Application of Artificial Intelligence in Abdominal Imaging

There is a wide range of abdominal diseases, and imaging examination plays an important role in the diagnosis, selection of treatment options and monitoring of efficacy in abdominal diseases. Advances in conventional imaging techniques based on morphological assessment have provided much valuable clinical information but have inherent limitations, although the accuracy of imaging assessment is influenced by the diagnostic level of the radiologist. In recent years, with the rapid development of artificial intelligence (AI) technology, especially extensive research on deep learning (DL) algorithms, AI has been applied to clinical work in areas such as lung nodule detection. However, the presence of low contrast between lesions and background images on abdominal images and the need for multiple phases make it relatively difficult for AI to detect and diagnose abdominal lesions. There are currently preliminary studies using AI to detect and diagnose pancreatic, gastric and colorectal cancers. Applying AI to the abdomen is more focused on pathological grading, prognosis and efficacy evaluation of abdominal neoplastic diseases. Many valuable results have been obtained in the study of diseases of the cavernous and parenchymal organs of the digestive system and diseases of the genitourinary system, showing the potential for clinical application.