Abstract Liver cancer is a major global health concern associated with high mortality, and accurate early diagnosis is essential for improving treatment outcomes. The latest rich strides in deep learning have enhanced liver cancer diagnosis accuracy to a great extent. This review is a comprehensive study of literature on the use of deep learning algorithms since 2016 with a focus on the convolutional neural network (CNN), recurrent neural network, generative adversarial network and other deep neural networks to detect benign and malignant liver neoplasms. The results indicate the better results of CNN-based models in correctly classifying liver lesions with the help of ultrasound, computed tomography and magnetic resonance imaging and positron emission tomography imaging modalities. Additional developments in diagnostic accuracy are transfer learning and ensemble. Multimodal methods using imaging and clinical information also demonstrate a high potential of early detection. The clinical potential of deep learning in the diagnosis of liver cancer is obvious, it becomes possible to decrease inter- reader variability, it is possible to make image interpretation faster and more accurate and it is possible to detect malignant lesions earlier and more precisely. To bring this potential into routine care, future work should focus on multicenter validation, better model readability and richer, better annotated data. Deep learning has the potential to benefit liver oncology diagnostic processes and patient outcomes with these advancements.
Ruba et al. (Tue,) studied this question.
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: