Predicting the Association Between Oral Hyperpigmentation and Pregnancy in Women Aged 18-45 Using Machine Learning and Directed Acyclic Graphs: A Cross-Sectional Study

Predicting the Association Between Oral Hyperpigmentation and Pregnancy in Women Aged 18-45 Using Machine Learning and Directed Acyclic Graphs: A Cross-Sectional Study | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Research Article Predicting the Association Between Oral Hyperpigmentation and Pregnancy in Women Aged 18-45 Using Machine Learning and Directed Acyclic Graphs: A Cross-Sectional Study Hifza Noor, Arooba Malik, Momina Zahra, Mahnoor Fatima, Rameesa Aymen, and 1 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-8368647/v1 This work is licensed under a CC BY 4.0 License Status: Under Review Version 1 posted 5 You are reading this latest preprint version Abstract Objectives The study aims to predict the association between oral hyperpigmentation and pregnancy in women between 18 and 45 years using machine learning (ML) algorithms. Directed acyclic graphs (DAGs) were constructed to explore the association between oral hyperpigmentation and demographic factors such as age, occupation, ethnicity, and number of pregnancies. Methods This cross-sectional study involved pregnant and non-pregnant participants aged between 18 and 45. The study was conducted at one public and one private sector hospital. DAGs were used to represent relationships between different variables. Three ML models, including Logistic Regression (LR), Random Forest (RF) and Gradient Boosting Machine (GBM) algorithms, were used to predict the association between oral hyperpigmentation and pregnancy based upon relevant theoretically backed predictors. Results The results show that labial mucosa and hard palate hyperpigmentation had a positive correlation with the number of pregnancies, with p-values of 0.003 and 0.018, respectively. All three ML models demonstrated excellent predictive power. However, the LR model with an accuracy rate of 96.67% was particularly effective in predicting the association between oral hyperpigmentation and pregnancy. Conclusions This study highlighted the effectiveness of LR, supported by DAGs, in predicting oral hyperpigmentation among women aged 18–45. Key factors such as the number of pregnancies, age, ethnicity, and occupation significantly influenced site-specific hyperpigmentation. The findings support the integration of ML into maternal oral healthcare for early detection and culturally tailored interventions. Simpler ML models proved most accurate in moderate-sized datasets with fewer covariates. Algorithms Cross-Sectional Studies Maternal Health Regression Analysis Oral Health Figures Figure 1 Figure 2 Figure 3 Figure 4 Introduction Oral hyperpigmentation is a physiological phenomenon due to the excess deposition of melanin and other pigments on the oral mucosal surfaces (Abati, 2024). It is mostly undetectable and presents as discoloration of the gingiva, buccal mucosa, or other intraoral sites (Agarwal et al., 2015 ). This hyperpigmentation, due to its resemblance with pathological lesions and aesthetic concerns, especially for pregnant women facing hormonal changes, serves as a diagnostic challenge (Abati, 2024; Cicek & Ertas, 2003; Pecci-Lloret et al., 2024 ). The prevalence of oral hyperpigmentation changes considerably amongst the population, occurring in approximately 20.8% of people worldwide (95% CI: 17.1–25.0%) (Rotbeh et al., 2022 ). Besides genetic, environmental, social, and cultural factors influencing oral hyperpigmentation, this condition is influenced by physiological factors like age, occupation, ethnicity, and the number of pregnancies (Gulati et al., n.d.). For example, people with darker skin tones, such as those of Africans and South Africans, tend to have more oral hyperpigmentation than people with lighter skin tones (Masilana et al., 2017 ; Ponnaiyan et al., 2024; Feller et al., 2014 ). A study carried out in South Africa, with a prevalence of 42.3% and 430 subjects from various ethnic backgrounds, reported a greater detection rate for orofacial hyperpigmentation (Masilana et al., 2017 ). Similarly, oral hyperpigmentation patterns in developing countries are strongly affected by genes due to the diverse population in terms of race and ethnicity (Almazrooa et al., 2025). Environmental exposures, including smoking, sun exposure, and air quality, play an important part in hyperpigmentation. It is known that ultraviolet (UV) radiation affects the synthesis of melanin (Abati, 2024). In areas of developing countries that get a lot of UV radiation, people typically experience greater levels of skin and oral hyperpigmentation (Shahzad et al., 2018 ). Similarly, brown to black discoloration of the oral mucosa in smoker’s melanosis, particularly in the buccal, hard palate, and gingiva, has a link with smoking (Abati, 2024). A precise diagnosis and patient reassurance, specifically for populations at risk, require an understanding of its patterns and causes. Hormonal changes during pregnancy, notably high levels of progesterone and estrogen, promote melanin to be produced in excess, which results in hyperpigmentation in different parts of the body (Pecci-Lloret et al., 2024 ). The areas involved included the face ( melasma ) and abdomen ( Linea nigra ) (Motosko et al., 2017 ). However, studies on the association of pregnancy and oral hyperpigmentation are extremely limited in developing countries, with no integrated models developed to predict these changes. This represents a major gap in women's oral health research (Sattar et al., 2017 ). While studying oral hyperpigmentation, it was observed that the traditional analysis methods frequently lack the ability to detect the complex nature of these interlinked environmental, physiological, and demographic factors. Oral hyperpigmentation research still underutilizes advanced techniques like Directed Acyclic Graphs (DAGs) and Machine Learning (ML), in spite of their increasing relevance in the healthcare domain (Dubuc et al., 2022 ). While DAGs enable the clear visualization of assumed causal relationships and enable researchers to differentiate between confounders, mediators, and colliders and avoid statistical bias, ML algorithms are capable of detecting non-linear associations and generating predictive insights. Together, ML and DAGs provide a more efficient and open method for simulating complex conditions (Piccininni et al., n.d.). Women from developing countries are not equally represented among AI-driven oral health datasets, which limits the development of appropriate and efficient healthcare solutions for a sizable portion of the population (Marko et al., 2025 ). This study aims to bridge that gap by investigating and predicting oral hyperpigmentation in pregnant and non-pregnant women aged 18 to 45, using DAG-supported ML models. This study contributes to more inclusive, data-driven maternal healthcare by developing predictive models tailored to the ethnically diverse population, aiming to improve patient counselling, improve early detection, and guide preventive measures in oral healthcare for females by addressing the relationship between age, occupation, ethnicity, and pregnancy status (Nyariro et al., 2023 ). Objectives: The objective of the study is to predict the association between oral hyperpigmentation and pregnancy utilizing multiple ML algorithms. This association will be explored using directed acyclic graphs to support theory-driven variable selection and improve model interpretability Purpose: We hypothesized that ML models, supported by DAGs, can more meaningfully predict oral hyperpigmentation and identify key contributing factors such as age, occupation, ethnicity, and number of pregnancies. Methods Sample Description and Study Framework : In this cross-sectional study, 447 women within the ages of 18 and 45 were included from a public and private hospital through convenience sampling. The study was conducted from September 2021 to September 2024. Sample Size: Based on a probable population of 20,000, a 50% response distribution, a 96% confidence level, and a 5% margin of error, a sample size of 414 was calculated using the Raosoft sample size calculator ( http://www.raosoft.com/samplesize.html ). To further reduce the margin of error to 4.8%, we took a sample size of 447 students. Inclusion criteria: Pregnant and non-pregnant participants aged 18 to 45 years who were not suffering from any comorbid conditions were included in the study (Rezazadeh et al., 2014 ). Exclusion criteria: The exclusion criteria of the current study included factors such as smoking history (Tadakamadla et al., 2012 ), systemic diseases (Addison's disease, melanoma), and medications known to cause hyperpigmentation (e.g., hydroxychloroquine) (Silva et al., 2022 ). DAGs Construction: DAGs depicted the relationship between oral hyperpigmentation and variables such as age, ethnicity, occupation, and number of pregnancies, as shown in Fig. 1 . DAG 1 shows the effect of ethnicity, age and number of pregnancies on oral hyperpigmentation, whereas DAG 2 shows the effect of occupation, age and ethnicity on oral hyperpigmentation. Mediating variables like occupation in DAG 2 link the primary exposure variable to the outcome variable. Demographic variables were treated as confounding variables, which were controlled for in ML models to prevent bias. Data Collection: A standardized data collection form was designed to record both demographic details and clinical indicators of oral hyperpigmentation. Prior to the participants fulfilling the questionnaire, their written and informed consent was obtained. The factors recorded were age, ethnicity, occupation, number of pregnancies, presence or absence of oral hyperpigmentation, and specific hyperpigmentation across various oral sites like labial mucosa, soft palate, buccal mucosa, hard palate, floor of the mouth, and gingiva. The Dummett Oral Hyperpigmentation Index (DOPI) was used to detect the presence and intensity of hyperpigmentation. The scoring pattern of DOPI (Muruppel et al., 2020 ) is as follows: No hyperpigmentation; indicated pink tissue, signifying the absence of clinical hyperpigmentation. Mild hyperpigmentation indicated light brown discoloration. Moderate hyperpigmentation indicated medium brown discoloration. Heavy hyperpigmentation indicated deep brown or blue-black discoloration. Trained dental professionals with proper lighting conditions conducted clinical examinations to ensure accuracy and reduce inter-examiner variability. The inter-rater agreement was recorded as 90%. The hyperpigmentation scores were recorded on-site during the assessment. At least three of the five researchers evaluated each participant, and if there was a difference of opinion, the results were scrutinised to reach a unanimous decision. The final decision regarding unsettled differences was to be made by a fourth, senior- most examiner. The variables of age and number of pregnancies were recorded as continuous variables, whereas ethnicity (balochi, kashmiri, pathan, punjabi, and sindhi) and occupation (student, housewife, and job) were taken as categorical variables, and oral hyperpigmentation as binary variables (present, absent). Oral hyperpigmentation of the gingiva, buccal mucosa, labial mucosa, hard palate, soft palate, and floor of the mouth were all assessed as ordinal variables (no hyperpigmentation, mild hyperpigmentation, moderate hyperpigmentation, and severe hyperpigmentation). To ensure data integrity and precise statistical analysis, both self-reported and clinically observed data were systematically collected. Statistical Analysis: Using Jamovi, Chi-square tests and Independent T-tests were applied to analyse the variations between categorical and continuous data, respectively (Kardys et al., 2013 ; Sundar & H., 2024). The results were shown with corresponding 95% confidence intervals where appropriate, and statistical significance was evaluated at p < 0.05. Three ML algorithms were employed: Gradient Boosting Machine (GBM), Logistic Regression (LR), and Random Forest (RF) to model the binary outcome variable—presence or absence of oral hyperpigmentation. These models were utilized because they have proven abilities to interpret high-dimensional data and detect complex associations between variables. Preprocessing methods, such as one-hot encoding of categorical data and normalization of continuous variables, were done for the implementation model. A training/test ratio of 80:20 was used on the dataset. GridSearch cross-validation (CV) was used for hyperparameter tuning. The precision, recall, accuracy, the F1 score, and the extent under the ROC curve (AUC-ROC) were among the performance criteria for evaluation. Correlation analyses were done after stratification of the data to analyze the link between continuous variables. Correlation matrices were generated and shown by heatmaps. Figure 2 shows heatmaps in cool-warm colour schemes to highlight positive and negative correlations. To ensure data integrity, missing values were handled using appropriate imputation techniques. If more than 10% of the participants’ data were missing, that entry was excluded from analysis to prevent bias. Statistical Software Used: The analyses utilized Python (version 3.12) libraries, such as Scikit-Learn , like Pandas for analyzing large datasets, Seaborn for statistical