- Email: [email protected]
Morphometric investigations to assess the compatibility of mandible and skull
Forensic Science International 286 (2018) 193–198
Contents lists available at ScienceDirect
Forensic Science International journal homepage: www.elsevier.com/locate/forsciint
Morphometric investigations to assess the compatibility of mandible and skull Sandra Preisslera , Marcel A. Verhoffa , Frank Ramsthalerb , Franziska Holza , Axel Gehlc, Sarah C. Koelzera,* a
Institute of Legal Medicine, Department of Forensic Medicine, Goethe University, Kennedyallee 104, 60596, Frankfurt/Main, Germany Institute of Legal Medicine, Department of Forensic Medicine, University of Saarland Medical School, Kirrberger Str. 100/Building 100, 66421, Homburg/Saar, Germany c Institute of Legal Medicine, Department of Forensic Medicine, University Hospital Hamburg-Eppendorf, Building North N81, Butenfeld 34, 22529, Hamburg, Germany b
A R T I C L E I N F O
Article history: Received 18 October 2017 Received in revised form 4 March 2018 Accepted 8 March 2018 Available online 16 March 2018 Keywords: Mandible Skull Compatibility Computed tomography Commingling Intra-observer reliability
A B S T R A C T
When a morphologically separated skull and mandible are found in the same case context, the possibility of a match arises. Two criteria with which to determine a match are the rough articulation between the mandibular condyles and cranial base itself and, most importantly, the fit of the teeth. However, when there has been intravital or postmortem tooth loss, this important criterion is not available. To date, only Reichs (1989) has investigated further compatibility criteria to solve the question of putative commingling in a case where a mandible seemed to originate from a female, while all other bones originated from a male individual. In a different reported case (Preißler et al. 2017), a mandible seemed too big for a skull; DNA analysis, however, confirmed that both originated from the same female individual. To investigate the metric relationship between mandible and skull we measured the postmortem CT data records of 223 corpses (virtual skulls) in OsiriX© MD for the following linear parameters: bicondylar breadth (KDB), biradicular breadth (AUB), and bizygomatic breadth (ZYB). The indices KDB/ZYB and KDB/ AUB were developed and used to define ranges for matches and mismatches. Furthermore, the intraobserver reliability for the method was assessed. An intraclass correlation coefficient of >0.99 for every parameter showed that the used measurements are highly reliable. The 2.5–97.5 percentile for the KDB/AUB index lay between 0.91 and 1.05, while the range for the KDB/ ZYB index was between 0.87 and 1.00. Within these ranges, it is possible to roughly assess whether or not a mandible and skull might be compatible, even if this can only be verified by forensic DNA analysis. If an index value lies outside these ranges, it can be assumed that skull and mandible do not match. Future studies should include more samples from a broader population spectrum so that these metric relationships can be used for different populations. © 2018 Published by Elsevier B.V.
1. Introduction Anthropometric measurements belong to the essential methods in anthropology used to scale the human body. They are performed with the aim of standardizing measurements and of obtaining reliable anthropometric values [1]. Anthropometric measurements of the mandible are performed not only to address oral surgical [2] or clinical [3,4] questions, but also to address forensic issues, such as the estimation of age [5,6] or sex [5,7–10] in different populations. De Oliveira et al., for example,
* Corresponding author. https://doi.org/10.1016/j.forsciint.2018.03.013 0379-0738/© 2018 Published by Elsevier B.V.
investigated the correlation of sex and of age with mandibular ramus length in a Brazilian population. They found that while “mandibular ramus length is effective to estimate chronological age with a high degree of accuracy,” it is not an accurate measurement with which to estimate sex [5]. In the forensic sciences, there is a dearth of adequate bone measurement series from recent populations. To alleviate this shortcoming, the use of imaging techniques like computed tomography (CT) has increased in the last few years. In particular, anthropometric measurements of the skull have been performed to assess sex in different populations [8–14]. Kölzer et al., for example, investigated the frontal inclination of the human skull as a trait of sexual dimorphism. Because these investigators did not
194
S. Preissler et al. / Forensic Science International 286 (2018) 193–198
find significant sex differences with the variables they had used in their pilot study, they concluded that additional variables, based strictly on considerations of validity and reliability, needed to be tested [13]. In yet another study, CT scans were used not only for statistical analysis, but also to generate a 3D atlas of the human skull for further reference [15]. In diverse publications, Adams and Byrd emphasize that different sorting techniques, such as articulation, taphonomy, and elimination, should be used conjunctively to obtain best resolution results for commingled human skeletal remains [16– 18]. With the exception of Reichs’ study, no systematic measurements pertaining to the compatibility of mandible and skull have been taken so far. In her case report, Reichs describes the discovery of a human skull along with other human bones, including a mandible. Because the mandible was small and gracile and did not articulate well with the skull, a total of 16 other skulls were measured for comparison. The results indicated that, in this case, the mandible was, in fact, not too small for the skull and that the bones in question could well have stemmed from the same individual [19]. In another reported case [20], in which the investigators were confronted with toothless jaws, a mandible appeared to be too big for the skull. Although the skull was obviously female, the mandible was thought to belong to a male individual. Nonetheless, subsequent DNA analysis showed that both bones originated from the same female individual. To investigate the metric relationship between mandible and skull, digital measurements using volume-rendered CT skulls were performed in the present study. To this purpose, three linear parameters were measured and two new indices developed. The primary question was whether there were values for both indices which, in case of a transgression below or above, would indicate that a mandible and skull do not belong together. The procedure was carried out for the entire collective, separately for both sexes. Additionally, the study tested whether the three linear parameters are applicable for sex discrimination.
parameters were collected: bicondylar breadth, biradicular breadth, and bizygomatic breadth. Definitions and abbreviations of the linear parameters are given in Table 1. Fig. 1 shows examples for each parameter. All parameters were measured a second time by the same examiner to review the intra-observer reliability. Prior to the start of the experiments, approval was obtained from the ethical committee of the faculty of medicine at the JustusLiebig University Gießen (file number: 63/09). Statistical analyses were performed using SPSS v 17.0 and MedCalc v 16.8.4. Apart from descriptive statistics of the linear parameters, the reliability of the used method was determined. Afterwards a canonical discriminant analysis was performed. For the descriptive statistics, two indices were developed from linear parameters: The bicondylar breadth was set in relation to the biradicular breadth and also in relation to the bizygomatic breadth. Intra-observer reliability was assessed by calculating the intraclass correlation coefficient and variation coefficient. To verify the variance of both indices for matches and mismatches between mandible and skull, a comparison group was created. In this comparison group, mandibles were randomly matched with skulls, separately for each sex.
