Using radiomic features of lumbar spine CT images to differentiate osteoporosis from normal bone density

Abstract

Objective This study aimed to develop a predictive model to detect osteoporosis using radiomic features from lumbar spine computed tomography (CT) images. Methods A total of 133 patients were included in this retrospective study, 41 men and 92 women, with a mean age of 65.45 ± 9.82 years (range: 31–94 years); 53 had normal bone mineral density, 32 osteopenia, and 48 osteoporosis. For each patient, the L1–L4 vertebrae on the CT images were automatically segmented using SenseCare and defined as regions of interest (ROIs). In total, 1,197 radiomic features were extracted from these ROIs using PyRadiomics. The most significant features were selected using logistic regression and Pearson correlation coefficient matrices. Using these features, we constructed three linear classification models based on the random forest (RF), support vector machine (SVM), and K-nearest neighbor (KNN) algorithms, respectively. The training and test sets were repeatedly selected using fivefold cross-validation. The model performance was evaluated using the area under the receiver operator characteristic curve (AUC) and confusion matrix. Results The classification model based on RF had the highest performance, with an AUC of 0.994 (95% confidence interval [CI]: 0.979–1.00) for differentiating normal BMD and osteoporosis, 0.866 (95% CI: 0.779–0.954) for osteopenia versus osteoporosis, and 0.940 (95% CI: 0.891–0.989) for normal BMD versus osteopenia. Conclusions The excellent performance of this radiomic model indicates that lumbar spine CT images can effectively be used to identify osteoporosis and as a tool for opportunistic osteoporosis screening.

Artificial Intelligence-Based Prediction of Lower Extremity Deep Vein Thrombosis Risk After Knee/Hip Arthroplasty

“…A comparative experiment was performed with the following models: eXtreme gradient boosting (XGBoost) model, 18 random forest (RF) model, 19 support vector machines (SVM) model, 20 logistic regression (LR) model, 21 and backpropagation neural network (BPNN) model. 22 In addition, each machine learning algorithm solve the problems in a slightly different way, and different algorithms may give different answers to the same problem, therefore, we adopted ensemble model by weighted voting (an ensemble of XGBoost, RF, SVM, and LR models), 23 and the highest probability among the four model predictions was defined as the final prediction result.…”

Development and validation of a machine learning-derived radiomics model for diagnosis of osteoporosis and osteopenia using quantitative computed tomography

“…In addition, features include both first-order features and grayscale features. Wavelet features could reflect the spatial heterogeneity of vertebral body synthetically cause they contain high-order image [36]. Studies have shown that both skewness and glrlm can reflect the heterogeneity of regions of interest [37][38][39].…”

Application of radiomics model based on lumbar computed tomography in diagnosis of elderly osteoporosis

“…This study utilizes deep learning networks to predict the risk of spinal fractures based on clinical features and image data. Xue et al 17 constructed three models to predict osteoporosis through segmentation and recognition of lumbar CT images. However, the CT images selected in these studies come from a mixed population of young and elderly people, and the extracted features are not representative enough.…”

“…16 This study utilizes deep learning networks to predict the risk of spinal fractures based on clinical features and image data. Xue et al 17 generation of vertebral osteophytes in elderly people is more severe, with different texture and filtering features. [16][17][18] Therefore, this study extracts the radiological characteristics of CT images of the elderly to evaluate the efficacy of the radiological model in the diagnosis of osteoporosis.…”