INTRODUCTION
Sex identification is an important forensic and archaeological procedure [1, 20]. Comparison of ante mortem and post mortem records and examination of bones have a crucial role for identification [20, 21]. Pelvis, skull and long bone measurements are usually used for sex estimation [20]. But these bones may be damaged — burned and fragmented — because of disasters. In such cases, determined sex is usually used as an estimation of other biological features like age and ancestry. Thus, another method using other bones should be established for cases where the pelvis, skull and long bones are unavailable [21].
The hyoid bone originates from the branchial arches and arch cartilages [7]. It is a solitary bone without any bony articulations [5, 9]. It has muscles, ligaments and fascia attachments from cranium, mandible and pharynx that provide physiological functions; including speaking, breathing, swallowing and keeping the upper airway open during sleep [6, 7, 11, 15]. The hyoid bone has distinct morphological features that allow forensic researchers to quickly identify it. U-shaped bone lies anteriorly to fourth cervical vertebra in adults and is composed of a body, two greater cornua and two lesser cornua [5, 7, 15]. These cornua which are directed postero-laterally are attached to the body by synchondrosis [15]. Some hyoid bones show a diarthrodial structure and have a gap between body/cornua bone articulation [10].
Measurements of hyoid bone and the degree of fusion between hyoid bone and greater cornua might be helpful in age and sex estimation. Several studies including cadavers, photographs, and radiographs have used hyoid bone for sex estimation [1, 7, 11, 15]. Among these techniques, radiological assessment is important for age calculation and sex estimation. It is simple, easily applicable and non-invasive method, and can be evaluated in living individuals as well as in the dead. Cone-beam computed tomography (CBCT) is a radiological imaging system that provides three-dimensional visualisation of the anatomical structures and gives realistic measurements. While there are studies evaluating hyoid bone for age and sex estimation, no study has been found that examine sexual dimorphism of hyoid bone measurements of living patients in a contemporary Turkish population with CBCT images.
The aim of this study was to evaluate the fusion between hyoid body and cornua, and hyoid bone measurements in sex estimation and aging using CBCT images.
MATERIALS AND METHODS
In this retrospective study, CBCT images, which were performed because of patients’ dental problems, from database of oral and maxillofacial radiology department of Aydın Adnan Menderes University were used. The study was approved by the local ethical committee (No: 2020/36). For this study, 650 CBCT scans from lower jaw that were performed between October 2014 and February 2019 were visually inspected. Fractures and acquired or congenital abnormalities of hyoid bone were excluded from study. Images with whole hyoid bone structure visible and without any artefacts were included. According to inclusion and exclusion criteria, a total of 130 CBCT scans were selected for the study. The CBCT scans were obtained using standard exposure parameters and patient positioning protocols (field of view: 20 × 10 cm and 20 × 15 cm, resolution: 0.2–0.4 mm, 80–84 kVp, 8–12 mA) with a CBCT unit (Planmeca promax 3D Max, Helsinki, Finland). Romexis version 4.6.2. software (Planmeca, Helsinki, Finland) was used for hyoid bone measurements. Inter- and intra-observer rates were calculated by re-evaluating 30 randomly selected subset data (15 males and 15 females). The original author examined and measured all scans and the subset data. Co-author examined and measured the subset data to assess interobserver error. Coefficient of reliability (R) and relative technical error of measurements (rTEM, %) were calculated. R value > 0.75 was considered sufficiently exact which reveals the quantity of the between-subject variance free of measurement errors [25]. The acceptance ranges of rTEM for intraobserver errors and interobserver errors were < 1.5% and < 2.0%, respectively [24].
Axial, sagittal, coronal and oblique views were investigated. All of the measurements were made from axial images that were obtained and corrected with slice-by-slice tracing of coronal and sagittal images. Eight measurements were performed for each hyoid bone as below:
Fusion conditions were assigned to each side of the hyoid bone separately. The smallest distance between the greater cornua and hyoid body on each side was assessed and following three grades defined as below:
Statistical analysis
Statistical analyses were done on a personal computer using SPSS (IBM Statistical Package for the Social Sciences Statistics; New York, ABD) version 21.0 computer software. Means, standard deviations, and ranges of all hyoid bone measurements were calculated according to age. The normality was assessed using the Shapiro-Wilk test. The mean differences between the sexes in the all hyoid measurement data were assessed using two-sample independent t-test. The association of fusion types with sex and age groups were analysed with c2 test and Pearson correlation. Statistically significant level was considered at a p value of < 0.05.
