URSL Angle for Predicting Intraoperative Conversion

Overview

This international multicenter retrospective cohort study aims to externally validate the URSL Predictive Angle, a novel CT-derived morphometric parameter, as an imaging biomarker for predicting intraoperative conversion risk during transurethral ureteroscopic lithotripsy (URSL) for ureteral calculi. Intraoperative conversion is defined as the unanticipated abandonment of the planned endoscopic approach due to unfavorable ureteral anatomy or stone impaction, necessitating recourse to open ureterolithotomy, laparoscopic ureterolithotomy, or ureteral stent placement with planned staged secondary intervention. Such conversions are associated with protracted operative duration, heightened anesthetic exposure, and increased physical, psychological, and financial burdens on patients. The URSL Predictive Angle is a geometric parameter derived from preoperative coronal CT reformations that quantifies the degree of local ureteral angulation induced by an impacted calculus. Preliminary single-center data demonstrated robust predictive fidelity of this metric with an area under the ROC curve of 0.861 and an optimal discriminatory threshold of 131.8° yielding a sensitivity of 82% and a specificity of 88%. This study will enroll eligible patients from multiple domestic and international tertiary care hospitals who underwent URSL for ureteral calculi between January 1, 2018, and December 31, 2025. The primary outcome is intraoperative conversion of surgical approach. The predictive performance of the URSL angle will be evaluated using ROC curve analysis with AUC calculation. Propensity score matching and multivariable logistic regression will be employed to address potential confounding. This study is registered as a patient registry with a cohort model. Data collection is retrospective, involving abstraction from electronic medical records, operative logs, and PACS. A waiver of informed consent has been requested based on the retrospective design, minimal risk determination, stringent anonymization procedures, and impracticability of obtaining individual consent.

