3D-3D superimposition of pubic bones: an addition to the personal identification toolkit

When dealing with highly decomposed human remains, some of the primary identifiers may fall short due to the extent of postmortem decay and dispersion of body parts. In such conditions, the investigators must rely on alternative strategies to reach a positive personal identification. Superimposition of 3D models of anatomical structures from CT scans have brought about an innovative and efficient approach for the identification of unknown individuals, mainly focusing on the uniqueness of frontal and sphenoid sinuses. This short study applies the 3D-3D superimposition of pubic bones to the personal identification procedure, and aims at expanding the anthropological identification toolkit. Eighty abdominal CT scans were selected from a hospital database, including 40 males and 40 females. From each CT scan, pubic bones were semi-automatically segmented twice with ITK-SNAP, to simulate an antemortem (AM) and a postmortem (PM) model. For the segmentation, the right and left lateral limits were set at the medial margin of the iliopubic eminence and the inferior limit was at the middle point of the ischiopubic ramus. The two 3D models of pubic bones from the same and different individuals were then automatically superimposed according to the least point-to-point difference between the two surfaces through the VECTRA Analysis Module (VAM), recording the root mean square (RMS) point-to-point distance. For each sex, 40 matches and 40 mismatches were so created. Differences in RMS distance values between matches and mismatches were investigated through a paired Student’s t-test (p<0.05). Both in the male and female groups, RMS distance values were significantly lower in matches than in mismatches (p<0.0001). In fact, males’ RMS distance values ranged between 0.03 mm and 0.96 mm (mean 0.37 ± 0.26 mm) for the matches and between 1.06 mm and 5.26 mm (mean 2.30 ± 0.89 mm) for the mismatches. Female’s RMS distances ranged between 0.01 mm and 0.97 (mean 0.28 ± 0.26 mm) for the matches and between 0.97 mm and 3.33 mm for the mismatches (mean 2.01 ± 0.54 mm). No overlaps between matches and mismatches were found in males; however, one superimposition of a mismatch in the female group produced a RMS distance values overlapping with the matches. This pilot study showed that 3D-3D registration of anatomical structures could provide a helpful addition to the identification procedures in the instance, for example, of unavailable ordinary anthropological indicators of identity, such as frontal sinuses. The quantitative outcome of the superimposition could then be used to strengthen a presumptive identifications or an exclusion, when appropriate AM CT images of this body portion are accessible.