data visualizations and correlations, and NumPy for numerical calculations (Wornow et al., 2023 ). Ethical Approval: The university's Research Ethics Committee provided ethical authorisation for the study. In order to minimise selection bias, participants who matched the inclusion criteria and provided written informed consent were included in the study. The datasets and code generated and/or analysed during the current study are available from the corresponding author on reasonable request. Results Descriptive statistics: According to the data shown in Table 1, 447 participants were selected for the study, with their numbers of pregnancies ranging between 0 and 5. Among them, 289 (64.7%) had no oral hyperpigmentation, while 158 (35.3%) had oral hyperpigmentation. The distribution of hyperpigmentation showed no statistically significant association with the number of pregnancies (p = 0.736), although a slight increase in hyperpigmentation was observed among those with multiple pregnancies. Labial mucosa hyperpigmentation demonstrated a statistically significant association with the number of pregnancies (p = 0.003). Most participants (388, 86.8%) had no labial hyperpigmentation, 56 had light hyperpigmentation (12.5%), and only 3 (0.7%) had moderate to severe hyperpigmentation. Additionally, there were significant differences in hard palate hyperpigmentation between pregnancy groups (p = 0.018). Only 7 people (1.6%) had insignificant hyperpigmentation on the hard palate, mostly in those with one to three pregnancies, while 440 people (98.4%) had no visible hyperpigmentation at all. Hyperpigmentation on the floor of the mouth was less frequent but still statistically significant (p-value= 0.044). Mild hyperpigmentation was reported in 3 participants (0.7%), all of whom had experienced 3 or more pregnancies. As shown in Table 1, the results showed a p-value of 0.736, which indicated an insignificant association between the oral hyperpigmentation and the number of pregnancies. Hyperpigmentation on the labial mucosa and hard palate, having p-values of 0.003 and 0.018, respectively, showed significance with the number of pregnancies. According to Figure 2, oral hyperpigmentation was more commonly observed among housewives and individuals with jobs compared to students. The higher hyperpigmentation levels in housewives and employed participants may be attributed to greater exposure to environmental or lifestyle-related factors. Correlation Analysis: The Pearson coefficients for the correlation (R) between the main factors affecting oral hyperpigmentation are shown in Table 2. Age and the number of pregnancies have been found to be slightly positively correlated (R = 0.60). There was a moderate correlation between gingival hyperpigmentation and buccal mucosa (R = 0.38) but a significant correlation between gingival hyperpigmentation and overall oral hyperpigmentation (R = 0.86). Figure 3 shows a correlation heatmap illustrating relationships between age, number of pregnancies, and oral hyperpigmentation . Values represent Pearson correlation coefficients, color-coded on a cool-warm scale, where red indicates stronger positive associations and blue indicates weaker or negative ones. ML Model Metrics : Three ML models were applied to predict oral hyperpigmentation according to age, occupation, ethnicity and pregnancy status. All three ML models demonstrated excellent predictive power. LR provided the best performance with an accuracy of 96.67%; RF and GBM came a close second (both 95.56%). For all three models, precision, recall and F1 scores were similarly high, with LR performing particularly well, as shown in Table 3. ROC-AUC Graphs: ROC-AUC scores indicated that LR had the strongest discriminatory ability with an AUC of 0.9601, compared to 0.9445 for RF and GBM. The ROC curves (Figure 4) show that all models performed well, though LR demonstrated marginally better performance at various thresholds. Discussion The current study reported an insignificant association between the total presence of oral hyperpigmentation and the number of pregnancies. However, localized hyperpigmentation at the hard palate, floor of the mouth, and labial mucosa showed significant correlation with pregnancy. These findings support the hormonal theories, especially those that link melanin synthesis to higher levels of progesterone and estrogen during pregnancy (Snell & Bischitz, 1960). Age and Ethnicity: The number of pregnancies was moderately correlated with age and ethnicity (Rezazadeh et al., 2014). These findings and the model calculations are in accordance with findings from other studies (Koppolu et al., 2024). For example, in South Africa, a hyperpigmentation prevalence of 42% among darker skin tone individuals is consistent with our findings that some ethnic groups in developing countries are more likely to have oral hyperpigmentation (Masilana et al., 2017). Environmental and Occupational Factors: Compared to students, housewives and working females showed higher levels of hyperpigmentation, which showed occupation as a possible contributing factor, maybe as a result of increased exposure to environmental pollutants or stressors associated with their lifestyles. Based on the literature, some occupational categories are particularly affected by environmental factors such as smoking (Moravej-Salehi et al., 2015) and sun exposure (Lim et al., 2025). Machine Learning and DAGs: All three ML models demonstrated excellent predictive power. However, the LR model with an accuracy rate of 96.67% was particularly effective, whereas GBM and RF with an AUC of 0.9601 slightly underperformed. The underperformance of RF and GBM may be due to overfitting, inadequate data complexity, or not having enough training data, which are factors that are consistent with previous studies on small-sample ML applications in health domains (Han et al., 2021). In contrast, LR benefitted from its interpretability and ability to generalize well on relatively smaller datasets (Nusinovici et al., 2020). Under-representation of females from developing countries in AI models: In comparison with global models based on datasets that mainly focus on Western populations, this study confronts the significant gap in representation of females from developing countries in AI-driven oral health research via a prediction tool that considered the genetic and socio-environmental characteristics. Studies from South Asian ethnicities, such as women from Bangladesh, Nepal, or India, demonstrated high hyperpigmentation variation associated with skin tone and reproductive status, signifying that it might be beneficial (Rijal et al., 2021; Ponnaiyan et al., 2014). To validate and adjust the model for wider South Asian use, comparative research needs to be encouraged (Rotbeh et al., 2022). Clinical and Public Health Implications and Recommendations: Real-time risk assessments can be obtained by integrating ML into electronic health records (EHRs), which can help direct interventions and referrals, especially for expectant mothers (Lipschuetz et al., 2024). Early diagnosis of pregnancy-related oral hyperpigmentation, such as on the labial mucosa and hard palate, may improve treatment compliance and minimise anxiety. While the LR method helps in identifying high-risk women and expedites education and referrals, culturally tailored explanations and community programs can improve involvement. Strengths and Limitations: The limitations of the study included its cross-sectional design and dependency on self-reported demographic data, which could result in reporting bias or recall error. Likewise, environmental factors, including smoking status as well as UV exposure, were not recorded, which would have led to the exclusion of significant hyperpigmentation contributors. The strengths of the study included the utilization of trained dentists for hyperpigmentation assessment, detailed information gathered using predefined clinical indicators, and an innovative blended utilization of ML and DAGs to further improve understanding and prediction effectiveness. In comparison to prior studies, this study resolves a number of significant limitations. We addressed the more generic age sample used in past research by focusing on women within the ages of 18 and 45, allowing us to evaluate pregnancy-related oral hyperpigmentation in a demographically relevant cohort. Overall, this study highlights important topics for further research while providing a more focused and methodologically rigorous examination of oral hyperpigmentation in women from developing countries. To improve reproducibility, the study applied explicit methods for analysis incorporating openly accessible software (Jamovi, Python-based ML libraries) and standard machine learning methods (normalization, encoding, GridSearch CV). Prospective research might look into open-access dissemination of data and software availability that would allow worldwide comparison and validation by other researchers. Conclusions This study shows how ML algorithms can potentially be applied for accurately assessing oral hyperpigmentation in women within the age range of 18 and 45 when reinforced by DAGs. The results emphasize the importance of targeted methods of diagnosis, primarily in areas with limited resources and highlight the complicated associations among tailored oral hyperpigmentation and pregnancy, ethnic background, age, and occupation. To improve generalizability, additional studies should include psychological and environmental factors, contrasts to comparable groups, and broader geographical aspects. Declarations Ethics approval and consent to participate: This study was conducted in accordance with the ethical principles of the Declaration of Helsinki (2013 revision). Ethical approval was obtained from the Institutional Review Board / Ethical Committee of Islamic International Dental College, Riphah International University, Islamabad, Pakistan (Ref. No. IIDC/IRC/2020/006/013). All participants were informed about the purpose, procedures, potential risks, and benefits of the study. Written informed consent was obtained from all participants prior to enrollment. Participation was voluntary, confidentiality and anonymity were ensured, and participants had the right to withdraw from the study at any stage without any consequences. Consent for publication: Not applicable Availability of data and material: The datasets and code generated and/or analysed during the current study are available from the corresponding author on reasonable request. Competing interests: The authors declare that they have no competing interests. Funding: This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors. Author’s contributions: All authors contributed substantially to the work. Hifza Noor and Arooba Malik contributed to the original draft, writing, methodology, data collection, conceptualization, references, review, and editing. Rameesa Aymen contributed in writing. Mahnoor Fatima contributed to interpretation of results. Faisal Shafiq and Momina Zahra contributed to the final review and editing of the manuscript. Acknowledgements: Javed Ashraf contributed to the statistical analysis. References Abati, S., Sandri, G. F., Finotello, L., & Polizzi, E. (2024). Differential Diagnosis of Pigmented Lesions in the Oral Mucosa: A Clinical Based Overview and Narrative Review. Cancers , 16 (13), 2487. https://doi.org/10.3390/cancers16132487 Agarwal, M. C., Agarwal, S., Madan, E., & Madan, R. (2015). GINGIVAL HYPERPIGMENTATION REVISITED . https://www.researchgate.net/publication/361585336 Ciçek, Y., & Ertaş, U. (2003). The normal and pathological hyperpigmentation of oral mucous membrane: a review. The Journal of Contemporary Dental Practice , 4 (3), 76–86. Pecci-Lloret, M. P., Linares-Pérez, C., Pecci-Lloret, M. R., Rodríguez-Lozano, F. J., & Oñate-Sánchez, R. E. (2024). Oral Manifestations in Pregnant Women: A Systematic Review. Journal of Clinical Medicine , 13 (3). https://doi.org/10.3390/jcm13030707 Rotbeh, A., Kazeminia, M., Kalantari, M., & Rajati, F. (2022). Global prevalence of oral hyperpigmentation and its related factors: a systematic review and meta-analysis. Journal of Stomatology, Oral and Maxillofacial Surgery , 123 (5), e411–e424. https://doi.org/10.1016/j.jormas.2022.01.009 Gulati, N., Dutt, P., Gupta, N., & Tyagi, P. (n.d.). GINGIVAL HYPERPIGMENTATION: REVISITED. In Journal of Advanced Medical and Dental Sciences Research |Vol (Vol. 4). Masilana, A., Khammissa, R. A. G., Lemmer, J., & Feller, L. (2017). Physiological oral melanin hyperpigmentation in a South African sample: A clinical study. Journal of Investigative and Clinical Dentistry , 8 (4). https://doi.org/10.1111/jicd.12258 Ponnaiyan, D., Jegadeesan, V., Perumal, G., & Anusha, A. (2014). Correlating skin color with gingival hyperpigmentation patterns in South Indians - a cross sectional study. Oral Health and Dental Management , 13 (1), 132–136. Feller, L., Masilana, A., Khammissa, R. A., Altini, M., Jadwat, Y., & Lemmer, J. (2014). Melanin: the biophysiology of oral melanocytes and physiological oral hyperpigmentation. Head & Face Medicine , 10 (1), 8. https://doi.org/10.1186/1746-160X-10-8 Soulafa Almazrooa, Amna Arooj, Sobia Hassan, & Amber Kiyani. (2025). JOHOE_Volume 14_Issue 1_Pages 1-5. Journal of Oral Health and Oral Epidemiology , 14 , 1–5. https://doi.org/10.34172/johoe.2403.1627 Shahzad, A., Kiyani, A., & Paiker, S. (2018). Prevalence