Univariate discriminant function analysis for sex classification was used for all measurements. Then stepwise discriminant function analysis was used to find most accurate sex classification. Function at group centroids were calculated, sectioning points for each measurements were calculated by adding group centroids and dividing their total by 2. Wilk’s lambda was calculated. Correct prediction rates were calculated by the leave-one-out cross validation test.
RESULTS
Table 1 demonstrates the R and rTEMs values. The R values ranged from 0.96 to 0.99. The rTEMs values for interobserver and intraobserver errors ranged from 1.03% to 1.88% and from 1.12% to 1.94%, respectively. The mean, standard deviation, and ranges of all hyoid bone measurements according to gender were shown in Table 2. There were significant differences between sexes for all measurements except the LD-D (p = 0.437). There is no significant difference in the mean age of genders (p = 0.123). No significant difference was found between RD-D and LD-D, RP-W and LP-W (p = 0.448 and p = 0.700, respectively).
Variable |
Intraobserver |
Interobserver |
||
R |
rTEM |
R |
rTEM |
|
Hyd-W |
0.96 |
1.54% |
0.96 |
1.24% |
B-L |
0.98 |
1.32% |
0.96 |
1.85% |
Hyd-L |
0.99 |
1.45% |
0.97 |
1.63% |
RD-D |
0.99 |
1.03% |
0.98 |
1.25% |
RP-W |
0.98 |
0.88% |
0.97 |
1.12% |
LD-D |
0.99 |
1.35% |
0.99 |
1.54% |
LP-W |
0.98 |
1.66% |
0.96 |
1.36% |
Stature |
0.96 |
1.08% |
0.96 |
1.94% |
|
Male |
Female |
P value |
Total |
|||
Mean ± SD |
Range |
Mean ± SD |
Range |
Mean ± SD |
Range |
||
Age [years] |
33.04 ± 13.42 |
15–55 |
35.29 ± 15.20 |
14–55 |
0.123 |
34.37 ± 14.49 |
14–55 |
Hyd-W |
45.20 ± 5.62 |
35.40–58.00 |
42.32 ± 4.64 |
28.60–54.50 |
0.001* |
43.50 ± 5.15 |
28.60–8.00 |
B-L |
23.75 ± 2.49 |
18.20–28.70 |
20.39 ± 2.12 |
13.20–24.70 |
< 0.001* |
21.77 ± 2.81 |
13.20–28.70 |
Hyd-L |
37.74 ± 4.00 |
30.20–46.40 |
32.28 ± 3.01 |
26.40–37.50 |
< 0.001* |
34.52 ± 4.36 |
26.40–46.40 |
RD-D |
3.38 ± 0.73 |
2.00–4.80 |
3.11 ± 0.55 |
2.10–4.50 |
0.017* |
3.22 ± 0.64 |
2.00–4.80 |
RP-W |
3.52 ± 0.77 |
2.10–4.90 |
2.83 ± 0.51 |
2.00–4.10 |
< 0.001* |
3.11 ± 0.72 |
2.00–4.90 |
LD-D |
3.32 ± 0.83 |
2.00–5.00 |
3.22 ± 0.71 |
2.00–5.20 |
0.437 |
3.26 ± 0.76 |
2.00–5.20 |
LP-W |
3.42 ± 0.79 |
2.00–5.10 |
2.94 ± 0.51 |
1.80–4.10 |
< 0.001* |
3.13 ± 0.68 |
1.80–5.10 |
Stature |
86.30 ± 7.38 |
73.00–103.00 |
76.73 ± 6.84 |
60.40–90.40 |
< 0.001* |
80.65 ± 8.48 |
60.40–103.00 |
The measurement means and standard deviations according to fusion types for males and females in the population are shown in Table 3. The mean age of subjects in all hyoid fusion types showed no difference with regard to sex. Outcomes of the two-sample independent t-test for sexual differences between genders are shown in Table 4. The Hyd-L, LP-W and stature variables showed sexual dimorphism in all hyoid fusion types.
