Algorithm and validation of a computer method for quantifying attachment locus of glenohumeral ligament in vivo

The aim of this work is to validate an algorithm that quantifies the locus of glenohumeral ligament (GHL) attachments on glenohumeral joint (GHJ) bones. A computed tomography scan of a GHJ was segmented to reconstruct the humerus, scapula, anatomical neck (AN) and glenoid rim (GR) into 3D meshes of interconnecting nodal vectors. These were applied to construct a 'clock face' coordinate system in which 3 o'clock points anteriorly. Based on the assigned clock face coordinate frame and the fitted plane, the error between the fitted plane and the actual bony node was quantified through manual data extraction. This was tested on 50 specimens. Mean algorithm quantification errors for GHL attachments were 4.8 (SD 2.2?mm) and 4.5?mm (1.7?mm) for the humerus and glenoid, respectively. Further studies would apply this to investigate GHL length changes during function and may suggest how these structures should be handled during surgical repairs.