of Oral Anomalies and Pathologies in the Pakistani Population – A Cross Sectional Study. Journal of The Pakistan Dental Association , 27 (1), 13–17. https://doi.org/10.25301/JPDA.271.13 Pecci-Lloret, M. P., Linares-Pérez, C., Pecci-Lloret, M. R., Rodríguez-Lozano, F. J., & Oñate-Sánchez, R. E. (2024). Oral Manifestations in Pregnant Women: A Systematic Review. Journal of Clinical Medicine , 13 (3). https://doi.org/10.3390/jcm13030707 Motosko, C. C., Bieber, A. K., Pomeranz, M. K., Stein, J. A., & Martires, K. J. (2017). Physiologic changes of pregnancy: A review of the literature. International Journal of Women’s Dermatology , 3 (4), 219–224. https://doi.org/10.1016/j.ijwd.2017.09.003 Sattar, U., Malik, U., Shireen, Z., Faiz Ahmad Khan, S., Health CenterQadirabad, R., & Ghazi Khan, D. (2017). Oral health challenges in pregnant women in Pakistan: A research analysis. International Journal of Advanced Biotechnology and Research (IJBR) , 8 , 1724–1730. http://www.bipublication.com Dubuc, A., Zitouni, A., Thomas, C., Kémoun, P., Cousty, S., Monsarrat, P., & Laurencin, S. (2022). Improvement of Mucosal Lesion Diagnosis with Machine Learning Based on Medical and Semiological Data: An Observational Study. Journal of Clinical Medicine , 11 (21), 6596. https://doi.org/10.3390/jcm11216596 Piccininni, M., Konigorski, S., Rohmann, J. L., & Kurth, T. (n.d.). Title: Directed Acyclic Graphs and causal thinking in clinical risk prediction modeling . Marko, J. G. O., Neagu, C. D., & Anand, P. B. (2025). Examining inclusivity: the use of AI and diverse populations in health and social care: a systematic review. BMC Medical Informatics and Decision Making , 25 (1), 57. https://doi.org/10.1186/s12911-025-02884-1 Nyariro, M., Emami, E., Caidor, P., & Abbasgholizadeh Rahimi, S. (2023). Integrating equity, diversity and inclusion throughout the lifecycle of AI within healthcare: a scoping review protocol. BMJ Open , 13 (9), e072069. https://doi.org/10.1136/bmjopen-2023-072069 Rezazadeh, F., Falsafi, N., Sarraf, Z., & Shahbazi, M. (2014). Oral Mucosal Disorders in Pregnant versus Non-Pregnant Women. Dentistry Journal , 2 (4), 134–141. https://doi.org/10.3390/dj2040134 Tadakamadla, J., Kumar, S., Nagori, A., Tibdewal, H., Duraiswamy, P., & Kulkarni, S. (2012). Effect of smoking on oral hyperpigmentation and its relationship with periodontal status. Dental Research Journal , 9 (Suppl 1), S112-4. Silva, P. U. J., Oliveira, M. B., Vieira, W., Cardoso, S. V., Blumenberg, C., Franco, A., Siqueira, W. L., & Paranhos, L. R. (2022). Oral hyperpigmentation as an adverse effect of chloroquine and hydroxychloroquine use. Medicine , 101 (11). https://doi.org/10.1097/MD.0000000000029044 Kardys, I., Hoeks, S., van Domburg, R., Lenzen, M., & Boersma, E. (2013). Tools and techniques - Statistics: Analysis of continuous data using the t-test and ANOVA. In EuroIntervention (Vol. 9, Issue 6, pp. 765–767). https://doi.org/10.4244/EIJV9I6A123 S., V., A.S, H., & S. Sundar, J. (2024). CHI-SQUARE TESTS: A QUICK GUIDE FOR HEALTH RESEARCHERS. International Journal of Advanced Research , 12 (10), 1214–1222. https://doi.org/10.21474/IJAR01/19746 Wornow, M., Gyang Ross, E., Callahan, A., & Shah, N. H. (2023). APLUS: A Python library for usefulness simulations of machine learning models in healthcare. Journal of Biomedical Informatics , 139 , 104319. https://doi.org/10.1016/j.jbi.2023.104319 Muruppel, A. M., Pai, B. S. J., Bhat, S., Parker, S., & Lynch, E. (2020). Laser-Assisted Dehyperpigmentation—An Introspection of the Science, Techniques, and Perceptions. Dentistry Journal , 8 (3), 88. https://doi.org/10.3390/dj8030088 Snell, R. S., & Bischitz, P. G. (1960). The Effect of Large Doses of Estrogen and Estrogen and Progesterone on Melanin Hyperpigmentation**From the Department of Anatomy, Medical School King’s College, Newcastle upon Tyne, University of Durham, England. Journal of Investigative Dermatology , 35 (2), 73–82. https://doi.org/10.1038/jid.1960.87 Rezazadeh, F., Falsafi, N., Sarraf, Z., & Shahbazi, M. (2014). Oral Mucosal Disorders in Pregnant versus Non-Pregnant Women. Dentistry Journal , 2 (4), 134-141. https://doi.org/10.3390/dj2040134 Koppolu, P., Almutairi, H., Yousef, S. al, Ansary, N., Noushad, M., Vishal, M. B., Swapna, L. A., Alsuwayyigh, N., Albalawi, M., Shrivastava, D., & Srivastava, K. C. (2024). Relationship of skin complexion with gingival tissue color and hyperhyperpigmentation. A multi-ethnic comparative study. BMC Oral Health , 24 (1), 451. https://doi.org/10.1186/s12903-024-04189-7 Moravej-Salehi, E., Moravej-Salehi, E., & Hajifattahi, F. (2015). Relationship of Gingival Hyperpigmentation with Passive Smoking in Women. Tanaffos , 14 (2), 107–114. Lim, H. W., Piquero-Casals, J., Schalka, S., Leone, G., Trullàs, C., Brown, A., Foyaca, M., Gilaberte, Y., Krutmann, J., & Passeron, T. (2025). Photoprotection in pregnancy: addressing safety concerns and optimizing skin health. Frontiers in Medicine , 12 . https://doi.org/10.3389/fmed.2025.1563369 Han, S., Williamson, B. D., & Fong, Y. (2021). Improving random forest predictions in small datasets from two-phase sampling designs. BMC Medical Informatics and Decision Making , 21 (1), 322. https://doi.org/10.1186/s12911-021-01688-3 Nusinovici, S., Tham, Y. C., Chak Yan, M. Y., Wei Ting, D. S., Li, J., Sabanayagam, C., Wong, T. Y., & Cheng, C.-Y. (2020). Logistic regression was as good as machine learning for predicting major chronic diseases. Journal of Clinical Epidemiology , 122 , 56–69. https://doi.org/10.1016/j.jclinepi.2020.03.002 Rijal, A. H., Dhami, B., Pandey, N., & Aryal, D. (2021). Prevalence of Gingival Hyperpigmentation and its Association with Gingival Biotype and Skin Colour. Journal of Nepalese Society of Periodontology and Oral Implantology , 5 (1), 19–25. https://doi.org/10.3126/jnspoi.v5i1.38178 Ponnaiyan, D., Jegadeesan, V., Perumal, G., & Anusha, A. (2014). Correlating skin color with gingival hyperpigmentation patterns in South Indians - a cross sectional study. Oral Health and Dental Management , 13 (1), 132–136. Lipschuetz, M., Guedalia, J., Cohen, S. M., Unger, R., Yagel, S., & Sompolinsky, Y. (2024). Machine Learning in Electronic Health Records: Identifying High-Risk Obstetric Patients Pre and During Labor . https://doi.org/10.3233/SHTI240096 Tables Table 1. Descriptive Statistics Number of Pregnancies 0 N (%) 1 N (%) 2 N (%) 3 N (%) 4 N (%) 5 N (%) Total N (%) P-value Oral Hyperpigmentation Absent 78 (27%) 54 (18.7%) 62 (21.5%) 58 (20.1%) 26 (9%) 11 (3.8%) 289 (100%) 0.736 Present 45 (28.5%) 21 (13.3%) 39 (24.7%) 34 (21.5%) 12 (7.6%) 7 (4.4%) 158 (100%) Total 123 (27.5%) 75 (16.8%) 101 (22.6%) 92 (20.6%) 38 (8.5%) 18 (4.0%) 447 (100%) Labial Mucosa No Hyperpigmentation 115 (29.6%) 68 (17.5%) 85 (21.9%) 78 (20.1%) 28 (7.2%) 14 (3.6%) 388 (100%) 0.003 Mild Hyperpigmentation 8 (14.3%) 7 (12.5%) 16 (28.6%) 13 (23.2%) 9 (16.1%) 3 (5.4%) 56 (100%) Moderate Hyperpigmentation 0 (0%) 0 (0%) 0 (0%) 1 (50%) 0 (0%) 1 (50%) 2 (100%) Heavy Hyperpigmentation 0 (0%) 0 (0%) 0 (0%) 0 (0%) 1 (100%) 0 (0%) 1 (100%) Total 123 (27.5%) 75 (16.8%) 101 (22.6%) 92 (20.6%) 38 (8.5%) 18 (4.0%) 447 (100%) Hard Palate No Hyperpigmentation 123 (28%) 73 (16.6%) 101 (23.0%) 88 (20.0%) 38 (8.6%) 17 (3.9%) 440 (100%) 0.018 Mild Hyperpigmentation 0 (0%) 2 (28.6%) 0 (0%) 4 (57.1%) 0 (0%) 1 (14.3%) 7 (100%) Total 123 (27.5%) 75 (16.8%) 101 (22.6%) 92 (20.6%) 38 (8.5%) 18 (4.0%) 447 (100%) Floor of The Mouth No Hyperpigmentation 123 (27.7%) 75 (16.9%) 101 (22.7%) 90 (20.3%) 38 (8.6%) 17 (3.8%) 444 (100%) 0.044 Mild Hyperpigmentation 0 (0%) 0 (0%) 0 (0%) 2 (66.7%) 0 (0%) 1 (33.3%) 3 (100%) Total 123 (27.5%) 75 (16.8%) 101 (22.6%) 92 (20.6%) 38 (8.5%) 18 (4.0%) 447 (100%) Table 2. Variable Associations Analysis Correlation of Key Variables Correlation Coefficient (R) Age and Number of Pregnancies 0.6 Total Hyperpigmentation and Gingiva 0.86 Total Hyperpigmentation and Labial Mucosa 0.47 Total Hyperpigmentation and Buccal Mucosa 0.38 Table 3. Machine Learning Model Metrics Model Accuracy Precision Recall F1 Score AUC-ROC Random Forest 0.9556 0.9667 0.9063 0.9355 0.9445 Gradient Boosting Machine 0.9556 0.9667 0.9063 0.9355 0.9445 Logistic Regression 0.9667 0.9677 0.9375 0.9524 0.9601 Additional Declarations No competing interests reported. Cite Share Download PDF Status: Under Review Version 1 posted Reviewers invited by journal 13 Jan, 2026 Editor invited by journal 19 Dec, 2025 Editor assigned by journal 18 Dec, 2025 Submission checks completed at journal 18 Dec, 2025 First submitted to journal 15 Dec, 2025 You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-8368647","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":573977458,"identity":"40c94278-d648-420b-938e-dd92df4e9025","order_by":0,"name":"Hifza Noor","email":"","orcid":"","institution":"Riphah International University","correspondingAuthor":false,"prefix":"","firstName":"Hifza","middleName":"","lastName":"Noor","suffix":""},{"id":573977459,"identity":"666bbf87-7679-4ae0-a692-d62b8b69418c","order_by":1,"name":"Arooba Malik","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAABCElEQVRIiWNgGAWjYJACZhDBx8zG+CDBQEIOxDnwgBgtbMxszAYfKmyMwVoSiNLCwMYmOeNMWmIDiIdPC//swwcfF9Rsk2djZ0uQ5m07nD4/7PBDoC12croN2LVInEtLNp5x7LZhGzPbAWOgltyNt9MMgFqSjc0O4LDmDI+ZNA/bbcY2ZvaGZLCW2QkgLQcSt+HQIg/W8u+2PUjLYZDDDGenf8CrxQCkhbftdiLQYQcbgd5PkJfOwW+L4Rm2ZOOZfbeTgVqSGYCBbLhBOqfgQIIBbr/InWEGhti327b9/MfMfwCjUl5+dvrmDx8q7ORweh/TqWCVBsQqBwH5BlJUj4JRMApGwUgAAI3aYGmh4fEYAAAAAElFTkSuQmCC","orcid":"","institution":"Riphah International University","correspondingAuthor":true,"prefix":"","firstName":"Arooba","middleName":"","lastName":"Malik","suffix":""},{"id":573977460,"identity":"56a24fa0-15a0-417f-833f-4e823ae8974b","order_by":2,"name":"Momina Zahra","email":"","orcid":"","institution":"Riphah International University","correspondingAuthor":false,"prefix":"","firstName":"Momina","middleName":"","lastName":"Zahra","suffix":""},{"id":573977461,"identity":"41377259-7f0b-45da-8c40-960575aaee2d","order_by":3,"name":"Mahnoor Fatima","email":"","orcid":"","institution":"Pakistan Institute of Medical Sciences","correspondingAuthor":false,"prefix":"","firstName":"Mahnoor","middleName":"","lastName":"Fatima","suffix":""},{"id":573977462,"identity":"1b589810-29aa-4967-92f3-e63d5f9181ad","order_by":4,"name":"Rameesa Aymen","email":"","orcid":"","institution":"Riphah International University","correspondingAuthor":false,"prefix":"","firstName":"Rameesa","middleName":"","lastName":"Aymen","suffix":""},{"id":573977463,"identity":"33aff277-9355-49c6-ada8-ca603da07685","order_by":5,"name":"Faisal Shafiq Malik","email":"","orcid":"","institution":"Riphah International University","correspondingAuthor":false,"prefix":"","firstName":"Faisal","middleName":"Shafiq","lastName":"Malik","suffix":""}],"badges":[],"createdAt":"2025-12-15 16:53:20","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-8368647/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-8368647/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":100595436,"identity":"09d4c77c-9b19-47af-b9e2-ad44ca469cfe","added_by":"auto","created_at":"2026-01-19 13:48:28","extension":"docx","order_by":0,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":372563,"visible":true,"origin":"","legend":"","description":"","filename":"ORALPIGMENTATIONOrgarticleRevised.docx","url":"https://assets-eu.researchsquare.com/files/rs-8368647/v1/25b3bb249ce74d979c8314ee.docx"},{"id":100594725,"identity":"f10370f6-2c0c-46e9-99dd-064aaa54aa72","added_by":"auto","created_at":"2026-01-19 13:44:24","extension":"json","order_by":1,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":7861,"visible":true,"origin":"","legend":"","description":"","filename":"0b3a672f406d40bfbc58397757f23b3c.json","url":"https://assets-eu.researchsquare.com/files/rs-8368647/v1/0285b52b22c7abb6ba651261.json"},{"id":100594908,"identity":"219e45cf-70bd-4646-826d-fb7a3ffe3608","added_by":"auto","created_at":"2026-01-19 13:46:22","extension":"xml","order_by":2,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":111123,"visible":true,"origin":"","legend":"","description":"","filename":"0b3a672f406d40bfbc58397757f23b3c1enriched.xml","url":"https://assets-eu.researchsquare.com/files/rs-8368647/v1/534e88ec91565f0061ea1b57.xml"},{"id":100565369,"identity":"43cbf00d-4f72-4cd9-8c8a-b653224af892","added_by":"auto","created_at":"2026-01-19 09:04:48","extension":"png","order_by":7,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":15284,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage1.png","url":"https://assets-eu.researchsquare.com/files/rs-8368647/v1/4ca9fab7e53e88ad50b73f98.png"},{"id":100565368,"identity":"c4bcd413-0ac1-4078-bcc7-bf07f10cd543","added_by":"auto","created_at":"2026-01-19 09:04:48","extension":"png","order_by":8,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":13344,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage2.png","url":"https://assets-eu.researchsquare.com/files/rs-8368647/v1/b76176944b8db38f0aed8542.png"},{"id":100594969,"identity":"ef02d6fa-349f-4a93-b866-361371744694","added_by":"auto","created_at":"2026-01-19 13:46:46","extension":"png","order_by":9,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":16524,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage3.png","url":"https://assets-eu.researchsquare.com/files/rs-8368647/v1/208162be588761e950967407.png"},{"id":100595262,"identity":"d40ae9e7-2f36-4aaf-bdbd-2d86d664469a","added_by":"auto","created_at":"2026-01-19 13:48:04","extension":"png","order_by":10,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":34555,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage4.png","url":"https://assets-eu.researchsquare.com/files/rs-8368647/v1/50ee01b82de9d0fcae92f2a1.png"},{"id":100565376,"identity":"7d125b17-3f2d-426c-9013-d25d1f035964","added_by":"auto","created_at":"2026-01-19 09:04:48","extension":"xml","order_by":11,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":110596,"visible":true,"origin":"","legend":"","description":"","filename":"0b3a672f406d40bfbc58397757f23b3c1structuring.xml","url":"https://assets-eu.researchsquare.com/files/rs-8368647/v1/ce1c421d3ad43cf50e03b2aa.xml"},{"id":100565375,"identity":"27ed2426-cf74-421e-b35b-15911a4b02b0","added_by":"auto","created_at":"2026-01-19 09:04:48","extension":"html","order_by":12,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":121024,"visible":true,"origin":"","legend":"","description":"","filename":"earlyproof.html","url":"https://assets-eu.researchsquare.com/files/rs-8368647/v1/6a991f705467f79e30359b52.html"},{"id":100595837,"identity":"06093f4f-321e-460d-97f4-1a70e94f76ce","added_by":"auto","created_at":"2026-01-19 13:49:30","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":76763,"visible":true,"origin":"","legend":"\u003cp\u003eDAG model showing assumed causal relationships influencing oral pigmentation in pregnancy.