Advances in Artificial Intelligence for the Radiological Diagnosis of Hepatocellular Carcinoma

Abstract

Hepatocellular carcinoma (HCC) is the third most common cause of cancer-related deaths globally. Unlike most other cancers, HCC can be diagnosed solely on imaging for high-risk patients. However, this is frequently complicated by atypical or indeterminate features necessitating biopsy or close follow-up with serial imaging. Artificial intelligence (AI) has the potential to allow for more accurate tumour classification and, thus, avoid unnecessary biopsies. Additionally, earlier diagnosis opens up the potential for curative therapies and improves patient outcomes. A number of artificial intelligence models, including machine learning, convolutional neural networks and radiomics-based models have been tested on ultrasound, CT and MRI images of liver lesions. The following review will outline the most impactful papers in this field.