2. Materials and methods
3.2. Match – mismatch
Altogether 223 virtual skulls (113 female, 106 male, 4 unknown), generated from anonymized postmortem CT data records from the Institute of Legal Medicine in Hamburg, were examined. The individuals were scanned in a Diamond Select MX 8000 Quad CT-Scanner (slice thickness 1.3 mm, 120 kV, 299– 349 mAs, pitch: 0.88, image matrix size: 512 512, Phillips, Amsterdam Netherlands). The associated metadata for each individual consisted of age, sex, weight, and height. Skulls with bone fractures and highly fragmented skulls were excluded from measurements. The CT data were imported into an OsiriX© MD database. The virtual skulls were aligned in the volume-rendering mode in basilar view and measured with the help of the measurement tool in OsiriX© MD (in cm). The following
To find out whether a mandible meshes with a skull, two indices were developed: KDB/AUB and KDB/ZYB. Descriptive statistics were performed for both indices and are summarized in Tables 3 and 4. The “notched box-and-whisker plot” (Fig. 3) displays a graphical statistical summary of the selected index values. The central notched box represents the values from the mid 50% of the data (25–75 percentile). Red horizontal bars indicate data distribution within the 2.5–97.5 percentile (95% of the data falls within this range). The plots on the right-hand sides of Fig. 3A–F illustrate the wider spread of value distribution in the control group (mismatching pairs of mandibles and skulls that were generated randomly, both separately for each sex and combined for both sexes).
3. Descriptive results 3.1. Reliability Table 2 shows the intraclass correlation coefficient (ICC) and the 95% confidence interval for both single measures and average measures. All three linear parameters show values >0.99 for ICC (p 0.01). Furthermore, mountain plots were created for each linear parameter. These mountain plots (Fig. 2A–C) visualize the difference between the first and the second measurement for each parameter. They illustrate the symmetric character of the linear parameters. ZYB is the best linear parameter, with a difference less than 0.15 cm when comparing both measurements, while KDB shows the highest differences with up to 0.45 cm.
Table 1 Definitions and abbreviations of the linear parameters. The abbreviations for bizygomatic breadth and biradicular breadth were taken from Martin and Knußmann [1]. For bicondylar breadth no abbreviation was found, so an abbreviation according to the corresponding landmark was chosen, in which B was substituted for L. Linear parameter
Abbreviation Definition (Martin and Knußmann [1])
Bicondylar breadth Biradicular breadth Bizygomatic breadth
KDB AUB
Linear distance, measured externally from the most lateral point on the left condyle to the most lateral point on the right (condylion to condylion) Linear distance between the lowest points of the zygomatic arc root (radiculare to radiculare)
ZYB
Linear distance, measured from the most lateral point on the left zygomatic arc to the most lateral point on the right (zygion to zygion)
S. Preissler et al. / Forensic Science International 286 (2018) 193–198
195
Fig. 1. Virtual skull in basilar view with the linear parameters: (A) bicondylar breadth (KDB), (B) biradicular breadth (AUB), and (C) bizygomatic breadth (ZYB).
Table 2 Results for the tested intra-observer reliability (ICC = intraclass correlation, CI = confidence interval). Linear parameter
KDB AUB ZYB
Single measures
Average measures
ICC
95% CI
ICC
95% CI
0.9956 0.9976 0.9983
[0.9943–0.9966] [0.9969–0.9982] [0.9977–0.9988]
0.9978 0.9988 0.9992
[0.9971–0.9983] [0.9985–0.9991] [0.9989–0.9994]
4. Discussion In the present study, 223 cranial CT scans were investigated to assess the compatibility of mandible and skull. The reason for the study was a case, previously reported by us, in which a mandible seemed to be too large for the skull and, therefore, did not seem to be compatible. This assumption was, however, disproved by forensic DNA analysis [20]. The case illustrates how difficult it may sometimes be to decide whether these bones articulate well and, therefore, whether they both belong to the same decedent or not. One reason for misjudgment may be found in the anatomy of the mandibular joint, where the caput mandibulae of the
processus condylaris articulates with the facies articularis of the fossa mandibularis and the tuberculum articulare at the os temporale. Between the caput mandibulae and fossa mandibularis, there is the discus articularis, which is fused with the surrounding joint capsule [21]. Because the fossa mandibularis is usually larger than the caput mandibulae, the articular disk has to balance the incongruity [22]. Conversely, it can imply that this disk, along with the comparatively flat subcranial joint socket, may allow the caput mandibulae to protrude both laterally and medially from the articulation. This protrusion may lead to a misjudgment of the “fit” between mandible and skull, as in the case we reported. After 3D reconstruction of the available virtual skulls, the linear parameters bicondylar breadth (KDB), biradicular breadth (AUB), and bizygomatic breadth (ZYB) were measured in the volumerendering modus of OsiriX© MD. The results of the reliability analysis (all values > 0.99) again confirm that the used digital method is highly reliable for linear parameters [23]. The prerequisite for this method is an adequate resolution of the 3D datasets, which is given at a slice thickness of 1.3 mm or less [23]. In the present study, slice thicknesses of 1.3 mm were used exclusively. Furthermore, the landmarks for the
Fig. 2. Mountain plots for the three measurements: (A) ZYB, (B) AUB, and (C) KDB. The differences for both measurements for each skull are given on the X-axis in [cm].