\u003c/p\u003e\n\u003cp\u003ea. DAG 1 represents the effect of ethnicity, age, and number of pregnancies on oral hyperpigmentation.\u003c/p\u003e\n\u003cp\u003eb. DAG 2 represents the effect of ethnicity, age, and occupation on oral hyperpigmentation.\u003c/p\u003e","description":"","filename":"floatimage1.png","url":"https://assets-eu.researchsquare.com/files/rs-8368647/v1/17c237c5abcdf34cd597f92f.png"},{"id":100595718,"identity":"45e16ed6-cdec-4cf1-aabb-947c180fcf0c","added_by":"auto","created_at":"2026-01-19 13:49:15","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":62810,"visible":true,"origin":"","legend":"\u003cp\u003eBar graph showing pigmentation status between three study groups based on occupation.\u003c/p\u003e","description":"","filename":"floatimage2.png","url":"https://assets-eu.researchsquare.com/files/rs-8368647/v1/ee8988e443daff89cd4d0e3d.png"},{"id":100565365,"identity":"ea5fc3dd-c7b3-4fa0-872f-c7fb54a73bcf","added_by":"auto","created_at":"2026-01-19 09:04:48","extension":"png","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":68736,"visible":true,"origin":"","legend":"\u003cp\u003eCorrelation Heatmap between age, number of pregnancies, and hyperpigmentation on labial and buccal mucosa.\u003c/p\u003e","description":"","filename":"floatimage3.png","url":"https://assets-eu.researchsquare.com/files/rs-8368647/v1/3432f431b8e4816af90f3acf.png"},{"id":100594904,"identity":"e056e011-0a29-496e-9aed-7281c063c925","added_by":"auto","created_at":"2026-01-19 13:46:21","extension":"png","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":103911,"visible":true,"origin":"","legend":"\u003cp\u003eROC curve showing model performance for pigmentation prediction.\u003c/p\u003e","description":"","filename":"floatimage4.png","url":"https://assets-eu.researchsquare.com/files/rs-8368647/v1/d804ce0ea9b88e00cbd32317.png"},{"id":100796084,"identity":"63f8818b-47ef-4b99-969a-95b2918521f0","added_by":"auto","created_at":"2026-01-21 13:39:11","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":1134142,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-8368647/v1/21cb1c66-73df-4d4b-b5fa-8fb0a5353606.pdf"}],"financialInterests":"No competing interests reported.","formattedTitle":"Predicting the Association Between Oral Hyperpigmentation and Pregnancy in Women Aged 18-45 Using Machine Learning and Directed Acyclic Graphs: A Cross-Sectional Study","fulltext":[{"header":"Introduction","content":"\u003cp\u003eOral hyperpigmentation is a physiological phenomenon due to the excess deposition of melanin and other pigments on the oral mucosal surfaces (Abati, 2024). It is mostly undetectable and presents as discoloration of the gingiva, buccal mucosa, or other intraoral sites (Agarwal et al., \u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2015\u003c/span\u003e). This hyperpigmentation, due to its resemblance with pathological lesions and aesthetic concerns, especially for pregnant women facing hormonal changes, serves as a diagnostic challenge (Abati, 2024; Cicek \u0026amp; Ertas, 2003; Pecci-Lloret et al., \u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e2024\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eThe prevalence of oral hyperpigmentation changes considerably amongst the population, occurring in approximately 20.8% of people worldwide (95% CI: 17.1\u0026ndash;25.0%) (Rotbeh et al., \u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e2022\u003c/span\u003e). Besides genetic, environmental, social, and cultural factors influencing oral hyperpigmentation, this condition is influenced by physiological factors like age, occupation, ethnicity, and the number of pregnancies (Gulati et al., n.d.). For example, people with darker skin tones, such as those of Africans and South Africans, tend to have more oral hyperpigmentation than people with lighter skin tones (Masilana et al., \u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e2017\u003c/span\u003e; Ponnaiyan et al., 2024; Feller et al., \u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e2014\u003c/span\u003e). A study carried out in South Africa, with a prevalence of 42.3% and 430 subjects from various ethnic backgrounds, reported a greater detection rate for orofacial hyperpigmentation (Masilana et al., \u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e2017\u003c/span\u003e). Similarly, oral hyperpigmentation patterns in developing countries are strongly affected by genes due to the diverse population in terms of race and ethnicity (Almazrooa et al., 2025).\u003c/p\u003e \u003cp\u003eEnvironmental exposures, including smoking, sun exposure, and air quality, play an important part in hyperpigmentation. It is known that ultraviolet (UV) radiation affects the synthesis of melanin (Abati, 2024). In areas of developing countries that get a lot of UV radiation, people typically experience greater levels of skin and oral hyperpigmentation (Shahzad et al., \u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e2018\u003c/span\u003e). Similarly, brown to black discoloration of the oral mucosa in smoker\u0026rsquo;s melanosis, particularly in the buccal, hard palate, and gingiva, has a link with smoking (Abati, 2024). A precise diagnosis and patient reassurance, specifically for populations at risk, require an understanding of its patterns and causes.\u003c/p\u003e \u003cp\u003eHormonal changes during pregnancy, notably high levels of progesterone and estrogen, promote melanin to be produced in excess, which results in hyperpigmentation in different parts of the body (Pecci-Lloret et al., \u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e2024\u003c/span\u003e). The areas involved included the face (\u003cem\u003emelasma\u003c/em\u003e) and abdomen (\u003cem\u003eLinea nigra\u003c/em\u003e) (Motosko et al., \u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e2017\u003c/span\u003e). However, studies on the association of pregnancy and oral hyperpigmentation are extremely limited in developing countries, with no integrated models developed to predict these changes. This represents a major gap in women's oral health research (Sattar et al., \u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e2017\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eWhile studying oral hyperpigmentation, it was observed that the traditional analysis methods frequently lack the ability to detect the complex nature of these interlinked environmental, physiological, and demographic factors. Oral hyperpigmentation research still underutilizes advanced techniques like Directed Acyclic Graphs (DAGs) and Machine Learning (ML), in spite of their increasing relevance in the healthcare domain (Dubuc et al., \u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e2022\u003c/span\u003e). While DAGs enable the clear visualization of assumed causal relationships and enable researchers to differentiate between confounders, mediators, and colliders and avoid statistical bias, ML algorithms are capable of detecting non-linear associations and generating predictive insights. Together, ML and DAGs provide a more efficient and open method for simulating complex conditions (Piccininni et al., n.d.).\u003c/p\u003e \u003cp\u003eWomen from developing countries are not equally represented among AI-driven oral health datasets, which limits the development of appropriate and efficient healthcare solutions for a sizable portion of the population (Marko et al., \u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e2025\u003c/span\u003e). This study aims to bridge that gap by investigating and predicting oral hyperpigmentation in pregnant and non-pregnant women aged 18 to 45, using DAG-supported ML models. This study contributes to more inclusive, data-driven maternal healthcare by developing predictive models tailored to the ethnically diverse population, aiming to improve patient counselling, improve early detection, and guide preventive measures in oral healthcare for females by addressing the relationship between age, occupation, ethnicity, and pregnancy status (Nyariro et al., \u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e2023\u003c/span\u003e).\u003c/p\u003e\n\u003ch3\u003eObjectives:\u003c/h3\u003e\n\u003cp\u003eThe objective of the study is to predict the association between oral hyperpigmentation and pregnancy utilizing multiple ML algorithms. This association will be explored using directed acyclic graphs to support theory-driven variable selection and improve model interpretability\u003c/p\u003e \u003cdiv id=\"Sec3\" class=\"Section2\"\u003e \u003ch2\u003ePurpose:\u003c/h2\u003e \u003cp\u003eWe hypothesized that ML models, supported by DAGs, can more meaningfully predict oral hyperpigmentation and identify key contributing factors such as age, occupation, ethnicity, and number of pregnancies.\u003c/p\u003e \u003c/div\u003e"},{"header":"Methods","content":"\u003cdiv id=\"Sec5\" class=\"Section2\"\u003e \u003ch2\u003e\u003cem\u003eSample Description and Study Framework\u003c/em\u003e:\u003c/h2\u003e \u003cp\u003eIn this cross-sectional study, 447 women within the ages of 18 and 45 were included from a public and private hospital through convenience sampling. The study was conducted from September 2021 to September 2024.\u003c/p\u003e \u003c/div\u003e\n\u003ch3\u003eSample Size:\u003c/h3\u003e\n\u003cp\u003eBased on a probable population of 20,000, a 50% response distribution, a 96% confidence level, and a 5% margin of error, a sample size of 414 was calculated using the Raosoft sample size calculator (\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttp://www.raosoft.com/samplesize.html\u003c/span\u003e\u003cspan address=\"http://www.raosoft.com/samplesize.html\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e). To further reduce the margin of error to 4.8%, we took a sample size of 447 students.\u003c/p\u003e\n\u003ch3\u003eInclusion criteria:\u003c/h3\u003e\n\u003cp\u003ePregnant and non-pregnant participants aged 18 to 45 years who were not suffering from any comorbid conditions were included in the study (Rezazadeh et al., \u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e2014\u003c/span\u003e).\u003c/p\u003e \u003cdiv id=\"Sec8\" class=\"Section2\"\u003e \u003ch2\u003eExclusion criteria:\u003c/h2\u003e \u003cp\u003eThe exclusion criteria of the current study included factors such as smoking history (Tadakamadla et al., \u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e2012\u003c/span\u003e), systemic diseases (Addison's disease, melanoma), and medications known to cause hyperpigmentation (e.g., hydroxychloroquine) (Silva et al., \u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e2022\u003c/span\u003e).\u003c/p\u003e \u003c/div\u003e\n\u003ch3\u003eDAGs Construction:\u003c/h3\u003e\n\u003cp\u003eDAGs depicted the relationship between oral hyperpigmentation and variables such as age, ethnicity, occupation, and number of pregnancies, as shown in Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003e. DAG 1 shows the effect of ethnicity, age and number of pregnancies on oral hyperpigmentation, whereas DAG 2 shows the effect of occupation, age and ethnicity on oral hyperpigmentation. Mediating variables like occupation in DAG 2 link the primary exposure variable to the outcome variable. Demographic variables were treated as confounding variables, which were controlled for in ML models to prevent bias.\u003c/p\u003e\n\u003ch3\u003eData Collection:\u003c/h3\u003e\n\u003cp\u003eA standardized data collection form was designed to record both demographic details and clinical indicators of oral hyperpigmentation. Prior to the participants fulfilling the questionnaire, their written and informed consent was obtained. The factors recorded were age, ethnicity, occupation, number of pregnancies, presence or absence of oral hyperpigmentation, and specific hyperpigmentation across various oral sites like labial mucosa, soft palate, buccal mucosa, hard palate, floor of the mouth, and gingiva.