Volumetric Analysis of Stone Burden: Comparative Evaluation of the Accuracy of Contemporary Volume Determination Tools
Aymon Ali, Seyed Amiryaghoub M. Lavasani, Yi Xi Wu, Yezan F. Hadidi, Grant Sajdak, Andrei D. Cumpanas, Jacob C. Tsai, Jaime Altamirano-Villarroel, Seyed Hossein H. Sharifi, Zhamshid Okhunov, Bruce M. Gao, Jonathan Badin-Castro, Sohrab N. Ali, Jaylen M. Lee, Pengbo Jiang, Roshan M. Patel, Duane Baldwin, Jaime Landman, Ralph V. ClaymanIntroduction:
Historically, ellipsoid formulae (EF) based on maximum linear measurements (MLMs) have been utilized for stone volume determination; however, the irregular shapes of renal calculi may introduce inaccuracy. Recent advances in CT-based volumetric assessment tools, including 3D Slicer ® (3DS), Enterprise Imaging (EI), and Visage Imaging ® (VI), offer the opportunity to estimate stone volume via segmentation-based methods. This study compares 3DS, EI, and four workflows of VI, against the industry standards for volume determination: Gas pycnometry (GP) and water displacement (WD).
Methods:
A total of 64 air-dried human stones (Group 1: MLM ≤1.5 cm; 0.4–1.5 cm) and 54 BegoStones (Group 2: MLM >1.5 cm; 1.6–6.1 cm) underwent CT scans in air and phantom mediums. Stone volumes were calculated using all aforementioned methodologies, including four VI subtools (balloon contouring, auto 3D segmentation, 3D segmentation with manual correction, and multimethod) compared with GP- and WD-determined volume based on their MLM. Mean-squared errors were compared using ANOVA with Dunnett’s
Results:
For stones with MLM <1.5 cm, Visage Balloon Contour demonstrated the highest accuracy (0.0003 cm³, 95% confidence interval [CI]: 0.0002–0.0004), followed by VI-Multi (0.0006 cm³, 95% CI: 0.0004–0.0008) and 3DS (0.0007 cm³, 95% CI: 0.0005–0.0011). EI and the EF showed comparable performance (0.0009 cm³ each). For stones with MLM ≥1.5 cm, EI achieved the highest accuracy (0.198 cm³, 95% CI: 0.147–0.267), followed by Visage Balloon Contour (0.274 cm³, 95% CI: 0.204–0.369). In contrast, the EF demonstrated a substantially higher error (3.514 cm³, 95% CI: 2.611–4.728). Visage 3D Segmentation exhibited the poorest performance across both size categories, with the highest errors observed for larger stones (29.572 cm³, 95% CI: 21.978–39.789).
Conclusion:
EI and VI’s Balloon Contour tool yielded volumetric measurements comparable with GP and WD.