Discriminant Function Analysis for Gender Determination Using Maxillary Sinus Volume and Index: A Cone-Beam Computer Tomography Study

Abstract

Objective

The purpose of the study was to establish baseline values for maxillary sinus volume (MSV) and maxillary sinus index (MSI) among a population of Saudi adults and to test the accuracy of using MSV and MSI for gender determination among the study samples.

Materials and Methods

Two hundred and seventy-five cone-beam computer tomographic (CBCT) images of adults that satisfy the study criteria were recruited as the study samples. The CBCT images were obtained from the patients' records in the CBCT database of the dental clinic of King Faisal University, Al-Ahsa, Saudi Arabia. Maxillary sinus height (H), width (W), and depth (D) were measured from the coronal, axial, and sagittal image views using a vision iCAT software. The MSV and MSI were obtained by the mathematical formula MSV = H*W*D*0.33 in cm³ and MSI = W/H. Data was analyzed using SPSS software, and a discriminant function test was used to test for gender determination.

Results

Mean MSV was 12.67 mm ± 3.46, and mean MSI was 0.88. No statistically significant difference was observed between the right and left (p > 0.05). There was strong gender dimorphism in both MSV and MSI in the samples. Males have statistically higher mean (p = 0.001) MSV and MSI (p = 0.002). The discriminant function test shows an overall gender predictive accuracy of 84% when MSV was used for gender determination and 77% when MSI was used.

Conclusion

Both MSV and MSI are good predictors for gender determination, but MSV is superior and has higher accuracy.

Keywords

cone-beam computer tomography - gender dimorphism - gender determination - maxillary sinus index - maxillary sinus volume

Introduction

Personal identification is a key component of forensic investigation, regarded as one of the most significant priorities in solving criminal cases or evaluating disasters that are of forensic concern.[1] It may include using pieces of evidence, including body parts or radiographic images, to make estimates or predictions related to individuals' parameters like stature, weight, or sex.[1] Anthropometric and radiographic measurements of body parts are therefore important in forensic or medical-related cases because they can serve as “pieces of evidence” to use for personal identification. Gender determination is the use of morphological, physiological, and biochemical differences between males and females in the living and human remains and makes predictions as to the sex of the individual.[2] It is used in the initial stages in forensics and can also help clinicians in formulating empirical treatment options.[2] The skull bone is one of the skeletal parts of the body that are used for gender determination. Maxillary bone and sinuses are known to remain fairly intact even when the skull bones are badly blemished in severe accidental injuries.[3] [4] Maxillary sinuses (MSs) are the largest pyramidal paranasal sinuses located in the maxillary bones.[5] [6] The location of the MS is below the orbit, and the sinus floor is composed of the alveolar bone and the apex facing laterally toward the zygomatic bone. Its development is initiated in the 10th week of intrauterine life and in the 11th week of development, the mucosa lining the further anterior end of the ethmoid infundibulum exhibits penetrations into the surrounding mesenchyme, resulting in an independent cavity that represents the primordium of the MS.[7] The sinus grows rapidly during two developmental periods: from the 17th to the 20th week and between the 25th and the 28th week.[7] The MS and ethmoidal sinuses begin to form when three horizontal extensions emerge from the nose's lateral wall and grow in directions. The inferior projection, known as the maxilla-turbinate, forms the MS.[8] The superior projection known as the ethmoid turbinate forms ethmoidal air cells and the drainage channels that correspond to them. At birth, the MS are primitive. After the ninth year, they expand, and they fully manifest throughout puberty, reaching the adult size around 18 to 20 years of age.[8] MS being the largest of the sinuses has many neurovascular relations like anterior superior alveolar nerve and arteries and structures in the pterygopalatine fossa.[9] Apart from humidification of inhaled air, reducing the weight of the skull, production of nitrogen monoxide by the MS is regarded an important function in local immunological defense mechanisms.[9]

The low frequencies of developmental defects like hypoplasia, etc., seen in the MS make it a reliable anatomical structure in forensic sciences.[5] In the MS, pneumatization varies from person to person and between the genders.[5] The pneumatization of the MS is the sinus extension toward the alveolar bone, which is affected by genetic and environmental factors.[6] Its anatomy has been used for identification of skeletal remains and determination of gender. Various studies have concluded that the measurements of the MS can help determine the gender in forensic investigation.[2] Individual variations due to the complex anatomy make the paranasal sinuses an area for exploration in forensics for determining individuals' age and gender.[2]

