A common and difficult orthopedic injury, proximal humerus fractures frequently require surgery to restore functional shoulder motion. For optimal patient outcomes and to inform treatment choices, accurate evaluation of fracture stability and healing progress is essential. A thorough combination of the use of computational loads through finite element modeling (FEM) with innovative imaging methods to evaluate proximal humerus fractures was used to predict their fracture locations based on volumetric bone mineral density (vBMD).

To quantify fracture characteristics, vBMD, and morphological changes related to proximal humerus fractures, 2 different imaging modalities were used. The effectiveness of single energy quantitative (SEQCT) and dual-energy quantitative computed tomography (DEQCT) was compared. BONE and STD reconstruction kernels, an algorithmic procedure, were parameters that were changed post-processing to alter the frequency content of images obtained from the SEQCT and DEQCT scanners. These imaging techniques enable a multidimensional study, giving rise to a deeper comprehension of fracture locations, bone quality, and provide preliminary steps to inform risk factors.

To achieve this objective, fourteen cadaveric shoulders (n = 7 left, n = 7 right) were scanned under each scanning modality and post-processed into both BONE and STD reconstruction kernel, respectively. Four different images were created per shoulder scan:​ SEQCT BONE/STD and DEQCT BONE/STD, creating a total of 56 different possible images. Density in volumetric bone mineral density (vBMD) was calculated in the humeral head, metaphysis, and diaphysis regions, for each imaging modality. Image processing software was utilized to create 3D models of the humerus and highlighting the selected regions. Specimen-specific slope/intercept was used to convert from native Hounsfield Units (HU) to equivalent vBMD [mgK2HPO4/cm3]. Strain measurements were calculated using FEMs derived from each proximal humerus model and are reported as maximum and minimum principal strain through visual representations and histograms according to each region, showcasing differences in them between BONE and STD reconstruction kernel.

Fracture location predictions are possible from the FEM visual representation in both maximum/minimum principal strain with the aid of the figure showcasing strain levels in the three regions. These qualitative and respective quantitative data are cadaver dependent as each cadaver was influenced according to anatomical differences such as bone quality, comorbidities, age, patient activity levels and more. The FE models generated showcased that 7 out of the 7 models generated from SEQCT predict fracture location in the metaphysis region and 7 out of the 7 models generated from DEQCT also predict fracture location in the metaphysis region.

As they clarify the significance of fracture morphology, bone quality, and loading circumstances on fracture location, the study's findings shed light on the relationship between imaging parameters and biomechanics of proximal humerus fractures. This research also seeks to assist physicians in choosing the best imaging approach for precise fracture characterization and treatment planning by comparing the performance of several imaging modalities.

Ultimately, the combination of innovative imaging methods with finite element modeling can increase our ability to understand proximal humerus fractures by providing us knowledge on the optimal imaging modality, with the long-term objective of enabling better clinical outcomes for patients with these injuries through improved diagnostic and treatment options.