A new scoring system to predict febrile urinary tract infection after retrograde intrarenal surgery

Abstract

The current study aimed to determine the risk factors and define a new scoring system for predicting febrile urinary tract infection (F-UTI) following retrograde intrarenal surgery (RIRS) by using machine learning methods. We retrospectively analyzed the medical records of patients who underwent RIRS and 511 patients were included in the study. The patients were divided into two groups: Group 1 consisted of 34 patients who developed postoperative F-UTI, and Group 2 consisted of 477 patients who did not. We applied feature selection to determine the relevant variables. Consistency subset evaluator and greedy stepwise techniques were used for attribute selection. Logistic regression analysis was conducted on the variables obtained through feature selection to develop our scoring system. The accuracy of discrimination was assessed using the receiver operating characteristic curve. Five of the 19 variables, namely diabetes mellitus, hydronephrosis, administration type, a history of post-ureterorenoscopy (URS) UTI, and urine leukocyte count, were identified through feature selection. Binary logistic regression analysis showed that hydronephrosis, a history of post-URS UTI, and urine leukocyte count were significant independent predictors of F-UTI following RIRS. These three factors demonstrated good discrimination ability, with an area under curve value of 0.837. In the presence of at least one of these factors, 32 of 34 patients (94.1%) who developed postoperative F-UTI were successfully predicted. This new scoring system developed based on hydronephrosis, a history of post-URS UTI, and urine leukocyte count can successfully discriminate patients at risk of F-UTI development after RIRS.

Keywords: Complication; Feature selection; Infection; Machine learning; Retrograde intrarenal surgery; Scoring system.