Transarterial chemoembolisation-based combination therapy as a potential new standard for intermediate-stage hepatocellular carcinoma in the targeted immunotherapy era Authors: Shuanggang Chen, Lujun Shen, Han Qi, Pan Hu, Fei Cao, Lin Xie, Yujia Wang, Weijun Fan Published: 2026 ABSTRACT AbstractTransarterial chemoembolisation (TACE) is the standard therapy for intermediate-stage hepatocellular carcinoma (HCC); however, its efficacy is limited by significant tumour heterogeneity. Recent landmark trials (EMERALD-1 and LEAP-012) demonstrated that adding targeted immunotherapy to TACE significantly improves progression-free survival compared with TACE alone. However, the inconsistent overall survival (OS) benefit, particularly in LEAP-012, highlights the necessity for precision patient stratification. Key determinants of outcomes include tumour burden, liver function, microvascular invasion (MVI), systemic inflammation, immune status and comorbidities. For patients with low tumour burden, optimised locoregional therapies can achieve higher complete response rates. For high tumour burden, multimodal strategies are required, with increasing evidence supporting TACE plus targeted therapy with or without immunotherapy. MVI-positive HCC derives added benefit from adjuvant tyrosine kinase inhibitors (TKIs) after TACE, while dynamic liver function monitoring is crucial for safety, especially in patients with Child–Pugh B. Future directions should focus on developing and validating robust multidimensional stratification models. These models should integrate key established determinants—such as tumour burden, liver function reserve, MVI, systemic inflammatory/immune status, alpha-fetoprotein (AFP) and comorbidities—with promising emerging biomarkers (eg, circulating tumour DNA). Concurrent efforts are needed to optimise combination regimens and conduct health-economic evaluations. Global collaboration and cost-effectiveness analyses are essential to prioritise high-benefit subgroups (eg, those with high tumour burden, MVI positivity and/or mild inflammation, active immune status alongside preserved liver function and limited comorbidities) for intensive combination therapy while avoiding overtreatment in lower-benefit populations.Keywords: HEPATOCELLULAR CARCINOMA, CANCER IMMUNOBIOLOGY, IMMUNOLOGY, IMMUNOLOGY IN HEPATOLOGY Introduction Transarterial chemoembolisation (TACE) remains the first-line standard treatment for intermediate-stage hepatocellular carcinoma (HCC) in suitable candidates (ie, those with well-defined nodules, preserved portal flow and selective access); however, this recommendation may not apply to patients with extremely high tumour burden or poor liver function reserve.1 However, this stage is now recognised as encompassing a profoundly heterogeneous population with wide variations in tumour burden, liver function and biology, leading to significantly divergent outcomes after TACE.2 This realisation has rendered a uniform TACE strategy insufficient and has driven the development of subclassification systems (eg, Up-to-Seven, Up-to-Eleven, Seven-Eleven, Bolondi criteria, Kinki Criteria, Yamakado Criteria and NTAA prognostic model) to identify prognostically distinct subgroups.2 3 Building on this stratified approach, the integration of TACE with systemic therapies—particularly targeted immunotherapy—has emerged as a pivotal strategy, especially for identified high-benefit subgroups. Two recent global, multicentre, randomised controlled trials—EMERALD-1 and LEAP-012, published back-to-back in The Lancet—demonstrated that TACE plus targeted immunotherapy significantly prolonged progression-free survival (PFS) compared with TACE alone in patients with Barcelona Clinic Liver Cancer (BCLC) stage B HCC (LEAP-012, HR=0.57, 95% CI 0.41 to 0.77; EMERALD-1, HR=0.71, 95% CI 0.52 to 0.95)4 5 (table 1). These findings provide important preliminary evidence for the efficacy and safety of TACE plus targeted immunotherapy. However, improvements in PFS do not consistently translate into meaningful OS benefits in oncology trials.6 This is exemplified by the LEAP-012 trial: while follow-up for OS is ongoing in EMERALD-1, mature OS data from LEAP-012 revealed no statistically significant difference (HR=0.78, 95% CI 0.52 to 1.18).4 5 This discrepancy highlights the need to identify specific subgroups of intermediate-stage HCC that may derive true survival benefit from TACE-targeted immunotherapy, whereas the true efficacy could be masked by confounding effects in heterogeneous populations. Table 1Comparative overview of two landmark phase 3 trials (LEAP-012 and EMERALD-1 trials) investigating TACE-based combination therapy for intermediate-stage HCCLEAP-012EMERALD-1Study designMulti-centre, double-blind, randomised, phase 3 study performed at 137 hospitals or medical centresMulti-regional, randomised, double-blind, placebo-controlled, phase 3 study performed at 157 medical centresInclusion criteriaAdults ≥18 years with AASLD-confirmed HCC; unresectable, non-metastatic (no PV invasion/extrahepatic disease, BICR-confirmed) and TACE-eligible; ≥1 measurable lesion per RECIST v1.1; life expectancy ≥3 months; HAP score A–C; ECOG PS 0/1; Child–Pugh AAdults ≥18 years with locally confirmed HCC (imaging/histopathology); unresectable, non-metastatic and TACE-eligible (≤4 TACEs within 16 weeks); no extrahepatic disease; ≥1 measurable lesion per mRECIST; ECOG PS 0/1; Child–Pugh A–B7; upper endoscopy within 6 months pre-randomisation for bevacizumab safety assessmentExclusion criteriaBICR-confirmed tumours: ≥10 cm in any dimension, >10 tumours or occupying ≥50% of the liver volume; prior systemic therapy for HCCPrior systemic therapy for HCC; prior embolisation (curative-intent embolisation permitted); imaging-confirmed major PVTT (Vp3/Vp4 tumour thrombus)Trial grouping and number of participants in each groupLenvatinib plus pembrolizumab plus TACE (n=237)Placebo plus TACE (n=243)Durvalumab plus bevacizumab plus TACE (n=204)Durvalumab plus placebo plus TACE (n=207)Placebo plus TACE (n=205)Number and proportion of patients with