Anthropology __ 2018 A1 Investigating the Accuracy of Additive Manufacturing Skeletal Samples for Evidence Reconstruction Rachael M. Carew, MSc*, Viale Muratori 185, Modena, MO 41124, ITALY; Ruth M. Morgan, PhD, University College London, UCL Centre for the Forensic Sciences, 35 Tavistock Square, London WC1H 9EZ, UNITED KINGDOM; and Carolyn Rando, PhD, University College London, Institute of Archaeology, 31-34 Gordon Square, London WC1H 0PY, UNITED KINGDOM After attending this presentation, attendees will be aware of the problems inherent within 3D modeling and additive manufacturing (3D printing) in forensic anthropology and will recognize that these issues are pertinent to forensic science reconstruction approaches and, in particular, the presentation of 3D prints as evidence in court. This presentation will impact the forensic science community by alerting attendees to the uses and limitations of additive manufacturing from Computed Tomography (CT) scans in a forensic context. Recommendations and thought-provoking findings could initiate discussions and collaborations toward further exploration into the use of models for reconstruction purposes in anthropology and wider forensic disciplines. This research investigates the metrology of 3D modeling and 3D printing osteological samples from CT scans. It is expected that such models will be sufficiently accurate for anthropological comparisons, but the extent of the effect of modeling parameters is currently unknown. There are two documented cases in which 3D prints of human remains have been used in United Kingdom courts of law.1,2 In one of these cases, lawyers cast doubt on the reliability of a 3D-printed cranium, as the process has not been validated in a forensic context.2 This type of evidence can be utilized after a virtual postmortem (avoiding disruption of remains), from antemortem clinical CT data, or in cases in which the subject survives. Studies have begun to validate these techniques, such as the reliability of obtaining accurate anthropological measurements from CT reconstructions of certain skeletal elements and the accuracy of 3D printing in medicine/anatomy, but further research is needed.3-6 Three dry osteological samples (a cranium, clavicle, and first metatarsal) were CT scanned using a multi-detector Toshiba Acquillon™ ONE. 3D models were segmented and converted into Stereolithic (STL) files for printing. First, the parameters applied in the virtual model generation were investigated for quality. Second, two additive manufacturing methods, Selective Laser Sintering (SLS) and Fused Deposition Modeling (FDM), were tested on three printers, and a statistical comparison of anthropometric measurements taken from the original samples, the virtual models, and the 3D prints was undertaken. The results of this study indicate that: (1) the threshold values (labels) used in virtual model segmentation affect the quality of the model, as does the amount of smoothing employed; (2) there was no statistically significant difference between the clavicle and first metatarsal 3D prints and the original dry bone (p values >0.05), except for one of the clavicle FDM prints; (3) the cranium 3D print was not statistically distinguishable from the virtual model (p values >0.05) following removal of inconsistent cranial measurements, but the cranium 3D print was statistically significantly different when compared to the source bone (p <0.05); (4) the virtual models had mean Absolute Errors (AE) of 1.6mm±0.9mm for the cranium, 1.5mm±0.7mm for the clavicle, and 1.1mm±0.7mm for the first metatarsal; and, (5) the 3D prints produced using SLS technology had smaller respective AE, with 1.4mm±0.9mm for the cranium, 1.2mm±0.2mm for the clavicle, and 0.7mm±0.5mm for the first metatarsal. This empirical research provides initial data to validate the process of additive manufacturing in forensic anthropology. The data demonstrated that accurate 3D prints can be produced from CT-scanned bones, but with limitations. Segmentation of the virtual model was found to be a crucial step for producing accurate models, and it is thought that applying additional smoothing could help. Further exploration of additive manufacturing and samples that exhibit trauma, pathology, and taphonomy will progress toward producing best practice guidelines and validation of the technique. Reference(s): 1. Baier W., D.G. Norman, J.M. Warnett, M. Payne, N.P. Harrison, N.C. Hunt, B.A. Burnett, and M.A. Williams. Novel Application of Three- Dimensional Technologies in a Case of Dismemberment. Forensic Sci Int. 270 (Jan 2017): 139-45. 2. Scott C. 3D Printed Skulls Presented as Evidence in Murder Trial, in a First for the British Legal System. 