Knee osteoarthritis (OA)   is a common joint disease that causes pain, deformity, and dysfunction in the bones, joint membranes, and surrounding ligaments that form joints due to gradual damage or degenerative changes in the articular cartilage. Knee OA is one of the leading causes of disability in the elderly and is the most common form of OA[1]. The standard method for measuring the severity of knee OA using X-ray images is the Kellgren – Lawrence (KL) grading system, which uses five grades ranging from 0 to 4 according to severity[2]. Source Review In recent years, deep learning has shown promising results in interpreting radiographic images for the diagnosis of knee OA[2]. Several studies have employed deep neural networks in automatically predicting KL grade from plain knee joint radiographs[1–6]. However, these studies treat KL grading as a multi-class classification task and ignore the inherent ordinal nature within the KL grades [6]. In this study, the authors proposed an ord