\u003c/p\u003e \u003cp\u003eThe Dummett Oral Hyperpigmentation Index (DOPI) was used to detect the presence and intensity of hyperpigmentation. The scoring pattern of DOPI (Muruppel et al., \u003cspan citationid=\"CR25\" class=\"CitationRef\"\u003e2020\u003c/span\u003e) is as follows:\u003c/p\u003e \u003cp\u003e \u003cul\u003e \u003cli\u003e \u003cp\u003eNo hyperpigmentation; indicated pink tissue, signifying the absence of clinical hyperpigmentation.\u003c/p\u003e \u003c/li\u003e \u003cli\u003e \u003cp\u003eMild hyperpigmentation indicated light brown discoloration.\u003c/p\u003e \u003c/li\u003e \u003cli\u003e \u003cp\u003eModerate hyperpigmentation indicated medium brown discoloration.\u003c/p\u003e \u003c/li\u003e \u003cli\u003e \u003cp\u003eHeavy hyperpigmentation indicated deep brown or blue-black discoloration.\u003c/p\u003e \u003c/li\u003e \u003c/ul\u003e \u003c/p\u003e \u003cp\u003eTrained dental professionals with proper lighting conditions conducted clinical examinations to ensure accuracy and reduce inter-examiner variability. The inter-rater agreement was recorded as 90%. The hyperpigmentation scores were recorded on-site during the assessment. At least three of the five researchers evaluated each participant, and if there was a difference of opinion, the results were scrutinised to reach a unanimous decision. The final decision regarding unsettled differences was to be made by a fourth, senior- most examiner.\u003c/p\u003e \u003cp\u003eThe variables of age and number of pregnancies were recorded as continuous variables, whereas ethnicity (balochi, kashmiri, pathan, punjabi, and sindhi) and occupation (student, housewife, and job) were taken as categorical variables, and oral hyperpigmentation as binary variables (present, absent). Oral hyperpigmentation of the gingiva, buccal mucosa, labial mucosa, hard palate, soft palate, and floor of the mouth were all assessed as ordinal variables (no hyperpigmentation, mild hyperpigmentation, moderate hyperpigmentation, and severe hyperpigmentation). To ensure data integrity and precise statistical analysis, both self-reported and clinically observed data were systematically collected.\u003c/p\u003e \u003cdiv id=\"Sec11\" class=\"Section2\"\u003e \u003ch2\u003eStatistical Analysis:\u003c/h2\u003e \u003cp\u003eUsing Jamovi, Chi-square tests and Independent T-tests were applied to analyse the variations between categorical and continuous data, respectively (Kardys et al., \u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003e2013\u003c/span\u003e; Sundar \u0026amp; H., 2024). The results were shown with corresponding 95% confidence intervals where appropriate, and statistical significance was evaluated at p\u0026thinsp;\u0026lt;\u0026thinsp;0.05. Three ML algorithms were employed: Gradient Boosting Machine (GBM), Logistic Regression (LR), and Random Forest (RF) to model the binary outcome variable\u0026mdash;presence or absence of oral hyperpigmentation. These models were utilized because they have proven abilities to interpret high-dimensional data and detect complex associations between variables. Preprocessing methods, such as one-hot encoding of categorical data and normalization of continuous variables, were done for the implementation model. A training/test ratio of 80:20 was used on the dataset. GridSearch cross-validation (CV) was used for hyperparameter tuning. The precision, recall, accuracy, the F1 score, and the extent under the ROC curve (AUC-ROC) were among the performance criteria for evaluation.\u003c/p\u003e \u003cp\u003eCorrelation analyses were done after stratification of the data to analyze the link between continuous variables. Correlation matrices were generated and shown by heatmaps. Figure\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e2\u003c/span\u003e shows heatmaps in cool-warm colour schemes to highlight positive and negative correlations.\u003c/p\u003e \u003cp\u003eTo ensure data integrity, missing values were handled using appropriate imputation techniques. If more than 10% of the participants\u0026rsquo; data were missing, that entry was excluded from analysis to prevent bias.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec12\" class=\"Section2\"\u003e \u003ch2\u003eStatistical Software Used:\u003c/h2\u003e \u003cp\u003eThe analyses utilized Python (version 3.12) libraries, such as \u003cem\u003eScikit-Learn\u003c/em\u003e, like \u003cem\u003ePandas\u003c/em\u003e for analyzing large datasets, \u003cem\u003eSeaborn\u003c/em\u003e for statistical data visualizations and correlations, and \u003cem\u003eNumPy\u003c/em\u003e for numerical calculations (Wornow et al., \u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e2023\u003c/span\u003e).\u003c/p\u003e \u003c/div\u003e\n\u003cp\u003e\u003cem\u003eEthical Approval:\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003eThe university\u0026apos;s Research Ethics Committee provided ethical authorisation for the study. In order to minimise selection bias, participants who matched the inclusion criteria and provided written informed consent were included in the study. The datasets and code generated and/or analysed during the current study are available from the corresponding author on reasonable request.\u003c/p\u003e"},{"header":"Results","content":"\u003cp\u003e\u003cem\u003eDescriptive statistics:\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003eAccording to the data shown in Table 1, 447 participants were selected for the study, with their numbers of pregnancies ranging between 0 and 5. Among them, 289 (64.7%) had no oral hyperpigmentation, while 158 (35.3%) had oral hyperpigmentation. The distribution of hyperpigmentation showed no statistically significant association with the number of pregnancies (p = 0.736), although a slight increase in hyperpigmentation was observed among those with multiple pregnancies.\u003c/p\u003e\n\u003cp\u003eLabial mucosa hyperpigmentation demonstrated a statistically significant association with the number of pregnancies (p = 0.003). Most participants (388, 86.8%) had no labial hyperpigmentation, 56 had light hyperpigmentation (12.5%), and only 3 (0.7%) had moderate to severe hyperpigmentation.\u003c/p\u003e\n\u003cp\u003eAdditionally, there were significant differences in hard palate hyperpigmentation between pregnancy groups (p = 0.018). Only 7 people (1.6%) had insignificant hyperpigmentation on the hard palate, mostly in those with one to three pregnancies, while 440 people (98.4%) had no visible hyperpigmentation at all.\u003c/p\u003e\n\u003cp\u003eHyperpigmentation on the floor of the mouth was less frequent but still statistically significant (p-value= 0.044). Mild hyperpigmentation was reported in 3 participants (0.7%), all of whom had experienced 3 or more pregnancies.\u003c/p\u003e\n\u003cp\u003eAs shown in Table 1, the results showed a p-value of 0.736, which indicated an insignificant association between the oral hyperpigmentation and the number of pregnancies. Hyperpigmentation on the labial mucosa and hard palate, having p-values of 0.003 and 0.018, respectively, showed significance with the number of pregnancies.\u003c/p\u003e\n\u003cp\u003eAccording to Figure 2, oral hyperpigmentation was more commonly observed among housewives and individuals with jobs compared to students. The higher hyperpigmentation levels in housewives and employed participants may be attributed to greater exposure to environmental or lifestyle-related factors.\u003c/p\u003e\n\u003cp\u003e\u003cem\u003eCorrelation Analysis:\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003eThe Pearson coefficients for the correlation (R) between the main factors affecting oral hyperpigmentation are shown in Table 2. Age and the number of pregnancies have been found to be slightly positively correlated (R = 0.60). There was a moderate correlation between gingival hyperpigmentation and buccal mucosa (R = 0.38) but a significant correlation between gingival hyperpigmentation and overall oral hyperpigmentation (R = 0.86). Figure 3 shows a correlation heatmap illustrating relationships between age, number of pregnancies, and oral hyperpigmentation\u003cem\u003e.\u003c/em\u003e Values represent Pearson correlation coefficients, color-coded on a cool-warm scale, where red indicates stronger positive associations and blue indicates weaker or negative ones.\u003c/p\u003e\n\u003cp\u003e\u003cem\u003eML Model Metrics\u003c/em\u003e:\u003c/p\u003e\n\u003cp\u003eThree ML models were applied to predict oral hyperpigmentation according to age, occupation, ethnicity and pregnancy status. All three ML models demonstrated excellent predictive power. LR provided the best performance with an accuracy of 96.67%; RF and GBM came a close second (both 95.56%). For all three models, precision, recall and F1 scores were similarly high, with LR performing particularly well, as shown in Table 3.\u003c/p\u003e\n\u003cp\u003e\u003cem\u003eROC-AUC Graphs:\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003eROC-AUC scores indicated that LR had the strongest discriminatory ability with an AUC of 0.9601, compared to 0.9445 for RF and GBM. The ROC curves (Figure 4) show that all models performed well, though LR demonstrated marginally better performance at various thresholds.\u003c/p\u003e"},{"header":"Discussion","content":"\u003cp\u003eThe current study reported an insignificant association between the total presence of oral hyperpigmentation and the number of pregnancies. However, localized hyperpigmentation at the hard palate, floor of the mouth, and labial mucosa showed significant correlation with pregnancy. These findings support the hormonal theories, especially those that link melanin synthesis to higher levels of progesterone and estrogen during pregnancy (Snell \u0026amp; Bischitz, 1960).\u003c/p\u003e\n\u003cp\u003e\u003cem\u003eAge and Ethnicity:\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003eThe number of pregnancies was moderately correlated with age and ethnicity (Rezazadeh et al., 2014).\u0026nbsp;These findings and the model calculations are in accordance with findings from other studies (Koppolu et al., 2024). For example, in South Africa, a hyperpigmentation prevalence of 42% among darker skin tone individuals is consistent with our findings that some ethnic groups in developing countries are more likely to have oral hyperpigmentation (Masilana et al., 2017).\u003c/p\u003e\n\u003cp\u003e\u003cem\u003eEnvironmental and Occupational Factors:\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003eCompared to students, housewives and working females showed higher levels of hyperpigmentation, which showed occupation as a possible contributing factor, maybe as a result of increased exposure to environmental pollutants or stressors associated with their lifestyles. Based on the literature, some occupational categories are particularly affected by environmental factors such as smoking (Moravej-Salehi et al., 2015) and sun exposure (Lim et al., 2025).\u003c/p\u003e\n\u003cp\u003e\u003cem\u003eMachine Learning and DAGs:\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003eAll three ML models demonstrated excellent predictive power. However, the LR model with an accuracy rate of 96.67% was particularly effective, whereas GBM and RF with an AUC of 0.9601 slightly underperformed. The underperformance of RF and GBM may be due to overfitting, inadequate data complexity, or not having enough training data, which are factors that are consistent with previous studies on small-sample ML applications in health domains (Han et al., 2021). In contrast, LR benefitted from its interpretability and ability to generalize well on relatively smaller datasets (Nusinovici et al., 2020).\u003c/p\u003e\n\u003cp\u003e\u003cem\u003eUnder-representation of females from developing countries in AI models:\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003eIn comparison with global models based on datasets that mainly focus on Western populations, this study confronts the significant gap in representation of females from developing countries in AI-driven oral health research via a prediction tool that considered the genetic and socio-environmental characteristics. Studies from South Asian ethnicities, such as women from Bangladesh, Nepal, or India, demonstrated high hyperpigmentation variation associated with skin tone and reproductive status, signifying that it might be beneficial (Rijal et al., 2021; Ponnaiyan et al., 2014). To validate and adjust the model for wider South Asian use, comparative research needs to be encouraged (Rotbeh et al., 2022).\u003c/p\u003e\n\u003cp\u003e\u003cem\u003eClinical and Public Health Implications and Recommendations:\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003eReal-time risk assessments can be obtained by integrating ML into electronic health records (EHRs), which can help direct interventions and referrals, especially for expectant mothers (Lipschuetz et al., 2024). Early diagnosis of pregnancy-related oral hyperpigmentation, such as on the labial mucosa and hard palate, may improve treatment compliance and minimise anxiety. While the LR method helps in identifying high-risk women and expedites education and referrals, culturally tailored explanations and community programs can improve involvement.