Automatic phantom-less calibration of routine CT scans for the evaluation of osteoporosis and hip fracture risk

Abstract

Background/Purpose. The diagnosis of osteoporosis remains a paramount concern for orthopedic surgeons worldwide. We aim to (1) evaluate the efficacy of automatic phantom-less quantitative computed tomography (PL-QCT) in diagnosing osteoporosis and (2) investigate its clinical value in predicting hip fracture risk. Methods: A cohort of 705 patients was included in the study. Hip CT scans from 310 patients and spinal CT scans from 315 patients were analyzed using automatic PL-QCT. The consistency of bone mineral density (BMD) measurement obtained by dual-energy X-ray absorptiometry (DXA), phantom-based QCT (PB-QCT), and automatic PL-QCT was examined through linear regression analysis and Bland-Altman plots. The ability of automatic PL-QCT to predict osteoporosis and hip fracture risk was assessed using ROC analysis. Results: Linear regression and Bland-Altman plots demonstrated a high level of agreement between BMD measurements from PL-QCT and those from hip DXA and lumbar PB-QCT. The AUC values for PL-QCT and PB-QCT in diagnosing osteoporosis were 0.903 (95 % CI 0.852–0.955) and 0.900 (95 % CI 0.847–0.953). The AUC values for predicting hip fracture risk, based on femoral neck BMD measured by PL-QCT and DXA, were 0.869 (95 % CI 0.823–0.915) and 0.831(95 % CI 0.778–0.885), respectively. When the femoral neck BMD was combined with the percentage of inter-muscular adipose tissue area, the AUC increased to 0.929 (95 % CI 0.897–0.961). Conclusion: Automatic PL-QCT has shown superior performance in predicting hip fracture risk compared to DXA. Furthermore, the novel PL-QCT demonstrates comparable predictive efficacy to that of PB-QCT, suggesting its potential as a valuable tool in clinical practice.