Inter-observer agreement of the estimated contact surface area system for renal masses

Abstract Background Nephrometry scores have been designed to assist in standardization of renal masses before nephron sparing surgery. Estimated contact surface area (eCSA) has been recently described. However, its reproducibility between raters has not been rigorously explored. Our aim was to examine the reproducibility of eCSA and RENAL (RS) nephrometry systems for renal masses by studying the inter-observer agreement between 4 raters. Materials and Methods We reviewed the cross-sectional images (computed tomography, magnetic resonance imaging) of patients who underwent nephron sparing surgery (open, laparoscopic and robotic). Four urologists independently scored the renal tumors according to the eCSA and RS systems. Scoring was done separately for each nephrometry system. Inter observer agreement was assessed for total scores and separate components of each system by calculating the intraclass correlation coefficient (ICC). Results Ninety patients were scored to achieve power of 83% with α=0.05. Both eCSA and RS demonstrated excellent inter-observer agreement (ICC = 0.89 both). The Radius component of eCSA and RS had the highest ICC (0.97 and 0.9 respectively) compared with the other components. Location and Anterior/Posterior components of RS demonstrated poor inter-observer correlation (ICC = 0.69 and 0.50 respectively). Maximal difference in RS complexity group assignment was 1 in 47.8% of cases. Conclusions eCSA has excellent inter-observer agreement, similar to RS. Directly measurable anatomic variables have better inter-observer correlation compared with qualitative variables. These results strengthen the applicability of eCSA as a surgical complexity metric for renal surgery.