\u003c/p\u003e\n\u003cp\u003e\u003cem\u003eStrengths and Limitations:\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003eThe limitations of the study included its cross-sectional design and dependency on self-reported demographic data, which could result in reporting bias or recall error. Likewise, environmental factors, including smoking status as well as UV exposure, were not recorded, which would have led to the exclusion of significant hyperpigmentation contributors.\u003c/p\u003e\n\u003cp\u003eThe strengths of the study included the utilization of trained dentists for hyperpigmentation assessment, detailed information gathered using predefined clinical indicators, and an innovative blended utilization of ML and DAGs to further improve understanding and prediction effectiveness. In comparison to prior studies, this study resolves a number of significant limitations. We addressed the more generic age sample used in past research by focusing on women within the ages of 18 and 45, allowing us to evaluate pregnancy-related oral hyperpigmentation in a demographically relevant cohort. Overall, this study highlights important topics for further research while providing a more focused and methodologically rigorous examination of oral hyperpigmentation in women from developing countries. To improve reproducibility, the study applied explicit methods for analysis incorporating openly accessible software (Jamovi, Python-based ML libraries) and standard machine learning methods (normalization, encoding, GridSearch CV). Prospective research might look into open-access dissemination of data and software availability that would allow worldwide comparison and validation by other researchers.\u003c/p\u003e"},{"header":"Conclusions","content":"\u003cp\u003eThis study shows how ML algorithms can potentially be applied for accurately assessing oral hyperpigmentation in women within the age range of 18 and 45 when reinforced by DAGs. The results emphasize the importance of targeted methods of diagnosis, primarily in areas with limited resources and highlight the complicated associations among tailored oral hyperpigmentation and pregnancy, ethnic background, age, and occupation. To improve generalizability, additional studies should include psychological and environmental factors, contrasts to comparable groups, and broader geographical aspects.\u003c/p\u003e"},{"header":"Declarations","content":"\u003cp\u003e\u003cstrong\u003eEthics approval and consent to participate:\u003c/strong\u003e\u003cstrong\u003e\u003cbr\u003e\u003c/strong\u003eThis study was conducted in accordance with the ethical principles of the Declaration of Helsinki (2013 revision). Ethical approval was obtained from the Institutional Review Board / Ethical Committee of Islamic International Dental College, Riphah International University, Islamabad, Pakistan (Ref. No. IIDC/IRC/2020/006/013). All participants were informed about the purpose, procedures, potential risks, and benefits of the study. Written informed consent was obtained from all participants prior to enrollment. Participation was voluntary, confidentiality and anonymity were ensured, and participants had the right to withdraw from the study at any stage without any consequences.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eConsent for publication:\u0026nbsp;\u003c/strong\u003eNot applicable\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAvailability of data and material:\u0026nbsp;\u003c/strong\u003eThe datasets and code generated and/or analysed during the current study are available from the corresponding author on reasonable request.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eCompeting interests:\u0026nbsp;\u003c/strong\u003eThe authors declare that they have no competing interests.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eFunding:\u0026nbsp;\u003c/strong\u003eThis research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAuthor’s contributions:\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eAll authors contributed substantially to the work. Hifza Noor and Arooba Malik contributed to the original draft, writing, methodology, data collection, conceptualization, references, review, and editing. Rameesa Aymen contributed in writing. Mahnoor Fatima contributed to interpretation of results. Faisal Shafiq and Momina Zahra contributed to the final review and editing of the manuscript.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAcknowledgements:\u0026nbsp;\u003c/strong\u003eJaved Ashraf contributed to the statistical analysis.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\n\u003cli\u003eAbati, S., Sandri, G. F., Finotello, L., \u0026amp; Polizzi, E. (2024). Differential Diagnosis of Pigmented Lesions in the Oral Mucosa: A Clinical Based Overview and Narrative Review. \u003cem\u003eCancers\u003c/em\u003e, \u003cem\u003e16\u003c/em\u003e(13), 2487. https://doi.org/10.3390/cancers16132487\u003c/li\u003e\n\u003cli\u003eAgarwal, M. C., Agarwal, S., Madan, E., \u0026amp; Madan, R. (2015). \u003cem\u003eGINGIVAL HYPERPIGMENTATION REVISITED\u003c/em\u003e. https://www.researchgate.net/publication/361585336\u003c/li\u003e\n\u003cli\u003eCi\u0026ccedil;ek, Y., \u0026amp; Ertaş, U. (2003). The normal and pathological hyperpigmentation of oral mucous membrane: a review. \u003cem\u003eThe Journal of Contemporary Dental Practice\u003c/em\u003e, \u003cem\u003e4\u003c/em\u003e(3), 76\u0026ndash;86.\u003c/li\u003e\n\u003cli\u003ePecci-Lloret, M. P., Linares-P\u0026eacute;rez, C., Pecci-Lloret, M. R., Rodr\u0026iacute;guez-Lozano, F. J., \u0026amp; O\u0026ntilde;ate-S\u0026aacute;nchez, R. E. (2024). Oral Manifestations in Pregnant Women: A Systematic Review. \u003cem\u003eJournal of Clinical Medicine\u003c/em\u003e, \u003cem\u003e13\u003c/em\u003e(3). https://doi.org/10.3390/jcm13030707\u003c/li\u003e\n\u003cli\u003eRotbeh, A., Kazeminia, M., Kalantari, M., \u0026amp; Rajati, F. (2022). Global prevalence of oral hyperpigmentation and its related factors: a systematic review and meta-analysis. \u003cem\u003eJournal of Stomatology, Oral and Maxillofacial Surgery\u003c/em\u003e, \u003cem\u003e123\u003c/em\u003e(5), e411\u0026ndash;e424. https://doi.org/10.1016/j.jormas.2022.01.009\u003c/li\u003e\n\u003cli\u003eGulati, N., Dutt, P., Gupta, N., \u0026amp; Tyagi, P. (n.d.). GINGIVAL HYPERPIGMENTATION: REVISITED. In \u003cem\u003eJournal of Advanced Medical and Dental Sciences Research |Vol\u003c/em\u003e (Vol. 4).\u003c/li\u003e\n\u003cli\u003eMasilana, A., Khammissa, R. A. G., Lemmer, J., \u0026amp; Feller, L. (2017). Physiological oral melanin hyperpigmentation in a South African sample: A clinical study. \u003cem\u003eJournal of Investigative and Clinical Dentistry\u003c/em\u003e, \u003cem\u003e8\u003c/em\u003e(4). https://doi.org/10.1111/jicd.12258\u003c/li\u003e\n\u003cli\u003ePonnaiyan, D., Jegadeesan, V., Perumal, G., \u0026amp; Anusha, A. (2014). Correlating skin color with gingival hyperpigmentation patterns in South Indians - a cross sectional study. \u003cem\u003eOral Health and Dental Management\u003c/em\u003e, \u003cem\u003e13\u003c/em\u003e(1), 132\u0026ndash;136.\u003c/li\u003e\n\u003cli\u003eFeller, L., Masilana, A., Khammissa, R. A., Altini, M., Jadwat, Y., \u0026amp; Lemmer, J. (2014). Melanin: the biophysiology of oral melanocytes and physiological oral hyperpigmentation. \u003cem\u003eHead \u0026amp; Face Medicine\u003c/em\u003e, \u003cem\u003e10\u003c/em\u003e(1), 8. https://doi.org/10.1186/1746-160X-10-8\u003c/li\u003e\n\u003cli\u003eSoulafa Almazrooa, Amna Arooj, Sobia Hassan, \u0026amp; Amber Kiyani. (2025). JOHOE_Volume 14_Issue 1_Pages 1-5. \u003cem\u003eJournal of Oral Health and Oral Epidemiology\u003c/em\u003e, \u003cem\u003e14\u003c/em\u003e, 1\u0026ndash;5. https://doi.org/10.34172/johoe.2403.1627\u003c/li\u003e\n\u003cli\u003eShahzad, A., Kiyani, A., \u0026amp; Paiker, S. (2018). Prevalence of Oral Anomalies and Pathologies in the Pakistani Population \u0026ndash; A Cross Sectional Study. \u003cem\u003eJournal of The Pakistan Dental Association\u003c/em\u003e, \u003cem\u003e27\u003c/em\u003e(1), 13\u0026ndash;17. https://doi.org/10.25301/JPDA.271.13\u003c/li\u003e\n\u003cli\u003ePecci-Lloret, M. P., Linares-P\u0026eacute;rez, C., Pecci-Lloret, M. R., Rodr\u0026iacute;guez-Lozano, F. J., \u0026amp; O\u0026ntilde;ate-S\u0026aacute;nchez, R. E. (2024). Oral Manifestations in Pregnant Women: A Systematic Review. \u003cem\u003eJournal of Clinical Medicine\u003c/em\u003e, \u003cem\u003e13\u003c/em\u003e(3). https://doi.org/10.3390/jcm13030707\u003c/li\u003e\n\u003cli\u003eMotosko, C. C., Bieber, A. K., Pomeranz, M. K., Stein, J. A., \u0026amp; Martires, K. J. (2017). Physiologic changes of pregnancy: A review of the literature. \u003cem\u003eInternational Journal of Women\u0026rsquo;s Dermatology\u003c/em\u003e, \u003cem\u003e3\u003c/em\u003e(4), 219\u0026ndash;224. https://doi.org/10.1016/j.ijwd.2017.09.003\u003c/li\u003e\n\u003cli\u003eSattar, U., Malik, U., Shireen, Z., Faiz Ahmad Khan, S., Health CenterQadirabad, R., \u0026amp; Ghazi Khan, D. (2017). Oral health challenges in pregnant women in Pakistan: A research analysis. \u003cem\u003eInternational Journal of Advanced Biotechnology and Research (IJBR)\u003c/em\u003e, \u003cem\u003e8\u003c/em\u003e, 1724\u0026ndash;1730. http://www.bipublication.com\u003c/li\u003e\n\u003cli\u003eDubuc, A., Zitouni, A., Thomas, C., K\u0026eacute;moun, P., Cousty, S., Monsarrat, P., \u0026amp; Laurencin, S. (2022). Improvement of Mucosal Lesion Diagnosis with Machine Learning Based on Medical and Semiological Data: An Observational Study. \u003cem\u003eJournal of Clinical Medicine\u003c/em\u003e, \u003cem\u003e11\u003c/em\u003e(21), 6596. https://doi.org/10.3390/jcm11216596\u003c/li\u003e\n\u003cli\u003ePiccininni, M., Konigorski, S., Rohmann, J. L., \u0026amp; Kurth, T. (n.d.). \u003cem\u003eTitle: Directed Acyclic Graphs and causal thinking in clinical risk prediction modeling\u003c/em\u003e.\u003c/li\u003e\n\u003cli\u003eMarko, J. G. O., Neagu, C. D., \u0026amp; Anand, P. B. (2025). Examining inclusivity: the use of AI and diverse populations in health and social care: a systematic review. \u003cem\u003eBMC Medical Informatics and Decision Making\u003c/em\u003e, \u003cem\u003e25\u003c/em\u003e(1), 57. https://doi.org/10.1186/s12911-025-02884-1\u003c/li\u003e\n\u003cli\u003eNyariro, M., Emami, E., Caidor, P., \u0026amp; Abbasgholizadeh Rahimi, S. (2023). Integrating equity, diversity and inclusion throughout the lifecycle of AI within healthcare: a scoping review protocol. \u003cem\u003eBMJ Open\u003c/em\u003e, \u003cem\u003e13\u003c/em\u003e(9), e072069. https://doi.org/10.1136/bmjopen-2023-072069\u003c/li\u003e\n\u003cli\u003eRezazadeh, F., Falsafi, N., Sarraf, Z., \u0026amp; Shahbazi, M. (2014). Oral Mucosal Disorders in Pregnant versus Non-Pregnant Women. \u003cem\u003eDentistry Journal\u003c/em\u003e, \u003cem\u003e2\u003c/em\u003e(4), 134\u0026ndash;141. https://doi.org/10.3390/dj2040134\u003c/li\u003e\n\u003cli\u003eTadakamadla, J., Kumar, S., Nagori, A., Tibdewal, H., Duraiswamy, P., \u0026amp; Kulkarni, S. (2012). Effect of smoking on oral hyperpigmentation and its relationship with periodontal status. \u003cem\u003eDental Research Journal\u003c/em\u003e, \u003cem\u003e9\u003c/em\u003e(Suppl 1), S112-4.\u003c/li\u003e\n\u003cli\u003eSilva, P. U. J., Oliveira, M. B., Vieira, W., Cardoso, S. V., Blumenberg, C., Franco, A., Siqueira, W. L., \u0026amp; Paranhos, L. R. (2022). Oral hyperpigmentation as an adverse effect of chloroquine and hydroxychloroquine use. \u003cem\u003eMedicine\u003c/em\u003e, \u003cem\u003e101\u003c/em\u003e(11). https://doi.org/10.1097/MD.0000000000029044\u003c/li\u003e\n\u003cli\u003eKardys, I., Hoeks, S., van Domburg, R., Lenzen, M., \u0026amp; Boersma, E. (2013). Tools and techniques - Statistics: Analysis of continuous data using the t-test and ANOVA. In \u003cem\u003eEuroIntervention\u003c/em\u003e (Vol. 9, Issue 6, pp. 765\u0026ndash;767). https://doi.org/10.4244/EIJV9I6A123\u003c/li\u003e\n\u003cli\u003eS., V., A.S, H., \u0026amp; S. Sundar, J. (2024). CHI-SQUARE TESTS: A QUICK GUIDE FOR HEALTH RESEARCHERS. \u003cem\u003eInternational Journal of Advanced Research\u003c/em\u003e, \u003cem\u003e12\u003c/em\u003e(10), 1214\u0026ndash;1222. https://doi.org/10.21474/IJAR01/19746\u003c/li\u003e\n\u003cli\u003eWornow, M., Gyang Ross, E., Callahan, A., \u0026amp; Shah, N. H. (2023). APLUS: A Python library for usefulness simulations of machine learning models in healthcare. \u003cem\u003eJournal of Biomedical Informatics\u003c/em\u003e, \u003cem\u003e139\u003c/em\u003e, 104319. https://doi.org/10.1016/j.jbi.2023.104319\u003c/li\u003e\n\u003cli\u003eMuruppel, A. M., Pai, B. S. J., Bhat, S., Parker, S., \u0026amp; Lynch, E. (2020). Laser-Assisted Dehyperpigmentation\u0026mdash;An Introspection of the Science, Techniques, and Perceptions. \u003cem\u003eDentistry Journal\u003c/em\u003e, \u003cem\u003e8\u003c/em\u003e(3), 88. https://doi.org/10.3390/dj8030088\u003c/li\u003e\n\u003cli\u003eSnell, R. S., \u0026amp; Bischitz, P. G. (1960). The Effect of Large Doses of Estrogen and Estrogen and Progesterone on Melanin Hyperpigmentation**From the Department of Anatomy, Medical School King\u0026rsquo;s College, Newcastle upon Tyne, University of Durham, England. \u003cem\u003eJournal of Investigative Dermatology\u003c/em\u003e, \u003cem\u003e35\u003c/em\u003e(2), 73\u0026ndash;82. https://doi.org/10.1038/jid.1960.87\u003c/li\u003e\n\u003cli\u003eRezazadeh, F., Falsafi, N., Sarraf, Z., \u0026amp; Shahbazi, M. (2014). Oral Mucosal Disorders in Pregnant\u003cem\u003e versus\u003c/em\u003e Non-Pregnant Women. \u003cem\u003eDentistry Journal\u003c/em\u003e, \u003cem\u003e2\u003c/em\u003e(4), 134-141. https://doi.org/10.3390/dj2040134\u003c/li\u003e\n\u003cli\u003eKoppolu, P., Almutairi, H., Yousef, S. al, Ansary, N., Noushad, M., Vishal, M. B., Swapna, L. A., Alsuwayyigh, N., Albalawi, M., Shrivastava, D., \u0026amp; Srivastava, K. C. (2024). Relationship of skin complexion with gingival tissue color and hyperhyperpigmentation. A multi-ethnic comparative study. \u003cem\u003eBMC Oral Health\u003c/em\u003e, \u003cem\u003e24\u003c/em\u003e(1), 451. https://doi.org/10.1186/s12903-024-04189-7\u003c/li\u003e\n\u003cli\u003eMoravej-Salehi, E., Moravej-Salehi, E., \u0026amp; Hajifattahi, F. (2015). Relationship of Gingival Hyperpigmentation with Passive Smoking in Women. \u003cem\u003eTanaffos\u003c/em\u003e, \u003cem\u003e14\u003c/em\u003e(2), 107\u0026ndash;114.