BCLC stage B in each groupLenvatinib plus pembrolizumab plus TACE (n=135, 57%)Placebo plus TACE (n=146, 60%)Durvalumab plus bevacizumab plus TACE (n=117, 57%)Durvalumab plus placebo plus TACE (n=114, 55%)Placebo plus TACE (n=122, 60%)Objective response rate (ORR) in the overall population of each group (RECIST version)47% (95% CI 40% to 53%) in the lenvatinib plus pembrolizumab plus TACE group33% (95% CI 27% to 40%) in the placebo plus TACE group44% in the durvalumab plus bevacizumab plus TACE group41% in the durvalumab plus placebo plus TACE group30% in the placebo plus TACE groupORR in the overall population of each group (mRECIST version)71% (95% CI 65% to 77%) in the lenvatinib plus pembrolizumab plus TACE group50% (95% CI 43% to 56%) in the placebo plus TACE group–––Median progression-free survival (PFS) of the overall population in each group (RECIST version)16.6 months (95% CI 14.6 to 18.7) in the lenvatinib plus pembrolizumab plus TACE group10.3 months (95% CI 8.2 to 12.5) in the placebo plus TACE group15.0 months (95% CI 11.1–18.9) in the durvalumab plus bevacizumab plus TACE group 10.0 months (95% CI 9.0 to 12.7) in the durvalumab plus placebo plus TACE group8.2 months (95% CI 6.9 to 11.1) in the placebo plus TACE groupHR for PFS in patients with BCLC stage B HCC across different groups (RECIST version)HR for PFS in the lenvatinib plus pembrolizumab plus TACE group vs placebo plus TACE group: HR=0.57 (95% CI 0.41 to 0.77)HR for PFS in the durvalumab plus bevacizumab plus TACE group vs placebo plus TACE group: HR=0.71 (95% CI 0.52 to 0.95)HR for OS in the overall population in each groupHR for OS in the lenvatinib plus pembrolizumab plus TACE group vs placebo plus TACE group: HR=0.80 (95% CI 0.57 to 1.11)Ongoing for the final overall survival analysisHR for OS in patients with BCLC stage B HCC across different groupsHR for OS in the lenvatinib plus pembrolizumab plus TACE group vs placebo plus TACE group: HR=0.78 (95% CI 0.52 to 1.18)Ongoing for the final overall survival analysisBoth studies demonstrated superior progression-free survival (PFS) with the addition of targeted immunotherapy (lenvatinib/pembrolizumab or durvalumab/bevacizumab) to TACE. The mature OS analysis from LEAP-012, however, did not reach statistical significance, highlighting the discrepancy between PFS and OS benefit and emphasising the necessity for precision stratification in this patient population.AASLD, American Association for the Study of Liver Diseases; BICR, blinded independent central review; ECOG, Eastern Cooperative Oncology Group; HAP, hepatoma arterial-embolisation prognosis; HCC, hepatocellular carcinoma; HR, hazard ratio; mRECIST, modified Response Evaluation Criteria in Solid Tumours; ORR, objective response rate; OS, overall survival; PFS, progression-free survival; PS, performance status; PV, portal vein; PVTT, portal vein tumour thrombus; RECIST, Response Evaluation Criteria in Solid Tumours; TACE, transarterial chemoembolisation. Comparative overview of two landmark phase 3 trials (LEAP-012 and EMERALD-1 trials) investigating TACE-based combination therapy for intermediate-stage HCC Both studies demonstrated superior progression-free survival (PFS) with the addition of targeted immunotherapy (lenvatinib/pembrolizumab or durvalumab/bevacizumab) to TACE. The mature OS analysis from LEAP-012, however, did not reach statistical significance, highlighting the discrepancy between PFS and OS benefit and emphasising the necessity for precision stratification in this patient population. AASLD, American Association for the Study of Liver Diseases; BICR, blinded independent central review; ECOG, Eastern Cooperative Oncology Group; HAP, hepatoma arterial-embolisation prognosis; HCC, hepatocellular carcinoma; HR, hazard ratio; mRECIST, modified Response Evaluation Criteria in Solid Tumours; ORR, objective response rate; OS, overall survival; PFS, progression-free survival; PS, performance status; PV, portal vein; PVTT, portal vein tumour thrombus; RECIST, Response Evaluation Criteria in Solid Tumours; TACE, transarterial chemoembolisation. Despite differences in combination regimens between the two studies, this contradiction underscores the urgent need for precision-based risk stratification in intermediate-stage HCC. Currently, stratification models based on prognostic factors such as tumour burden, liver function, microvascular invasion (MVI), systemic inflammation or immune status, tumour markers and comorbidities are still being explored. Identifying potential responders to combination therapy has become an important challenge. By integrating evidence from recent years, this review discusses the key stratification factors for intermediate-stage HCC and their impact on treatment strategies while identifying optimal candidate populations for TACE plus targeted immunotherapy (figure 1). We also propose our perspectives on the application of TACE plus targeted immunotherapy in intermediate-stage HCC. Conceptual framework for multidimensional stratification in intermediate-stage HCC. This schematic illustrates the core thesis of this review: due to the significant heterogeneity of intermediate-stage hepatocellular carcinoma (HCC), a uniform treatment approach is inadequate. The integration of transarterial chemoembolisation (TACE) with targeted immunotherapy has shown improved progression-free survival (PFS), but not consistently overall survival (OS). To identify the patients most likely to achieve a meaningful OS benefit from intensive combination therapy, treatment decisions must be guided by a multidimensional stratification model. Key determinants for this stratification include tumour burden, liver function reserve, microvascular invasion and systemic inflammatory/immune status. BCLC, Barcelona Clinic Liver Cancer; HCC, hepatocellular carcinoma; OS, overall survival; PFS, progression-free survival. Prognosis stratification and treatment decision-making for intermediate-stage hepatocellular carcinoma Tumour burden Extensive evidence underscores tumour burden as an important determinant of prognosis in