Abstract: Radiomics can convert medical images into minable data, which extracts high-throughput features from the images. By machine learning or statistical methods, key radiomics features were selected and used for model development to aid in clinical decision making on precise diagnosis, treatment evaluation, outcome prediction, etc. Though a lot of studies have demonstrated the useful potential of radiomics in clinical management, radiomics is still facing a lot of challenges such as data standardization, model validation before it could be implemented in clinical practice. In this short-review, clinical values and challenges will be briefly addressed.