Preoperative CT-based markers and inflammatory parameters as predictors of failed ureteroscopic access in unstented patients with ureteral stones: a prospective single-center cohort study.

Failed ureteroscopic access (FA) in unstented patients remains a significant clinical challenge. We aimed to identify preoperative clinical, laboratory, and CT-based predictors of FA, determine optimal cutoff values, and develop a clinically applicable risk-stratification model. This prospective cohort study enrolled 141 consecutive patients with ureteral stones undergoing unstented ureteroscopy (83 semi-rigid, 58 flexible) at a tertiary center (January 2024-January 2025). Preoperative demographics, pain characteristics, inflammatory markers (TLC, ESR, CRP, procalcitonin), and NCCT parameters (stone size, density, ureteral wall thickness [UWT], perinephric fat stranding [PFS] grade, hydronephrosis grade, HU of hydronephrosis, renal parenchymal thickness) were compared between successful access (SA, n = 112) and failed access (FA, n = 29) groups. Firth's penalized logistic regression, ROC analysis with DeLong comparison, and decision curve analysis (DCA) were performed. FA occurred in 29 patients (20.6%). UWT demonstrated the highest discriminatory power (AUC 0.965; cutoff ≥ 2.7 mm; sensitivity 96.6%, specificity 99.1%), followed by PFS grade (AUC 0.963; cutoff ≥ 2; sensitivity 100%, specificity 89.3%) and procalcitonin (AUC 0.957; cutoff ≥ 0.2 ng/mL). On Firth's penalized multivariate analysis, the primary two-variable Firth model identified UWT (OR 62.20 per mm; p < 0.001) and upper ureteral location (OR 14.27; p = 0.040) as independently associated with failed access (apparent AUC 0.986; 5-fold cross-validated AUC 0.967). The combined model achieved an AUC of 0.997 as an exploratory model with adequate calibration (Hosmer-Lemeshow test, p = 0.936). Exploratory DCA suggested potential net benefit across clinically relevant threshold probabilities. UWT ≥ 2.7 mm and PFS grade ≥ 2 on preoperative NCCT may be useful predictors of FA during unstented ureteroscopy, although these findings require external validation before clinical application. Incorporating these readily measurable CT parameters into a proposed preoperative risk-stratification algorithm may optimise surgical planning and patient counselling.