The Era of Immunotherapy in Hepatocellular Carcinoma: The New Mission and Challenges of Magnetic Resonance Imaging

SIMPLE SUMMARY: Clinicians should be aware that the immune microenvironment not only influences biological characteristics of HCC but also provides a reliable theoretical basis for comprehensively evaluating tumor responses after immunotherapy. Combining multi-parameter MRI with radiomics and deep learning can help to explore the biological characteristics of HCC. Future artificial intelligence technologies may extract biologically interpretable multidimensional structures from MRI images more effectively, thereby promoting precise and visual classification of HCC immune microenvironments. This advancement of MRI combined with multi-omics could potentially assist in individualized treatment planning. ABSTRACT: In recent years, significant advancements in immunotherapy for hepatocellular carcinoma (HCC) have shown the potential to further improve the prognosis of patients with advanced HCC. However, in clinical practice, there is still a lack of effective biomarkers for identifying the patient who would benefit from immunotherapy and predicting the tumor response to immunotherapy. The immune microenvironment of HCC plays a crucial role in tumor development and drug responses. However, due to the complexity of immune microenvironment, currently, no single pathological or molecular biomarker can effectively predict tumor responses to immunotherapy. Magnetic resonance imaging (MRI) images provide rich biological information; existing studies suggest the feasibility of using MRI to assess the immune microenvironment of HCC and predict tumor responses to immunotherapy. Nevertheless, there are limitations, such as the suboptimal performance of conventional MRI sequences, incomplete feature extraction in previous deep learning methods, and limited interpretability. Further study needs to combine qualitative features, quantitative parameters, multi-omics characteristics related to the HCC immune microenvironment, and various deep learning techniques in multi-center research cohorts. Subsequently, efforts should also be undertaken to construct and validate a visual predictive tool of tumor response, and assess its predictive value for patient survival benefits. Additionally, future research endeavors must aim to provide an accurate, efficient, non-invasive, and highly interpretable method for predicting the effectiveness of immune therapy.