Texture analysis of mandibular cortical bone on digital dental panoramic radiographs for the diagnosis of osteoporosis in Korean women

Abstract

To determine whether individual measurements or a combination of textural features and mandibular cortical width (MCW) derived from digital dental panoramic radiographs (DPRs) are more useful in assessment of osteoporosis. Textural features were obtained by using fractal dimension (FD) and gray-level co-occurrence matrix (GLCM). Digital DPRs and bone mineral densities (BMDs) of the lumbar spine and the femoral neck were obtained from 141 female patients. A naïve Bayes classifier, a k-nearest neighbor (k-NN) algorithm, and a support vector machine were assessed for classifying osteoporosis. The combinations of FD plus MCW (95.3%, 92.1%, 96.8%) and GLCM plus MCW (93.7%, 89.5%, 94.2%) for femoral neck BMD showed the highest diagnostic accuracy with the use of the naïve Bayes, k-NN, and support vector machine classifiers, respectively. The combination of textural features and MCW contributed a better assessment of osteoporosis compared with the use of only individual measurements.