Background: The role of altered joint mechanics on cartilage degeneration in in vivo models has not been studied successfully due to a lack of pre-injury information. We aimed 1) to develop an accurate in vivo canine model to measure the changes in joint loading and T2 star (T2*) relaxation time before and after unilateral supraspinatus tendon resections, and 2) to find the relationship between regional variations in articular cartilage loading patterns and T2* relaxation time distributions. Methods: Rigid markers were implanted in the scapula and humerus of tested dogs. The movement of the shoulder bones were measured by a motion tracking system during normal gaits. In vivo cartilage contact strain was measured by aligning 3D shoulder models with the motion tracking data. Articular cartilage T2* relaxation times were measured by quantitative MRI scans. Articular cartilage contact strain and T2* relaxation time were compared in the shoulders before and 3 months after the supraspinatus tendon resections. Results: Excellent accuracy and reproducibility were found in our in vivo contact strain measurements with less than 1% errors. Changes in articular cartilage contact strain exhibited similar patterns with the changes in the T2* relaxation time after resection surgeries. Regional changes in the articular cartilage T2* relaxation time exhibited positive correlations with regional contact strain variations 3 months after the supraspinatus resection surgeries. Conclusion: This is the first study to measure in vivo articular cartilage contact strains with high accuracy and reproducibility. Positive correlations between contact strain and T2* relaxation time suggest that the articular cartilage extracellular matrix may responds to mechanical changes in local areas.