The Evolving Role of Imaging in Hepatocellular Carcinoma: From Pathomolecular Profiling to Prognostic Decision-Making

Abstract

Background: Hepatocellular carcinoma (HCC), which represents the most common type of primary liver cancer, is increasingly perceived as a complex ecosystem with marked spatial and temporal heterogeneity as well as varied sensitivities to treatment. Recent revolutions in the management of HCC, particularly the introduction of several immune checkpoint inhibitor-based regimens, are placing pressing needs on more tailored imaging-aided prognostic decision-making for individualized treatment selection beyond the current "one-size-fits-all" tumor burden measurement.

Summary: With an accumulating number of advanced imaging and artificial intelligence techniques bridging the gap between preclinical and clinical applications, the role of imaging in HCC is rapidly expanding from conventional surveillance, diagnosis, staging, and treatment-response evaluation to personalized pathomolecular profiling, prognostication, and therapeutic decision-making. Ultimately, imaging may direct the selection of treatment modalities precisely tailored to individual patients and tumors.

Key messages: In this review, we describe the evolving role of imaging in the noninvasive assessment of key pathomolecular drivers of outcomes in HCC, outline the applications of imaging in prognostication, risk stratification, and selection of major treatment approaches, as well as discuss unmet needs and potential future directions.

Keywords: Carcinoma, hepatocellular; Imaging; Individualized treatment; Pathomolecular profiling; Prognostication.

Fig. 1.

Summary of the primary applications and challenges of imaging for individualized management in HCC. AI, artificial intelligence; EBRT, external beam radiation therapy; HCC, hepatocellular carcinoma; LT, liver transplantation; MVI, microvascular invasion; SR, surgical resection; TACE, transarterial chemoembolization; TARE, transarterial radioembolization; TLSs, tertiary lymphoid structures; tx, treatment; VETC, vessels encapsulating tumor clusters.

Fig. 2.

Imaging-aided treatment decision-making algorithm modified from the 2022 BCLC system. ^aUnfavorable imaging findings include semantic (e.g., presence of rim APHE), quantitative (e.g., decreased ADC value of diffusion-weighted MRI), metabolic (e.g., elevated standard uptake value_max on PET), radiomic, and DL features. Unfavorable imaging findings may be associated with various characteristics, such as poor histoprognostic factor (e.g., tumor grade, pathological subtype), local invasiveness (e.g., MVI, satellite nodules), molecular profile (e.g., TP53 mutation), or metastatic risk (e.g., vessels encapsulating tumor clusters). They may be different for curative-intent treatments, IAT, and systemic treatment. Solid arrows, LoE 3 (retrospective cohort studies); dashed arrows, LoE 5 (expert opinions). ^bUnfavorable imaging findings have shown promise to serve as a decision-making tool in this context, but consensus has not been reached across institutions and the role of imaging is largely affected by local practices. ^cIAT primarily referred to TACE and transarterial radioembolization. Dashed boxes, optional treatment (LoE 2 [randomized-controlled trials] but not yet incorporated into the BCLC system). ADC, apparent diffusion coefficient; AFP, α-fetoprotein; APHE, arterial phase hyperenhancement; BCLC, Barcelona Clinic Liver Cancer; CT, computed tomography; DL, deep learning; IAT, intra-arterial therapy; LoE, level of evidence based on the Oxford Centre for Evidence-Based Medicine; LR, liver resection; LT, liver transplantation; MRI, magnetic resonance imaging; MVI, microvascular invasion; PET, positron-emission tomography; PIVKA-II, protein induced by vitamin K absence or antagonist-II; PS, performance status; TACE, transarterial chemoembolization.