196
S. Preissler et al. / Forensic Science International 286 (2018) 193–198
Table 3 Summarized statistics for the KDB/ZYB index. Empty fields indicate that the sample size was too low. Stat. parameter
All cases
Female
Male
Unknown sex
Sample size Lowest value Highest value Mean 95% CI mean Median 95% CI median Variance Standard deviation Standard error mean 2.5–97.5 percentile
223 0.8448 1.0367 0.9338 [0.9294–0.9381] 0.9357 [0.9316–0.9395] 0.001104 0.03322 0.002225
113 0.8448 1.0367 0.9398 [0.9332–0.9464] 0.9417 [0.9338–0.9507] 0.001247 0.03531 0.003322 0.87–1.00
106 0.8488 0.9885 0.9285 [0.9228–0.9342] 0.9330 [0.9236–0.9371] 0.0008702 0.02950 0.002865 0.87–0.98
4 0.8650 0.9379 0.9025 [0.8519–0.9532] 0.9036 0.001013 0.03183 0.01591
Table 4 Summarized statistics for the KDB/AUB index. Empty fields indicate that the sample size was too low. Stat. parameter
All cases
Female
Male
Unknown sex
Sample size Lowest value Highest value Mean 95% CI mean Median 95% CI median Variance Standard deviation Standard error mean 2.5–97.5 percentile
223 0.8828 1.0833 0.9785 [0.9738–0.9832] 0.9783 [0.9727–0.9858] 0.001265 0.03557 0.002382
113 0.8987 1.0620 0.9782 [0.9715–0.9848] 0.9806 [0.9694–0.9894] 0.001273 0.03567 0.003356 0.91–1.05
106 0.9100 1.0833 0.9799 [0.9732–0.9866] 0.9786 [0.9694–0.9873] 0.001218 0.03491 0.003390 0.91–1.05
4 0.8828 0.9915 0.9491 [0.8737–1.0246] 0.9611
three linear parameters used in this study are well defined and easy to determine in the basilar view of the virtual skulls. The mountain plots underpin the results of the reliability test. Overall, the majority of measurement differences lay within 0.2 cm, “which is considered an acceptable amount of error in forensic anthropology” [24]. With a deviation of less than 0.15 cm, the best result was achieved when measuring ZYB, while KDB showed the highest deviation, of up to 0.45 cm. These measurement discrepancies can be explained by the image quality of the reconstructed skulls, which differs among the data records. Especially the contrast of some skull reconstructions was very dark, which occasionally made it difficult to localize the external borders of the condyles. In recent years, the use of CT data for anthropometric measurements has been investigated in comparison with the use of classical methods. Kragh et al. [25], for example, compared different forensic anthropometrical techniques. They took 50 human skulls and measured them manually as well as by a digital image method and found that 31 of the investigated dimensions, including bizygomatic breadth, basion-bregma-height, and biasterionic breadth, are reliable for both techniques. Stull et al. [24] investigated the accuracy and reliability of measurements of 3D volume-rendered whole body images and arrived at the same assessment. Furthermore, Verhoff et al. [23,26] performed osteometric measurements using 3Dreconstructed skulls as well as macerated skulls and showed that both techniques provide similar results. For the present study, two new indices were developed to obtain values for the relationship between mandible and skull. These values were supposed to help decide whether or not a mandible meshes with a skull. The range for the KDB/AUB index lay between 0.8828 and 1.0833. The range for the KDB/ZYB index was between 0.8448 and 1.0367. It should, however, be kept in mind that these values are only valid if the individuals in question originate from the same population from which the data were obtained [27,28].
0.002248 0.04742 0.02371
The notched-box-and whisker-plots in Fig. 3 visualize the data distribution for the two index values. In the graphs, match and mismatch of mandible and skull are juxtaposed for each index and sex, and additionally for all cases combined. Although the ranges for both indices are overall comparable for both sexes (Tables 3 and 4), there is a slight difference in the range of the KDB/ZYB index between the sexes, in the sense that the upper limit is slightly higher for females than for males. Although this observation is unlikely to be relevant in practice, it is, nevertheless, worth noting. From these data distributions, it can be deduced that if a skull and a mandible are found and their index values are within the 95% range of the boxplots, they might belong together, though an affiliation cannot be proved. The boxplots for the randomly matched skulls and mandibles show a considerably larger index value range than the matching ones. If the calculated index is greater or less than the 2.5–97.5 percentile (Tables 3 and 4, last row) in a concrete case, it can, on the basis of the two-sided analysis, be assumed with a probability of approximately 97.5% that the mandible and skull do not match. In her case report, Reichs [19] measured six parameters including fossa breadth and bifossae breadth to verify the compatibility of a mandible and a skull. However, these parameters could not be measured with the 3D reconstructions used in our study because the condyles of the mandibles were sitting in their joint sockets. For this reason, we chose to measure linear parameters with an anatomical relation to the condyles of the mandible. Although, lastly, only DNA analysis can provide conclusive evidence for a match, the two indices we developed indicate a range within which a match between a skull and a mandible cannot be excluded. However, perhaps more importantly, the true value of these indices appears to be their power to exclude a match, rather than to prove one. Future studies should include more samples from a broader population spectrum so that these metric relationships can be used for different populations.
S. Preissler et al. / Forensic Science International 286 (2018) 193–198
197
Fig. 3. “Notched box-and-whisker plots” of the selected index values. The central notched box represents the values from the mid 50% of the data (25–75 percentile). The horizontal middle black line represents the median, whereas the notches display the 95% CI of the median. The horizontal line extends from the minimum to the maximum value. Outlier values are displayed as separate points (defined as a value that is smaller than the lower quartile minus 3 times the interquartile range, or larger than the upper quartile plus 3 times the interquartile range). Red horizontal bars display data distribution within the 2.5–97.5 percentile (95% of the data falls within this range). The plot illustrates the wider spread of value distribution in cases of mismatching mandible-skull pairs (right side). A and B show the boxes for the KDB/ZYB index and the KDB/AUB index for both sexes, C and D show the distribution of females for the KDB/ZYB index and for the KDB/AUB index, respectively. Finally, E and F show the distribution of males for the KDB/ZYB index and for the KDB/AUB index, respectively.
198
S. Preissler et al. / Forensic Science International 286 (2018) 193–198
5. Conclusion for practice If a skull and a mandible are found morphologically separated, but in the same case context, a possible match should first be morphologically checked by seeing if the teeth fit. If this is impossible because of intravital or postmortal tooth loss, the use of the more questionable “fit” of the mandibular joint becomes relevant. Should the mandibular condyles seem too big or too small in comparison to the skull, examinations regarding sex discrimination should be performed. When the sex accords for both the skull and the mandible, the indices proposed in this study can be determined. For values lying within the sex-specific 2.5–97.5 percentile ranges of the indices it is possible to roughly assess whether or not the skull and mandible might be compatible, even if this can only be verified by forensic DNA analysis. If, on the other hand, an index value lies outside these ranges, it can be assumed that the skull and mandible do not match. Even if sex cannot be determined for the skull and the mandible in question, the indices can still be used in view of the fact that the 2.5–97.5 percentile ranges for the KDB/AUB index are the same for both sexes anyway. It would, however, be advisable to use the somewhat wider range (0.91–1.05) of the KDB/ZYB index for female sex, just to be on the safe side. Conflict of interest The authors declare no conflict of interest. References [1] R. Martin, R. Knußmann, Anthropologie, Handbuch der vergleichenden Biologie des Menschen. Band I: Wesen und Methoden der Anthropologie, Gustav Fischer Verlag, Stuttgart, New York, 1988, pp. 129–182. [2] C.J. Našel, M. Pretterklieber, A. Gahleitner, C. Czerny, M. Breitenseher, H. Imhof, Osteometry of the mandible performed using dental MR imaging, Am. J. Neuroradiol. 20 (7) (1999) 1221–1227. [3] Y.L. Zhang, J.L. Song, X.C. Xu, L.L. Zheng, Q.Y. Wang, Y.B. Fan, Z. Liu, Morphologic analysis of the temporomandibular joint between patients with facial asymmetry and asymptomatic subjects by 2D and 3D evaluation: a preliminary study, Medicine 95 (13) (2016)e3052. [4] L.F. Merigue, A.C.D.C.F. Conti, P.V.P. Oltramari-Navarro, R.D.L. Navarro, M.R.D. Almeida, Tomographic evaluation of the temporomandibular joint in malocclusion subjects: condylar morphology and position, Braz. Oral Res. 30 (1) (2016), doi:http://dx.doi.org/10.1590/1807-3107BOR-2016.vol30.0017. [5] F.T. de Oliveira, M.Q.S. Soares, V.A. Sarmento, C.M.F. Rubira, J.R.P. Lauris, I.R.F. Rubira-Bullen, Mandibular ramus length as an indicator of chronological age and sex, Int. J. Legal Med. 129 (1) (2015) 195–201. [6] D. Franklin, A. Cardini, Mandibular morphology as an indicator of human subadult age: interlandmark approaches, J. Forensic Sci. 52 (5) (2007) 1015– 1019. [7] M.A.A. Kharoshah, O. Almadani, S.S. Ghaleb, M.K. Zaki, Y.A.A. Fattah, Sexual dimorphism of the mandible in a modern Egyptian population, J. Forensic Legal Med. 17 (4) (2010) 213–215.