|
Sex |
Unfused hyoid |
Partially fused hyoid |
Fused hyoid |
|||
Mean |
SD |
Mean |
SD |
Mean |
SD |
||
Age (years) |
Total |
35.96 |
14.65 |
33.7 |
15.17 |
39.75 |
16.01 |
Hyd-W |
Male Female Total |
46.86 39.54 42.59 |
3.05 5.45 5.83 |
44.93 42.99 43.76 |
6.46 4.17 5.25 |
44.13 42.33 43.23 |
1.50 2.05 1.96 |
B-L |
Male Female Total |
22.82 17.71 19.84 |
1.46 2.10 3.15 |
24.34 20.97 22.30 |
2.46 1.69 2.61 |
21.60 21.06 21.33 |
2.63 0.90 1.89 |
Hyd-L |
Male Female Total |
37.38 29.97 33.05 |
2.23 2.32 4.35 |
38.28 32.94 35.06 |
4.48 2.97 4.47 |
34.93 31.33 33.13 |
0.63 2.06 2.37 |
RD-D |
Male Female Total |
3.26 2.92 3.06 |
0.50 0.47 .50 |
3.42 3.17 3.27 |
0.82 0.56 0.69 |
3.30 2.96 3.13 |
0.23 0.50 0.41 |
RP-W |
Male Female Total |
3.18 2.47 2.76 |
0.61 0.34 .58 |
3.63 2.92 3.20 |
0.78 0.53 0.72 |
3.43 2.83 3.13 |
0.91 0.36 0.73 |
LD-D |
Male Female Total |
3.82 2.51 3.05 |
0.91 0.34 0.91 |
3.24 3.36 3.31 |
0.80 0.64 0.71 |
3.03 3.46 3.25 |
0.59 1.03 0.83 |
LP-W |
Male Female Total |
3.48 2.65 3.00 |
0.29 0.13 .46 |
3.57 2.93 3.18 |
0.79 0.51 0.71 |
2.36 3.7 3.05 |
0.49 0.22 0.80 |
Stature |
Male Female Total |
82.38 72.78 76.78 |
6.32 8.99 9.20 |
88.10 77.77 81.86 |
7.49 6.21 8.41 |
81.50 75.90 78.70 |
2.72 3.74 4.27 |
|
Un-fused hyoids |
Partially fused hyoids |
Fused hyoids |
|||||||||
MD [mm] |
t-value |
DF |
p-value |
MD [mm] |
t-value |
DF |
p-value |
MD [mm] |
t-value |
DF |
p-value |
|
Hyd-W |
7.13 |
3.82 |
22 |
0.001* |
1.93 |
1.78 |
94 |
0.078 |
1.80 |
1.72 |
10 |
0.115 |
B-L |
5.10 |
6.60 |
22 |
< 0.001* |
3.37 |
7.94 |
94 |
< 0.001* |
0.53 |
0.46 |
10 |
0.649 |
Hyd-L |
7.40 |
7.81 |
22 |
< 0.001* |
5.33 |
7.01 |
94 |
< 0.001* |
3.60 |
4.08 |
10 |
0.002* |
RD-D |
0.33 |
1.64 |
22 |
0.114 |
0.25 |
1.78 |
94 |
0.078 |
0.33 |
1.45 |
10 |
0.176 |
RP-W |
0.70 |
3.61 |
22 |
0.002* |
0.71 |
5.29 |
94 |
< 0.001* |
0.60 |
1.49 |
10 |
0.167 |
LD-D |
1.30 |
4.90 |
22 |
< 0.001* |
–0.12 |
–0.82 |
94 |
0.410 |
-0.43 |
–0.89 |
10 |
0.394 |
LP-W |
0.82 |
9.27 |
22 |
< 0.001* |
0.64 |
4.80 |
94 |
< 0.001* |
–1.36 |
–6.18 |
10 |
< 0.001* |
Stature |
9.59 |
2.89 |
22 |
0.008* |
10.32 |
7.33 |
94 |
< 0 .001* |
5.60 |
2.96 |
10 |
0.014* |
Types of fusion according to sex and age groups are presented in Table 5. In young adult age group, a significant difference was found between sex and fusion types (p = 0.025), although there were no significant difference in adults and < 20 years groups (p = 0.068 and p = 0.150, respectively). The results of univariate discriminant function analyses for the hyoid bone measurements are shown in Table 6. The B-L (81.35%) and stature measurements (76.25%) contributed most significantly to sex estimation.
