solid Tumor (RECIST), modified RECIST, and the European Association for the Study of Liver (EASL) guidelines and correlated them with survival in patients with metastatic colorectal cancer (mCRC) treated with locoregional therapy (LRT). Our LRT registry was evaluated from 2008 to 2013. 228 mCRC patients were treated with LRT (91% drug-eluting beads, 9% radioembolization) were evaluated. Cox regression and Kaplan-Meier (KM) statistics were utilized for survival analysis. Excellent inter-rater agreement between EASL/mRECIST (κ = 905) was seen. Correlations between RECIST/mRECIST (κ = 0.638) and EASL/RECIST were weaker (κ = 0.638 and 0.598, respectively). There were significant differences in KM and Cox regression survivals between responders and nonresponders with all three methods (all P < 0.0001). Multivariate analysis identified RECIST response, tumor extent, performance status, concomitant chemotherapy, and prior surgery/ablation as independent prognostic factors. EASL response and mRECIST response were not found to be independent prognostic factors. Imaging biomarkers are not efficient and do not represent ideal surrogates for survival; however, they all display prognostic significance. RECIST is superior to mRECIST/EASL given its ability to stratify survival benefit according to response category and demonstrate independent prognostic significance. J. Surg. Oncol. 2016;113:443-448. © 2016 Wiley Periodicals, Inc.