The tuberculum marginale as a demographic marker: Morphometric insights from an Azerbaijani cohort

The tuberculum marginale (TM), a subtle protuberance of the zygomatic bone, has been largely overlooked in morphometric research despite its potential anthropological, surgical, and forensic relevance. Its role as a dynamic craniofacial landmark remains insufficiently characterized.

We conducted a CT-based morphometric analysis in an Azerbaijani population (n = 408; 207 males, 201 females) divided into early-adulthood (females 21-55 years, n = 103; males 22-60 years, n = 111) and elderly (females 56-74 years, n = 98; males 61-74 years, n = 96) cohorts. Bilateral TM dimensions and related craniofacial parameters were measured using a standardized protocol; intra-observer reproducibility was confirmed in a 120-subject subsample (pooled ICC = 0.955). We used Wilcoxon signed-rank and Mann-Whitney U tests, Spearman and point-biserial correlations, ordinary least-squares regression, and leave-one-out cross-validated (LOOCV) ROC and multivariate classification analyses (logistic regression, linear discriminant function analysis, PCA, K-means clustering); classification models were then tested in an independently collected external validation cohort (n = 111).

TM morphology showed consistent sexual dimorphism - most robust on the left side at both ages (Cohen's d = 0.74-0.85) - alongside age-related remodeling and a sex-dependent pattern of lateral asymmetry: females showed a significantly larger right than left TM at both ages, while males were essentially symmetric bilaterally. TM size decreased with age in both sexes, whereas TM-zygomaticofrontal suture distance increased markedly with age in males only. Cross-validated classification models achieved fair-to-good discriminatory power for sex estimation (LOOCV AUC 0.695-0.830; DFA accuracy 62-77%), strongest in the elderly group and on the left side; performance in the independent external validation cohort (58-77% accuracy) closely matched these LOOCV estimates, confirming that the models generalize. PCA and K-means clustering did not reveal a small number of dominant morphometric axes or distinct morphotypes.

TM morphology is not merely a stable anatomical landmark but a demographically sensitive, reproducibly measurable structure. Its cross-validated, sex- and age-dependent variation supports a role as a supplementary marker in forensic sex estimation, particularly in populations with limited osteometric reference databases, though its discriminatory power alone is modest and is best combined with other craniofacial indicators. Clinically, TM offers a reproducible surgical landmark for zygomatic reconstruction and orbital procedures. Broader, multi-cohort comparative studies are still needed to establish its population-specific and cross-ethnic applicability.