intermediate-stage HCC. Key variables of tumour burden, including tumour size and number, are closely associated with MVI risk, intrahepatic metastatic potential, reduced residual normal liver volume and impaired liver function reserve.7 These factors contribute to the marked heterogeneity in treatment outcomes of TACE in intermediate-stage HCC. Consequently, diversified treatment strategies tailored to distinct tumour burdens are imperative. Low tumour burden Low tumour burden is typically defined by the ‘up-to-seven’ criteria,8 wherein the therapeutic goal should focus on optimising local treatment modalities to improve complete response rates and long-term prognosis (figure 2). For resectable intermediate-stage HCC, surgical resection still demonstrates a significant prognostic advantage over traditional standalone TACE.9 10 For partially unresectable cases, resection combined with ablation represents an effective strategy to enhance prognosis, though it is often limited to ≤2 ablation sites, ablation lesion size ≤2 cm and achievement of complete ablation.11 Additionally, liver transplantation may serve as a viable option for intermediate-stage HCC exceeding the Milan criteria, such as those meeting the Hangzhou criteria, University of California San Francisco criteria or the ‘up-to-seven’ criteria.12,15 Furthermore, for select patients with intermediate-stage HCC who achieve downstaging through aggressive treatment, liver transplantation, resection or ablation can be attempted, although with strict patient selection.15,18 Treatment stratification based on the up-to-seven criteria in intermediate-stage hepatocellular carcinoma. This decision pathway illustrates the application of the up-to-seven criteria—where the sum of the diameter (in cm) of the largest tumour and the number of tumours is calculated—to stratify patients into two distinct management pathways. Patients within the criteria (sum ≤7) typically possess a lower tumour burden and may be candidates for curative-intent locoregional therapies. Patients beyond the criteria (sum >7) are classified as having a higher tumour burden, prompting consideration for multimodal or systemic combination strategies. This model aids in tailoring treatment intensity within a heterogeneous patient population. Currently, effective downstaging strategies for low tumour burden include TACE, hepatic arterial infusion chemotherapy (HAIC) and transarterial radioembolisation (TARE).1719,21 Specifically, for HAIC utilising regimens such as the modified FOLFOX (mFOLFOX, which includes oxaliplatin, fluorouracil and leucovorin) regimen, meta-analysis evidence indicates a significantly higher rate of successful downstaging to surgical resection compared with TACE (RR=2.37, 95% CI 1.54 to 3.66).20 Our prior research also demonstrated that TACE-based combination therapy can achieve imaging-confirmed ‘no evidence of disease’ in nearly 35% of patients with intermediate-stage HCC, significantly reducing mortality risk in this population.22 However, for most unresectable intermediate-stage HCC cases, TACE remains a valid therapeutic option. With advancing research, some studies have shown that for BCLC stage B HCC with limited tumour burden (≤ 5 lesions, ≤ 7 cm in size), the efficacy of TACE plus ablation markedly surpasses standalone TACE.23,25 Some investigations even advocate TACE–ablation combinations for all patients with intermediate-stage HCC, reporting superior outcomes compared with TACE alone.26,28 These findings highlight the important role of TACE–ablation synergy in unresectable and non-transplantable intermediate-stage HCC. Currently, the widely used ablation techniques in clinical practice include radiofrequency ablation (RFA), microwave ablation (MWA) and cryoablation (CA).29 Specifically, the mechanism of RFA primarily involves an electrode probe emitting RF waves, which induce ionic oscillation in tumour tissues to generate heat, ultimately leading to tumour cell necrosis.29 30 For MWA, its mechanism mainly relies on microwave electromagnetic fields to heat tumour tissues, rapidly destroying cancer cells through the rotation of polar molecules.29 30 In contrast, the core mechanism of CA is the formation of an ice ball under low temperatures, resulting in tumour cell death due to cryoinjury.29 To date, high-quality randomised controlled trials defining the optimal ablation modality in combination with TACE for HCC are still lacking. Nevertheless, findings from a meta-analysis have indicated that TACE plus MWA yields a significantly higher OR for long-term OS and objective response rate (ORR) than TACE plus RFA or TACE plus CA. Notably, young patients with tumour lesions ≤3 cm derive the most prominent clinical benefits from the TACE–MWA combination therapy.31 However, standardised protocols for patient selection, treatment dosing, ablation modalities, lesion size or number thresholds and endpoint definitions remain unestablished. Addressing these gaps is imperative to avoid overtreatment and ensure guideline-adherent care (figure 3). Evidence-based stratification and treatment flowchart of intermediate-stage HCC. BCLC, Barcelona Clinic Liver Cancer; BST, basic supportive treatment; CP, Child–Pugh; HAIC, hepatic artery infusion chemotherapy; HCC, hepatocellular carcinoma; MDT, multi-disciplinary treatment; PD-1, programmed cell death-1; SBRT, stereotactic body radiotherapy; TACE, transarterial chemoembolisation; TARE, transarterial radioembolisation; TKI, tyrosine kinase inhibitors. High tumour burden For patients with tumour burden exceeding the up-to-7 criteria,8 the cornerstone of treatment lies in the synergy between systemic therapy and local interventions, while prioritising downstaging and conversion therapy potential (figure 2). Targeted therapy, particularly TKIs, forms an important backbone. Evidence confirms that TACE plus TKIs (eg, lenvatinib or sorafenib) and/or ablation can significantly improve outcomes compared with TACE alone,32,36 establishing TACE plus targeted therapy as a validated foundational regimen. The negative results of TACE plus