Digital pair-matching of iliac bones: pilot study on a three-dimensional approach with models acquired through stereophotogrammetry

Three-dimensional (3D) pair-matching has brought about an innovative approach for the analysis of commingled skeletal remains, and it has been tested on bone models acquired through CT and laser scans. Here, 3D models of 40 innominate bones (20 left and 20 right) of 20 documented male individuals from a cemeterial skeletal collection were acquired through a stereophotogrammetric device (VECTRA M3, Canfield Scientific, Inc.). The ventral iliac surface was chosen as the anatomical region of interest (ROI) for the analysis. Each left ROI was mirrored and superimposed on the matching right ROI (contralateral element from the same individual) and mismatching ROIs (contralateral elements from different individuals). The point-to-point distance between models was calculated through the Vectra Analysis Module (VAM) software and the root mean square (RMS) point-to-point distance value was used to evaluate the sorting performance of the method, in terms of sensitivity and specificity rates. Differences in RMS between matches and mismatches were investigated through a Student's t test (p < 0.05). The state of preservation of the remains was assessed following an index of anatomical completeness and differences in RMS distances of true matches according to different anatomical completeness were assessed through the Mann-Whitney test (p < 0.05). RMS point-to-point distances of matches and mismatches were significantly different (p < 0.01), being the matches lower than mismatches. The RMS threshold of 2.9 mm identified all the true pairs; the test was 100% sensitive and 51% specific. The RMS of matches with a better state of preservation are significantly lower than the less preserved matches (p < 0.05). In general, a low RMS distance value may indicate a true match, being it to be further verified. The 3D approach for sorting innominate bones provides a valid screening test that could complete subjective and osteometric methods with numerical evidence of the match. Preliminary data suggest a possible relation between RMS distance values and taphonomic condition, which would benefit from further research.