Summary Chronic shoulder problems involving pain and loss of function affect 20% of adults (Pope, Croft et al., 1997). The proportion is even higher in adults over 50 years of age. In Québec, la Commission de la santé et de la sécurité au travail du Québec (CSST) reports that, for the years 2003 to 2006, shoulder injuries represented 39% of the 32,549 inflammatory injuries caused by repetitive work or trauma. Shoulder injuries required an average of 75.4 days of sick leave and $4,533 of income replacement indemnities (IRIs) per worker (CSST, 2007). By way of comparison, lower limb injuries represented less than 10% of the reported injuries and required an average of 39.5 days of sick leave and $2,439 of IRIs per worker (CSST, 2007). These figures indicate not only the magnitude of the problem but also that shoulder injuries require proportionally more time for the restoration of adequate function and for achieving a sustainable return to work. They also highlight the importance of being able to quickly identify individuals at risk of developing shoulder function problems in order to steer them toward effective adapted rehabilitation programs. The aim of this project was to develop a radiological index that correlates with shoulder function in workers with rotator cuff tears. The first part of the work involved developing a radiographic image acquisition method at different arm elevations in the scapular plane. We used the EOS™ low-dose medical imaging system equipped with ultra-sensitive sensors for this purpose. After generating radiographs, image analysis and 3D-reconstruction techniques were applied to the image sequence in order to reconstruct a morphologically realistic model of the scapula and a simplified model of the humerus and to study their relative movement.Validation of the method on 10 dry scapulae, five cadaveric shoulders, and five workers showed that glenohumeral (GH) translations could be evaluated with accuracy better than 1.6 mm and repeatability ranging from 1.8 to 3.6 mm. The application of this method to 41 workers with rotator cuff tears showed that there was a weak correlation between range of GH translation motions and different aspects of shoulder function. We were able to show, for example, a weak but significant correlation (R=0.322, p<0.02) between maximum upward displacement of the humeral head relative to the glenoid and the age- and sex-adjusted Constant score. We were also able to demonstrate that workers whose Constant score was lower than 50/100 had significantly different infero-superior and medio-lateral displacement (p<0.01) from pathological workers whose Constant score was higher than 50/100. In tandem with this work, we conducted an exploratory study of the MRI images acquired during the project. We adapted a method previously developed by Billuart et al. (2008) and came up with a method for calculating a ratio of coaptation/elevating forces in the fibres of the middle deltoid, using a 3D reconstruction of the muscle and bone structures of the shoulder complex. We were able to demonstrate, in 11 pathological workers and five healthy workers, that the anterior fibers of the middle deltoid had a bigger coaptation role than did the posterior fibres, which suggests that the anterior portion of the middle deltoid could potentially offset the rotator cuff deficiency and maintain good shoulder function despite the presence of a tear in the cuff. This project thus achieved a major advance in knowledge from both the technical and scientific points of view. First, it resulted in the development of an accurate, 3D, and minimally invasive morpho-functional analysis method for the shoulder. In the longer term, this method could be used in the clinical setting for morphological studies of the shoulder or pseudo-cinematic studies of the scapula. In addition, the project produced a morpho-muscular index that could also help explain loss of shoulder function in some workers. This study therefore represents another step forward in the morpho-functional study of the shoulder, which will undoubtedly enhance our understanding of shoulder pathologies and eventually improve the diagnosis, management, and quality of life of workers who suffer from shoulder pain.