Sex estimation using Magnetic Resonance Imaging measurements of hand and foot in Egyptian population

Abstract

Background

The identification of skeletal remains begins with a sex evaluation since other biological profile elements, such as size and age, are sex-specific. The pelvis is the favored element for sex assessment since it is very sexually dimorphic.

Aim of the study

Delivering hand and foot bones sex estimate criteria for Egyptian population identification was the goal of this work.

Subject and methods

Eighty-two Egyptian adult subjects (41 males and 41 females) aged ≥ 21 years old were subjected to MRI scan on the right hand and foot to assess 9 measurements: hand length, hand width, four measurements of middle metacarpal bone (length, breadth, head breadth and base breadth), first metatarsal bone length, first metatarsal bone width and (first metatarsal bone length x first metatarsal bone width).

Results

All the measured parameters in the right hand and foot were higher in males than females. First metatarsal bone length x first metatarsal bone width in the foot and middle metacarpal bone breadth in the hand were the most sexually dimorphic parameters, with accuracy of 86.6 percent and 84.1 percent, respectively. In stepwise discriminant analysis, three of nine measurements were selected: First metatarsal bone length x first metatarsal bone width, middle metacarpal bone breadth and hand width for the sex prediction equation in the Egyptian population, with a cross-validated sex classification accuracy of 90.2%.

Conclusion

Sex can be assessed from hand and foot metric parameters measured by MRI with high accuracy.

Background

Identification of human remains is considered one of the most important forensic investigations, especially in mass disasters like earthquakes, forest fires, and plane crashes (Maalman et al. 2021).

Regarding legal investigations, the examination of human skeletal remains is important to estimate the sex, age, stature, and ancestry of the remains to establish the biological profile used in personal identification (Pickering and Bachman 2009). Sex estimation is the first and most important factor to be evaluated, as it forms the basis for subsequent estimations that are related to sex, e.g., age and stature (Franklin 2010).

Sex estimation can be done by comparing the anthropological measurements of the skeleton with standard measurements (Ibeachu et al. 2011). The pelvis is the preferred bone in sex estimation because it is considered a highly sexually distinguishing bone, but other bones were also investigated in this field, like the hands and feet (Spradley and Jantz 2011).

Hands and feet consist of small, compact bones that often remain intact after death due to resistance to postmortem changes and less exposure to environmental changes. Therefore, metatarsals, metacarpals, and phalanges are often kept well-preserved in many forensic cases (Krishan 2008; Mountrakis et al. 2010). In the last few years, forensic imaging has rapidly developed and broadened to be used in many forensic disciplines, like forensic anthropology, forensic odontology, forensic ballistics, and wildlife forensics (Zhang 2022).

Several studies reported the use of imaging techniques (e.g., X-ray and Computed tomography) to estimate the sex in different populations by using hand or foot measurements. El Morsi and Al Hawary (2013) used the length of all metacarpals and phalanges of the right and left hands from X-ray radiographs to develop a discriminant formula that can be used to estimate sex in the Egyptian population. De Silva et al. (2014) also used digital right-hand measurements to estimate sex in the Western Australian population.

Magnetic resonance imaging (MRI) is a technique with relatively low radiation effects that can examine all types of tissue and produce detailed images of the internal body (Hori et al. 2021). The combination of more than one parameter in one bone or parameters from different bones may enhance the accuracy of sex estimation (Kharuhadetch et al. 2022). In this study, we used a combination of MRI radiological measurements of the right hand and right foot bones of living adults to formulate a combined equation and test its reliability to estimate the sex of individuals in the contemporary Egyptian population.

Methods

Study strategy

This research was carried out in the radio-diagnosis department of the Zagazig University Hospitals in Sharkia governorate, Egypt, from December 2021 to December 2022. All participants provided their written informed consent, and the research was authorized by the Zagazig University Faculty of Medicine’s Human Research Ethics Council, Egypt (ZU-IRB#6563/2-12-2020).

The study was carried out on eighty-two subjects (41 males and 41 females) with an age range of 21 to 66 years.

Patients who were ineligible for MRI, such as those with a cardiac pacemaker or cochlear implant, were excluded from the study. Other exclusion criteria included individuals under the age of 21, those with skeletal immaturity, a history of fractures, bone tumors, or arthritis, as well as those with pathological lesions such as congenital and developmental dysplasia, metabolic bone diseases, connective tissue diseases, and previous orthopedic surgery, were excluded. Pregnant or breastfeeding women, as well as patients with chronic liver, kidney, and heart disorders.

