RESUMO
Esophageal cancer is a tumor with high morbidity and mortality in China, which is generally diagnosed at late stage and yields poor prognosis. Early diagnosis and correct staging are the basis, and reasonable treatment is the most important. Radiomics can make use of existing imaging resources for deeper mining, and make secondary use of its potential high-throughput data through deep learning or machine learning, thereby establishing a radiomics prediction model. This may become an essential marker of tumor prognosis to predict overall survival or tumor progression, thus stratifying patients at different risk for individualized treatment. In this article, the basic concepts of radiomics, its application in prognostic prediction of esophageal cancer and its combination with clinical and genetic studies were reviewed.