[8] H. Dong, M. Deng, W. Wang, J. Zhang, J. Mu, G. Zhu, Sexual dimorphism of the mandible in a contemporary Chinese Han population, Forensic Sci. Int. 255 (2015) 9–15. [9] T. Kano, S. Oritani, T. Michiue, T. Ishikawa, A.M. Hishmat, N. Sogawa, O. Inamori-Kawamoto, H. Maeda, Postmortem CT morphometry with a proposal of novel parameters for sex discrimination of the mandible using Japanese adult data, Legal Med. 17 (3) (2015) 167–171. [10] T.S. Tunis, R. Sarig, H. Cohen, B. Medlej, N. Peled, H. May, Sex estimation using computed tomography of the mandible, Int. J. Legal Med. 131 (6) (2017) 1691– 1700. [11] S.S. Tambawala, F.R. Karjodkar, K. Sansare, N. Prakash, A.C. Dora, Sexual dimorphism of foramen magnum using cone beam computed tomography, J. Forensic Legal Med. 44 (2016) 29–34. [12] O. Ekizoglu, E. Hocaoglu, E. Inci, I.O. Can, D. Solmaz, S. Aksoy, C.F. Buran, I. Sayin, Assessment of sex in a modern Turkish population using cranial anthropometric parameters, Legal Med. 21 (2016) 45–52. [13] S.C. Kölzer, I.V. Kümmel, J.T. Kölzer, F. Ramsthaler, S. Plenzig, A. Gehl, M.A. Verhoff, Die Stirnneigung am knöchernen Schädel als Geschlechtsdiskriminierungsmerkmal – Begriffsdefinition und Quantifizierung, Arch. Kriminol. 235 (2015) 11–21. [14] F. Ramsthaler, M. Kettner, A. Gehl, M.A. Verhoff, Digital forensic osteology: morphological sexing of skeletal remains using volume-rendered cranial CT scans, Forensic Sci. Int. 195 (1) (2010) 148–152. [15] E.E. Abdel Fatah, N.R. Shirley, R.L. Jantz, M.R. Mahfouz, Improving sex estimation from crania using a novel three-dimensional quantitative method, J. Forensic Sci. 59 (3) (2014) 590–600. [16] B.J. Adams, J.E. Byrd, Resolution of small-scale commingling: a case report from the Vietnam war, Forensic Sci. Int. 156 (1) (2006) 63–69. [17] J.E. Byrd, B.J. Adams, Osteometric sorting of commingled human remains, J. Forensic Sci. 48 (4) (2003) 717–724. [18] B.J. Adams, J.E. Byrd, Recovery, Analysis, and Identification of Commingled Human Remains, Humana Press, Totowa, NJ, 2008, pp. 199–220. [19] K.J. Reichs, Cranial suture eccentricities: a case in which precocious closure complicated determination of sex and commingling, J. Forensic Sci. 34 (1) (1989) 263–273. [20] S. Preißler, S. Kauferstein, C. Niess, M.A. Verhoff, S.C. Kölzer, Über die morphologische Möglichkeit zur Prüfung der Kompatibilität von Mandibula und Cranium – Ein Fallbericht, Arch. Kriminol. 239 (2017) 99–108. [21] T.H. Schiebler, W. Schmidt (Hrsg.), Anatomie, 4. Auflage, Springer, Berlin, Heidelberg, New York, London, Paris, Tokyo, 1987, pp. 432–433. [22] J. Waschke, T.M. Böckers, F. Paulsen, Anatomie. Das Lehrbuch, Sobotta, 1. Auflage, Elsevier Urban & Fischer, München, 2015, pp. 522–525. [23] M.A. Verhoff, F. Ramsthaler, J. Krähahn, U. Deml, R.J. Gille, S. Grabherr, M.J. Thali, K. Kreutz, Digital forensic osteology – possibilities in cooperation with the Virtopsy1 project, Forensic Sci. Int. 174 (2) (2008) 152–156. [24] K.E. Stull, M.L. Tise, Z. Ali, D.R. Fowler, Accuracy and reliability of measurements obtained from computed tomography 3D volume rendered images, Forensic Sci. Int. 238 (2014) 133–140. [25] T. Kragh, S. Holley, S. Kirchhoff, F. Fischer, G. Grupe, M. Graw, Forensischanthropometrischer Methodenvergleich. Reliabilität eines bildgebenden Verfahrens, Rechtsmedizin 21 (2011) 445–451. [26] M.A. Verhoff, F. Ramsthaler, J. Krähahn, R.J. Gille, P. Kage, S. Kage, M. Durschnabel, J. Strunk, U. Deml, S. Grabherr, C. Jackowski, S. Ross, L. Oesterhelweg, M. Thali, K. Kreutz, Digitale forensische Osteologie, Rechtsmedizin 17 (1) (2007) 29–34. [27] B. Herrmann, G. Gruppe, S. Hummel, H. Piepenbrink, H. Schutkowski, Prähistorische Anthropologie – Leitfaden der Feld- und Labormethoden, 91, Springer Verlag, 1990, pp. 335–336. [28] T.D. White, P.A. Folkens, The Human Bone Manual, Elsevier Academic Press, 2005, pp. 31–33.
