應用於骨骼肌肉超音波之膝關節軟骨電腦輔助量化評估系統

Abstract

膝蓋退化性關節炎是年長者中最常見的關節炎疾病，會造成膝蓋疼痛、腫脹以及大大的降低病人的活動範圍。臨床診斷上，膝蓋軟骨的變異時常是退化性關節炎的主要因素。目前雖然X光是最常用來診斷退化性關節炎的工具，但其中也存在了許多限制以及X光診斷無法獲得足夠有關於膝軟骨的資訊。但是超音波卻可以輕易地對膝軟骨的厚度做量化計算，並透過評估軟骨附近的軟組織來判斷是否有發炎狀態，發炎可以視為是膝蓋退化性關節炎的前期徵兆。因此，這篇論文提出了一套應用於骨骼肌肉超音波並可自動偵測膝關節軟骨厚度的電腦輔助量化評估系統，此系統包含了對膝軟骨區域的全自動切割、更完善地找出膝軟骨上下邊界的處理，以及在膝軟骨區域的上下邊界間做精準的厚度量測，以協助骨科醫師評估病患膝軟骨的完善程度以及膝蓋退化性關節炎的嚴重程度。在量化評估系統的精確度及可靠度驗證中，共使用了67張影像進行測試與驗證，分別來自11個男性及6個女性病患。測試方法通過驗證比較系統自動產出的量化結果及骨科醫師所量測的正確結果，得知系統在厚度上的量測可達到83.69%的準確率，在量測點上的定位僅有0.842mm的誤差。因此，根據實驗結果，此系統對骨骼肌肉超音波之膝關節軟骨提供了可靠及穩定的軟骨厚度量測結果，並可如預期地克服了一些X光及超音波先天上的限制。

Knee osteoarthritis is the most common type of arthritis in the elderly, which causes knee pain, swelling, joint movement restriction to patients. Cartilage abnormalities are primary features of osteoarthritis. Although X-ray is the most common radiological examination for diagnosis of knee osteoarthritis, the radiographic diagnosis has limitations and few information of cartilage condition. Musculoskeletal ultrasound (US) is a useful technique to quantify the cartilage thickness and detect inflammation in the knee joint by evaluating the cartilage periarticular soft tissue, and inflammation was considered contribute to the progression of knee osteoarthritis. In this study, we proposed an automatic computer-aided quantitative assessment system for thickness measurement of knee cartilage in the US image. In the proposed system, cartilage area segmentation was presented to extract the cartilage area, and the boundary delineation was used to automatically detect and refine the upper and lower boundaries of knee cartilage. Then, thickness measurement was performed to measure the distance between the upper and lower boundaries of the cartilage area. In the experiment, the proposed assessment system was tested with a dataset of 67 cases, and the overall accuracy of thickness measurement achieved 83.69%. Also, the target position located by the proposed system had the overall mean difference of 0.842 mm by comparing to the ground truth measured by orthopedist. In conclusion, based on the experiment result, our system has ability to provide a confident and reliable thickness measurement of knee cartilage in musculoskeletal US, and to overcome the limitations of X-ray and ultrasound.