sorafenib in the TACE 2 trial represent a special case, which may be associated with trial-specific design factors, such as the timing of sorafenib initiation, definition of PFS, identification of MVI-positive populations and active monitoring and management of liver function.36 37 In contrast, the integration of immunotherapy has shown a more complex efficacy profile. The failure of TACE plus durvalumab (anti-programmed death-ligand 1 (PD-L1)) to improve outcomes suggests that simply adding an immune checkpoint inhibitor to TACE is insufficient. Similarly, the systemic dual-combination of atezolizumab (anti-PD-L1) plus bevacizumab (anti-vascular endothelial growth factor)—a regimen not containing TACE—also did not improve PFS or OS in BCLC stage B HCC.4 38 These results indicate that neither TACE plus immunotherapy nor TKI plus immunotherapy (in a TACE-free setting) alone are reliably effective strategies for this heterogeneous population. Therefore, the emerging paradigm supported by trials, like EMERALD-1 and LEAP-012, is not the replacement of the TACE-TKI backbone, but its strategic intensification. The combination of TACE, a TKI, and an immunotherapy agent (a triple-therapy regimen) leverages the established synergy of TACE with targeted therapy, while it attempts to amplify the anti-tumour immune response. This approach underscores that immunotherapy may yield the greatest benefit not as a substitute or simple add-on but as a third component integrated into a potent locoregional-systemic foundation. Future research should focus on identifying the patient subgroups most likely to benefit from this intensive triple combination. The emergence of HAIC offers a novel pathway for high-burden tumours. The mFOLFOX-based HAIC shows superior efficacy over conventional TACE in intermediate-stage HCC with tumours ≥7 cm,39 and drug-eluting bead transarterial chemoembolisation (DEB-TACE) combined with HAIC significantly improves local control rates for large tumours.40 HAIC coupled with targeted immunotherapy expands therapeutic dimensions, enabling long-term survival in select patients after downstaging and conversion.41,43 Explorations of TACE–HAIC (a treatment strategy combining TACE with HAIC) combined with targeted and immunotherapy regimens have shown preliminary efficacy, delivering potential survival benefits for high-burden intermediate-stage HCC.44 45 With the continuous advancement of locoregional therapeutic techniques, the clinical value of TARE, also known as selective internal radiation therapy, has attracted increasing attention in the management of intermediate-stage HCC with high tumour burden. This therapy involves the infusion of yttrium-90–labelled microspheres via tumour-feeding arteries to deliver targeted brachytherapy to tumour tissues while occluding the tumour blood supply through the embolic effect of microspheres, which leads to cancer cell death due to the deprivation of nutritional support. Its efficacy and safety in the treatment of intermediate-stage HCC have been verified in multiple clinical studies.46,49 Notably, the TRACE single-centre randomised controlled trial specifically compared the safety and efficacy of TARE and DEB-TACE (typically using doxorubicin-loaded beads) in patients with unresectable HCC and relatively low tumour burden. The vast majority of the enrolled patients were at BCLC stage B, with the median tumour diameters of 4.2 and 4.7 cm in the TARE and DEB-TACE groups, respectively. The results confirmed that TARE was significantly superior to DEB-TACE in terms of time to tumour progression and OS, with a comparable incidence of adverse events between the two groups.49 In contrast, a retrospective exploratory study comparing TARE and TACE for unresectable HCC with relatively high tumour burden showed that for unresectable HCC lesions with diameter >8 cm, TARE achieved therapeutic efficacy comparable with that of TACE but with a significantly lower incidence of adverse events (72% in the TACE group vs 5% in the TARE group).50 However, this study was a retrospective analysis with a relatively low proportion of intermediate-stage HCC cases in the enrolled population; thus, its conclusions require further validation by well-designed clinical studies. With the advent of targeted and immunotherapy, the strategy of TARE plus systemic therapy has also been explored in HCC with high tumour burden. For instance, a single-arm trial conducted by Lee et al demonstrated that TARE plus durvalumab in the treatment of patients with locally advanced HCC yielded an ORR of 83.3% and an 18-month OS rate of 58.3%, which indicated that TARE plus immunotherapy holds promising clinical potential.51 Furthermore, Huang et al explored and compared the safety and efficacy of TARE plus targeted-immunotherapy vs TACE plus targeted-immunotherapy in patients with high tumour burden HCC. The results showed that TARE plus targeted-immunotherapy was significantly superior to the TACE-based combination regimen in terms of ORR, PFS and OS.52 Collectively, these findings further corroborate the potential advantages of TARE in the management of high tumour burden HCC: TARE not only exhibits a more favourable safety profile as a monotherapy but also exerts a synergistic anti-tumour effect when combined with targeted and immunotherapeutic agents. This not only provides a more optimal therapeutic option for patients with high tumour burden HCC, a population presenting great clinical treatment challenges, but also lays a solid foundation for the subsequent optimisation and exploration of clinical treatment strategies for this patient cohort. Despite these promising advances, inherent tumour and patient heterogeneity in high-burden populations remains a challenge. Future priorities include (1) identifying populations amenable to downstaging/conversion therapy (eg, biomarkers and radiomics models) and optimising strategies, (2) exploring multimodal synergy for non-downstaging candidates and (3) balancing efficacy and safety, mitigating hepatic toxicity from multisystemic combinations and immune-related