Focussed classification of scapula fractures: failure of the lateral scapula suspension system

Following an increase in the incidence of scapular fractures and interest in the outcome of their treatment, a basic classification system was developed for ease of use in the emergency setting. It has been expanded to a comprehensive system to allow for more in-depth classification of scapular fractures for clinical research and surgical decision making. It focusses on three specific regions of the scapula: the scapular body, the glenoid fossa and the lateral scapular suspension system (LSSS). This article presents a classification of the LSSS involvement to better characterise the injuries of this region and to emphasise its relevance to evaluation of the position of the scapula, hence the glenoid fossa, and so the centre of rotation of the shoulder joint.

An iterative consensus and evaluation process comprising an international group of seven experienced shoulder specialist and orthopaedic trauma surgeons was used to specify and evaluate the failure of the LSSS associated with scapula fractures. This was supported by a series of agreement studies. The system considered lack of involvement (S0), incomplete (S1) and complete (S2) failure of the LSSS. The last evaluation was conducted on a consecutive collection of 120 scapula fractures documented by three-dimensional computed tomography (3D CT) reconstruction videos.

Surgeons agreed on the involvement/failure of the LSSS in 47% of the 120 cases with an overall Kappa of 0.54. The sample most likely included 70 S0, 29 S1 and 21 S2 cases, where surgeons showed median classification accuracies of 93%, 71% and 80% for these categories, respectively. While two surgeons showed some uncertainty about their classification, the remaining surgeons only failed to identify LSSS failure in <20% of the cases. Kappa coefficients of reliability for classification of incomplete and complete LSSS involvement according to subcategories were 0.85 and 0.82, respectively.

While LSSS involvement can be reliably identified, its characterisation regarding complexity is problematic even with 3D CT images. The proposed LSSS system is considered clinically relevant and sufficient to further assess its role in treatment-decision processes and outcome prognosis.