The conventional radiograph is the most used technique to analyze the MS. However, it has its limitations when it comes to three-dimensional (3D) structure research. Cone-beam computer tomography (CBCT), followed by a 3D reconstructions provides valuable measurements like craniocaudal, transverse, and anteroposterior (AP) diameters; this has demonstrated consistency with manual measurements of the MS obtained from dry human skulls.[7] The advantages of CBCT include its providing 3D multiplanar images, a lower radiation dose than conventional computed tomography (CT), lower costs, and a shorter acquisition time.[7] The CBCT is now used as a reliable method for evaluating MS volumes (MSVs).[5] The MS index (MSI), determined using the MS height and width, is a relatively new parameter in the MS evaluation, although not many studies used it for such purpose.[1] The MSI is easier to determine as is calculated by dividing the sinus height with depth and can be important in the sinus evaluation during implant insertions, diagnosing, and treating cases of malocclusion, among other pathologies.[1]

Although studies were conducted to provide volumetric parameters of MS, to the best of our knowledge, only very few studies were conducted that presented the morphometry of the MS in terms of MSI and MSV among Saudi population. The present study is therefore aimed at establishing baseline data for these MS parameters (MSI and MSV) among a population of Saudi adults to supplement the few existing data. More importantly, the purpose was to investigate the reliability of using MSI, which is a relatively new and easy-to-use tool used in MS morphometry. Additionally, we want to compare the degree of accuracy of gender prediction using MSI with that of MSV, which was well studied for gender determination. This may come handy because obtaining MSI is much easier since it requires only two-dimensional images, unlike MSV, which would require 3D images.

Materials and Methods

Study Setting and Design

The study was conducted at the Dental Clinic of the King Faisal University, Saudi Arabia. A retrospective cross-section design was employed to obtain the CBCT images from year 2020 to 2024.

Samples Selection

Two hundred and seventy-five CBCT images were randomly selected from the image databank of the dental clinic. Images were of male and female adult (20–60 years) Saudi population. Any obvious anatomical abnormality was excluded before the final selection of images. Ethical approval (number: KFU:REC:2025-ETHICS:3018) was obtained from the institute's deanship of scientific and research unit.

CBCT Image Analysis and Measurements

The CBCT machine was a 3D Accuitomo 170 (MORITA, Japan) with 90 Kv, 5 to 8 mA. It uses an exposure time of 17.5 seconds and a 0.25-mm voxel size. A Vision iCAT imaging software, downloaded in the personal computer, was used in processing MS images. Axial, coronal, and sagittal multiplanar image views were used in measuring different MS dimensions using standard measurements. The image analyses were performed by a radiologist (with more than 8 years of experience) who helped in the calibrations and obtaining the images, and a dentist (with 15 years' experience) who did the measurements. [Fig. 1] shows the measurements of the MS height, depth, and width. All measurements were taken using a line ruler in the software, measurements to the nearest 1.5 mm as follows:

Fig. 1 Cone-beam computer tomography (CBCT) images of maxillary sinus (MS) height (A), width, and depth (B).

1. MS height (H): maximum vertical height from the lowest point on the floor to the roof, seen on coronal images.

2. MS width (W): maximum distance between the medial wall and the outermost point, seen on coronal images.

3. MS depth—AP (D): distance from the base to the apex seen in axial images.

4. The MSV was calculated using the formula H*W*D*0.33 mm³.

5. MSI was calculated using the formula H/W.

Statistical Analyses

The variables used for the analysis are the right and left MSV and MSI and the values obtained were expressed in terms of minimum, maximum, and mean ± standard deviation (SD). A Kolmogorov–Smirnoff test was performed to test whether or not the variables were normally distributed. Student's t-test was used for mean comparison between genders and to test for the strength of gender dimorphism. Discriminant function (DF) analysis was performed for the determination of gender using MSV and MSI as predictor variables. A DF equation model was obtained with gender as a classifying variable and MSV and MSI as independent variables. The analyses were performed using SPSS version 23 (IBM Corporation, for Windows), and 5% (p < 0.05) level of error was considered significant for statistical inference.

