890P - Prospective comparison of RECIST and alternative response assessment criteria in the evaluation of metastatic renal cell cancer patients from phase II of the multi-centre randomised STAR trial

Abstract

Background: Combined size and enhancement criteria show potential to predict earlier disease response or progression in metastatic renal cancer (mRCC). We assessed categorisation differences using RECIST 1.1, Choi and modified Choi (mChoi) criteria with normalised rather than absolute enhancement values in phase II of the STAR trial (comparing conventional continuation versus drug free intervals of tyrosine kinase inhibitor treatment in mRCC). Methods: 44 patients underwent contrast-enhanced computed tomography (CE-CT) at baseline, 12 and 24 weeks post therapy. Automated software was used by 2 independent readers to evaluate 104 target lesions. Target lesion sum of longest diameter, normalised enhancement values (relative to aortic attenuation) and subsequent percentage change at 12 and 24-week CT were measured. Response categorisation into stable disease (SD), partial response (PR) or progressive disease (PD) was undertaken by RECIST 1.1, Choi and mChoi response criteria, and discrepant cases scored. Reader agreement was assessed by Cohen’s kappa test. Results: By RECIST 1.1, patients were 68%(n = 30)/41% (18)SD, 27%(12)/45%(20) PR, 2%(1)/9%(4) PD and 2%(1)/2%(1) CR at 12 and 24 weeks respectively. At 12 weeks 27 patients had discrepant categorisation: PR by both Choi/mChoi criteria but SD by RECIST in 17 and PR by CHOI, SD by mCHOI in 10. With absolute versus normalised enhancement values, 3 further patients would have remained as SD by mChoi at 12 weeks. At 24 weeks 14 remained discrepant: both Choi/mChoi PR but SD by RECIST in 10, PR by Choi but SD in mCHOI/RECIST in 4 patients. 11 previously discrepant patients by RECIST versus Choi/mChoi became concordant (8 PR, 3 PD) at 24 weeks. The concordance was excellent for RECIST (k 0.9, k 0.8) and mChoi (k 0.9 k 0.79), and good/excellent reader for Choi criteria (k 0.76 k1.0) at 12 and 24 weeks respectively. Conclusions: Early response, confirmed at 24 weeks, was more frequent for Choi/mChoi than RECIST. Substantial/excellent agreement was noted in response categorisation with normalised versus absolute enhancement indicating this is a potentially robust approach. Clinical trial identification: This project was funded by the National Institute for Health Research Health Technology Assessment Programme (project number 09/91/21) The views and opinions expressed therein are those of the authors and do not necessarily reflect those of the HTA, NIHR, NHS or the Department of Health. Legal entity responsible for the study: Leeds institute of clinical trials research Funding: NHS National Institute for Health Research Disclosure: F. Thistlethwaite: Previous ESMO travel and accommodation reimbursement by Ipsen and Bristol Meyers Squibb, consultant for Pfizer and Bristol-Meyers Squibb and received research funding from Pfizer, Novartis and Aveo. All other authors have declared no conflicts of interest.