I-Cheng Lee

Hepatocellular carcinoma (HCC) remains a leading cause of cancer-related mortality worldwide, largely due to high recurrence rates after curative treatment and poor prognosis in advanced-stage disease. Multiphasic computed tomography (CT) and magnetic resonance imaging (MRI) play central roles in the noninvasive diagnosis and management of HCC; however, diagnostic variability and limited sensitivity for early-stage and atypical lesions continue to pose major clinical challenges. In recent years, artificial intelligence (AI) has emerged as a powerful tool for enhancing medical image analysis and improving clinical decision-making in liver oncology. This review summarizes recent advances in AI-based applications for HCC imaging using CT and MRI. We discuss the roles of AI in lesion detection and segmentation, differential diagnosis and tumor classification, imaging-based biological characterization, and prediction of treatment response and prognosis. In particular, radiomics and deep learning approaches have enabled automated extraction of high-dimensional imaging features, facilitating noninvasive assessment of tumor aggressiveness and heterogeneity. The integration of multimodal data, including imaging, clinical, and molecular information, further enhances predictive performance and personalized risk stratification. Despite encouraging results, most existing studies are limited by retrospective designs, heterogeneous methodologies, and insufficient external validation. Challenges related to model interpretability, data standardization, and clinical integration also remain unresolved. We highlight current limitations and future research directions, emphasizing the need for multi-center collaboration, prospective validation, and development of transparent and regulatory-compliant systems. With continued technological innovation and interdisciplinary collaboration, AI-driven imaging analysis is expected to play an increasingly important role in precision medicine and outcome optimization for patients with HCC.

Introduction: Evidence supporting the combination of immunotherapy with intra-arterial therapy (IAT) for unresectable hepatocellular carcinoma (HCC) remains limited. This study aimed to evaluate whether the addition of IAT to first-line immunotherapy improves survival outcomes. Methods: Using the Taiwan Liver Cancer Association (TLCA) Research Group database, we identified patients with unresectable HCC who initiated first-line immunotherapy between May 2017 and January 2024. Patients with Barcelona Clinic Liver Cancer (BCLC) stage A or D disease, prior liver transplantation, or follow-up of 1 (aHR 1.58, 95% CI: 1.11–2.27; p = 0.012), tumor burden beyond the up-to-7 criteria (aHR 1.67, 95% CI: 1.02–2.74; p = 0.043), and portal vein invasion (aHR 1.67; 95% CI: 1.19–2.35; p = 0.003). Five-year OS was higher in the IAT group than in the control group (28.7% [95% CI: 17.8–46.2] vs. 16.4% [95% CI: 10.7–25.2]; p = 0.007). Landmark and subgroup analyses consistently supported the OS benefit of IAT. Conclusion: In unresectable HCC, the addition of IAT to first-line immunotherapy improves response rates, prolongs the duration of response, and enhances OS.

Background: Recurrence rates following thermal ablation for hepatocellular carcinoma (HCC) remain high, but the patterns of recurrence and post-recurrence outcomes are not well characterized. This study aimed to investigate the recurrence patterns and long-term post-recurrence survival (PRS) in patients with HCC after thermal ablation to inform post-recurrence treatment strategy. Methods: A retrospective analysis was conducted on 824 patients who underwent thermal ablation for HCC between 2007 and 2023. Recurrence patterns and factors influencing PRS in patients with recurrence within and beyond the Milan criteria were evaluated. An independent cohort of 198 patients served as an external validation cohort for the prognostic models. Results: During a median follow-up of 54.5 months, 536 patients experienced HCC recurrence, with 83.8% within and 16.2% beyond Milan criteria. For patients with recurrence within Milan criteria, early recurrence, recurrent tumor size, AFP, ALBI grade, and FIB-4 score were independent predictors of PRS. In patients with recurrence beyond Milan criteria, diabetes mellitus, macrovascular invasion, AFP, ALBI grade, and FIB-4 score independently predicted PRS. Based on PRS predictors, a risk model stratified patients with recurrence within Milan criteria into four risk groups, with median PRS of 103.7, 65, 48.7, and 28.6 months, respectively (p < 0.001). For recurrence beyond Milan criteria, a separate risk model classified patients into three risk groups, showing median PRS of 70.1, 24.7, and 8.1 months, along with probabilities of successful downstaging of 66.7%, 39.3%, and 0%, respectively (p < 0.001). The external validation results showed that both the Milan-in and Milan-out models demonstrated significant discriminative performance in the validation cohort. Conclusions: PRS in patients with recurrent HCC after thermal ablation is significantly influenced by recurrence patterns, tumor characteristics, and host factors. These findings may guide post-recurrence treatment strategies and optimize the timing of salvage liver transplantation.