This international multicenter retrospective cohort study aims to externally validate the URSL Predictive Angle, a novel CT-derived morphometric parameter, for predicting intraoperative conversion risk during transurethral ureteroscopic lithotripsy (URSL) for ureteral calculi. BACKGROUND Transurethral ureteroscopic lithotripsy is the first-line minimally invasive treatment for ureteral calculi. However, in a subset of patients, intraoperative stone impaction, severe ureteral tortuosity, or anatomic variants may prevent successful ureteroscope advancement. In such cases, the surgeon must abandon the planned endoscopic approach and convert to open ureterolithotomy, laparoscopic ureterolithotomy, or temporizing stent placement with staged intervention. These unanticipated conversions increase operative duration, anesthetic exposure, and patient burden. Currently, preoperative conversion risk assessment relies largely on subjective surgeon judgment, with no objective, reproducible predictive tool available. Existing literature focuses on stone attributes (size, location, Hounsfield units) or ureteral wall thickness, but ureteral geometry remains unexplored. PRELIMINARY DATA Our group previously conducted a single-center study developing the URSL Predictive Angle, measured on preoperative coronal CT. This angle quantifies local ureteral angulation caused by an impacted calculus. Preliminary results showed an AUC of 0.877, with an optimal cutoff of 132.1° (sensitivity 80.3%, specificity 89.8%). However, external validation across diverse populations and institutions is needed. STUDY OBJECTIVES Primary Objective: To externally validate the predictive performance of the URSL angle as an imaging biomarker for intraoperative conversion risk, assessing AUC, sensitivity, specificity, and optimal threshold. Secondary Objectives: (1) To evaluate inter-observer agreement and test-retest reliability of URSL angle measurements across international centers; (2) To investigate the incremental predictive value of the URSL angle when combined with stone size, location, and Hounsfield units. STUDY DESIGN This is an international multicenter retrospective cohort study. Eligible patient records will be retrieved from January 1, 2018 to December 31, 2025 (eight-year period) from multiple domestic and international tertiary care hospitals with accredited urology services. Consecutive enrollment will be applied. PARTICIPANT SELECTION Inclusion Criteria: (1) Age \> 18 years; (2) Preoperative imaging confirmation of unilateral ureteral calculi (solitary or multiple, with index stone designated); (3) Planned and attempted URSL as initial treatment; (4) Available preoperative thin-slice CT (≤1.25 mm) in DICOM format with sufficient quality for multiplanar reconstruction. Exclusion Criteria: (1) No preoperative CT or poor image quality; (2) Incomplete medical records or insufficient operative detail; (3) Prior ipsilateral percutaneous nephrostomy or ureteral stent; (4) Calculi at ureteropelvic junction or ureteral orifice; (5) Prior ipsilateral ureteral surgery or urinary diversion. GROUP ALLOCATION Patients will be classified into Conversion Group or Success Group based on intraoperative conversion criteria. Two investigators will independently review operative records, with arbitration by a third senior urologist if disagreement occurs. Indeterminate cases will be excluded. OUTCOME VARIABLE Primary Outcome: Intraoperative conversion, defined as unanticipated abandonment of URSL due to technical impediments from unfavorable ureteral anatomy or stone impaction, requiring open ureterolithotomy, laparoscopic ureterolithotomy, or stent placement with staged intervention. Conversions due to equipment failure or non-anatomic causes are excluded. PREDICTOR VARIABLE Primary Predictor: URSL Predictive Angle. Standardized measurement protocol: (1) Select coronal plane with maximal stone cross-sectional area; (2) Measure longest stone axis, marking superior and inferior poles as Point A and Point B; (3) Trace 5mm cephalad from Point A along ureteral centerline to Point A'; (4) Trace 5mm caudad from Point B to Point B'; (5) Identify Point C on stone contour farthest from line AB; (6) Calculate ∠A'CB'. The 5mm offset distance balances avoidance of periureteral inflammation while preserving stone-induced deformation. Dual independent measurement with arbitration for discrepancies \>5° will be performed. COVARIATES Demographic: age, sex, BMI. Stone-related: laterality, multiplicity, location, length, Hounsfield units. Comorbidities: hypertension, diabetes, coronary artery disease. Quality control variables include center code, data abstractor ID, and measurer ID. SAMPLE SIZE Sample size was calculated using PASS 2025 based on ROC curve analysis. Parameters: anticipated AUC = 0.80, null AUC = 0.50, α = 0.05 (two-tailed), power = 0.90, event rate = 4%, attrition = 10-15%, multicenter design effect = 1.25. Unadjusted sample size: 1,700 patients (68 Conversion, 1,632 Success). Final target: 2,125 patients (85 expected in Conversion Group). DATA MANAGEMENT Data will be collected using a standardized electronic case report form in Excel. Each center will employ dual independent data entry with reconciliation. The coordinating center (The First Hospital of Jilin University) will perform centralized logic checks, range validation, and missing data analysis. All data will be anonymized before transmission. The database will be locked after verification and archived for 10 years. STATISTICAL ANALYSIS Analyses will be performed using SPSS 27.0 and R software. P \< 0.05 (two-tailed) is considered significant. Descriptive statistics: Mean ± SD for normal continuous variables; median (IQR) for non-normal; frequencies and percentages for categorical variables. Propensity score matching (1:3 nearest-neighbor, caliper 0.02) will balance age, sex, BMI, stone laterality, stone location, stone multiplicity, hypertension, diabetes, and coronary artery disease. Balance will be assessed using standardized mean differences (SMD \< 0.10). Univariate analysis in the matched cohort will examine associations with conversion status. Variables with P \< 0.10 will enter multivariable logistic regression, with the URSL angle as the primary independent variable. ROC curve analysis will evaluate discriminatory capacity, with AUC and 95% CI reported using the DeLong method. The Youden index will identify the optimal cutoff, with corresponding sensitivity, specificity, PPV, and NPV. Missing data: Patients missing primary outcome or predictor will be excluded. Complete-case analysis will be used for covariates. BIAS MITIGATION Selection bias: Consecutive enrollment, explicit eligibility criteria, and multicenter participation. Information bias: Standardized eCRF, uniform imaging protocol, dual independent measurement, outcome blinding during measurement, centralized data verification. Confounding: Propensity score matching and multivariable regression in the matched cohort. Center effects: Uniform protocol training, regular inter-center communication, and cross-center measurement consistency audits. QUALITY MANAGEMENT Each center will designate a site PI, two data abstractors, and two imaging measurers. Center-level internal audits will verify data fidelity. The coordinating center will perform completeness checks, logical consistency verification, and outlier detection. Inter-observer agreement for imaging measurements will be assessed using ICC, with ICC \> 0.80 considered satisfactory. ETHICAL CONSIDERATIONS IRB approval will be obtained from the coordinating center and each participating center prior to study initiation. A waiver of informed consent will be requested based on: (1) retrospective design with secondary use of clinical data; (2) minimal risk determination; (3) stringent data anonymization; (4) impracticability of obtaining individual consent across an eight-year international multicenter cohort. The study will comply with the Declaration of Helsinki and applicable privacy regulations (China PIPL, EU GDPR, US HIPAA). All data will be de-identified, with no direct personal identifiers in the eCRF. Data transmission will be encrypted, and the central database will be password-protected with restricted access. No individual patient or institution will be identifiable in any publication.