The impact of alignment strategy on the calculation of cross‐sectional properties from high‐resolution <scp>CT</scp> scans of the non‐adult clavicle

Recent advances in the study of cross-sectional geometry have highlighted the value of ontogenetic data for understanding skeletal function, development, and biomechanics. While multiple protocols for extracting bone cross-sections have been validated in limb long bones, the clavicle (despite its unique curvature and developmental trajectory) has received little methodological scrutiny. Critically, two main alignment strategies can be used to extract clavicular midshaft cross-sections from 3D data, yet their comparability remains untested. This lack of standardization compromises the reproducibility and interpretability of clavicular cross-sectional property (CSP) data, particularly in developmental studies. This study evaluates two published alignment protocols for clavicular CSP extraction in non-adults: one that follows the natural curvature of the bone and one based on a standardized transverse plane. We analyzed high-resolution computed tomography scans of 81 individuals (ages 0.5-20 years) from a medieval osteological collection. Midshaft cross-sections were generated using both methods for each clavicle, and CSP variables were calculated. Comparative analysis involved linear regressions, Bland-Altman plots, and percent prediction error (%PE) to assess both systematic bias and the magnitude and direction of error between methods. While total area showed a significant proportional bias and %PE values exceeding the 5% acceptable error threshold, cortical subperiosteal area exhibited a consistent but non-proportional bias, indicating uniform overestimation across specimens. Z_p and J showed the largest directional errors, reflecting greater sensitivity to alignment differences. These discrepancies are not merely numerical but reflect fundamental geometric divergences introduced by the alignment strategy. As a result, CSP values obtained from these two protocols are not directly interchangeable. Together, Bland-Altman and %PE analyses provide a robust framework for evaluating agreement and highlight where pooling data across alignment strategies may be justified and where it is not. Our findings underscore the importance of methodological consistency in clavicular cross-sectional analyses. Given the anatomical complexity of the clavicle and its developmental plasticity, alignment strategy exerts a significant influence on derived CSP values. We recommend that future studies clearly report alignment criteria and avoid direct comparisons between datasets processed with different protocols.