Retrospective validation of 3D Slicer against CTA source images for accurate intracerebral hematoma volumetry: a comparative study with the Tada formula

Abstract

Background: Accurate measurement of intracerebral hemorrhage (ICH) volume is critical for clinical decisions and prognosis. While the Tada formula (ABC/2) is commonly used, its accuracy—particularly for irregular hematomas—remains limited. This study retrospectively validates the free, open-source software 3D Slicer against manual segmentation of CTA source images as the gold standard, and compares its performance with the ABC/2 method. Methods: This study retrospectively analyzed imaging data from 118 patients with spontaneous ICH who underwent both non-contrast CT (NCCT) and CTA. The CTA images were acquired with thin-slice reconstruction (0.625 mm slice thickness). Hematoma volume was measured by two blinded evaluators using the following methods: (1) manual segmentation on thin-slice CTA source images (gold standard); (2) semi-automatic segmentation on NCCT images using 3D Slicer software; and (3) calculation on NCCT images using the Tada formula (ABC/2). Intraclass correlation coefficient (ICC) was used to assess intra- and inter-observer agreement. Bland-Altman analysis and linear regression were employed to evaluate the agreement and bias between 3D Slicer, the ABC/2 method, and the gold standard. Results: Analysis of 118 spontaneous intracerebral hemorrhage patients revealed that the 3D Slicer method achieved near-perfect agreement (ICC = 0.997) with the gold standard of CTA manual segmentation, with a minimal mean bias of + 0.77 mL and narrow limits of agreement. In contrast, the Tada formula demonstrated significantly poorer agreement (ICC = 0.887), a substantial systematic overestimation (mean bias = + 4.65 mL), and wide limits of agreement (-10.8 to + 20.2 mL). Subgroup analyses further confirmed the superior robustness of 3D Slicer. The reliability of the Tada formula was markedly lower for irregularly shaped hematomas (ICC = 0.835) and deteriorated drastically for large hematomas > 60 mL (ICC = 0.516). Conclusion: In conclusion, this study validates 3D Slicer as a highly accurate tool for ICH volumetry using CTA source images, demonstrating superior performance over the Tada formula, especially for irregular hematomas. As a freely available platform, it represents a superior alternative for precise ICH assessment in both clinical and research settings.