Developing a nomogram for predicting intravesical recurrence after radical nephroureterectomy: a retrospective cohort study of mainland Chinese patients.

Abstract

To evaluate the role of Ki-67 in predicting subsequent intravesical recurrence following radical nephroureterectomy and to develop a predictive nomogram for upper tract urothelial carcinoma patients. This retrospective analysis involved 489 upper tract urothelial carcinoma patients who underwent radical nephroureterectomy with bladder cuff excision. The data set was randomly split into a training cohort of 293 patients and a validation cohort of 196 patients. Immunohistochemical analysis was used to assess the immunoreactivity of the biomarker Ki-67 in the tumor tissues. A multivariable Cox regression model was utilized to identify independent intravesical recurrence predictors after radical nephroureterectomy before constructing a nomographic model. Predictive accuracy was quantified using time-dependent receiver operating characteristic curve. Decision curve analysis was performed to evaluate the clinical benefit of models. With a median follow-up of 54months, intravesical recurrence developed in 28.2% of this sample (n=137). Tumor location, multifocality, pathological T stage, surgical approach, bladder cancer history and Ki-67 expression levels were independently associated with intravesical recurrence (all P<0.05). The full model, which intercalated Ki-67 with traditional clinicopathological parameters, outperformed both the basic model and Xylinas' model in terms of discriminative capacity (all P<0.05). Decision-making analysis suggests that the more comprehensive model can also improve patients' net benefit. This new model, which intercalates the Ki-67 biomarker with traditional clinicopathological factors, appears to be more sensitive than nomograms previously tested across mainland Chinese populations. The findings suggest that Ki-67 could be useful for determining risk-stratified surveillance protocols following radical nephroureterectomy and in generating an individualized strategy based around intravesical recurrence predictions.