The Development of a Quantitative Method for the Detection of Periarticular Osteoporosis Using Density Features within ROIs from Computed Radiography Images of the Hand
Periarticular osteoporosis of the hands and feet is one of the major diagnostic criteria for rheumatoid arthritis (RA). However, a quantitative method to detect periarticular osteoporosis using radiographs has not been reported. In this chapter, the authors propose a quantitative method for the detection of periarticular osteoporosis using density features of regions of interest (ROIs) from computed radiography (CR) images of the hand. The proposed method measures the density features of ROIs using histogram analysis, co-occurrence matrices, Fourier analysis, and the extraction of line components. Periarticular osteoporosis is detected using a discernment function based on these measurements. The sensitivity and specificity of the proposed method was assessed using 188 joints from 17 cases, including 6 normal cases (without periarticular osteoporosis) and 11 abnormal cases (with periarticular osteoporosis). The sensitivity of the method was 88.9%, and the specificity was 98.1%. Therefore, the authors consider this method to be potentially useful to radiologists for detecting periarticular osteoporosis in the hands.