Magnetic Resonance Imaging Compared With Computed Tomography for Segmentation-Based Cavity Volume Measurements in Gelatin Head Phantoms

Gelatin-based phantoms are commonly used as soft-tissue simulants in wound ballistics research, yet MRI has not previously been applied for volumetric analysis in this setting. This study compared permanent cavity volumes measured on CT using soft and hard-kernel reconstructions (CTsoft, CThard) and MRI using 3D T1 and T2-weighted sequences (MRIT1, MRIT2) in 20 gelatin head phantoms shot with different ammunition types. Volumes were segmented manually and evaluated using correlation and Bland-Altman analyses. CT-MRI comparisons showed small mean biases (-1.07 to 3.03 cm3) but wide 95% limits of agreement (LoA), ranging from -22.1 to 19.9 cm3 depending on the modality pair, corresponding to potential deviations of approximately ±90% to 100% of a typical cavity volume. The closest agreement was observed for CThard-MRIT1 and CTsoft-MRIT2, while MRIT1-CTsoft exhibited the widest LoA. Metallic fragments did not impair segmentation, whereas air-filled regions produced susceptibility-related signal voids on MRI. Overall, CT and MRI both allow reliable visualization and measurement of permanent wound channels, but variability between modalities limits their interchangeability for precise volumetry in individual cases.