adverse events (figure 3). Liver function reserve For intermediate-stage HCC with liver function Child–Pugh (CP) scores of 5–7, preserving liver function reserve is important, as its loss may lead to hepatic failure, treatment interruption or increased mortality; therefore, selecting effective treatment regimens must prioritise this goal to improve long-term prognosis. As noted previously, there are various combination therapy approaches for intermediate-stage HCC. However, research on the dynamic correlation is limited between different treatments and frequencies and changes in liver function reserves, especially the need to carefully evaluate the impact of local treatment combined with targeted immunotherapy on liver function reserves in intermediate-stage HCC. Although EMERALD-1 and LEAP-012 have shown that TACE plus targeted immunotherapy does not significantly increase adverse events related to liver function changes compared with TACE alone, neither study dynamically assessed the long-term effects of this combination on liver function.4 5 The significance of assessing dynamic changes in liver function lies in effectively elucidating the impact of this local combination of targeted immunotherapy on the long-term prognosis of intermediate-stage HCC. Subgroup analysis results from LEAP-012 indicate that TACE plus lenvatinib and pembrolizumab can improve PFS in intermediate-stage HCC, but it does not significantly improve OS, which may also suggest that this combination leads to a decline in long-term liver function reserves, thereby weakening its efficacy.4 5 Cancer-specific survival (CSS) as a primary endpoint may be more beneficial for evaluating the effectiveness of this combination regimen. However, the decline in liver function reserves during treatment may also act as a promoter of tumour progression, leading to no significant improvement in CSS. For patients with mid-stage HCC who have a liver function CP score of 8 or 9, multi-disciplinary treatment (MDT) is necessary. Laparoscopic liver resection can serve as a reference for MDT; however, the incidence of postoperative complications significantly increases, especially for patients with a CP score of 8.53 54 Additionally, for patients with a CP score of 8, adjuvant TACE after resection is a safe and effective treatment option, particularly for patients with intermediate-stage HCC with multiple lesions, tumour size >5 cm and margins <1 cm.54 However, whether this treatment regimen is safe and effective for patients with a CP score of 9 has not been assessed, so careful selection of the benefit population is required. Compared with symptomatic supportive care, capecitabine-based chemotherapy appears to be safe and effective for patients with intermediate-stage HCC with liver function CP score of 8 or 9. However, attention should be paid to adverse reactions such as severe thrombocytopenia and hand-foot syndrome.55 Anti-angiogenic combined with anti-programmed cell death protein 1 (PD-1) antibody therapy is also safe and effective for patients with a CP score of 8 or 9; however, these patients are more likely to experience dose reduction or discontinuation of targeted drugs.56 For carefully selected patients with intermediate-stage HCC with a CP score of 8 or 9, stereotactic body radiotherapy seems to be safe and effective, but special attention should be paid to the potential risk of death due to radiation hepatitis57 (figure 3). Microvascular invasion In recent years, the treatment strategy for patients with MVI-positive intermediate-stage HCC has become an important area of investigation, as this population highlights the limitations of TACE alone. TACE therapy upregulates vascular endothelial growth factor expression in tumour tissues, potentially exacerbating vascular invasion and metastatic potential in these patients.58 Consequently, whether combining targeted therapy with TACE is necessary for patients with MVI-positive intermediate-stage HCC has emerged as a key research focus. A recent randomised controlled trial demonstrated that sorafenib plus TACE significantly improved OS, PFS and ORR in patients with recurrent MVI-positive intermediate-stage HCC compared with TACE alone.59 60 This positive result contrasts with the negative finding of the TACE 2 trial, in which TACE plus sorafenib did not improve prognosis in a broader, unselected population of patients with unresectable HCC.37 The discrepancy likely stems from critical differences in the patient cohorts: the former focused on a biomarker-defined, high-benefit subgroup (MVI-positive), whereas the latter included a heterogeneous mix of patients without MVI selection. Importantly, the regimen did not enhance outcomes in the MVI-negative group,60 which collectively underscores MVI as a pivotal biomarker for stratifying patients who may benefit from the addition of sorafenib to TACE. Additionally, HAIC has shown preliminary efficacy in patients with MVI-positive intermediate-stage HCC.61 Studies have also indicated that HAIC plus sorafenib significantly improves outcomes in patients with HCC and portal vein tumour thrombi.62 63 Therefore, whether HAIC plus sorafenib further enhances survival in patients with MVI-positive intermediate-stage HCC warrants exploration in future research. MVI highlights the importance of pathological examination. However, current major guidelines, such as the European Association for the Study of the Liver, American Association for the Study of Liver Diseases and National Comprehensive Cancer Network guidelines, have not listed tissue biopsy as a routine diagnostic or pre-treatment essential step for unresectable intermediate-stage HCC.7 64 65 Therefore, obtaining tissue specimens in these patients is challenging, which hinders the development of personalised treatment for intermediate-stage HCC. However, some studies are attempting to use liquid biopsies to obtain peripheral blood circulating tumour cells and DNA to predict the efficacy of intermediate-stage HCC, suggesting that low levels of peripheral blood circulating tumour DNA (ctDNA) are associated