Development of a toolkit to analyze lower limb radiographic imaging with knee osteoarthritis

摘要

Quantification of clinical parameters in osteoarthritis (OA) diagnosis is important for the health professionals' intervention. The purpose of this study was to develop a computational toolkit that quantifies morphologic, anthropomorphic and mathematical parameters of the lower limb to help those professional on periodical evaluation of patients. The developed toolkit is reliable, semiautomatic and reproductive. This toolkit considers the influence of image amplification and distortions. Its functions were validated in two different ways: through a non-anthropomorphic and an anthropomorphic calibration system of known dimensions. Accuracy and reproducibility were determined with the calibration systems prior to the use of the toolkit to analyze human X-rays images of OA patients. The tool allows a safe evaluation of the patient for two-dimensional images. Today the follow-up of OA is accomplished with computerized tomography, which increases the patient's absorbed dose. This is potentially harmful due to the increased radiosensitive effects of this radiation (possibly affecting blood production) increasing the risk of those patients that need periodical evaluations. The developed toolkit was able to quantify morphologic, anthropomorphic and mathematical parameters of the lower limb, reducing the subjectivity of the evaluation of bi-dimensional X-rays images.