Volumetric stone size assessment did not demonstrate improved prediction of treatment success compared with...

To evaluate whether volumetric assessment of stone burden improves prediction of treatment success compared with cumulative stone diameter (CSD) and whether performance varies by surgical modality or stone size.

We analyzed a prospectively maintained cohort undergoing PCNL or ureteroscopy (URS) from October 2020 to September 2024. Stone burden was assessed by CSD, ellipsoid volume, and software-derived three-dimensional volume. Treatment success was defined as no residual fragment ≥ 2 mm on postoperative CT. ROC analyses and three parallel multivariable logistic regression models, each containing one stone-burden metric plus stone multiplicity, HU, and surgical modality, were performed. Calibration was assessed using 500 bootstrap resamples, and modality interactions were tested.

Among 150 patients (81 PCNL, 69 URS), 91 (61%) achieved treatment success. Univariable AUCs were 0.632 (95% CI 0.540-0.724) for CSD, 0.602 (0.511-0.693) for ellipsoid volume, and 0.595 (0.504-0.687) for software-derived volume. Adjusted-model AUCs were 0.703 (0.615-0.791), 0.690 (0.602-0.779), and 0.687 (0.599-0.776), respectively, with no significant pairwise differences. Stone multiplicity was associated with lower odds of success in both volume-based models. No metric-by-modality interaction was significant. Exploratory PCNL analyses by a 2-cm CSD threshold had wide confidence intervals.

Volumetric assessment did not demonstrate improved prediction of treatment success compared with CSD. The three metrics showed similar, modest discrimination in parallel adjusted models. Stone multiplicity may capture clinically relevant complexity not represented by stone size alone.