Every individual underwent an MRI scan utilizing a restricted plan exclusively tailored for the research. A 1.5-T Philips Gyroscan Achieva (Best, Netherlands) enclosed-arrangement full-body scanner with an extremity coil was used to conduct an MRI of the right hand and right foot. The participants were educated about the kind and duration of the scan (which lasted from 15 to 30 minutes) and were directed to eliminate any metallic items and maintain stillness.

Right-hand MRI measurement parameters

Participants were positioned in a position called the "superman position" (Pazahr et al. 2021). In this position, the subject is put in a prone stance with their arm raised above their head. A molded holder was used to immobilize their wrist in a neutral position, and the radiocarpal joint was utilized as a reference point. The hand was covered with a circular coil (C 200) that was secured with rubber bands. The axial, coronal, and sagittal imaging planes were determined based on the finger rather than the hand, and an adjacent finger was included in the field of view (FOV) for comparison purposes. The MR imaging protocol involved three planes: coronal (T1WI, T2WI, proton density with fat saturation, and STIR), and sagittal (T2WI), along with specific parameters.

Beside the right-hand length and width, the third metacarpal bone in the right hand was used because it was proven to be one of the best bones that can be used in correct sex estimation in the Egyptian population, according to El Morsi and Al Hawary (2013). We included both the length and widths of the third metacarpal bone according to De Silva et al. (2014).

Right hand length, width, and four measurements of the middle metacarpal bone were assessed in centimeters as follows:

• 1- Length of the hand (HL): The measurement is taken from the tip of the middle finger to the center point of a line that joins the distal styloid processes of the radius and ulna bones (Fig. 1).

• 2- Hand width (HW): Distance from the most medial point of the second metacarpophalangeal joint to the most lateral point of the fifth metacarpophalangeal joint (Fig. 1).

• 3- Middle metacarpal bone length (MML): It is taken from the most proximal to the most distal points of the middle metacarpal bone (Fig. 2).

• 4- Middle metacarpal bone breadth (MMB): The measurement is taken from the most lateral point to the most medial point of the mid-shaft region of the middle metacarpal bone (Fig. 2).

• 5- Breadth of the head of the middle metacarpal bone (MMHB): The measurement is taken from the most lateral point to the most medial point of the head of the middle metacarpal bone (Fig. 2).

• 6- Breadth of the base of the middle metacarpal bone (MMBB): The measurement is taken from the most lateral point to the most medial point of the base of the middle metacarpal bone (Fig. 2).

Fig. 1

Representative MRI images showing right hand length and width which were measured in centimeters in a female (A) and a male (B)

Fig. 2

Representative MRI images showing right hand middle metacarpal bone measurements in centimeters in a female (A) and a male (B)

Right foot MRI measurement parameters

Participants were positioned in a supine posture with their feet first. An extremity surface coil was positioned on the surface of the foot. The MRI protocol included: a) capturing Sagittal T1-weighted images (TR 500–650 msec, TE 20 msec, slice thickness 2.5 mm, gap 0.3 mm, FOV 170 mm, and NSA =2). b) Capturing Sagittal T2 fat saturation images (TR 4900 msec, TE 100 msec, slice thickness 3 mm, gap 0.3 mm, FOV 170 mm, and NSA =1) c) Capturing Sagittal stir images (TR 4000-5000 msec, TE 100 msec, TI 130 msec, slice thickness 3 mm, gap 0.3 mm, FOV 170 mm, and NSA =1). The first metatarsal bone in the right foot was proven to be the best bone that can be used for correct sex estimation in the Iranian population (Akhlaghi et al. 2017). For the right foot, the following parameters in centimeters (Fig. 3):

• 1- Length of the first metatarsal bone (IML): The measurement is taken from the center of the caput to the center of the basis.

• 2- Width of the first metatarsal bone (IMW): The measurement is taken from the most lateral point to the most medial point of the mid-shaft region.

• 3- First metatarsal bone length x width (IML x IMW).

Fig. 3

Representative MRI images showing length and width of the first metatarsal bone of right foot which were measured in centimeters in a female (A) and a male (B)

Statistical analyses

The SPSS software (Statistical Package for the Social Sciences) version 28 was utilized to conduct data analysis. The means and standard deviations were employed to describe quantitative variables. To confirm the assumptions for parametric tests, the Shapiro-Wilk test was administered. In normally distributed data, the Independent T test was performed to determine the mean difference in quantitative variables between males and females. An ROC curve was used to identify the optimal cutoff for a specific quantitative parameter in forecasting sex. The statistical significance level was established at P<0.05, and a highly significant difference was observed at p≤0.001. A stepwise Discriminant Analysis was executed to develop a formula for the most precise sex classifications. The Wilk's lambda values were computed to assess how accurately each equation classified the samples into males and females.

