Hepatocellular carcinoma (HCC) is the sixth most common cancer globally and the third leading cause of cancer-related mortality (1). Surgical resection is potentially curative for patients with early-stage HCC, but about 50-70% of HCC patients are unfit for surgical resection due to advanced tumor stage (2). For patients with advanced HCC, multitarget tyrosine kinase inhibitors (TKIs) sorafenib and lenvatinib used to be the first line treatment, but the efficacy was modest and only conferred limited survival benefits. Recently, immune checkpoint inhibitors (ICIs) including anti-programmed death-1 (PD-1)/programmed death-ligand 1 (PD-L1) monoclonal antibodies have shown encouraging results in multiple cancers (3, 4). The IMbrave150 trial demonstrated a better tumor response and survival with the combination of atezolizumab and bevacizumab versus sorafenib (5), resulting in the accelerated approval of this combination for advanced HCC, which also started a new tide of clinical studies leading by ICI-based combination therapy in HCC. Among these, the combination of TKI and ICI was also investigated and promising efficacy was reported in the KEYNOTE 524 trial with a median progression‐free survival (PFS) of 9.3 months and a median overall survival (OS) of 22 months for patients treated with pembrolizumab plus lenvatinib (6). But subsequent phase III study LEAP-002 which compared pembrolizumab plus lenvatinib with lenvatinib failed to meet its primary endpoint (7). However, TKIs and their combination with ICIs are still important treatment options for advanced HCC.

On the other hand, although atezolizumab plus bevacizumab is now the preferred first-line therapy for advanced HCC, more than 50% of patients still do not respond to treatment. To further increase the objective response rate (ORR) and improve patient survival, loco-regional treatments (LRTs) including transcatheter arterial chemoembolization (TACE), radiotherapy (RT) and hepatic arterial infusion chemotherapy (HAIC) have been combined with ICI-based systemic therapies for unresectable HCC (uHCC) (8, 9). TACE, which is the first therapeutic modality to provide survival benefits for patients with uHCC (10), is the standard treatment for Barcelona Clinic Liver Cancer (BCLC) stage B HCC in western countries (11) and BCLC stage B/C HCC in Asia‐Pacific (12). The results of LAUNCH and TACTICS trial showed that TACE combined with TKI significantly prolonged the survival of HCC patients (13, 14). Mechanistically, the upregulated expression of vascular endothelial growth factor (VEGF) and fibroblast growth factor (FGF) by TACE (15) could be effectively inhibited by TKIs (16), leading to better clinical outcomes in patients treated with TACE plus TKIs (17). In addition, the potential benefit of TACE plus PD‐1 inhibitor has also been revealed. A retrospective study has reported that TACE can be safely integrated with PD-1 inhibitor and lead to significant delay in tumor progression and disease downstaging in selected patients (18). Meanwhile, TACE with or without lenvatinib plus pembrolizumab is under investigation for intermediate-stage HCC not amenable to curative treatment in phase 3 LEAP-012 study (19). However, currently, whether the addition of TACE to the combination of TKIs and PD-1 inhibitor (ICI+TKI+TACE) is superior to ICI+TKI combination therapy is still not clear. This study leverages a retrospective cohort of HCC patients treated with ICI+TKI+TACE triple therapy or ICI+TKI doublet therapy to examine the differences of efficacy and patient outcome between the two regimens.