Contents lists available at ScienceDirect
Forensic Science International journal homepage: www.elsevier.com/locate/forsciint
Morphometric investigations to assess the compatibility of mandible and skull Sandra Preisslera , Marcel A. Verhoffa , Frank Ramsthalerb , Franziska Holza , Axel Gehlc, Sarah C. Koelzera,* a
Institute of Legal Medicine, Department of Forensic Medicine, Goethe University, Kennedyallee 104, 60596, Frankfurt/Main, Germany Institute of Legal Medicine, Department of Forensic Medicine, University of Saarland Medical School, Kirrberger Str. 100/Building 100, 66421, Homburg/Saar, Germany c Institute of Legal Medicine, Department of Forensic Medicine, University Hospital Hamburg-Eppendorf, Building North N81, Butenfeld 34, 22529, Hamburg, Germany b
A R T I C L E I N F O
Article history: Received 18 October 2017 Received in revised form 4 March 2018 Accepted 8 March 2018 Available online 16 March 2018 Keywords: Mandible Skull Compatibility Computed tomography Commingling Intra-observer reliability
A B S T R A C T
When a morphologically separated skull and mandible are found in the same case context, the possibility of a match arises. Two criteria with which to determine a match are the rough articulation between the mandibular condyles and cranial base itself and, most importantly, the fit of the teeth. However, when there has been intravital or postmortem tooth loss, this important criterion is not available. To date, only Reichs (1989) has investigated further compatibility criteria to solve the question of putative commingling in a case where a mandible seemed to originate from a female, while all other bones originated from a male individual. In a different reported case (Preißler et al. 2017), a mandible seemed too big for a skull; DNA analysis, however, confirmed that both originated from the same female individual. To investigate the metric relationship between mandible and skull we measured the postmortem CT data records of 223 corpses (virtual skulls) in OsiriX© MD for the following linear parameters: bicondylar breadth (KDB), biradicular breadth (AUB), and bizygomatic breadth (ZYB). The indices KDB/ZYB and KDB/ AUB were developed and used to define ranges for matches and mismatches. Furthermore, the intraobserver reliability for the method was assessed. An intraclass correlation coefficient of >0.99 for every parameter showed that the used measurements are highly reliable. The 2.5–97.5 percentile for the KDB/AUB index lay between 0.91 and 1.05, while the range for the KDB/ ZYB index was between 0.87 and 1.00. Within these ranges, it is possible to roughly assess whether or not a mandible and skull might be compatible, even if this can only be verified by forensic DNA analysis. If an index value lies outside these ranges, it can be assumed that skull and mandible do not match. Future studies should include more samples from a broader population spectrum so that these metric relationships can be used for different populations. © 2018 Published by Elsevier B.V.
1. Introduction Anthropometric measurements belong to the essential methods in anthropology used to scale the human body. They are performed with the aim of standardizing measurements and of obtaining reliable anthropometric values [1]. Anthropometric measurements of the mandible are performed not only to address oral surgical [2] or clinical [3,4] questions, but also to address forensic issues, such as the estimation of age [5,6] or sex [5,7–10] in different populations. De Oliveira et al., for example,
* Corresponding author. https://doi.org/10.1016/j.forsciint.2018.03.013 0379-0738/© 2018 Published by Elsevier B.V.
investigated the correlation of sex and of age with mandibular ramus length in a Brazilian population. They found that while “mandibular ramus length is effective to estimate chronological age with a high degree of accuracy,” it is not an accurate measurement with which to estimate sex [5]. In the forensic sciences, there is a dearth of adequate bone measurement series from recent populations. To alleviate this shortcoming, the use of imaging techniques like computed tomography (CT) has increased in the last few years. In particular, anthropometric measurements of the skull have been performed to assess sex in different populations [8–14]. Kölzer et al., for example, investigated the frontal inclination of the human skull as a trait of sexual dimorphism. Because these investigators did not
194
S. Preissler et al. / Forensic Science International 286 (2018) 193–198
find significant sex differences with the variables they had used in their pilot study, they concluded that additional variables, based strictly on considerations of validity and reliability, needed to be tested [13]. In yet another study, CT scans were used not only for statistical analysis, but also to generate a 3D atlas of the human skull for further reference [15]. In diverse publications, Adams and Byrd emphasize that different sorting techniques, such as articulation, taphonomy, and elimination, should be used conjunctively to obtain best resolution results for commingled human skeletal remains [16– 18]. With the exception of Reichs’ study, no systematic measurements pertaining to the compatibility of mandible and skull have been taken so far. In her case report, Reichs describes the discovery of a human skull along with other human bones, including a mandible. Because the mandible was small and gracile and did not articulate well with the skull, a total of 16 other skulls were measured for comparison. The results indicated that, in this case, the mandible was, in fact, not too small for the skull and that the bones in question could well have stemmed from the same individual [19]. In another reported case [20], in which the investigators were confronted with toothless jaws, a mandible appeared to be too big for the skull. Although the skull was obviously female, the mandible was thought to belong to a male individual. Nonetheless, subsequent DNA analysis showed that both bones originated from the same female individual. To investigate the metric relationship between mandible and skull, digital measurements using volume-rendered CT skulls were performed in the present study. To this purpose, three linear parameters were measured and two new indices developed. The primary question was whether there were values for both indices which, in case of a transgression below or above, would indicate that a mandible and skull do not belong together. The procedure was carried out for the entire collective, separately for both sexes. Additionally, the study tested whether the three linear parameters are applicable for sex discrimination.
parameters were collected: bicondylar breadth, biradicular breadth, and bizygomatic breadth. Definitions and abbreviations of the linear parameters are given in Table 1. Fig. 1 shows examples for each parameter. All parameters were measured a second time by the same examiner to review the intra-observer reliability. Prior to the start of the experiments, approval was obtained from the ethical committee of the faculty of medicine at the JustusLiebig University Gießen (file number: 63/09). Statistical analyses were performed using SPSS v 17.0 and MedCalc v 16.8.4. Apart from descriptive statistics of the linear parameters, the reliability of the used method was determined. Afterwards a canonical discriminant analysis was performed. For the descriptive statistics, two indices were developed from linear parameters: The bicondylar breadth was set in relation to the biradicular breadth and also in relation to the bizygomatic breadth. Intra-observer reliability was assessed by calculating the intraclass correlation coefficient and variation coefficient. To verify the variance of both indices for matches and mismatches between mandible and skull, a comparison group was created. In this comparison group, mandibles were randomly matched with skulls, separately for each sex.
