LAUSR

OBJECTIVE In complicated Neer three- and four-part proximal humerus fracture (PHF), greater tuberosity (GT) fragments are often comminuted, and the currently widely used locking plate may not fix GT fragments effectively. A further understanding of morphological characteristics of the GT fragments may help explore new fixation devices. This study aimed to determine the fracture line morphology of the GT fragment of Neer three- or four-part PHF and analyze the location relationship between the locking plate and the GT fragment. METHODS Seventy-one three-dimensional computed tomography scans of Neer three- and four-part PHF were retrospectively reviewed between January 2014 and June 2019. Fracture fragments were reconstructed and virtually reduced in the Mimics software, and fracture lines of GT fragments were depicted on a humerus template in the 3-matic software and then were superimposed altogether. The common sites of the GT fracture were identified, and the location relationship between the locking plate and GT fragments was analyzed in a computer-simulated scenario. RESULTS The fracture line morphology of GT fragments was similar between Neer three- and four-part PHF. The overall morphology of GT fragments was in a fan shape, which could be summarized as anterior, superior, posterior, and middle lines. Of these, we identified 51 split and 29 avulsion type GT fragments based on the Mutch classification, and they could occur simultaneously in a PHF. The overall morphology of split type fragments was in a fan shape, and avulsion type fragments showed a quite distinguishable distribution pattern. A GT fragment could be classified as anterior-split, posterior-split, complete-split, anterior -avulsion, and posterior-avulsion type based on its morphology and location. The median percentage of fragment area covered by the plate was 32.3% in all of the fragments, and it was 69.4%, 23.0%, 37.2%, 21.8%, 0.0% in anterior-split, posterior-split, complete-split, anterior-avulsion, and posterior-avulsion type GT fragments. We defined the posterior-split, anterior-avulsion, and posterior-avulsion type GT fragments as the risky GT fragments, and they occurred in 43 (60.6%) Neer three- and four-part PHFs. CONCLUSION The fracture line morphology of GT fragments of Neer three- and four-part PHF was in a fan shape. GT fragments could be classified based on their location and morphology. The extent of GT fragment coverage provided by the locking plate differed in various fragment types, and we identified the anterior-avulsion, posterior-avulsion, and posterior-split type fragments as the risky GT fragments with a high incidence rate in Neer three- and four-part PHFs.