Objective To simulate anterosuperior instabil ity of the shoulder by a combination of massive irreparable rotator cuff tears and coracoacromial arch disruption in cadaveric specimens, use proximally based conjoined tendon transfer forcoracoacromial l igament (CAL) reconstruction to restrain against superior humeral subluxation, and investigate its feasibility and biomechanics property. Methods Nine donated male-adult and fresh-frozen cadaveric glenohumeral joints were applied to mimic a massive irreparable rotator cuff tear in each shoulder. The integrity of the rotator cuff tendons and morphology of the CAL were visually inspected in the course of specimen preparation. Cal ipers were used to measure the length of the CAL’s length of the medial and the lateral bands, the width of coracoid process and the acromion attachment, and the thickness in the middle, as well as the length, width and thickness of the conjoined tendon and the lateral half of the removed conjoined tendon. The glenohumeral joints were positioned in a combination of 30° extension, 0° abduction and 30° external rotation. The value of anterosuperior humeral head translation was measured after the appl ication of a 50 N axial compressive load to the humeral shaft under 4 sequential scenarios: intact CAL, subperiosteal CAL release, CAL anatomic reattachment, entire CAL excision after lateral half of the proximally based conjoined tendon transfer for CAL reconstruction. Results All specimens had an intact rotator cuff on gross inspection. CAL morphology revealed 1 Y-shaped, 4 quadrangular, and 4 broad l igaments. The length of the medial and lateral bands of the CAL was (28.91 ± 5.56) mm and (31.90 ± 4.21) mm, respectively; the width of coracoid process and acromion attachment of the CAL was (26.80 ± 10.24) mm and (15.86 ± 2.28) mm, respectively; and the thickness of middle part of the CAL was (1.61 ± 0.36) mm. The length, width, and thickness of the proximal part of the proximally based conjoined tendon was (84.91 ± 9.42), (19.74 ± 1.77), and (2.09 ± 0.45) mm, respectively. The length and width of the removed lateral half of the proximally conjoined tendon was (42.67 ± 3.10) mm and (9.89 ± 0.93) mm, respectively. The anterosuperior humeral head translation was intact CAL (8.13 ± 1.99) mm, subperiosteal CAL release (9.68 ± 1.97) mm, CAL anatomic reattachment (8.57 ± 1.97) mm, and the lateral half of the proximally conjoined tendon transfer for CAL reconstruction (8.59 ± 2.06) mm. A significant increase in anterosuperior migration was found after subperiosteal CAL release was compared with intact CAL (P lt; 0.05). The translation after CAL anatomic reattachment and lateral half of the proximally conjoined tendon transfer for CAL reconstruction increased over intact CAL, though no significance was found (P gt; 0.05); when they were compared with subperiosteal CAL release, the migration decreased significantly (P lt; 0.05). The translation of lateral half of the proximally conjoined tendon transfer for CAL reconstruction increased over CAL anatomic reattachment, but no significance was evident (P gt; 0.05). Conclusion The CAL should be preserved or reconstructed as far as possible during subacromial decompression, rotator cuff tears repair, and hemiarthroplasty for patients with massive rotator cuff deficiency. If preservation or the insertion reattachment after subperiosteal release from acromion of the CAL of the CAL is impossible, or CAL is entirely resected becauseof previous operation, the use of the lateral half of the proximally based conjoined tendon transfer for CAL reconstruction isfeasible.