Background and Aims: Gut microbiota could modulate the outcome of cancer immunotherapy, including HCC. Although metabolic dysfunction–associated steatotic liver disease-related HCC (MASLD-HCC) and viral hepatitis-related HCC (V-HCC) exhibit different tumor microenvironments and gut microbiome composition, both respond to combination immunotherapy. This study aimed to investigate whether distinct microbiota/metabolomic signatures are associated with outcomes of combination immunotherapy in MASLD-HCC and V-HCC, separately. Approach and Results: Between January 2021 and April 2024, 77 V-HCC patients and 25 with MASLD-HCC initiating first-line combination immunotherapy were prospectively enrolled. Pre-treatment fecal microbiota/metabolites and serum cytokines/chemokines were analyzed in association with durable tumor response, overall survival (OS) and progression-free survival (PFS). MASLD-HCC patients showed predominant fecal Bacteroides ovatus, Kluyvera georgiana, Klebsiella oxytoca, and Enterococcus faecium, with lower levels of short-chain fatty acids (SCFAs) and ursodeoxycholic acid (UDCA) compared to V-HCC. Durable responders (DRs) of MASLD-HCC had enriched Mediterraneibacter gnavus ATCC 29149 and significantly higher SCFAs (acetate, propionate, butyrate, and isobutyrate) and UDCA. In contrast, V-HCC-DRs were characterized by predominant Bifidobacterium and similarly enriched SCFAs. In both MASLD-HCC and V-HCC, fecal acetate level was a common significant predictor of DR, PFS, and OS. Patients with higher acetate levels had significantly longer OS (median: 25.2 vs. 11.3 mo; pp<0.001). Conclusions: Distinct gut microbiota, but shared beneficial metabolites, particularly fecal acetate, are associated with durable response to combination immunotherapy and improved survival in both MASLD-HCC and V-HCC. Fecal acetate may serve as a potential biomarker and therapeutic target for optimizing HCC treatment.

Background: Notable advances have been made in immune checkpoint inhibitors (ICIs) for cancer treatment. However, the adverse effects of ICIs, especially hepatotoxicity, remain a challenging problem. Whether patients in hepatitis B virus (HBV)‐endemic areas are prone to developing hepatic adverse events during ICI treatment warrants further exploration.Methods: From 2014 to 2020, the data of all patients with cancer who received ICI treatment at Taipei Veterans General Hospital were retrospectively reviewed. The incidence of and risk factors for hepatic adverse events, including hepatitis flare, immune‐related hepatitis (irHepatitis) and HBV reactivation (HBVr), were analysed through a Cox proportional hazard regression model.Results: A total of 1283 patients with cancer (190 hepatocellular carcinoma [HCC] patients and 1093 patients with non‐HCC malignancies) were eligible for analysis, of whom 283 (22.1%) were HBsAg‐positive. The incidence of hepatitis flare events of any grade was significantly higher in HCC patients than in non‐HCC patients (45.8% vs. 25.6%, p < 0.001). HCC and baseline alanine aminotransferase (ALT) > 40 U/L were independent risk factors for ≥ grade 3 hepatitis flare events. No difference was observed in irHepatitis risk between HCC patients and non‐HCC patients. ALT > 40 U/L was an independent risk factor for irHepatitis. Among 283 HBsAg‐positive patients, six patients (2.1%) experienced HBVr. HCC patients had a higher risk of HBVr than non‐HCC patients (4.4% vs. 0.6%). No specific risk factor for HBVr could be identified. However, none of the patients under nucleos/tide analogue (NUC) prophylaxis experienced HBVr in this study.Conclusions: Under ICI treatment, HCC patients had a higher risk of hepatitis flare events than non‐HCC patients. Abnormal baseline ALT levels are a risk factor for hepatic adverse events. NUC prophylaxis can minimise the risk of HBVr.

