A novel method for estimating sternoclavicular posterior rotation with promising accuracy: A validity comparison with minimizing acromioclavicular rotation approach.

The human shoulder complex's motion is modeled by nine rotational degrees of freedom (DoF) at the sternoclavicular (SC), acromioclavicular (AC), and glenohumeral joints. Non-invasive measurement of these rotations is desirable for shoulder kinematic assessment or musculoskeletal modeling. Accuracy of the conventional method for estimating SC posterior rotation is unclear and might be overestimated because it assumes no rotation in the AC joint. We aimed to explore whether our new method, allowing AC rotation, provides a more accurate estimation of SC posterior rotation than the conventional method. We compared estimates by both methods, in 18 postures among 8 healthy men, with those measured by the registration method from magnetic resonance images. Posthoc analyses showed significant differences between the registration and conventional methods in all 18 postures and in only one posture when compared to our method. While the conventional method tended toward overestimation and showed a 22.7° root-mean-square error for all postures, the new method had greater accuracy (6.8° root-mean-square error). By combining this method with the scapulothoracic rotation measurement method and other traditional methods, it should be possible to indirectly measure 3-DoF AC rotation, implying that non-invasive measurement of all 9-DoF rotations of the shoulder complex would now be possible.