To investigate the risk factors for postoperative fever following ureteroscopic laser lithotripsy for ureteral stones and to develop a predictive model utilizing the Random Forest algorithm.

A retrospective analysis was performed on 270 patients who underwent ureteroscopic laser lithotripsy for ureteral stones at our institution. Univariate and multivariate logistic regression analyses were conducted to identify independent risk factors associated with postoperative fever, followed by the construction of a Random Forest machine learning model.

Out of the 270 patients, 28 (10.37%) experienced fever after undergoing ureteroscopic laser lithotripsy. The multivariate analysis revealed that female gender (OR=1.356, 95% CI: 1.206–1.905, P=0.023), elevated preoperative urinary white blood cell count (OR=2.306, 95% CI: 1.866–5.175, P<0.01), positive preoperative urine culture (OR=1.103, 95% CI: 1.001–1.535, P<0.01), prolonged irrigation time (OR=1.125, 95% CI: 1.005–1.674, P=0.029), and extended surgical duration (OR=1.308, 95% CI: 1.127–1.766, P=0.029) were significant risk factors. The Random Forest model constructed based on these variables achieved an area under the curve (AUC) of 0.860 (95% CI: 0.759–0.934), with sensitivity, specificity, and accuracy rates of 77.2%, 89.3%, and 84.6%, respectively.

The machine learning model developed using the Random Forest algorithm demonstrates considerable potential for predicting postoperative fever following ureteroscopic laser lithotripsy.