\u003c/li\u003e\n\u003cli\u003eLim, H. W., Piquero-Casals, J., Schalka, S., Leone, G., Trull\u0026agrave;s, C., Brown, A., Foyaca, M., Gilaberte, Y., Krutmann, J., \u0026amp; Passeron, T. (2025). Photoprotection in pregnancy: addressing safety concerns and optimizing skin health. \u003cem\u003eFrontiers in Medicine\u003c/em\u003e, \u003cem\u003e12\u003c/em\u003e. https://doi.org/10.3389/fmed.2025.1563369\u003c/li\u003e\n\u003cli\u003eHan, S., Williamson, B. D., \u0026amp; Fong, Y. (2021). Improving random forest predictions in small datasets from two-phase sampling designs. \u003cem\u003eBMC Medical Informatics and Decision Making\u003c/em\u003e, \u003cem\u003e21\u003c/em\u003e(1), 322. https://doi.org/10.1186/s12911-021-01688-3\u003c/li\u003e\n\u003cli\u003eNusinovici, S., Tham, Y. C., Chak Yan, M. Y., Wei Ting, D. S., Li, J., Sabanayagam, C., Wong, T. Y., \u0026amp; Cheng, C.-Y. (2020). Logistic regression was as good as machine learning for predicting major chronic diseases. \u003cem\u003eJournal of Clinical Epidemiology\u003c/em\u003e, \u003cem\u003e122\u003c/em\u003e, 56\u0026ndash;69. https://doi.org/10.1016/j.jclinepi.2020.03.002\u003c/li\u003e\n\u003cli\u003eRijal, A. H., Dhami, B., Pandey, N., \u0026amp; Aryal, D. (2021). Prevalence of Gingival Hyperpigmentation and its Association with Gingival Biotype and Skin Colour. \u003cem\u003eJournal of Nepalese Society of Periodontology and Oral Implantology\u003c/em\u003e, \u003cem\u003e5\u003c/em\u003e(1), 19\u0026ndash;25. https://doi.org/10.3126/jnspoi.v5i1.38178\u003c/li\u003e\n\u003cli\u003ePonnaiyan, D., Jegadeesan, V., Perumal, G., \u0026amp; Anusha, A. (2014). Correlating skin color with gingival hyperpigmentation patterns in South Indians - a cross sectional study. \u003cem\u003eOral Health and Dental Management\u003c/em\u003e, \u003cem\u003e13\u003c/em\u003e(1), 132\u0026ndash;136.\u003c/li\u003e\n\u003cli\u003eLipschuetz, M., Guedalia, J., Cohen, S. M., Unger, R., Yagel, S., \u0026amp; Sompolinsky, Y. (2024). \u003cem\u003eMachine Learning in Electronic Health Records: Identifying High-Risk Obstetric Patients Pre and During Labor\u003c/em\u003e. https://doi.org/10.3233/SHTI240096\u003c/li\u003e\n\u003c/ol\u003e"},{"header":"Tables","content":"\u003cp\u003e\u003cstrong\u003eTable 1. Descriptive Statistics\u003c/strong\u003e\u003c/p\u003e\n\u003ctable border=\"0\" cellspacing=\"0\" cellpadding=\"0\" width=\"663\"\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd colspan=\"9\" valign=\"top\" style=\"width: 663px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eNumber of Pregnancies\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 105px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 68px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e0\u003c/strong\u003e\u003c/p\u003e\n \u003cp\u003eN (%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 67px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e1\u003c/strong\u003e\u003c/p\u003e\n \u003cp\u003eN (%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 67px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e2\u003c/strong\u003e\u003c/p\u003e\n \u003cp\u003eN (%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 67px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e3\u003c/strong\u003e\u003c/p\u003e\n \u003cp\u003eN (%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 67px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e4\u003c/strong\u003e\u003c/p\u003e\n \u003cp\u003eN (%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 81px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e5\u003c/strong\u003e\u003c/p\u003e\n \u003cp\u003eN (%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 79px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eTotal\u003c/strong\u003e\u003c/p\u003e\n \u003cp\u003eN (%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 62px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eP-value\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd colspan=\"9\" valign=\"top\" style=\"width: 663px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eOral Hyperpigmentation\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 105px;\"\u003e\n \u003cp\u003eAbsent\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 68px;\"\u003e\n \u003cp\u003e78 (27%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 67px;\"\u003e\n \u003cp\u003e54 (18.7%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 67px;\"\u003e\n \u003cp\u003e62 (21.5%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 67px;\"\u003e\n \u003cp\u003e58 (20.1%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 67px;\"\u003e\n \u003cp\u003e26\u003c/p\u003e\n \u003cp\u003e(9%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 81px;\"\u003e\n \u003cp\u003e11\u003c/p\u003e\n \u003cp\u003e(3.8%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 79px;\"\u003e\n \u003cp\u003e289\u003c/p\u003e\n \u003cp\u003e(100%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd rowspan=\"3\" valign=\"top\" style=\"width: 62px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003cp\u003e\u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp;0.736\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 105px;\"\u003e\n \u003cp\u003ePresent\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 68px;\"\u003e\n \u003cp\u003e45 (28.5%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 67px;\"\u003e\n \u003cp\u003e21 (13.3%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 67px;\"\u003e\n \u003cp\u003e39 (24.7%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 67px;\"\u003e\n \u003cp\u003e34 (21.5%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 67px;\"\u003e\n \u003cp\u003e12 (7.6%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 81px;\"\u003e\n \u003cp\u003e7\u0026nbsp;\u003c/p\u003e\n \u003cp\u003e(4.4%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 79px;\"\u003e\n \u003cp\u003e158 (100%)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 105px;\"\u003e\n \u003cp\u003eTotal\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 68px;\"\u003e\n \u003cp\u003e123 (27.5%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 67px;\"\u003e\n \u003cp\u003e75 (16.8%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 67px;\"\u003e\n \u003cp\u003e101 (22.6%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 67px;\"\u003e\n \u003cp\u003e92 (20.6%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 67px;\"\u003e\n \u003cp\u003e38 (8.5%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 81px;\"\u003e\n \u003cp\u003e18\u0026nbsp;\u003c/p\u003e\n \u003cp\u003e(4.0%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 79px;\"\u003e\n \u003cp\u003e447 (100%)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd colspan=\"9\" valign=\"top\" style=\"width: 663px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eLabial Mucosa\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 105px;\"\u003e\n \u003cp\u003eNo Hyperpigmentation\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 68px;\"\u003e\n \u003cp\u003e115 (29.6%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 67px;\"\u003e\n \u003cp\u003e68 (17.5%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 67px;\"\u003e\n \u003cp\u003e85 (21.9%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 67px;\"\u003e\n \u003cp\u003e78 (20.1%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 67px;\"\u003e\n \u003cp\u003e28 (7.2%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 81px;\"\u003e\n \u003cp\u003e14\u003c/p\u003e\n \u003cp\u003e(3.6%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 79px;\"\u003e\n \u003cp\u003e388 (100%)\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd rowspan=\"5\" valign=\"top\" style=\"width: 62px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003cp\u003e0.003\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 105px;\"\u003e\n \u003cp\u003eMild Hyperpigmentation\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 68px;\"\u003e\n \u003cp\u003e8 (14.3%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 67px;\"\u003e\n \u003cp\u003e7 (12.5%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 67px;\"\u003e\n \u003cp\u003e16 (28.6%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 67px;\"\u003e\n \u003cp\u003e13 (23.2%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 67px;\"\u003e\n \u003cp\u003e9 (16.1%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 81px;\"\u003e\n \u003cp\u003e3\u0026nbsp;\u003c/p\u003e\n \u003cp\u003e(5.4%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 79px;\"\u003e\n \u003cp\u003e56\u0026nbsp;\u003c/p\u003e\n \u003cp\u003e(100%)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 105px;\"\u003e\n \u003cp\u003eModerate Hyperpigmentation\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 68px;\"\u003e\n \u003cp\u003e0\u0026nbsp;\u003c/p\u003e\n \u003cp\u003e(0%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 67px;\"\u003e\n \u003cp\u003e0\u0026nbsp;\u003c/p\u003e\n \u003cp\u003e(0%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 67px;\"\u003e\n \u003cp\u003e0\u0026nbsp;\u003c/p\u003e\n \u003cp\u003e(0%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 67px;\"\u003e\n \u003cp\u003e1\u0026nbsp;\u003c/p\u003e\n \u003cp\u003e(50%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 67px;\"\u003e\n \u003cp\u003e0\u003c/p\u003e\n \u003cp\u003e\u0026nbsp;(0%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 81px;\"\u003e\n \u003cp\u003e1\u003c/p\u003e\n \u003cp\u003e(50%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 79px;\"\u003e\n \u003cp\u003e2\u003c/p\u003e\n \u003cp\u003e\u0026nbsp;(100%)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 105px;\"\u003e\n \u003cp\u003eHeavy Hyperpigmentation\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 68px;\"\u003e\n \u003cp\u003e0\u0026nbsp;\u003c/p\u003e\n \u003cp\u003e(0%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 67px;\"\u003e\n \u003cp\u003e0\u0026nbsp;\u003c/p\u003e\n \u003cp\u003e(0%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 67px;\"\u003e\n \u003cp\u003e0\u0026nbsp;\u003c/p\u003e\n \u003cp\u003e(0%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 67px;\"\u003e\n \u003cp\u003e0\u0026nbsp;\u003c/p\u003e\n \u003cp\u003e(0%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 67px;\"\u003e\n \u003cp\u003e1 (100%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 81px;\"\u003e\n \u003cp\u003e0\u0026nbsp;\u003c/p\u003e\n \u003cp\u003e(0%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 79px;\"\u003e\n \u003cp\u003e1\u0026nbsp;\u003c/p\u003e\n \u003cp\u003e(100%)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 105px;\"\u003e\n \u003cp\u003eTotal\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 68px;\"\u003e\n \u003cp\u003e123 (27.5%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 67px;\"\u003e\n \u003cp\u003e75 (16.8%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 67px;\"\u003e\n \u003cp\u003e101 (22.6%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 67px;\"\u003e\n \u003cp\u003e92 (20.6%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 67px;\"\u003e\n \u003cp\u003e38 (8.5%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 81px;\"\u003e\n \u003cp\u003e18\u0026nbsp;\u003c/p\u003e\n \u003cp\u003e(4.0%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 79px;\"\u003e\n \u003cp\u003e447 (100%)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd colspan=\"9\" valign=\"top\" style=\"width: 663px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eHard Palate\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 105px;\"\u003e\n \u003cp\u003eNo Hyperpigmentation\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 68px;\"\u003e\n \u003cp\u003e123 (28%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 67px;\"\u003e\n \u003cp\u003e73 (16.6%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 67px;\"\u003e\n \u003cp\u003e101 (23.0%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 67px;\"\u003e\n \u003cp\u003e88 (20.0%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 67px;\"\u003e\n \u003cp\u003e38 (8.6%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 81px;\"\u003e\n \u003cp\u003e17\u0026nbsp;\u003c/p\u003e\n \u003cp\u003e(3.9%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 79px;\"\u003e\n \u003cp\u003e440 (100%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd rowspan=\"3\" valign=\"top\" style=\"width: 62px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003cp\u003e0.018\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 105px;\"\u003e\n \u003cp\u003eMild