Tumour architecture, grade and location remain predictors of non-organ-confined upper tract urothelial carcinoma at time of radical nephroureterectomy: results from a multicenter Norwegian external validation study

Abstract

Purpose Selecting patients for intensified treatment for upper tract urothelial carcinoma can be challenging, partly due to the lack of accurate preoperative staging tools. Several preoperative staging models for upper tract urothelial carcinoma have been presented, but none have been externally validated. The aim of the current study was to perform an external validation of the Margulis nomogram for predicting non-organ-confined upper tract urothelial carcinoma at time of nephroureterectomy. Methods 209 patients from two high-volume centres in Norway were treated with radical nephroureterectomy during the period 2005–2017. 163 patients with complete data necessary for external validation of the Margulis nomogram were included in the study. All relevant covariates were analysed with uni- and multivariate regression analysis to assess their ability to predict non-organ-confined disease. The Margulis nomogram was applied on the present cohort to calculate predicted risk of non-organ-confined disease. This was compared to the observed risk to assess model calibration. The Margulis nomogram accuracy was measured as the area under the curve in a receiver operator characteristics curve to evaluate model discrimination. Results Tumour grade (OR 28.1, p\u2009 =\u20090.001) and architecture (OR 4.72, p\u2009 <\u20090.001) were independent predictors of non-organ-confined disease. There was a high concordance between predicted and observed risk quantified with a Cronbach alpha of 0.96. The Margulis nomogram had an area under the curve of 0.83 in predicting non-organ-confined disease when applied on the current cohort. Conclusions We consider the Margulis nomogram validated for clinical use.