Can mandibular lingual canals be used as a forensic fingerprint?

Abstract

Objectives: This study aimed to identify whether the lingual canals of the mandible can be used as a unique fingerprint when dealing with forensic victim identification.

Materials and methods: The study consisted of two parts; an observational part and an objective image analysis part. In the observational part a total of 100 in vivo high resolution CBCT datasets of human mandibles were included in the process of simulated matching of ante-mortem (AM) and post-mortem (PM) data. For the objective image analysis part 10 dry human mandibles were scanned with 2 different Cone Beam Computed tomography (CBCT) machines. In the observational part of the study trained observers attempted to correctly identify matching pairs of images taken from the same mandible out of a series of 100 mandibles. The aim was to simulate matching of the neurovascular structures on AM and PM mandibular midline images and determine the percentage of mandibles identified correctly. In the objective image analysis part, simulated matching was carried out using a specific CBCT dataset acquired to mimic a PM dataset and 10 datasets acquired from a different CBCT device which served as the source of potential AM cases. Comparison between AM and PM datasets resulted in the matching of the AM data and PM data obtained from the same mandible, leading to an assumed correct identification.

Results: The observational part of the study showed an average 95% correct identification of the mandibular midline neurovascular structures. Registration of mandibles resulted in perfect overlap of the same mandible from 2 different CBCT machine with an error distance equalling zero, while the registration of different mandibles deviated on average error distance 0.13 mm to 0.18 mm.

Conclusion: The percentage of fit for the simulated AM and PM data of the same mandible was 100%. This finding together with the significant deviations noted for the non-matching cases, may have a potential role in forensic identification in the same way that fingerprints are recognised as being a unique identifying feature.

Fig.1

Observational chart with 12 views of 11 different mandibles, with only 2 cross-sectional views on the plate deriving from one and the same mandibular midline.

Fig. 2

Manual registration for the same mandible with two different machine showing perfect overlap where 3D model shows a yellow color represent the post mortem case and the pink representing the ante mortem case.

Fig. 3

Cross-sectional views of different ante mortem mandibles. The contour around the bone shows the mask inside the ROI. The second contour is from the post mortem case which is manually registered at the level of the lingual canals. It is clearly seen that there is no perfect overlap.

Fig. 4

Results of the observation of the matched mandibles of the 100 in vivo cases of the mandible.

Fig. 5

Mandible A: The 3D color coded model after part comparison analysis showing values located around 0mm deviation, which is represented by the green color.The histogram ranges from -0.0005mm to 0.0005mm and the sliders fixed at -0.0005mm and 0.0005 mm.Mandible B and C: 3D models showing the minimum and maximum deviation between different mandibles and the histogram range is going from the minimum and maximum deviation found in the part analysis however the threshold sliders for defining the color code are fixed at -2.00mm to 2.00mm.

Fig. 6

Cross sectional view of different mandibles showing multiple shapes and directions of the lingual canals.