2016. Last accessed July 25, 2017, from https://3dprint.com/133715/ellie-butler-murder-trial/. 3. Brough A.L., J. Bennett, B. Morgan, S. Black, and G.N. Rutty. Anthropological Measurement of the Juvenile Clavicle Using Multi-Detector Computed Tomography-Affirming Reliability." J Forensic Sci. 58, no. 4 (Jul 2013): 946-51. 4. Fouri, Z., J. Damstra, P.O. Gerrits, and Y. Ren. Evaluation of Anthropometric Accuracy and Reliability Using Different Three-Dimensional Scanning Systems. Forensic Sci Int. 207, no. 1-3 (Apr 15 2011): 127-34. 5. Smith E.J., J.A. Anstey, G. Venne, and R.E. Ellis. Using Additive Manufacturing in Accuracy Evaluation of Reconstructions from Computed Tomography. Proc Inst Mech Eng. H 227, no. 5 (May 2013): 551-9. 6. McMenamin P.G., M.R. Quayle, C.R. McHenry, and J.W. Adams. The Production of Anatomical Teaching Resources Using Three- Dimensional (3D) Printing Technology. Anat Sci Educ. 7, no. 6 (Nov-Dec 2014): 479-86. Additive Manufacturing, Metrology, Evidence Reconstruction Copyright 2018 by the AAFS. Permission to reprint, publish, or otherwise reproduce such material in any form other than photocopying must be obtained by the AAFS. ______________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________ *Presenting Author - 38 - Anthropology __ 2018 A2 Radiographic Image Analysis and the Estimation of Age at Death in Adult Males Andrew C. Seidel, MA*, Arizona State University, Human Evolution/Social Change, PO Box 872402, Tempe, AZ 85287-2402; Laura C. Fulginiti, PhD, Forensic Science Center, 701 W Jefferson, Phoenix, AZ 85007; and Joel K. Simcoe, Maricopa County Office of the Medical Examiner, 701 W Jefferson Street, Phoenix, AZ 85007 After attending this presentation, attendees will understand how the analysis of digital radiographs of the pubic bone can be used to estimate age at death for male decedents. This presentation will impact the forensic science community by providing data indicating that digital radiographic images of the male os pubis can be reliably used to place individuals within intervals of age at death comparable to those of more commonly employed methods based on morphological changes of the pubic bone.1,2 The techniques developed in this research can be used to estimate age at death from pubic bones where details of the symphyseal face have been damaged or otherwise obscured and, with refinement, may eventually enable age-at-death estimation without excision and maceration of the pubic bone. This research was conducted using the Hartnett-Fulginiti collection curated at the Forensic Science Center in Maricopa County, AZ. This collection is comprised of more than 600 specimens of pubic symphyses from decedents of known sex, age at death, and ancestry. A training sample, composed of 100 pubic bones from male decedents ranging in age from 18 to 91 years, was selected to provide roughly equal representation from each of the seven morphological phases of the Hartnett-Fulginiti method for the estimation of age at death.2 Although preference was given to the left os pubis, the right was used when the left was damaged or otherwise unsuitable for analysis. Digital radiographs of each specimen were taken from a fixed distance and with constant settings. The resulting images were subjected to analysis in which characteristics of the distribution of gray values (mean, standard deviation, minimum, maximum, median, mode, skew, and kurtosis) were recorded. In an attempt to control for individual variation in size and shape, image analysis was constrained to the triangular area defined by the upper and lower bounds of the symphyseal face and the most medial point on the margin of the obturator foramen. Recorded characteristics were evaluated for their correlation to age at death, and a subset of six variables (comprised of standard deviation, maximum value, skew, kurtosis, and the constructed metrics of range and signed difference between mean and median) was selected for a k-means clustering analysis. Although Hartigan’s Rule suggested that 13 clusters was the optimal clustering solution for this data, this resulted in several clusters that were based around a small number of individuals, and a clustering solution of nine was adopted instead. Each of the nine clusters was defined by the location of its centroid and described by the mean, standard deviation, and range of the within-cluster age-at-death distribution. To test whether radiographic image analysis could produce viable age-at-death estimates, the six variables listed above were recorded for a secondary sample of 57 randomly selected males ranging in age from 22 to 84 years. The six-dimensional Euclidean