Age groups |
Sex |
Fusion types |
P-value |
||
1 (non-fusion) |
2 (partial fusion) |
3 (fusion) |
|||
< 20 years |
Male Female Total |
24 (85.0%) 22 (55.0%) 59 (71%) |
6 (15.0%) 14 (35.0%) 20 (25.0%) |
– 4 (10.0%) 4 (5.0%) |
0.150 |
Young adult (20–30 years) |
Male Female Total |
24 (75%) 48 (92.4%) 72 (85.7%) |
2 (6.2%) 2 (3.8%) 4 (4.8%) |
6 (18.8%) 2 (3.8%) 8 (9.5%) |
0.025* |
Adult (31–55 years) |
Male Female Total |
26 (72.2%) 58 (90.6%) 84 (84.0%) |
4 (11.1%) – 4 (4.0%) |
6 (16.7%) 6 (9.4%) 12 (12.0%) |
0.068 |
|
Group centroid |
Sectioning point |
Wilks’ lambda |
Eigenvalue |
Canonical correlation |
Correct prediction rates (%) |
|||
Male |
Female |
Male |
Female |
Overall |
|||||
Hyd-W |
0.342 |
–0.237 |
0.052 |
0.924 |
0.082 |
0.275 |
29.6 |
84.6 |
57.1 |
B-L |
0.871 |
–0.603 |
0.134 |
0.652 |
0.533 |
0.590 |
70.4 |
92.3 |
81.35 |
Hyd-L |
0.933 |
–0.646 |
0.143 |
0.620 |
0.612 |
0.616 |
59.3 |
87.2 |
73.25 |
RD-D |
0.252 |
–0.174 |
0.036 |
0.957 |
0.045 |
0.207 |
22.2 |
89.7 |
55.95 |
RP-W |
0.643 |
–0.445 |
0.099 |
0.775 |
0.291 |
0.475 |
51.9 |
87.2 |
69.55 |
LP-W |
0.441 |
–0.305 |
0.068 |
0.880 |
0.136 |
0.347 |
40.7 |
79.5 |
60.1 |
Stature |
0.800 |
–0.554 |
0.123 |
0.690 |
0.450 |
0.557 |
70.4 |
82.1 |
76.25 |
DISCUSSION
In forensic science, it is important that measurements made on any structure be reproducible and reliable. Additionally, measurements should be identified and explained clearly to understand definition [14]. Several studies have used hyoid bone measurements for sex and age estimation. Most of these studies included direct bone measurements or pictures of hyoid bones extracted from autopsies or anatomical collections [1, 11, 12, 14, 15, 22, 23]. Some of studies focused on two-dimensional computed tomography images from databases [7, 21] or X-ray films with various projections [17]. In this study, measurements of hyoid bone were performed on CBCT images which were made by two observers and intra-observer and inter-observer reliability was found to be high. This shows that the measurements made on the CBCT image of the hyoid bone are reproducible and reliable.
In this study, the mean age of the patients with fused hyoid bone was found to be higher than the unfused and partial fused hyoid bone. In previous studies [8, 18, 21], it was reported that the ossification of the hyoid bone increased with age. This finding is consistent with the results of this study. Consistent with the results of previous studies [5, 14, 19, 21, 23], the mean values of all measurements of hyoid bone in all fusion types except LD-D were higher in males than females. Similarly, Logar et al. [14] reported that there was no significant difference between the sexes on the average of the maximum height of the upper end of the both left and right cornua.
All measurements except RD-D of unfused hyoid bone measurements showed sexual dimorphism in present study. B-L, Hyd-L, RP-W, LP-W and stature measurements of partially fused hyoid bone and Hyd-L, LP-W and stature measurements of fused hyoid bone showed sexual dimorphism. Although Reesink et al. [17] did not find any significant sex difference in the length of the cornu major, in our study Hyd-W of unfused hyoid bone measurements showed sexual dimorphism consisted with D’Anastasio et al. [2] and Shimizu et al. [18]. In the studies of Leksan et al. [13] and Shimizu et al. [18] B-L in males was significantly longer than that in females in all age groups which is concordant with our unfused and partially fused hyoid bone measurements. This means that discriminant function equations obtained from parameters showing sexual dimorphism can be used for sex determination.
In this study, B-L and Hyd-L showed a high prediction rate. Similarly, in the study of Torimitsu et al. [21] and Urbanová et al. [23], Hyd-L and B-L showed the highest level of sexual dimorphism. In addition, Pollard et al. [16] and Mukhopadhyay [15] previously reported that the female hyoid bones were generally large; therefore, the total anteroposterior length was reported to be more useful in predicting sex than Hyd-W. In this study, B-L achieved a successful sexual prediction rate of > 80%. Considering these results, it can be said that sex can be predicted even when there is no bone other than hyoid bone.
With the introduction of CBCT in dentistry, there has been an improvement in the ability to analyse bone structures in three-dimensional imaging [3]. Multiplanar images obtained from all planes were used to evaluate bone morphology. One-to-one measurements of bone structures can be made on sections formed in desired thickness and desired plane. Today, CBCT is an easily accessible and easily acquired imaging system with a relatively low radiation dose. It can be obtained from individuals so that measurements in the images can be used to estimate age and gender in forensic cases [4].
CONCLUSIONS
In this study, hyoid bone showed important morphologic changes and differences according to age and sex which can be useful in forensic science for age and sex estimation with high predictive rates. Imaging hyoid bone with CBCT is simple, easily applicable and very reproducible methodology. So, in forensic medicine, accurate measurements and evaluation can be performed by CBCT in living individuals with lower radiation dose. To support the results of the study, new studies are needed in which the distribution of sex is equal and the age groups are similar and with volume measurements.