Results

The study was conducted over 275 image samples of adult males and females from indigenous population; the male-to-female ratio was 1.1:1, and the age range was 20 to 60 years with a mean of 41years ± 10.98 SD. [Table 1] shows the means and SD of the right and left variables. The overall mean of the MSV was found to be 12.64 cm³ ± 3.47 SD for the right sinus, and 12.04 cm³ ± 3.10 SD for the left. No significant mean difference was found between the right and left MSVs (p = 0.33). The mean MSI in the right and left were 0.86 ± 0.10 and 0.90 ± 0.12, respectively, with no statistical difference (p = 0.21).

Table 1 Descriptive statistics of the right and left maxillary sinus volume and index

Variable	n	Min	Max	Mean	Standard deviation	p-Value
RsV (cm3)	275	9.61	21.05	12.67	3.46	0.33
LsV (cm3)	275	9.08	17.51	12.04	3.10
RMSI	275	0.58	1.14	0.86	0.10	0.21
LMSI	275	0.58	1.24	0.90	0.12

Abbreviations: LMSI, left maxillary sinus index; LsV, left sinus volume; RMSI, right maxillary sinus index; RsV, right sinus volume.

[Table 2] shows the independent samples t-test done to compare the mean difference of MSV and MSI between the males and females. The findings showed significant gender dimorphism in all the variables, with males having significantly higher MSV and MSI (p < 0.001). The trend was similar on both right and left sides.

Table 2 Independent t-test comparison of MSV and MSI between males and females

Variables	Males			Females			t-Test	p-Value
	n	Mean	SD	n	Mean	SD
RsV (cm3)	144	13.64	3.22	131	10.03	1.96	20.14	0.001
LsV (cm3)	144	12.14	2.90	131	9.25	1.69	30.21	0.001
RMSI	144	0.919	0.08	131	0.80	0.10	10.74	0.001
LMSI	144	0.961	0.10	131	0.84	0.11	9.10	0.021

Abbreviations: LMSI, left maxillary sinus index; LsV, left sinus volume; MSI, maxillary sinus index; MSV, maxillary sinus volume; RMSI, right maxillary sinus index; RsV, right sinus volume; SD, standard deviation.

The DF test performed using MSVs (right and left) to predict gender is summarized in [Table 3]. One hundred and twenty-four of the 144 males were correctly classified by the DF model as males, and 107 of the 131 females were correctly classified as females. This gives a sensitivity of 85% and a specificity of 84%. Conversely, when the DF was performed using MSI as predictors ([Table 4]), 118 out of the 144 males and 92 out of the females were correctly classified. This gives a sensitivity of 83% and specificity of 70% of the function using sinus index, which is slightly lower than that for sinus volume.

Table 3 Discriminant function test using maxillary sinus volume as predictors for gender

	Predicted group		Total
Original count	Male	Female
Male	122	22	144
Female	24	107	131
%
Male	84	16	100
Female	15	85	100

Table 4 Discriminant function table using maxillary sinus index as predictor for gender

	Predicted group		Total
Original count	Male	Female
Male	119	25	144
Female	39	92	131
%
Male	83	17	100
Female	30	70	100

[Table 5] summarizes the effect sizes and the strength of prediction power by using both MSV and MSI as predictors. The Wilks' lambda shows a lower value (0.507) when MSV is used as predictors for gender determination than the value obtained (0.696) when MSI was used. This indicates that MSV gives a better model with stronger discriminating power for gender prediction. Both models are, however, statistically significant (p < 0.001).

Table 5 Effect sizes and discriminating power of the DFT for gender prediction

Variable	Wilks' lambda	Standardized coefficient	Canonical coefficient	p-Value
RsV	0.507	0.481	0.177	0.0001
LsV		0.564	0.233
RMSI	0.696	0.455	8.733	0.0001
LMSI		0.346	2.271

Abbreviations: DFT, discriminant function test; LMSI, left maxillary sinus index; LsV, left sinus volume; RMSI, right maxillary sinus index; RsV, right sinus volume.

