Osteoarthritis (OA) is a prevalent chronic degenerative joint disease affecting loading-bearing joints like the knee and hip, with a significant global health burden. Current diagnostic methods are limited in detecting early OA or predicting its progression. Radiomics, a method that extracts high-dimensional imaging features from medical images, offers a promising solution by transforming qualitative image interpretation into quantitative data. This review summarizes the current development of radiomics in the intersection of engineering and medicine, discussing its applications and potential in OA diagnosis and prognosis. Radiomics involves extracting thousands of quantitative features from medical images, which are then analyzed using machine learning to distinguish OA types and stages. It has shown potential in improving early detection, tracking progression, and predicting treatment efficacy. While radiomics has been extensively studied in oncology, its application in musculoskeletal diseases, including OA, is still in its early stages. Radiomics analysis typically involves image acquisition, preprocessing, segmentation, feature extraction, dimension reduction, and model development. It has been applied to various OA joints, including the knee, hip, and temporomandibular joint (TMJ), with studies demonstrating its effectiveness in diagnosing and predicting OA. For example, a study using radiomics features from CT scans of 92 patients showed high accuracy in diagnosing TMJ OA. Another study combined radiomics features with clinical biomarkers to achieve a high area under the receiver operating characteristic curve (AUC) for diagnosing TMJ OA. Despite its potential, radiomics faces challenges such as limited sample sizes, variability in image acquisition parameters, and the need for standardized protocols. Future research should focus on improving the reproducibility and generalizability of radiomics studies, as well as exploring its applications in early OA detection and personalized treatment planning. The development of large-scale OA image databases and standardized feature definitions will be crucial for advancing radiomics in clinical practice. Overall, radiomics holds promise for enhancing the accuracy and efficiency of OA diagnosis and prognosis, contributing to better patient outcomes.Osteoarthritis (OA) is a prevalent chronic degenerative joint disease affecting loading-bearing joints like the knee and hip, with a significant global health burden. Current diagnostic methods are limited in detecting early OA or predicting its progression. Radiomics, a method that extracts high-dimensional imaging features from medical images, offers a promising solution by transforming qualitative image interpretation into quantitative data. This review summarizes the current development of radiomics in the intersection of engineering and medicine, discussing its applications and potential in OA diagnosis and prognosis. Radiomics involves extracting thousands of quantitative features from medical images, which are then analyzed using machine learning to distinguish OA types and stages. It has shown potential in improving early detection, tracking progression, and predicting treatment efficacy. While radiomics has been extensively studied in oncology, its application in musculoskeletal diseases, including OA, is still in its early stages. Radiomics analysis typically involves image acquisition, preprocessing, segmentation, feature extraction, dimension reduction, and model development. It has been applied to various OA joints, including the knee, hip, and temporomandibular joint (TMJ), with studies demonstrating its effectiveness in diagnosing and predicting OA. For example, a study using radiomics features from CT scans of 92 patients showed high accuracy in diagnosing TMJ OA. Another study combined radiomics features with clinical biomarkers to achieve a high area under the receiver operating characteristic curve (AUC) for diagnosing TMJ OA. Despite its potential, radiomics faces challenges such as limited sample sizes, variability in image acquisition parameters, and the need for standardized protocols. Future research should focus on improving the reproducibility and generalizability of radiomics studies, as well as exploring its applications in early OA detection and personalized treatment planning. The development of large-scale OA image databases and standardized feature definitions will be crucial for advancing radiomics in clinical practice. Overall, radiomics holds promise for enhancing the accuracy and efficiency of OA diagnosis and prognosis, contributing to better patient outcomes.