2. Materials and methods
3.2. Match – mismatch
Altogether 223 virtual skulls (113 female, 106 male, 4 unknown), generated from anonymized postmortem CT data records from the Institute of Legal Medicine in Hamburg, were examined. The individuals were scanned in a Diamond Select MX 8000 Quad CT-Scanner (slice thickness 1.3 mm, 120 kV, 299– 349 mAs, pitch: 0.88, image matrix size: 512 512, Phillips, Amsterdam Netherlands). The associated metadata for each individual consisted of age, sex, weight, and height. Skulls with bone fractures and highly fragmented skulls were excluded from measurements. The CT data were imported into an OsiriX© MD database. The virtual skulls were aligned in the volume-rendering mode in basilar view and measured with the help of the measurement tool in OsiriX© MD (in cm). The following
To find out whether a mandible meshes with a skull, two indices were developed: KDB/AUB and KDB/ZYB. Descriptive statistics were performed for both indices and are summarized in Tables 3 and 4. The “notched box-and-whisker plot” (Fig. 3) displays a graphical statistical summary of the selected index values. The central notched box represents the values from the mid 50% of the data (25–75 percentile). Red horizontal bars indicate data distribution within the 2.5–97.5 percentile (95% of the data falls within this range). The plots on the right-hand sides of Fig. 3A–F illustrate the wider spread of value distribution in the control group (mismatching pairs of mandibles and skulls that were generated randomly, both separately for each sex and combined for both sexes).
3. Descriptive results 3.1. Reliability Table 2 shows the intraclass correlation coefficient (ICC) and the 95% confidence interval for both single measures and average measures. All three linear parameters show values >0.99 for ICC (p 0.01). Furthermore, mountain plots were created for each linear parameter. These mountain plots (Fig. 2A–C) visualize the difference between the first and the second measurement for each parameter. They illustrate the symmetric character of the linear parameters. ZYB is the best linear parameter, with a difference less than 0.15 cm when comparing both measurements, while KDB shows the highest differences with up to 0.45 cm.
Table 1 Definitions and abbreviations of the linear parameters. The abbreviations for bizygomatic breadth and biradicular breadth were taken from Martin and Knußmann [1]. For bicondylar breadth no abbreviation was found, so an abbreviation according to the corresponding landmark was chosen, in which B was substituted for L. Linear parameter
Abbreviation Definition (Martin and Knußmann [1])
Bicondylar breadth Biradicular breadth Bizygomatic breadth
KDB AUB
Linear distance, measured externally from the most lateral point on the left condyle to the most lateral point on the right (condylion to condylion) Linear distance between the lowest points of the zygomatic arc root (radiculare to radiculare)
ZYB
Linear distance, measured from the most lateral point on the left zygomatic arc to the most lateral point on the right (zygion to zygion)
S. Preissler et al. / Forensic Science International 286 (2018) 193–198
195
Fig. 1. Virtual skull in basilar view with the linear parameters: (A) bicondylar breadth (KDB), (B) biradicular breadth (AUB), and (C) bizygomatic breadth (ZYB).
Table 2 Results for the tested intra-observer reliability (ICC = intraclass correlation, CI = confidence interval). Linear parameter
KDB AUB ZYB
Single measures
Average measures
ICC
95% CI
ICC
95% CI
0.9956 0.9976 0.9983
[0.9943–0.9966] [0.9969–0.9982] [0.9977–0.9988]
0.9978 0.9988 0.9992
[0.9971–0.9983] [0.9985–0.9991] [0.9989–0.9994]
4. Discussion In the present study, 223 cranial CT scans were investigated to assess the compatibility of mandible and skull. The reason for the study was a case, previously reported by us, in which a mandible seemed to be too large for the skull and, therefore, did not seem to be compatible. This assumption was, however, disproved by forensic DNA analysis [20]. The case illustrates how difficult it may sometimes be to decide whether these bones articulate well and, therefore, whether they both belong to the same decedent or not. One reason for misjudgment may be found in the anatomy of the mandibular joint, where the caput mandibulae of the
processus condylaris articulates with the facies articularis of the fossa mandibularis and the tuberculum articulare at the os temporale. Between the caput mandibulae and fossa mandibularis, there is the discus articularis, which is fused with the surrounding joint capsule [21]. Because the fossa mandibularis is usually larger than the caput mandibulae, the articular disk has to balance the incongruity [22]. Conversely, it can imply that this disk, along with the comparatively flat subcranial joint socket, may allow the caput mandibulae to protrude both laterally and medially from the articulation. This protrusion may lead to a misjudgment of the “fit” between mandible and skull, as in the case we reported. After 3D reconstruction of the available virtual skulls, the linear parameters bicondylar breadth (KDB), biradicular breadth (AUB), and bizygomatic breadth (ZYB) were measured in the volumerendering modus of OsiriX© MD. The results of the reliability analysis (all values > 0.99) again confirm that the used digital method is highly reliable for linear parameters [23]. The prerequisite for this method is an adequate resolution of the 3D datasets, which is given at a slice thickness of 1.3 mm or less [23]. In the present study, slice thicknesses of 1.3 mm were used exclusively. Furthermore, the landmarks for the
Fig. 2. Mountain plots for the three measurements: (A) ZYB, (B) AUB, and (C) KDB. The differences for both measurements for each skull are given on the X-axis in [cm].
