Diagnostic performance of Demirjian's third molar maturation stages G, G1, and H for identifying the 18-year age...

Determining attainment of 18 years is a critical task in forensic age estimation. This study evaluated and compared the sex-specific diagnostic performance of conventional and refined third-molar maturation cut-offs, Demirjian stages G and H, and the refined stage G1 for identifying individuals aged ≥ 18 years in a South Indian population. Panoramic radiographs of 599 individuals aged 14-22 years were retrospectively assessed. Mandibular left third molars were staged using the conventional Demirjian system and the Solari and Abramovitch modification. Diagnostic performance was evaluated separately for males and females using sensitivity, specificity, predictive values, likelihood ratios, accuracy, and post-test probabilities. Intra- and inter-examiner agreement was assessed using kappa statistics. Third-molar maturation effectively discriminated attainment of 18 years in both sexes, with stage-dependent trade-offs between sensitivity and specificity. Stages G and above and stages G1 and above demonstrated high sensitivity and strong exclusion of adulthood when negative, whereas stage H showed higher specificity and greater confirmatory value when present, particularly among females. Diagnostic performance of the refined G1 stage was comparable to that of stage G, with modest improvement in specificity and post-test probability in females. Inter-examiner agreement was excellent for stages G and H but only moderate for G1, indicating reduced reproducibility for the refined stage. In conclusion, conventional Demirjian stages G and H remain reliable and reproducible indicators for assessing attainment of 18 years. Although the refined G1 stage offers increased developmental resolution and supports probabilistic interpretation, its forensic utility is constrained by reduced inter-observer agreement.