The standardized DF coefficient shows that MSV has more explanatory power with overall coefficients of 0.53 than MSI with a coefficient of 0.401. This may indicate that MSV is a better predictor than MSI for gender prediction. The DF equations for the two models are found to be statistically significant (p = 0.001).

The DF equation for MSV was determined to be D = 0.177*RV + 0.233*LV – 4.390, with values toward 0.759 (> 0) to be classified as males and values toward −0.880 (< 0) to be classified as females. On the other hand, the DF equation for MSI was determined to be D = 8.73*RMSI (right MSI) + 2.271*LMSI (left MSI) – 9.27, with values toward 0.607 (> 0) classified as male and −0.604 (< 0) to be classified as female.

Discussion

Examination of the skull, pelvic bone and long bones can give up to 90.0 to 95.0% accuracy in sex determination.[9] [10] Both conventional CT scan and CBCT have been reported to be reliable methods in the estimation of different dimensions of the MS.[11] Studies have also shown that measurements of the MS sizes obtained from skull measurements were similar and consistent to those obtained by CT scans.[12] [13] [14] It may, therefore, be reasonable to assume that CBCT is a reliable method for measuring the dimensions of the MS. Apart from being an easy tool to be used in forensic analysis, knowing the MSI and MSV may be useful in planning sinus graft procedures, which are frequently conducted to rebuild and restore the edentulous posterior maxilla that has been affected by alveolar bone loss, following extraction and sinus pneumatization.[15] [16] During surgical procedures an elevated risk of hemorrhage is caused by the varied position of the vascularization of the Schneiderian membrane—which can be perforated—and the posterior-lateral projection of the sinus walls.[17] A thorough understanding of the MS vascular network is therefore necessary to prevent any difficulties during sinus lift surgery.[17] In approximately 10 to 30% of cases, the posterior superior alveolar artery, which is situated on the MS lateral wall, is present where the lateral window is formed for sinus floor elevating procedures.[18]

In this study, we used CBCT images to obtain MSV and MSI, provided reference ranges for both, and used statistical methods to predict gender of the subjects using these dimensions. We found a statistically significant difference between males' and females' MSV and MSI. The study also found that DF tests can be used to predict gender using MSV and MSI with a moderate accuracy, and that using MSV gives a better prediction.

Studies have differed in their findings concerning gender dimorphism in MS dimensions. While we found a statistically significant gender difference in the mean MSV and MSI, many have found no such difference. Najem et al found no gender dimorphism in MS height and width among a population of adult Egyptians, using CBCT.[19] They examined a sample of only 86 images and measured only the height and width. Similarly, Hettiarachchi et al[20] and Ghada et al[21] found no gender difference in MSV among CBCT images of adults. Both these studies were also conducted on a relatively small sample of 20 and 67, respectively, and the latter study used conventional CT rather than CBCT, which was used for this study. On the other hand, there are studies that show significant gender differences similar to our findings.[22] [23] [24] Vidya et al in a study on 30 dry skulls of South Indians, found that the MSV of the males was slightly more than those in the females.[25] Furthermore, although not many studies evaluated the effect of MSI on gender determination, Khaitan et al[1] and Chalkoo and Salma[26] observed that MSI is significantly different in males than in females. Both of these studies used lateral X-ray cephalogram in measuring the sinus dimensions. In Europe, a study about gender determination found that the volume of MS was significantly larger and wider in males than in females.[27]

The DF test we performed on the samples was able to predict and classify 122 males correctly out of 144 and 107 out of 131 of females correctly, giving a sensitivity of 85%. This is when MSV was used as a predictor. A similar study conducted by Uthman et al found that 74.4% of males and 73.3% of females were correctly classified using MS dimensions.[28] Similarly, Chandra et al have found that MS area and perimeter by lateral cephalogram could be used accurately in gender determination. The correct predictive accuracy they found was 70.8% in males and 62.5% in females.[29] Some investigators in Egypt conducted a study to predict gender from some radiologic measurements of the MS using multidetector CT and came up with a predictive accuracy of 70.8% in males and 62.5% in females.[4]

