Modeling stature estimation based on pelvic dimensions and lower extremity bone lengths.

OBJECTIVE: This study aimed to investigate the relationship between pelvic and lower extremity measurements obtained from orthoroentgenography images and stature, and to develop sex-specific stature estimation models based on these measurements. MATERIALS AND METHODS: This retrospective study included 114 adult individuals (67 females and 47 males) with recorded stature. Sex-related differences were evaluated using the independent samples t-test. Associations between measurements were assessed separately for females and males using Pearson correlation analysis. Multiple linear regression, LASSO regression and Multivariate Adaptive Regression Splines (MARS) models were constructed for stature estimation. Statistical analyses were performed using SPSS (version 31.0) and R software (version 4.3.0), with statistical significance set at p < 0.05. RESULTS: Most pelvic and lower extremity measurements showed statistically significant sex-based differences. Femur length emerged as a consistent and important predictor of stature in both sexes. In multiple linear regression analysis, the model including FL, SIJD, LTD, and CVH yielded an R2 value of 0.723 in males, whereas the model including FL and SIJD produced an R2 value of 0.40 in females. The MARS model yielded R2 values of 0.729 in males based on FL, SIJD, LTD, and CVH, and 0.420 in females based on FL and ICL. CONCLUSION: Stature estimation based on pelvic and lower extremity measurements shows clear sex-specific differences. Both multiple linear regression and MARS models demonstrated higher explanatory performance in males than in females. These findings highlight the importance of sex-specific and multivariate modeling strategies and suggest that the proposed models may serve as complementary tools in forensic anthropology and clinical applications.