A computer aided diagnosis system for measurement of mandibular cortical thickness on dental panoramic radiographs in prediction of women with low bone mineral density.

Abstract

Osteoporosis detection at earlier stages can enhance the life span of an elderly individual. The aim of the study is to perform semi-automated measurement of mandibular cortical thickness (MCT) on a dental panoramic radiograph (DPR) and thereby to predict the risk of low BMD among the studied population. The study involved 76 women (mean age: 57.2 ± 12.6 years). The DPR was obtained using KODAK 8000C system. The BMD of right total hip (T-BMD) was obtained using DPX Prodigy Dual-energy X-ray absorptiometry (DXA) Scanner. The DPR obtained were subjected to image processing techniques to perform MCT measurement. The region of interest was manually selected around the mental foramen and enhanced using a median filter. The Ostu segmentation was performed and connected component labelling operation was performed to determine the lower boundary by finding the contour with maximum area. Subsequently, the haar wavelet operation was carried out to find the magnitude and thereby select the upper delineating cortical boundary. The Pearson test results revealed (r = 0.96, p < 0.01) for the standard (manual) MCT measurement against the MCT measured using the proposed semi-automated scheme. ROC analysis revealed that MCT = 2.5 mm could be an optimal threshold in spotting individuals at risk of low BMD. The results of the study revealed that the MCT measured on a DPR using the proposed approach could be helpful for identifying individuals at risk of low BMD.