Similar to our study, Prabhat et al examined the MS of a sample of 30 adults' CT scans in the Indian population. They conducted DF tests and the method was able to predict gender with an accuracy of 80.0% in males and 86.7% in females, with an overall accuracy rate of 83.3%.[30]

This study found that MSV was a better predictor of gender than MSI with both the right and left MS dimensions added in the DF equation. Not many studies were conducted on comparing the two variables for gender determination in the same population. Prabhat et al conducted DF test to obtain an overall accuracy rate of 83.3% with three MS parameters entered into the model.[30] Our DF equation is simple and fairly accurate (85%) with only two variables entered. This could be because of a larger sample size (275) in our population, the use of CBCT, and/or the variable volume we used in the equation.

Conclusion

The study provided a normative value for MSV and MSI in Saudi Arabian population using CBCT. It concluded that the two variables are sexually dimorphic and that both are good predictors for gender determination, with MSV being a better predictor. The DF equation performed provided a high predictive accuracy, which was comparable if not higher than many studies.

Recommendations

The probability of surgical problems could be significantly decreased with accurate diagnosis, appropriate treatment planning, and familiarity with the surrounding anatomy. It is important to understand that the size of the MS varies between men and women. Surgeons may provide patients with more individualized care as a result of this knowledge. When diagnosing sinusitis, contemplating orthodontic procedures, carrying maxillofacial surgeries, or implanting dental prosthesis, understanding these distinctions is essential to getting good results. Future research might examine the surgical consequences of MS measurements in greater detail, in relations to neighboring anatomical structures. It is therefore possible to look into the effects of these variations on associated surgical specialties such as ophthalmology and neurosurgery.

Conflict of Interest

None declared.