Results

There was no difference between the studied groups regarding the mean value of age (p > 0.05). The mean age of males was 42.07 ± 13.69, and the mean age of females was 41.56 ± 10.63.

When analyzing certain MRI radiological measurements of the right hand and foot, males showed significantly (p<0.001) higher measurements than females for all the parameters investigated (HL, HW, MML, MMB, MMHB, MMBB, IML, IMW, and IML×IMW) (Table 1).

Table 1 Comparison between males and females as regard age and MRI parameters of right hand and foot in Egyptian population

The best cutoff of HL in the prediction of male sex is ≥18.84 with an area under the curve of 0.857, a sensitivity of 75.6%, a specificity of 75.6%, and an overall accuracy of 75.6%. At cutoff ≥7.255, HW can refer to the male gender with a sensitivity of 85.4%, a specificity of 70.7%, and an overall accuracy of 78%. MML≥6.785 can predict male sex with a sensitivity of 80.5%, a specificity of 73.2%, and an overall accuracy of 76.8%. The best cutoff of MMB in the prediction of male sex is ≥0.755 with an area under the curve of 0.919, a sensitivity of 85.4%, a specificity of 82.9%, and an overall accuracy of 84.1%. The best cutoff of MMHB in the prediction of male sex is ≥1.455 with an area under the curve of 0.796, a sensitivity of 80.5%, a specificity of 70.7%, and an overall accuracy of 75.6% (Table 2, Figs. 4 and 5).

Table 2 Performance of hand and foot parameters in identification of sex in Egyptian population

Fig. 4

ROC curve showing performance of MML, MMB, MMHB, MMBB, HL and HW in identification of sex in Egyptian population (AUC for MML, MMB, MMHB and MMBB, HL and HW were 0.81,0.755, 0.796, 0.829, 0.857 and 0.866 respectively)

Fig. 5

ROC curve showing performance of IMW and IML ×IMW in identification of sex in Egyptian population (AUC for IMW, and IML ×IMW were 0.839 and 0.905 respectively)

The best cutoff of MMBB in the prediction of male sex is ≥1.485 with an area under the curve of 0.829, a sensitivity of 82.9%, a specificity of 70.7%, and an overall accuracy of 76.8%. The best cutoff of IML in the prediction of male sex is ≥6.005 with an area under the curve of 0.856, a sensitivity of 80.6%, a specificity of 75.6%, and an overall accuracy of 78%. The best cutoff of IML in the prediction of male sex is ≥1.175 with an area under the curve of 0.839, a sensitivity of 78%, a specificity of 75.6%, and an overall accuracy of 78%. The best cutoff of IML*IMW in the prediction of male sex is ≥7.265 with an area under the curve of 0.905, a sensitivity of 87.8%, a specificity of 85.4%, and an overall accuracy of 86.6% (Table 2, Figs. 4 and 5).

The most sexually dimorphic individual measurements were those that yielded the highest expected sex classification accuracy: MI ×WI (86.6%), followed by MMB (84.1%), HW (78%), and MI (78%). In the stepwise discriminant analysis of the right hand and foot parameters, 3 measurements (IML ×IMW, MMB, and HW) from 9 were selected because they yielded the highest expected sex classification accuracy, as previously proved. Sex can be predicted from this equation:

Sex = 6.544 * MMB + 1.095 * HW + 0.415 * IML × IMW -16.137.

If the result is above or equal to 0, the bones belong to males, and if <0, they belong to females, with a cross-validation of 90.2% (Table 3).

Table 3 Stepwise discriminant analysis of hand and foot parameters to detect sex in Egyptian population

Discussion

Regarding the accuracy of sex estimation using skeletal measurements, the pelvis is the most specific dimorphic bone (due to the childbirth factor in females), and the accuracy was reported to be 95% when complete (Abdallah et al. 2021). The next bone with high accuracy was the skull, with accuracy in sex estimation up to 94% (Yang et al. 2020). The long bones were reported to have a high degree of accuracy as follows: The humerus (between 83% and 96%), the radius (between 85% and 94%), the femur (between 76% and 97%), and the tibia (91%) (Robinson and Bidmos 2009). Hand and foot bones were used also in sex estimation with high accuracy (Mountrakis et al. 2010).

In this study, some MRI radiological measurements of the right hand (hand length, hand width, and four measurements of middle metacarpal bone) and right foot (first metatarsal bone length, first metatarsal bone width, and first metatarsal bone length x first metatarsal bone width) were used to estimate the sex of individuals and establish a sex-predicting equation in the Egyptian population. Comparing both sex measurements, the results of this study showed that males were significantly (p<0.001) higher in all right hand and right foot measurements than females.