196
S. Preissler et al. / Forensic Science International 286 (2018) 193–198
Table 3 Summarized statistics for the KDB/ZYB index. Empty fields indicate that the sample size was too low. Stat. parameter
All cases
Female
Male
Unknown sex
Sample size Lowest value Highest value Mean 95% CI mean Median 95% CI median Variance Standard deviation Standard error mean 2.5–97.5 percentile
223 0.8448 1.0367 0.9338 [0.9294–0.9381] 0.9357 [0.9316–0.9395] 0.001104 0.03322 0.002225
113 0.8448 1.0367 0.9398 [0.9332–0.9464] 0.9417 [0.9338–0.9507] 0.001247 0.03531 0.003322 0.87–1.00
106 0.8488 0.9885 0.9285 [0.9228–0.9342] 0.9330 [0.9236–0.9371] 0.0008702 0.02950 0.002865 0.87–0.98
4 0.8650 0.9379 0.9025 [0.8519–0.9532] 0.9036 0.001013 0.03183 0.01591
Table 4 Summarized statistics for the KDB/AUB index. Empty fields indicate that the sample size was too low. Stat. parameter
All cases
Female
Male
Unknown sex
Sample size Lowest value Highest value Mean 95% CI mean Median 95% CI median Variance Standard deviation Standard error mean 2.5–97.5 percentile
223 0.8828 1.0833 0.9785 [0.9738–0.9832] 0.9783 [0.9727–0.9858] 0.001265 0.03557 0.002382
113 0.8987 1.0620 0.9782 [0.9715–0.9848] 0.9806 [0.9694–0.9894] 0.001273 0.03567 0.003356 0.91–1.05
106 0.9100 1.0833 0.9799 [0.9732–0.9866] 0.9786 [0.9694–0.9873] 0.001218 0.03491 0.003390 0.91–1.05
4 0.8828 0.9915 0.9491 [0.8737–1.0246] 0.9611
three linear parameters used in this study are well defined and easy to determine in the basilar view of the virtual skulls. The mountain plots underpin the results of the reliability test. Overall, the majority of measurement differences lay within 0.2 cm, “which is considered an acceptable amount of error in forensic anthropology” [24]. With a deviation of less than 0.15 cm, the best result was achieved when measuring ZYB, while KDB showed the highest deviation, of up to 0.45 cm. These measurement discrepancies can be explained by the image quality of the reconstructed skulls, which differs among the data records. Especially the contrast of some skull reconstructions was very dark, which occasionally made it difficult to localize the external borders of the condyles. In recent years, the use of CT data for anthropometric measurements has been investigated in comparison with the use of classical methods. Kragh et al. [25], for example, compared different forensic anthropometrical techniques. They took 50 human skulls and measured them manually as well as by a digital image method and found that 31 of the investigated dimensions, including bizygomatic breadth, basion-bregma-height, and biasterionic breadth, are reliable for both techniques. Stull et al. [24] investigated the accuracy and reliability of measurements of 3D volume-rendered whole body images and arrived at the same assessment. Furthermore, Verhoff et al. [23,26] performed osteometric measurements using 3Dreconstructed skulls as well as macerated skulls and showed that both techniques provide similar results. For the present study, two new indices were developed to obtain values for the relationship between mandible and skull. These values were supposed to help decide whether or not a mandible meshes with a skull. The range for the KDB/AUB index lay between 0.8828 and 1.0833. The range for the KDB/ZYB index was between 0.8448 and 1.0367. It should, however, be kept in mind that these values are only valid if the individuals in question originate from the same population from which the data were obtained [27,28].
0.002248 0.04742 0.02371
The notched-box-and whisker-plots in Fig. 3 visualize the data distribution for the two index values. In the graphs, match and mismatch of mandible and skull are juxtaposed for each index and sex, and additionally for all cases combined. Although the ranges for both indices are overall comparable for both sexes (Tables 3 and 4), there is a slight difference in the range of the KDB/ZYB index between the sexes, in the sense that the upper limit is slightly higher for females than for males. Although this observation is unlikely to be relevant in practice, it is, nevertheless, worth noting. From these data distributions, it can be deduced that if a skull and a mandible are found and their index values are within the 95% range of the boxplots, they might belong together, though an affiliation cannot be proved. The boxplots for the randomly matched skulls and mandibles show a considerably larger index value range than the matching ones. If the calculated index is greater or less than the 2.5–97.5 percentile (Tables 3 and 4, last row) in a concrete case, it can, on the basis of the two-sided analysis, be assumed with a probability of approximately 97.5% that the mandible and skull do not match. In her case report, Reichs [19] measured six parameters including fossa breadth and bifossae breadth to verify the compatibility of a mandible and a skull. However, these parameters could not be measured with the 3D reconstructions used in our study because the condyles of the mandibles were sitting in their joint sockets. For this reason, we chose to measure linear parameters with an anatomical relation to the condyles of the mandible. Although, lastly, only DNA analysis can provide conclusive evidence for a match, the two indices we developed indicate a range within which a match between a skull and a mandible cannot be excluded. However, perhaps more importantly, the true value of these indices appears to be their power to exclude a match, rather than to prove one. Future studies should include more samples from a broader population spectrum so that these metric relationships can be used for different populations.
S. Preissler et al. / Forensic Science International 286 (2018) 193–198
197
Fig. 3. “Notched box-and-whisker plots” of the selected index values. The central notched box represents the values from the mid 50% of the data (25–75 percentile). The horizontal middle black line represents the median, whereas the notches display the 95% CI of the median. The horizontal line extends from the minimum to the maximum value. Outlier values are displayed as separate points (defined as a value that is smaller than the lower quartile minus 3 times the interquartile range, or larger than the upper quartile plus 3 times the interquartile range). Red horizontal bars display data distribution within the 2.5–97.5 percentile (95% of the data falls within this range). The plot illustrates the wider spread of value distribution in cases of mismatching mandible-skull pairs (right side). A and B show the boxes for the KDB/ZYB index and the KDB/AUB index for both sexes, C and D show the distribution of females for the KDB/ZYB index and for the KDB/AUB index, respectively. Finally, E and F show the distribution of males for the KDB/ZYB index and for the KDB/AUB index, respectively.
198
S. Preissler et al. / Forensic Science International 286 (2018) 193–198
5. Conclusion for practice If a skull and a mandible are found morphologically separated, but in the same case context, a possible match should first be morphologically checked by seeing if the teeth fit. If this is impossible because of intravital or postmortal tooth loss, the use of the more questionable “fit” of the mandibular joint becomes relevant. Should the mandibular condyles seem too big or too small in comparison to the skull, examinations regarding sex discrimination should be performed. When the sex accords for both the skull and the mandible, the indices proposed in this study can be determined. For values lying within the sex-specific 2.5–97.5 percentile ranges of the indices it is possible to roughly assess whether or not the skull and mandible might be compatible, even if this can only be verified by forensic DNA analysis. If, on the other hand, an index value lies outside these ranges, it can be assumed that the skull and mandible do not match. Even if sex cannot be determined for the skull and the mandible in question, the indices can still be used in view of the fact that the 2.5–97.5 percentile ranges for the KDB/AUB index are the same for both sexes anyway. It would, however, be advisable to use the somewhat wider range (0.91–1.05) of the KDB/ZYB index for female sex, just to be on the safe side. Conflict of interest The authors declare no conflict of interest. References [1] R. Martin, R. Knußmann, Anthropologie, Handbuch der vergleichenden Biologie des Menschen. Band I: Wesen und Methoden der Anthropologie, Gustav Fischer Verlag, Stuttgart, New York, 1988, pp. 129–182. [2] C.J. Našel, M. Pretterklieber, A. Gahleitner, C. Czerny, M. Breitenseher, H. Imhof, Osteometry of the mandible performed using dental MR imaging, Am. J. Neuroradiol. 20 (7) (1999) 1221–1227. [3] Y.L. Zhang, J.L. Song, X.C. Xu, L.L. Zheng, Q.Y. Wang, Y.B. Fan, Z. Liu, Morphologic analysis of the temporomandibular joint between patients with facial asymmetry and asymptomatic subjects by 2D and 3D evaluation: a preliminary study, Medicine 95 (13) (2016)e3052. [4] L.F. Merigue, A.C.D.C.F. Conti, P.V.P. Oltramari-Navarro, R.D.L. Navarro, M.R.D. Almeida, Tomographic evaluation of the temporomandibular joint in malocclusion subjects: condylar morphology and position, Braz. Oral Res. 30 (1) (2016), doi:http://dx.doi.org/10.1590/1807-3107BOR-2016.vol30.0017. [5] F.T. de Oliveira, M.Q.S. Soares, V.A. Sarmento, C.M.F. Rubira, J.R.P. Lauris, I.R.F. Rubira-Bullen, Mandibular ramus length as an indicator of chronological age and sex, Int. J. Legal Med. 129 (1) (2015) 195–201. [6] D. Franklin, A. Cardini, Mandibular morphology as an indicator of human subadult age: interlandmark approaches, J. Forensic Sci. 52 (5) (2007) 1015– 1019. [7] M.A.A. Kharoshah, O. Almadani, S.S. Ghaleb, M.K. Zaki, Y.A.A. Fattah, Sexual dimorphism of the mandible in a modern Egyptian population, J. Forensic Legal Med. 17 (4) (2010) 213–215.
