Aims. The purpose of this study was to develop a convolutional neural network (CNN) for fracture detection, classification, and identification of greater tuberosity displacement ≥ 1 cm, neck-shaft angle (NSA) ≤ 100°, shaft translation, and articular fracture involvement, on plain radiographs. Methods. The CNN was trained and tested on radiographs sourced from 11 hospitals in Australia and externally validated on radiographs from the Netherlands. Each radiograph was paired with corresponding CT scans to serve as the reference standard based on dual independent evaluation by trained researchers and attending orthopaedic surgeons. Presence of a fracture, classification (non- to minimally displaced; two-part, multipart, and glenohumeral dislocation), and four characteristics were determined on 2D and 3D CT scans and subsequently allocated to each series of radiographs. Fracture characteristics included greater tuberosity displacement ≥ 1 cm, NSA ≤ 100°, shaft translation (0% to < 75%, 75% to 95%, > 95%), and the extent of articular involvement (0% to < 15%, 15% to 35%, or > 35%). Results. For detection and classification, the algorithm was trained on 1,709 radiographs (n = 803), tested on 567 radiographs (n = 244), and subsequently externally validated on 535 radiographs (n = 227). For characterization, healthy shoulders and glenohumeral dislocation were excluded. The overall accuracy for fracture detection was 94% (area under the receiver operating characteristic curve (AUC) = 0.98) and for classification 78% (AUC 0.68 to 0.93). Accuracy to detect greater tuberosity fracture displacement ≥ 1 cm was 35.0% (AUC 0.57). The CNN did not recognize NSAs ≤ 100° (AUC 0.42), nor fractures with ≥ 75% shaft translation (AUC 0.51 to 0.53), or with ≥ 15% articular involvement (AUC 0.48 to 0.49). For all objectives, the model’s performance on the external dataset showed similar accuracy levels. Conclusion. CNNs proficiently rule out proximal humerus fractures on plain radiographs. Despite rigorous training methodology based on CT imaging with multi-rater consensus to serve as the reference standard, artificial intelligence-driven classification is insufficient for clinical implementation. The CNN exhibited poor diagnostic ability to detect greater tuberosity displacement ≥ 1 cm and failed to identify NSAs ≤ 100°, shaft translations, or articular fractures. Cite this article: Bone Joint J 2024;106-B(11):1348–1360

In this study, we describe a morphological classification for greater tuberosity fractures of the proximal humerus. We divided these fractures into three types: avulsion, split and depression. We retrospectively reviewed all shoulder radiographs showing isolated greater tuberosity fractures in a Level I trauma centre between July 2007 and July 2012. We identified 199 cases where records and radiographs were reviewed and included 79 men and 120 women with a mean age of 58 years (23 to 96). The morphological classification was applied to the first 139 cases by three reviewers on two occasions using the Kappa statistic and compared with the AO and Neer classifications. The inter- and intra-observer reliability of the morphological classification was 0.73 to 0.77 and 0.69 to 0.86, respectively. This was superior to the Neer (0.31 to 0.35/0.54 to 0.63) and AO (0.30 to 0.32/0.59 to 0.65) classifications. The distribution of avulsion, split and depression type fractures was 39%, 41%, and 20%, respectively. This classification of greater tuberosity fractures is more reliable than the Neer or AO classifications. These distinct fracture morphologies are likely to have implications in terms of pathophysiology and surgical technique.

Cite this article: Bone Joint J 2014;96-B:646–51.

Our aim was to determine the effect of the initial pattern of fracture and the displacement of fragments on the outcome of proximal humeral fractures treated conservatively. We followed 93 consecutive patients prospectively for one year. Final movement and strength were compared with those of the contralateral side. The final American Shoulder and Elbow Society score and the Disabilities of Arm, Shoulder and Hand and Short-Form 36 questionnaires were compared with those provided by the patient on the day of the injury. Radiographs and CT scans with three-dimensional reconstruction were obtained in all patients. The pattern of the fracture and the displacement of individual fragments were analysed and correlated with the final outcome. There were two cases of nonunion and six of avascular necrosis. The majority of the fractures (84 patients; 90%) followed one of the following four patterns: posteromedial (varus) impaction in 50 patients (54%), lateral (valgus) impaction in 13 (14%), isolated greater tuberosity in 15 (16%), and anteromedial impaction fracture in six (6%). Head orientation, impaction of the surgical neck and displacement of the tuberosity correlated strongly with the outcome.

In fractures with posteromedial impaction, a poor outcome was noted as the articular surface displaced inferiorly increasing its distance from the acromion. A poorer outcome was noted as a fractured greater tuberosity displaced medially overlapping with the posterior articular surface. Lateral impaction fractures had a worse outcome than other patterns of fracture.

The generally-accepted treatment for large, displaced fractures of the glenoid associated with traumatic anterior dislocation of the shoulder is operative repair. In this study, 14 consecutive patients with large (> 5 mm), displaced (> 2 mm) anteroinferior glenoid rim fractures were treated non-operatively if post-reduction radiographs showed a centred glenohumeral joint.

After a mean follow-up of 5.6 years (2.8 to 8.4), the mean Constant score and subjective shoulder value were 98% (90% to 100%) and 97% (90% to 100%), respectively. There were no redislocations or subluxations, and the apprehension test was negative. All fragments healed with an average intra-articular step of 3.0 mm (0.5 to 11). No patient had symptoms of osteoarthritis, which was mild in two shoulders and moderate in one.

Traumatic anterior dislocation of the shoulder, associated with a large displaced glenoid rim fracture can be successfully treated non-operatively, providing the glenohumeral joint is concentrically reduced on the anteroposterior radiograph.