ObjectiveTo compare the difference of rotator cuff healing between different types of injury and between different repair methods, and to explore the animal model to accurately simulate the restorative process after repair of rotator cuff injury. MethodTwelve adult male beagle dogs (weighing, 10-15 kg) were divided into 3 groups (n=4) according to different processing methods:acute rotator cuff injury+Mason-Allen suture repair (group A), huge rotator cuff injury+Mason-Allen suture repair (group B), and huge rotator cuff injury+Mason-Allen combined with autogenous semitendinosus expansion suture repair (group C). The external fixation was used for immobilization after repair. After operation, the general situation of the animals was observed, and the infraspinatus tendon was harvested for gross observation at 6 weeks after operation. The biomechanical test of limit load and histological observation of tendon fibers were carried out. ResultsAll the animals survived to the end of the experiment. All incisions healed well and no infection occurred. Gross observation showed more scar tissues at the end of infraspinatus muscle tendon than normal tendon in group A; no obvious tendon tissue was observed at the end of infraspinatus muscle tendon in group B; the infraspinatus muscle tendon was covered with some white scar tissue, but the tendon and the general direction could be observed in group C. The limit load of groups A, B, and C were (223.75±24.28) , (159.25±34.87) , and (233.25±14.24) N respectively, group B was significantly lower than groups A and C (P<0.05) , and no significant differnce was found between group A and group C (P>0.05) . Histological observation showed normal arrangement of tendon fibers in group A; tendon fibers arranged disorderly in group B and tendon cells were significantly less than those of group A; tendon fibers arranged in neat in group C and tendon cells were more than those of group B. ConclusionsCanine autologous semitendinosus expansion repair of massive rotator cuff injury immobilization model can better simulate the clinical rotator cuff injury healing process, so it can be used as an ideal animal model for related research.