Going hand in hand with this work, many studies also used some hand, foot, or both hand and foot measurements to estimate the sex by direct measurement or imaging techniques and concluded that these measurements were significantly higher in males than females and could be used for sex estimation.

Some studies used both hand and foot measurements to estimate sex in different populations by direct measurements e.g., the metacarpals and metatarsals measurements in the Mexican population (Torres et al. 2020); foot length, foot breadth, hand length, and hand breadth in the Egyptian population (Mohamed 2013); hand (length and breadth), index/ring finger ratio, foot (length and breadth), and ankle breadth in the Nigerian population (Iroanya and Egwuatu 2020).

Other studies used only measurements of hand bones to estimate sex e.g. El Morsi and Al Hawary (2013) used the length of all metacarpals and phalanges of the right and left hands from X-ray radiographs in the Egyptian population. De Silva et al. (2014) used right-hand measurements from radiographic images (metacarpals and proximal phalanges) in the Western Italian population. Dey and Kapoor (2015) and Maalman et al. (2021) used direct measurements of the hand breadth and length from the Indian and Ghanaian populations, respectively.

Other studies used only measurements from the foot bones to estimate sex. Rodríguez et al. (2014) used computerized tomography to estimate sex from the measurements of the first metatarsal in a Portuguese population. Akhlaghi et al. (2017) used a radiographic examination of all metatarsal bones in the Iranian population. Bindurani et al. (2017) used direct measurements of foot length, foot breadth, and bimalleolar breadth in the Indian population. Lalwani et al. (2019) also used direct measurement of the foot length in the Central Indian population.

On the contrary, other studies documented some hand measurements to be higher in females compared to male measurements. Alicioglu et al. (2009) documented greater radiographic hand measurements in males more than in females except for the distal phalanges in the Turkish population. Eshak et al. (2011) reported greater measurements of the length of the distal phalanges of all fingers, 1st and 3rd proximal metacarpals, and all metacarpals in males more than in females but there was no significant difference in the length of the middle, 2nd, 4th or 5th proximal phalanges between males and females on Egyptian population by computerized tomography.

By determining the accuracy of each bone in sex estimation, the present study showed that the hand width, middle metacarpal bone breadth in the right hand, and first metatarsal bone length x first metatarsal bone width in the right foot were the most accurate measurements.

Going hand in hand with our results, Akhlaghi et al. (2017) found that the first metatarsal bone length x first metatarsal bone width was the most accurate foot measurement in estimating sex.

On the contrary, other measurements were used in previous studies for sex estimation in different populations with high accuracy, like the second metacarpal and the first metatarsal (Torres et al. 2020); the first distal phalanx, proximal phalanx, 3rd metacarpal, and 4th metacarpal lengths (El Morsi and Al Hawary 2013); the widths of the third metacarpal bone (De Silva et al. 2014); foot breadth (Bindurani et al. 2017); and foot length (Lalwani et al. 2019).

The differences in results between studies in this field may be attributed to the difference in population, sample size, measurements used, methods used in measurements, and statistical methods, which are all reported to affect sexual dimorphism (El Morsi and Al Hawary 2013; Dayarathne et al. 2021).

Research has proven the global variation in the skeletal measurements used for sex estimation due to genetic and environmental factors affecting skeletal growth (Ubelaker and DeGaglia 2017). Population-specific influences affect not only the growth in both sexes but also the expression and degree of skeletal dimorphism (Swift et al. 2023).

In the present study, a discriminant function equation from these three measurements (HW, MMB, and IML x IMW) was formulated to estimate sex in the Egyptian population with an accuracy of up to 90%. This formula could be used to develop a software program in future work to estimate the sex of the Egyptian population. Many software programs offer a reproducible method for sex estimation with no human bias (Bewes et al. 2019). Fordisc 3 is a software program that can aid in the estimation of sex, population affinity, and stature (Ousley and Jantz 2013). Diagnose Sexuelle Probabiliste V2 is another software that can be used in sex estimation from hipbone measurements (Machado et al. 2018).

Conclusions

Hand length and width, middle metacarpal bone, and first metatarsal bone measurements can be used for sex estimation in the contemporary Egyptian population with high accuracy, ranging from 75.6% to 86.6%. The most sexually dimorphic individual measurements are first metatarsal bone length x first metatarsal bone width 86.6% and middle metacarpal bone breadth 84.1%. Combining measurements by stepwise discriminant analysis yielded a higher accuracy of 90.2%. It is recommended that this study be conducted with a bigger sample size in order to corroborate the findings and improve accuracy. The application of more hand and foot measurements and the incorporation of the results into a software program for sex estimation in the Egyptian population might provide better results and yield more accuracy in sex dimorphism in future studies.