Semi-automatic proximal humeral trabecular bone density assessment tool: technique application and clinical validation.

PURPOSE: This study aimed to apply a newly developed semi-automatic phantom-less QCT (PL-QCT) to measure proximal humerus trabecular bone density based on chest CT and verify its accuracy and precision. METHODS: Subcutaneous fat of the shoulder joint and trapezius muscle were used as calibration references for PL-QCT BMD measurement. A self-developed algorithm based on a convolution map was utilized in PL-QCT for semi-automatic BMD measurements. CT values of ROIs used in PL-QCT measurements were directly used for phantom-based quantitative computed tomography (PB-QCT) BMD assessment. The study included 376 proximal humerus for comparison between PB-QCT and PL-QCT. Two sports medicine doctors measured the proximal humerus with PB-QCT and PL-QCT without knowing each other's results. Among them, 100 proximal humerus were included in the inter-operative and intra-operative BMD measurements for evaluating the repeatability and reproducibility of PL-QCT and PB-QCT. RESULTS: A total of 188 patients with 376 shoulders were involved in this study. The consistency analysis indicated that the average bias between proximal humerus BMDs measured by PB-QCT and PL-QCT was 1.0 mg/cc (agreement range - 9.4 to 11.4; P > 0.05, no significant difference). Regression analysis between PB-QCT and PL-QCT indicated a good correlation (R-square is 0.9723). Short-term repeatability and reproducibility of proximal humerus BMDs measured by PB-QCT (CV: 5.10% and 3.41%) were slightly better than those of PL-QCT (CV: 6.17% and 5.64%). CONCLUSIONS: We evaluated the bone quality of the proximal humeral using chest CT through the semi-automatic PL-QCT system for the first time. Comparison between it and PB-QCT indicated that it could be a reliable shoulder BMD assessment tool with acceptable accuracy and precision. This study developed and verify a semi-automatic PL-QCT for assessment of proximal humeral bone density based on CT to assist in the assessment of proximal humeral osteoporosis and development of individualized treatment plans for shoulders.

A Deep-Learning Model for Diagnosing Fresh Vertebral Fractures on Magnetic Resonance Images.

BACKGROUND: The accurate diagnosis of fresh vertebral fractures (VFs) was critical to optimizing treatment outcomes. Existing studies, however, demonstrated insufficient accuracy, sensitivity, and specificity in detecting fresh fractures using magnetic resonance imaging (MRI), and fall short in localizing the fracture sites. METHODS: This prospective study comprised 716 patients with fresh VFs. We obtained 849 Short TI Inversion Recovery (STIR) image slices for training and validation of the AI model. The AI models employed were yolov7 and resnet50, to detect fresh VFs. RESULTS: The AI model demonstrated a diagnostic accuracy of 97.6% for fresh VFs, with a sensitivity of 98% and a specificity of 97%. The performance of the model displayed a high degree of consistency when compared to the evaluations by spine surgeons. In the external testing dataset, the model exhibited a classification accuracy of 92.4%, a sensitivity of 93%, and a specificity of 92%. CONCLUSIONS: Our findings highlighted the potential of AI in diagnosing fresh VFs, offering an accurate and efficient way to aid physicians with diagnosis and treatment decisions.