| Objective: Osteoporosis vertebral compressive fracture(OVCF)is one of the sever complication in patients with osteoporosis.There was limitation in finding the risk factor of OVCF in previous evaluation technique.In this study,we developed a radiomics model(R-model)based on lumbar CT scan to assess the risk of OVCF in postmenopausal women.。Methods: Radiographic data of postmenopausal women patients in our hospital from January 2021 to August 2022 was collected.All the patients received both dual-energy X-ray absorptiometry(DEXA)and lumbar CT scan.329 vertebral bodies without compressive fracture in lumbar 1 to 4 were extracted by 3D slicer and randomly divided into train group(n=230)and test group(n=99).129 radiomics features was automatically calculated by pyradiomics module and minimum-Redundancy Maximum-Relevancy(m RMR),least absolute shrinkage and selection operator(LASSO)were used to shrink features for R-model construction.The sensitivity,specificity,accuracy,and area under the receiver operating characteristic curve(AUC)of the T-score and R-score were calculated.The AUCs of the two models were compared using the De Long test.Decision curve analysis(DCA)shew the clinical usefulness of R-model.Results: Eight features were chosen to construct the R-model.The AUCs of the T-score and R-score in training group were 0.845 and 0.945 respectively and were 0.818 and 0.914 respectively in test group.There was significant difference(p<0.001)in the AUC of two models and the decision curve analysis(DCA)shown the R-model has better overall net benefit than T-score model.Conclusions:The radiomics model based on lumbar CT scan can identify and predict high risk patients of OVCF with better sensitivity and accuracy even if in the patients with the same T-score in DEXA. |
