Interobserver and intraobserver variability in the response evaluation of cancer therapy according to RECIST and WHO-criteria

Abstract

Background. Response Evaluation Criteria In Solid Tumors (RECIST) and WHO-criteria are used to evaluate treatment effects in clinical trials. The purpose of this study was to examine interobserver and intraobserver variations in radiological response assessment using these criteria. Material and methods. Thirty-nine patients were eligible. Each patient's series of CT images were reviewed. Each patient was classified into one of four categories according RECIST and WHO-criteria. To examine interobserver variation, response classifications were independently obtained by two radiologists. One radiologist repeated the procedure on two additional different occasions to examine intraobserver variation. Kappa statistics was applied to examine agreement. Results. Interobserver variation using RECIST and WHO-criteria were 0.53 (95% CI 0.33–0.72) and 0.60 (0.39–0.80), respectively. Response rates (RR) according to RECIST obtained by reader A and reader B were 33% and 21%, respectively. RR according to WHO-criteria obtained by reader A and reader B were 33% and 23% respectively. Intraobserver variation using RECIST and WHO-criteria ranged between 0.76–0.96 and 0.86–0.91, respectively. Conclusion. Radiological tumor response evaluation according to RECIST and WHO-criteria are subject to considerable inter- and intraobserver variability. Efforts are necessary to reduce inconsistencies from current response evaluation criteria.