with better responses to intermediate-stage HCC.66,69 For example, in a phase Ib trial of sintilimab plus bevacizumab followed by resection for intermediate-stage HCC, Sun et al reported that lower baseline levels of ctDNA were associated with better tumour response.69 Biomarkers related to systemic inflammatory response and immune status Existing studies have confirmed that systemic inflammatory response and immune status are important biological factors influencing treatment stratification in patients with intermediate-stage HCC.70,74 Systemic inflammation scoring systems based on serological markers, such as the platelet-to-lymphocyte ratio (PLR), neutrophil-to-lymphocyte ratio (NLR), lymphocyte-to-monocyte ratio (LMR) and other chronic inflammation-related scores including the Glasgow Prognostic Score (GPS), C-reactive protein-to-albumin ratio (CAR) and lymphocyte-to-C-reactive protein ratio (LCR), have been validated in multiple studies as independent prognostic factors for patients with intermediate-stage HCC undergoing TACE or TACE plus ablation.70,78 For example, PLR has been demonstrated as an independent predictor of PFS after TACE,70 wheras GPS, CAR and LCR scores are significantly associated with OS.71 72 74 Further studies have shown that NLR, LMR, PLR and CAR can effectively stratify survival differences among patients receiving TACE plus ablation.7375,78 These findings strongly suggest that exacerbated inflammatory responses and immunosuppressive states are closely associated with poor prognosis. Notably, these biomarkers not only hold prognostic value but also offer potential therapeutic targets for stratified treatment. For patients with high inflammatory burden and immune deficiency, combination therapies including antivirals, anti-inflammatory drugs and immunomodulators can significantly enhance treatment outcomes.79 Future research should focus on identifying the optimal systemic inflammatory and immune markers for precise stratification to optimise individualised treatment strategies for patients with intermediate-stage HCC. Tumour markers and comorbidities AFP is a widely used yet complex biomarker in HCC, a complexity compounded by its lack of absolute disease specificity, as serum levels can also be elevated in response to hepatic inflammation and regeneration. Its prognostic utility in intermediate-stage HCC is not defined by a single universal cut-off but is highly context-dependent, influenced by the specific clinical scenario, treatment modality and patient population. This context-dependent profile is reflected in the range of validated cut-offs applicable to different settings. For example, in broad, mixed-stage cohorts, 20 and 400 ng/mL have been established as independent prognostic factors.80 Notably, the 400 ng/mL cut-off is the Food and Drug Administration–approved predictive biomarker for selecting patients with advanced HCC who may benefit from ramucirumab (an anti-vascular endothelial growth factor receptor 2 therapy).81 However, these thresholds, derived from heterogeneous populations, may not be optimised for specific subgroups. Indeed, in studies focusing specifically on intermediate-stage (BCLC B) HCC, the prognostic significance of the 400 ng/mL cut-off has been inconsistent.82 83 Furthermore, the frequently cited 200 ng/mL cut-off often relies on historical benchmarks rather than contemporary, data-driven analysis.9 In contrast, our previous study focused precisely on the specific context of patients with intermediate-stage HCC undergoing TACE. Employing data-driven optimisation via survival ROC analysis, we identified a prognostic cut-off of 156.5 ng/mL, which has subsequently undergone internal and multi-centre external validation.3 These observations collectively underscore that the appropriate AFP cut-off should be aligned with the treatment pathway (eg, systemic therapy vs TACE) and disease stage. Therefore, for risk-stratifying intermediate-stage HCC within TACE-based strategies, employing a context-specific cut-off such as 156.5 ng/mL is likely to provide more accurate prognostication than applying generic thresholds. This reinforces the imperative to interpret biomarkers within a multidimensional stratification framework. Future large-scale, global collaborative studies will be valuable to further refine optimal AFP cut-offs across diverse clinical scenarios and aetiologies. Additionally, comorbidities in patients with HCC may contribute to the heterogeneity of prognosis, often closely related to underlying liver disease, metabolic disorders and decompensated liver function. These comorbidities significantly shorten the OS of patients with intermediate-stage HCC by limiting tumour-tolerant treatment, increasing treatment-related adverse events and exacerbating the malignant biological behaviour of tumours.84 Therefore, establishing a multi-dimensional stratification model that includes tumour burden, liver function, MVI, inflammatory–immune status, tumour markers and comorbidities can reduce clinical heterogeneity and improve the overall efficacy of patients with intermediate-stage HCC. Perspectives Overall, the combination of TACE with targeted immunotherapy marks a significant paradigm shift in the treatment of intermediate-stage HCC. This evolution is not only supported by pivotal global clinical trials but is also increasingly reflected in the principles endorsed by regional authoritative consensus statements. Notably, the recently published Chinese Expert Consensus on Sequential Surgery Following Conversion Therapy Based on Immune Checkpoint Inhibitors and Antiangiogenic Targeted Drugs for Advanced Hepatocellular Carcinoma (2024 Edition) systematically elaborates on a modern therapeutic principle centred on aggressive combination therapy underpinned by refined patient selection.85 Although this consensus primarily focuses on advanced-stage disease, the proactive treatment paradigm it establishes—‘achieving tumour control through intensive regimens such as TACE plus targeted immunotherapy to create opportunities for potential curative intervention