Femoral medullary composition in advanced post mortem decomposition: agreement of attenuation-based volumetric fractions across two PMCT reconstruction protocols

Abstract In advanced decomposition, conventional post mortem markers become progressively less informative, whereas bone and medullary content remain comparatively preserved. This retrospective study assessed whether femoral medullary composition can be quantified consistently on post mortem computed tomography (PMCT) using a semi-automated segmentation approach and whether attenuation-based volumetric fractions are associated with cadaveric alteration. PMCT examinations from 100 decomposed forensic cases (199 femora) were analysed with the SegMed module in 3D Slicer. Eight medullary fractions were quantified using soft-tissue (ST) and bone (BN) reconstruction protocols. Agreement was evaluated using intraclass correlation coefficients (ICCs), paired Wilcoxon tests and Bland–Altman analysis; associations with the Radiological Alteration Index (RAI) and estimated post mortem interval (PMI) were assessed using Spearman correlations with Benjamini–Hochberg correction. Agreement between ST and BN reconstructions was excellent for gas, fat-like marrow, fluid and intermediate-density fractions (ICCs > 0.94), despite small systematic offsets, whereas agreement was lower for residual medullary content on the right (ICC = 0.52) and unstable for debris. Gas fractions were positively associated with RAI, whereas fluid and intermediate-density fractions were negatively associated (|ρ| approximately 0.25–0.37). No volumetric fraction was significantly associated with estimated PMI after correction. Semi-automated PMCT segmentation therefore yielded consistent measurements of the principal femoral medullary fractions across the two reconstruction protocols evaluated. These fractions are candidate quantitative markers of radiological cadaveric alteration, but their value for elapsed-time estimation requires validation in cohorts with more accurately known PMIs.