Anatomical variation in humeri: gender and side comparison using statistical shape modelling

Purpose: The surgical management of proximal humeral fractures remains challenging. Anatomical reduction of the fracture has been reported as the keystone for a sufficient surgical fixation and successful outcome. However, mostly there is no example of its premorbid state. Literature suggests that the mirrored contralateral side can be used as a reconstruction template. But is this a correct technique to use? The purpose of this study is to define anatomical variation between humeri based on gender and side comparison. Methods: Two different statistical shape models of the humerus were created and their modes of variation were described. One model contained 110 unpaired humeri. The other model consisted of 65 left and corresponding right humeri. Results: The compactness of the statistical shape model containing 110 humeri showed that two principal components explain more than 95% of the variation and the generalization showed that a random humerus can be described with an accuracy of 0.39 mm. For only three parameters, statistically significant differences were observed between left and right. However, comparing the mean of the different metrics on the humeri of men and women, almost all were significant. Conclusion: Since there were only small differences between left and right humeri, using the mirrored contralateral side as a reconstruction template for fracture reduction can be defended. The variable anatomy between men and women could explain why locking plates not always fit to the bone.