in suitable patients’—resonates strongly with the management rationale for intermediate-stage HCC subgroups possessing high-risk features (eg, high tumour burden). Furthermore, the principle of ‘precise patient selection based on multidisciplinary team evaluation’ emphasised throughout the consensus aligns perfectly with the core thesis of ‘multidimensional risk stratification’ advocated in this review. This paradigm shift is concretely driven by landmark trials such as LEAP-012 and EMERALD-1, which confirmed that integrating systemic therapies, including PD-1/PD-L1 inhibitors and TKIs, may overcome the limitations of monotherapy, such as suboptimal response rates post-TACE and insufficient immune activation. Subgroup analyses from LEAP-012 have revealed the significant PFS advantage of TACE plus targeted immunotherapy; however, the absence of an OS benefit serves as an important reminder that this intensive combination is not universally applicable to all patients with intermediate-stage HCC; its true efficacy is likely masked by confounding effects within a highly heterogeneous population. Consequently, the paramount clinical challenge now is to precisely identify the true beneficiaries amidst this profound heterogeneity. Based on current evidence, we propose the following framework of recommendations for future research and clinical practice regarding TACE plus targeted immunotherapy in intermediate-stage HCC: Establish a stratified research framework with OS as the gold standard: future high-quality clinical trials should prioritise evaluating the efficacy and safety of combination regimens in high-benefit subgroups preselected via multidimensional models (integrating tumour burden, liver function, MVI, etc.), with OS as the primary endpoint. PFS as a surrogate endpoint requires cautious interpretation, as its improvement does not invariably translate into a meaningful OS benefit. As an illustrative example (but not an exclusive definition), a candidate profile for high benefit warranting evaluation of intensive combination therapy could include patients presenting with a combination of features such as high tumour burden (eg, exceeding the up-to-seven criteria), preserved liver function (CP A) and the presence of MVI. The integration of exploratory biomarkers like ctDNA or circulating tumour cells (CTCs) may further refine risk assessment in the context of clinical trials. Establish a stratified research framework with OS as the gold standard: future high-quality clinical trials should prioritise evaluating the efficacy and safety of combination regimens in high-benefit subgroups preselected via multidimensional models (integrating tumour burden, liver function, MVI, etc.), with OS as the primary endpoint. PFS as a surrogate endpoint requires cautious interpretation, as its improvement does not invariably translate into a meaningful OS benefit. As an illustrative example (but not an exclusive definition), a candidate profile for high benefit warranting evaluation of intensive combination therapy could include patients presenting with a combination of features such as high tumour burden (eg, exceeding the up-to-seven criteria), preserved liver function (CP A) and the presence of MVI. The integration of exploratory biomarkers like ctDNA or circulating tumour cells (CTCs) may further refine risk assessment in the context of clinical trials. Promote the exploratory integration of pathology and liquid biopsy: for patients with accessible tissue, pathological examination should be performed to clarify key information such as MVI status, histological subtype and differentiation grade. For patients without tissue availability, exploratory use of investigational liquid biopsy technologies (eg, ctDNA and CTCs) may be considered in the context of clinical trials or research settings. It should be explicitly acknowledged that these biomarkers currently lack standardised assays and prospective validation for clinical decision-making. Promote the exploratory integration of pathology and liquid biopsy: for patients with accessible tissue, pathological examination should be performed to clarify key information such as MVI status, histological subtype and differentiation grade. For patients without tissue availability, exploratory use of investigational liquid biopsy technologies (eg, ctDNA and CTCs) may be considered in the context of clinical trials or research settings. It should be explicitly acknowledged that these biomarkers currently lack standardised assays and prospective validation for clinical decision-making. Caution in populations with low anticipated benefit: for patients with low tumour burden and MVI-negative status (and potentially negative exploratory biomarkers like ctDNA/CTCs), treatment should prioritise optimising local therapies to improve complete response rates and long-term outcomes. The routine addition of TACE plus targeted immunotherapy in this group is not supported by current evidence and risks overtreatment. Caution in populations with low anticipated benefit: for patients with low tumour burden and MVI-negative status (and potentially negative exploratory biomarkers like ctDNA/CTCs), treatment should prioritise optimising local therapies to improve complete response rates and long-term outcomes. The routine addition of TACE plus targeted immunotherapy in this group is not supported by current evidence and risks overtreatment. Address high inflammation response and immune dysfunction: in patients with marked chronic systemic inflammation or impaired immune status, adjuvant strategies such as antivirals, anti-inflammatory agents or immunomodulators may be considered to potentially improve the therapeutic environment and long-term prognosis. Address high inflammation response and immune dysfunction: in patients with marked chronic systemic inflammation or impaired immune status, adjuvant strategies such as antivirals, anti-inflammatory agents or immunomodulators may be considered to potentially improve the therapeutic environment and long-term prognosis. Dynamic liver function