Purpose: Diabetes mellitus (DM) negatively impacts chronic hepatitis B patients, but its role in those with HBV-related hepatocellular carcinoma (HCC) undergoing ablation remains unclear. This study aims to evaluate the influence of DM on recurrence patterns and overall survival (OS) among patients with HBV-related HCC undergoing ablation. Patients and Methods: We retrospectively enrolled 372 patients receiving thermal ablation for HBV-related HCC, including 96 (25.8%) patients with DM. Factors associated with local tumor progression (LTP), distant recurrence, and OS were analyzed. The prognostic value of DM in IMbrave050-defined high-risk population was validated. Results: DM did not correlate with LTP, whereas patients with DM had significantly higher risk of distant recurrence (median time to recurrence 23.7 versus 46.2 months, p=0.032), poorer OS (median OS 75.6 versus 106 months, p=0.011), and poorer post-recurrence survival (70.7 versus 106 months, p=0.009). In multivariate analysis, DM (hazard ratio (HR)=1.466, p=0.012), FIB-4 score, multiple tumors, and AFP level were independent predictors of distant recurrence, while DM (HR=1.424, p=0.028), ALBI score, tumor size, AFP and creatinine levels were significantly associated with OS. A DM-based risk score effectively discriminated the risk of distant recurrence. The IMbrave050 criteria could stratify the risk of LTP but not distant recurrence. DM status further discriminated the risk of distant recurrence and mortality in the IMbrave050-defined high-risk population. Conclusion: Patients with DM had an increased risk of distant recurrence and mortality after thermal ablation for HBV-related HCC, highlighting the importance of increasing awareness of DM and implementing rigorous post-ablation monitoring for diabetic HCC patients.

Introduction: Novel biomarkers reflecting liver fibrosis and the immune microenvironment may correlate with the risk of hepatocellular carcinoma (HCC) recurrence. This study aimed to evaluate the prognostic value of serum biomarkers in predicting HCC recurrence. Methods: Serum biomarkers, including M2BPGi, IL-6, IL-10, CCL5, VEGF-A, soluble PD-1, PD-L1, TIM-3, and LAG-3, were measured in 247 patients with HCC undergoing surgical resection. Factors associated with recurrence-free survival (RFS) and overall survival (OS) were evaluated. The ERASL-post model and IMbrave050 criteria were used to define HCC recurrence risk groups. Results: Serum M2BPGi levels significantly correlated with FIB-4 score, aspartate transaminase-to-platelet ratio index, ALBI score, alpha-fetoprotein (AFP), alanine transaminase, aspartate transaminase, IL-10, CCL5, VEGF-A, soluble PD-1, PD-L1, TIM-3, and LAG-3 levels. M2BPGi, VEGF-A, soluble PD-1, and TIM-3 levels significantly correlated with RFS. In multivariate analysis, M2BPGi >1.5 COI (hazard ratio [HR] = 2.100, p < 0.001), tumor size >5 cm (HR = 1.859, p = 0.002), multiple tumors (HR = 2.562, p < 0.001), AFP >20 ng/mL (HR = 2.141, p < 0.001), and microvascular invasion (HR = 1.954, p = 0.004) were independent predictors of RFS. M2BPGi levels significantly stratified the recurrence risk in ERASL-post and IMbrave050 risk groups. An M2BPGi-based model could significantly discriminate RFS in the overall cohort as well as in the IMbrave050 low- and high-risk groups. M2BPGi >1.5 COI was also an independent predictor of OS after resection (HR = 2.707, p < 0.001). Conclusion: Serum M2BPGi levels significantly correlated with surrogate markers of liver fibrosis, liver function, and immunology. M2BPGi is a significant predictor of HCC recurrence and survival after resection and could be incorporated into recurrence-prediction models.