T1-Weighted MRI of the proximal humerus: Toward a non-ionizing approach to forensic age estimation in a French cohort.

Forensic age estimation of living individuals is increasingly performed using magnetic resonance imaging (MRI) to avoid ionizing radiation exposure. The proximal humeral epiphysis remains comparatively understudied relative to the hand-wrist and knee. This retrospective study evaluated 346 shoulder MRI examinations from a French population aged 7.18-33.82 years (204 males, 142 females), staged according to the combined Schmeling-Kellinghaus system. Two raters independently assessed ossification stage, blinded to age and sex. Spearman's rank correlation, Mann-Whitney U tests, Cohen's kappa, and an ordinal probit transition analysis were performed. A strong, statistically significant correlation was found between chronological age and ossification stage (ρ = 0.878 overall; 0.880 in males, 0.878 in females; p < 0.001). No statistically significant sex-related difference in age was observed at any stage (all p > 0.05). Intra- and interobserver reliability were very good (κ = 0.911 and 0.890, respectively). The minimum age for stage 4 (visible epiphyseal scar) was 18.33 years in males and 18.06 years in females; stage 5 was not observed despite ages extending to 33.82 years. Transition analysis provided sex-specific ages at which the probability of progression to the next stage reached 50%, offering a measure less dependent on sampling extremes than the conventional minimum-age approach. A previously undescribed interpretive pitfall - a trabecular hypointense pattern potentially mimicking a persistent epiphyseal-metaphyseal gap - is also reported, with recommendations for its recognition. These findings extend the combined Schmeling-Kellinghaus staging system to a Western European cohort for the first time and support its applicability as a research tool, while underscoring that minimum-age and transition-age estimates from any single cohort should not be applied as forensic thresholds without independent population-specific validation.