monitoring: the impact of TACE plus targeted immunotherapy on short-term and long-term liver function reserve should be dynamically assessed, particularly in patients with CP scores 7–9. Novel biomarkers with high sensitivity and specificity for real-time liver function monitoring should be explored. Dynamic liver function monitoring: the impact of TACE plus targeted immunotherapy on short-term and long-term liver function reserve should be dynamically assessed, particularly in patients with CP scores 7–9. Novel biomarkers with high sensitivity and specificity for real-time liver function monitoring should be explored. Develop predictive models to guide patient selection for conversion therapy: developed and validated predictive models are urgently needed to identify patients with intermediate-stage HCC who are most likely to achieve successful downstaging and pathological complete response following intensive combination therapy. Such models would enable more precise patient selection, maximising the chance of curative outcomes while avoiding futile overtreatment. Develop predictive models to guide patient selection for conversion therapy: developed and validated predictive models are urgently needed to identify patients with intermediate-stage HCC who are most likely to achieve successful downstaging and pathological complete response following intensive combination therapy. Such models would enable more precise patient selection, maximising the chance of curative outcomes while avoiding futile overtreatment. Compare against other advanced treatment standards: future trials should move beyond comparing TACE plus targeted immunotherapy against TACE alone. To answer the most relevant clinical questions, the new combination should be directly compared against other contemporary, effective treatment options, such as TACE plus TKIs (eg, lenvatinib) or HAIC-based combination targeted immunotherapy. Compare against other advanced treatment standards: future trials should move beyond comparing TACE plus targeted immunotherapy against TACE alone. To answer the most relevant clinical questions, the new combination should be directly compared against other contemporary, effective treatment options, such as TACE plus TKIs (eg, lenvatinib) or HAIC-based combination targeted immunotherapy. Recognise the limitation of PFS and explore composite endpoints: consistent with the principle of establishing OS as the gold standard (as outlined in Perspective 1), it is important to recognise that PFS is an imperfect surrogate endpoint in HCC, as it does not capture treatment-related liver function deterioration—an important determinant of prognosis. Therefore, future trials of TACE combination therapy should explore composite endpoints, such as time to progression or time to liver function deterioration, to provide a more holistic assessment of net clinical benefit. Recognise the limitation of PFS and explore composite endpoints: consistent with the principle of establishing OS as the gold standard (as outlined in Perspective 1), it is important to recognise that PFS is an imperfect surrogate endpoint in HCC, as it does not capture treatment-related liver function deterioration—an important determinant of prognosis. Therefore, future trials of TACE combination therapy should explore composite endpoints, such as time to progression or time to liver function deterioration, to provide a more holistic assessment of net clinical benefit. Foster global collaboration to validate and discover prognostic biomarkers: international consortia are crucial to discover and validate robust prognostic and predictive biomarkers across diverse aetiologies (eg, hepatitis B virus, hepatitis C virus and non-alcoholic fatty liver disease) and populations. This will ensure that stratification models are globally applicable and can identify universal principles of treatment response. Foster global collaboration to validate and discover prognostic biomarkers: international consortia are crucial to discover and validate robust prognostic and predictive biomarkers across diverse aetiologies (eg, hepatitis B virus, hepatitis C virus and non-alcoholic fatty liver disease) and populations. This will ensure that stratification models are globally applicable and can identify universal principles of treatment response. Conduct context-specific health economic evaluations: cost-effectiveness analyses across varied healthcare systems are essential to guide the sustainable and equitable implementation of these advanced, often costly, therapies. Such studies will inform resource allocation and access policies, ensuring patients who stand to benefit most can receive treatment regardless of geographic or economic setting. Conduct context-specific health economic evaluations: cost-effectiveness analyses across varied healthcare systems are essential to guide the sustainable and equitable implementation of these advanced, often costly, therapies. Such studies will inform resource allocation and access policies, ensuring patients who stand to benefit most can receive treatment regardless of geographic or economic setting. In conclusion, the integration of TACE with targeted therapy and immunotherapy represents a paradigm shift, a strategy that is most compellingly indicated for patients with intermediate-stage HCC who are predicted to derive high benefit through multi-dimensional stratification. Future efforts must focus on refining these stratification tools and addressing the practical challenges of global implementation to fully realise the potential of this promising strategy. Footnotes Funding: The work was supported by the National Key Research and Development Program of China (Grant number: 2023YFC2414000) and the National Natural Science Foundation of China (Grant number: 82372060 and 82572330). Patient consent for publication: Not applicable. Ethics approval: Not applicable. Provenance and peer review: Not commissioned; externally peer reviewed. Data availability statement Not applicable. References Associated Data Data Availability Statement Not applicable.