Hyperpigmentation\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 68px;\"\u003e\n \u003cp\u003e0\u0026nbsp;\u003c/p\u003e\n \u003cp\u003e(0%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 67px;\"\u003e\n \u003cp\u003e2 (28.6%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 67px;\"\u003e\n \u003cp\u003e0\u0026nbsp;\u003c/p\u003e\n \u003cp\u003e(0%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 67px;\"\u003e\n \u003cp\u003e4 (57.1%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 67px;\"\u003e\n \u003cp\u003e0\u0026nbsp;\u003c/p\u003e\n \u003cp\u003e(0%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 81px;\"\u003e\n \u003cp\u003e1\u0026nbsp;\u003c/p\u003e\n \u003cp\u003e(14.3%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 79px;\"\u003e\n \u003cp\u003e7\u0026nbsp;\u003c/p\u003e\n \u003cp\u003e(100%)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 105px;\"\u003e\n \u003cp\u003eTotal\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 68px;\"\u003e\n \u003cp\u003e123 (27.5%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 67px;\"\u003e\n \u003cp\u003e75 (16.8%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 67px;\"\u003e\n \u003cp\u003e101 (22.6%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 67px;\"\u003e\n \u003cp\u003e92 (20.6%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 67px;\"\u003e\n \u003cp\u003e38 (8.5%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 81px;\"\u003e\n \u003cp\u003e18\u0026nbsp;\u003c/p\u003e\n \u003cp\u003e(4.0%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 79px;\"\u003e\n \u003cp\u003e447 (100%)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd colspan=\"9\" valign=\"top\" style=\"width: 663px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eFloor of The Mouth\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 105px;\"\u003e\n \u003cp\u003eNo Hyperpigmentation\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 68px;\"\u003e\n \u003cp\u003e123 (27.7%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 67px;\"\u003e\n \u003cp\u003e75 (16.9%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 67px;\"\u003e\n \u003cp\u003e101 (22.7%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 67px;\"\u003e\n \u003cp\u003e90 (20.3%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 67px;\"\u003e\n \u003cp\u003e38 (8.6%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 81px;\"\u003e\n \u003cp\u003e17\u0026nbsp;\u003c/p\u003e\n \u003cp\u003e(3.8%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 79px;\"\u003e\n \u003cp\u003e444 (100%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd rowspan=\"3\" valign=\"top\" style=\"width: 62px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003cp\u003e0.044\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 105px;\"\u003e\n \u003cp\u003eMild Hyperpigmentation\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 68px;\"\u003e\n \u003cp\u003e0\u0026nbsp;\u003c/p\u003e\n \u003cp\u003e(0%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 67px;\"\u003e\n \u003cp\u003e0\u0026nbsp;\u003c/p\u003e\n \u003cp\u003e(0%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 67px;\"\u003e\n \u003cp\u003e0\u0026nbsp;\u003c/p\u003e\n \u003cp\u003e(0%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 67px;\"\u003e\n \u003cp\u003e2 (66.7%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 67px;\"\u003e\n \u003cp\u003e0\u0026nbsp;\u003c/p\u003e\n \u003cp\u003e(0%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 81px;\"\u003e\n \u003cp\u003e1\u0026nbsp;\u003c/p\u003e\n \u003cp\u003e(33.3%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 79px;\"\u003e\n \u003cp\u003e3\u0026nbsp;\u003c/p\u003e\n \u003cp\u003e(100%)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 105px;\"\u003e\n \u003cp\u003eTotal\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 68px;\"\u003e\n \u003cp\u003e123 (27.5%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 67px;\"\u003e\n \u003cp\u003e75 (16.8%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 67px;\"\u003e\n \u003cp\u003e101 (22.6%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 67px;\"\u003e\n \u003cp\u003e92 (20.6%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 67px;\"\u003e\n \u003cp\u003e38 (8.5%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 81px;\"\u003e\n \u003cp\u003e18\u0026nbsp;\u003c/p\u003e\n \u003cp\u003e(4.0%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 79px;\"\u003e\n \u003cp\u003e447 (100%)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n\u003c/table\u003e\n\u003cp\u003e\u003cstrong\u003e\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eTable 2. Variable Associations Analysis\u003c/strong\u003e\u003c/p\u003e\n\u003ctable border=\"0\" cellspacing=\"0\" cellpadding=\"0\" width=\"632\"\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 349px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eCorrelation of Key Variables\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 283px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eCorrelation Coefficient (R)\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 349px;\"\u003e\n \u003cp\u003eAge and Number of Pregnancies\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 283px;\"\u003e\n \u003cp\u003e0.6\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 349px;\"\u003e\n \u003cp\u003eTotal Hyperpigmentation and Gingiva\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 283px;\"\u003e\n \u003cp\u003e0.86\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 349px;\"\u003e\n \u003cp\u003eTotal Hyperpigmentation and Labial Mucosa\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 283px;\"\u003e\n \u003cp\u003e0.47\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 349px;\"\u003e\n \u003cp\u003eTotal Hyperpigmentation and Buccal Mucosa\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 283px;\"\u003e\n \u003cp\u003e0.38\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n\u003c/table\u003e\n\u003cp\u003e\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eTable 3. Machine Learning Model Metrics\u003c/strong\u003e\u003c/p\u003e\n\u003ctable border=\"0\" cellspacing=\"0\" cellpadding=\"0\" width=\"636\"\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 148px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eModel\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 106px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eAccuracy\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 100px;\"\u003e\n \u003cp\u003e\u003cstrong\u003ePrecision\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 94px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eRecall\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 90px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eF1 Score\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 98px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eAUC-ROC\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 148px;\"\u003e\n \u003cp\u003eRandom Forest\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 106px;\"\u003e\n \u003cp\u003e0.9556\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 100px;\"\u003e\n \u003cp\u003e0.9667\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 94px;\"\u003e\n \u003cp\u003e0.9063\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 90px;\"\u003e\n \u003cp\u003e0.9355\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 98px;\"\u003e\n \u003cp\u003e0.9445\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 148px;\"\u003e\n \u003cp\u003eGradient Boosting Machine\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 106px;\"\u003e\n \u003cp\u003e0.9556\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 100px;\"\u003e\n \u003cp\u003e0.9667\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 94px;\"\u003e\n \u003cp\u003e0.9063\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 90px;\"\u003e\n \u003cp\u003e0.9355\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 98px;\"\u003e\n \u003cp\u003e0.9445\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 148px;\"\u003e\n \u003cp\u003eLogistic Regression\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 106px;\"\u003e\n \u003cp\u003e0.9667\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 100px;\"\u003e\n \u003cp\u003e0.9677\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 94px;\"\u003e\n \u003cp\u003e0.9375\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 90px;\"\u003e\n \u003cp\u003e0.9524\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 98px;\"\u003e\n \u003cp\u003e0.9601\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n\u003c/table\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":false,"highlight":"","institution":"","isAcceptedByJournal":false,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"bmc-oral-health","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"ohea","sideBox":"Learn more about [BMC Oral Health](http://bmcoralhealth.biomedcentral.com/)","snPcode":"","submissionUrl":"https://www.editorialmanager.com/ohea/default.aspx","title":"BMC Oral Health","twitterHandle":"BMC_series","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"em","reportingPortfolio":"BMC Series","inReviewEnabled":true,"inReviewRevisionsEnabled":true},"keywords":"Algorithms, Cross-Sectional Studies, Maternal Health, Regression Analysis, Oral Health","lastPublishedDoi":"10.21203/rs.3.rs-8368647/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-8368647/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003ch2\u003eObjectives\u003c/h2\u003e \u003cp\u003eThe study aims to predict the association between oral hyperpigmentation and pregnancy in women between 18 and 45 years using machine learning (ML) algorithms. Directed acyclic graphs (DAGs) were constructed to explore the association between oral hyperpigmentation and demographic factors such as age, occupation, ethnicity, and number of pregnancies.\u003c/p\u003e\u003ch2\u003eMethods\u003c/h2\u003e \u003cp\u003eThis cross-sectional study involved pregnant and non-pregnant participants aged between 18 and 45. The study was conducted at one public and one private sector hospital. DAGs were used to represent relationships between different variables. Three ML models, including Logistic Regression (LR), Random Forest (RF) and Gradient Boosting Machine (GBM) algorithms, were used to predict the association between oral hyperpigmentation and pregnancy based upon relevant theoretically backed predictors.\u003c/p\u003e\u003ch2\u003eResults\u003c/h2\u003e \u003cp\u003eThe results show that labial mucosa and hard palate hyperpigmentation had a positive correlation with the number of pregnancies, with p-values of 0.003 and 0.018, respectively. All three ML models demonstrated excellent predictive power. However, the LR model with an accuracy rate of 96.67% was particularly effective in predicting the association between oral hyperpigmentation and pregnancy.\u003c/p\u003e\u003ch2\u003eConclusions\u003c/h2\u003e \u003cp\u003eThis study highlighted the effectiveness of LR, supported by DAGs, in predicting oral hyperpigmentation among women aged 18\u0026ndash;45. Key factors such as the number of pregnancies, age, ethnicity, and occupation significantly influenced site-specific hyperpigmentation. The findings support the integration of ML into maternal oral healthcare for early detection and culturally tailored interventions. Simpler ML models proved most accurate in moderate-sized datasets with fewer covariates.\u003c/p\u003e","manuscriptTitle":"Predicting the Association Between Oral Hyperpigmentation and Pregnancy in Women Aged 18-45 Using Machine Learning and Directed Acyclic Graphs: A Cross-Sectional Study","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2026-01-19 09:04:39","doi":"10.21203/rs.3.rs-8368647/v1","editorialEvents":[{"type":"communityComments","content":0},{"type":"reviewersInvited","content":"","date":"2026-01-13T13:34:33+00:00","index":"","fulltext":""},{"type":"editorInvited","content":"","date":"2025-12-19T10:05:08+00:00","index":"","fulltext":""},{"type":"editorAssigned","content":"","date":"2025-12-18T09:49:06+00:00","index":"","fulltext":""},{"type":"checksComplete","content":"","date":"2025-12-18T09:48:18+00:00","index":"","fulltext":""},{"type":"submitted","content":"BMC Oral Health","date":"2025-12-15T16:43:59+00:00","index":"","fulltext":""}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"bmc-oral-health","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"ohea","sideBox":"Learn more about [BMC Oral Health](http://bmcoralhealth.biomedcentral.com/)","snPcode":"","submissionUrl":"https://www.editorialmanager.com/ohea/default.aspx","title":"BMC Oral Health","twitterHandle":"BMC_series","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"em","reportingPortfolio":"BMC Series","inReviewEnabled":true,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"438351e7-df67-4c88-bcd0-8389720ffa4f","owner":[],"postedDate":"January 19th, 2026","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"under-review","subjectAreas":[],"tags":[],"updatedAt":"2026-01-19T09:04:39+00:00","versionOfRecord":[],"versionCreatedAt":"2026-01-19 09:04:39","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-8368647","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-8368647","identity":"rs-8368647","version":["v1"]},"buildId":"XKTyCvWXoU3ODBz1xrDgd","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}