[8] H. Dong, M. Deng, W. Wang, J. Zhang, J. Mu, G. Zhu, Sexual dimorphism of the mandible in a contemporary Chinese Han population, Forensic Sci. Int. 255 (2015) 9–15. [9] T. Kano, S. Oritani, T. Michiue, T. Ishikawa, A.M. Hishmat, N. Sogawa, O. Inamori-Kawamoto, H. Maeda, Postmortem CT morphometry with a proposal of novel parameters for sex discrimination of the mandible using Japanese adult data, Legal Med. 17 (3) (2015) 167–171. [10] T.S. Tunis, R. Sarig, H. Cohen, B. Medlej, N. Peled, H. May, Sex estimation using computed tomography of the mandible, Int. J. Legal Med. 131 (6) (2017) 1691– 1700. [11] S.S. Tambawala, F.R. Karjodkar, K. Sansare, N. Prakash, A.C. Dora, Sexual dimorphism of foramen magnum using cone beam computed tomography, J. Forensic Legal Med. 44 (2016) 29–34. [12] O. Ekizoglu, E. Hocaoglu, E. Inci, I.O. Can, D. Solmaz, S. Aksoy, C.F. Buran, I. Sayin, Assessment of sex in a modern Turkish population using cranial anthropometric parameters, Legal Med. 21 (2016) 45–52. [13] S.C. Kölzer, I.V. Kümmel, J.T. Kölzer, F. Ramsthaler, S. Plenzig, A. Gehl, M.A. Verhoff, Die Stirnneigung am knöchernen Schädel als Geschlechtsdiskriminierungsmerkmal – Begriffsdefinition und Quantifizierung, Arch. Kriminol. 235 (2015) 11–21. [14] F. Ramsthaler, M. Kettner, A. Gehl, M.A. Verhoff, Digital forensic osteology: morphological sexing of skeletal remains using volume-rendered cranial CT scans, Forensic Sci. Int. 195 (1) (2010) 148–152. [15] E.E. Abdel Fatah, N.R. Shirley, R.L. Jantz, M.R. Mahfouz, Improving sex estimation from crania using a novel three-dimensional quantitative method, J. Forensic Sci. 59 (3) (2014) 590–600. [16] B.J. Adams, J.E. Byrd, Resolution of small-scale commingling: a case report from the Vietnam war, Forensic Sci. Int. 156 (1) (2006) 63–69. [17] J.E. Byrd, B.J. Adams, Osteometric sorting of commingled human remains, J. Forensic Sci. 48 (4) (2003) 717–724. [18] B.J. Adams, J.E. Byrd, Recovery, Analysis, and Identification of Commingled Human Remains, Humana Press, Totowa, NJ, 2008, pp. 199–220. [19] K.J. Reichs, Cranial suture eccentricities: a case in which precocious closure complicated determination of sex and commingling, J. Forensic Sci. 34 (1) (1989) 263–273. [20] S. Preißler, S. Kauferstein, C. Niess, M.A. Verhoff, S.C. Kölzer, Über die morphologische Möglichkeit zur Prüfung der Kompatibilität von Mandibula und Cranium – Ein Fallbericht, Arch. Kriminol. 239 (2017) 99–108. [21] T.H. Schiebler, W. Schmidt (Hrsg.), Anatomie, 4. Auflage, Springer, Berlin, Heidelberg, New York, London, Paris, Tokyo, 1987, pp. 432–433. [22] J. Waschke, T.M. Böckers, F. Paulsen, Anatomie. Das Lehrbuch, Sobotta, 1. Auflage, Elsevier Urban & Fischer, München, 2015, pp. 522–525. [23] M.A. Verhoff, F. Ramsthaler, J. Krähahn, U. Deml, R.J. Gille, S. Grabherr, M.J. Thali, K. Kreutz, Digital forensic osteology – possibilities in cooperation with the Virtopsy1 project, Forensic Sci. Int. 174 (2) (2008) 152–156. [24] K.E. Stull, M.L. Tise, Z. Ali, D.R. Fowler, Accuracy and reliability of measurements obtained from computed tomography 3D volume rendered images, Forensic Sci. Int. 238 (2014) 133–140. [25] T. Kragh, S. Holley, S. Kirchhoff, F. Fischer, G. Grupe, M. Graw, Forensischanthropometrischer Methodenvergleich. Reliabilität eines bildgebenden Verfahrens, Rechtsmedizin 21 (2011) 445–451. [26] M.A. Verhoff, F. Ramsthaler, J. Krähahn, R.J. Gille, P. Kage, S. Kage, M. Durschnabel, J. Strunk, U. Deml, S. Grabherr, C. Jackowski, S. Ross, L. Oesterhelweg, M. Thali, K. Kreutz, Digitale forensische Osteologie, Rechtsmedizin 17 (1) (2007) 29–34. [27] B. Herrmann, G. Gruppe, S. Hummel, H. Piepenbrink, H. Schutkowski, Prähistorische Anthropologie – Leitfaden der Feld- und Labormethoden, 91, Springer Verlag, 1990, pp. 335–336. [28] T.D. White, P.A. Folkens, The Human Bone Manual, Elsevier Academic Press, 2005, pp. 31–33.