• References

• 1 Khaitan T, Kabiraj A, Ginjupally U, Jain R. Cephalometric analysis for gender determination using maxillary sinus index: a novel dimension in personal identification. Int J Dent 2017; 2017: 7026796
• 2 Zarate-Reyes A, Chavez-Lazo Y, Guerrero ME, Wahjudianto N, Wahjuningrum DA. Sexual dimorphism of maxillary sinuses in adult population using cone-beam computed tomography. A pilot study. Saudi Dent J 2024; 36 (04) 528-532
• 3 Kanthem RK, Guttikonda VR, Yeluri S, Kumari G. Sex determination using maxillary sinus. J Forensic Dent Sci 2015; 7 (02) 163-167
• 4 Amin MF, Hassan EI. Sex identification in Egyptian population using multidetector computed tomography of the maxillary sinus. J Forensic Leg Med 2012; 19 (02) 65-69
• 5 Kalaskar R, Ijalkar R, Kalaskar A. Volumetric evaluation of maxillary and frontal sinuses in 9–14 year old children—a cone-beam computed tomography pilot study. J Forensic Dent Sci 2023; 13 (02) 70-77
• 6 Gulec M, Tassoker M, Magat G, Lale B, Ozcan S, Orhan K. Three-dimensional volumetric analysis of the maxillary sinus: a cone-beam computed tomography study. Folia Morphol (Warsz) 2020; 79 (03) 557-562
• 7 Okşayan R, Sökücü O, Yeşildal S. Evaluation of maxillary sinus volume and dimensions in different vertical face growth patterns: a study of cone-beam computed tomography. Acta Odontol Scand 2017; 75 (05) 345-349
• 8 Pohunek P. Development, structure and function of the upper airways. Paediatr Respir Rev 2004; 5 (01) 2-8
• 9 Uthman AT, Al-Rawi NH, Al-Timimi JF. Evaluation of foramen magnum in gender determination using helical CT scanning. Dentomaxillofac Radiol 2012; 41 (03) 197-202
• 10 Gülekon IN, Turgut HB. The external occipital protuberance: can it be used as a criterion in the determination of sex?. J Forensic Sci 2003; 48 (03) 513-516
• 11 Lee JH, Park JT. Three-dimensional CBCT based evaluation of the maxillary sinus by facial index. Int J Environ Res Public Health 2022; 19 (09) 5040
• 12 Sahlstrand-Johnson P, Jannert M, Strömbeck A, Abul-Kasim K. Computed tomography measurements of different dimensions of maxillary and frontal sinuses. BMC Med Imaging 2011; 11: 8
• 13 Ariji Y, Ariji E, Yoshiura K, Kanda S. Computed tomographic indices for maxillary sinus size in comparison with the sinus volume. Dentomaxillofac Radiol 1996; 25 (01) 19-24
• 14 Kawarai Y, Fukushima K, Ogawa T. et al. Volume quantification of healthy paranasal cavity by three-dimensional CT imaging. Acta Otolaryngol Suppl 1999; 540: 45-49
• 15 Cavalcanti MC, Guirado TE, Sapata VM. et al. Maxillary sinus floor pneumatization and alveolar ridge resorption after tooth loss: a cross-sectional study. Braz Oral Res 2018; 32: e64
• 16 Stern A, Green J. Sinus lift procedures: an overview of current techniques. Dent Clin North Am 2012; 56 (01) 219-233 , x
• 17 Guerrero JS. Lateral window sinus augmentation: complications and outcomes of 101 consecutive procedures. Implant Dent 2015; 24 (03) 354-361
• 18 Testori T, Weinstein T, Taschieri S, Wallace SS. Risk factors in lateral window sinus elevation surgery. Periodontol 2000 2019; 81 (01) 91-123
• 19 Najem S, Siraj S, Wael M, Rania A, Elaziz A, Yousria S. Maxillary sinus assessment for gender and age determination using cone beam computed tomography in an Egyptian sample. Alex Dent J 2020; 46 (02) 63-69
• 20 Hettiarachchi PVKS, Gunathilake PMPC, Jayasinghe RM. et al. Linear and volumetric analysis of maxillary sinus pneumatization in a Sri Lankan population using cone beam computer tomography. BioMed Res Int 2021; 2021: 6659085
• 21 Ghada IL, Fathy AI, Adel AA, Rasha AE. The correlation of the maxillary sinus volume with age and mid face parameters using computed tomography. Med J Cairo Univ 2022; 90 (01) 201-208
• 22 Etemadi S, Seylavi G, Yadegari A. Correlation of the maxillary sinus volume with gender and some of craniofacial indices using cone beam computed tomography. Biosci Biotech Res Comm 2017; 10 (03) 580-586
• 23 Pirner S, Tingelhoff K, Wagner I. et al. CT-based manual segmentation and evaluation of paranasal sinuses. Eur Arch Otorhinolaryngol 2009; 266 (04) 507-518
• 24 Sonone J, Nagpure PS, Puttewar M, Garg D. Changes in maxillary sinus volume and it's walls thickness due to chronic rhinosinusitis: a prospective study. Indian J Otolaryngol Head Neck Surg 2019; 71 (Suppl. 03) 2182-2185
• 25 Vidya CS, Shamasundar NM, Manjunatha B. Raichurkar K. Evaluation Of size and volume of maxillary sinus to determine gender by 3D computerized tomography scan method using dry skulls of South Indian origin. Int J Curr Res Rev 2013; 5 (03) 97-100
• 26 Chalkoo AH, Salma R. Lateral cepahalogram- a tool for gender determination using maxillary sinus. J Adv Med Dent Sci Res 2023; 11 (02) 46-49
• 27 Fernandes CL. Forensic ethnic identification of crania: the role of the maxillary sinus–a new approach. Am J Forensic Med Pathol 2004; 25 (04) 302-313
• 28 Uthman AT, Al-Rawi NH, Al-Naaimi AS, Al-Timimi JF. Evaluation of maxillary sinus dimensions in gender determination using helical CT scanning. J Forensic Sci 2011; 56 (02) 403-408
• 29 Chandra S, Devi P, Taneja N, Sah K, Kaur N. Forensic importance of maxillary sinus in gender determination: a morphometric analysis from Western Uttar Pradesh, India. Eur J Gen Dent 2014; 3 (01) 53-56
• 30 Prabhat M, Rai S, Kaur M, Prabhat K, Bhatnagar P, Panjwani S. Computed tomography based forensic gender determination by measuring the size and volume of the maxillary sinuses. J Forensic Dent Sci 2016; 8 (01) 40-46

Publication History

Article published online:
28 April 2025