The success of IMbrave150 trial has launched a new era of ICI-based combination therapies in HCC. So far, the ORR was 29.8% with the combination of atezolizumab plus bevacizumab in the IMbrave 150 study (5), and the highest ORR of systemic therapies for HCC was reported as 36% with pembrolizumab plus lenvatinib in the Keynote 524 study (6). Currently, the efficacy improvement with ICI-based systemic therapies has clearly hit a plateau for HCC. In addition to systemic therapies, LRTs including TACE were also utilized in combination with ICIs to further increase the efficacy (8, 9). Except the conventional role of inducing tumor necrosis, TACE was also found to promote T-cell activation via abscopal effects (22). The tumor necrosis caused by TACE increased the release of tumor-associated antigens (23, 24), which has been proven to recruit DCs and increase AFP-specific CD4⁺T-cell response (25), thus synergizing with ICIs to increase cytotoxic T lymphocytes and decrease tumor-infiltrating Treg cells (26). Lenvatinib was reported to effectively inhibit the angiogenic growth factors triggered by the extensive ischemic necrosis in preclinical (27) or clinical studies (17), which was also an important factor associated with T-cell activation. Recently, several studies have investigated the efficacy of TKI plus ICI in combination with TACE and demonstrated an ORR of 46.7%-69.3% (28–35). The LEAP-012 trial demonstrated a significant and clinically meaningful improvement in PFS for patients with unresectable HCC compared to TACE plus placebo. Similarly, the combination of durvalumab, bevacizumab, and TACE in the EMERALD-1 study also showed potential to establish a new standard of care (36, 37). Nonetheless, the studies comparing the efficacy of ICI+TKI+TACE with ICI+TKI were still limited. Herein, we leveraged a retrospective cohort of HCC patients treated with triple therapy or doublet therapy and firstly revealed that the outcomes were not significantly improved after the addition of TACE to ICI plus TKI combination therapy.

The efficacy of TKI or ICI monotherapy was limited in clinical studies, with an ORR of 6.5% with sorafenib (3), 15-17% with PD-1 inhibitors (4), and 18.8% with lenvatinib (3). The anti-angiogenesis function of TKIs including lenvatinib was involved in several steps of T-cell activation, including the restoration of antigen presentation, the priming and activation of T-cell responses, and the modulation of the tumor immune microenvironment (38, 39). Furthermore, lenvatinib was also found to regulate pathways modulating antitumor immunity, including the reduction of tumor PD-L1 expression levels and Treg differentiation by blocking Fibroblast growth factor receptor-4 (FGFR4) (40) and reducing the Treg proportion via TGF-b pathway inhibition (41). Indeed, the combination therapy of lenvatinib and pembrolizumab initially displayed encouraging efficacy for HCC (6). However, subsequent phase III study LEAP-002 only achieved an ORR of 26.1% and failed to meet its primary endpoint on OS and PFS (7). Similar negative results were also observed in the Cosmic-312 study (42). However, subgroup analysis of LEAP-002 revealed that patients with higher AFP levels, extrahepatic spread or MVI may significantly benefit from lenvatinib plus pembrolizumab, indicating the efficacy of this combination in selected patients (7). Likewise, our data also showed that the outcomes of patients treated with ICI+TKI+TACE triple therapy were not significantly improved versus ICI+TKI doublet therapy. But subgroup and interaction analysis of PFS identified that patients with no preLRT or treated with lenvatinib as combination are more likely to benefit from triple therapy, further highlighting the importance of patient selection in the combination therapy for HCC, which indeed merits further investigation.

LRTs are common regimens to treat patients with uHCC (22). The combinations of LRTs and ICI-based systemic therapies are promising therapeutic strategies and are currently under investigation in HCC (8). Mechanistically, other than tumor elimination, LRT can promote antigenic or immunogenic cell death, thereby augmenting tumor immunogenicity and synergizing with ICIs (8, 43). However, the optimal patient population which is suited for combined locoregional treatments is still not clear. With the recent overwhelming trend of ICI-based combinational therapy in HCC, the triple therapy of ICI+TKI+TACE has increasingly been utilized in clinical practice. However, our real-world data revealed no significant improvement in PFS and OS with triple therapy compared to doublet therapy, suggesting that the addition of TACE to the ICI+TKI combination therapy did not necessarily improve outcomes in the overall population, thereby challenging the notion that “more is better” in HCC combination therapy. Furthermore, the benefit of triple therapy was only observed in selected patients (without preLRT or treated with lenvatinib as combination), emphasizing the importance of tumor heterogeneity and personalized therapy in HCC.

The REFLECT trial and subsequent re-analysis of data showed non-inferiority of lenvatinib versus sorafenib in terms of OS, as well as statistically significant and clinically meaningful improvement in PFS, time to progression, and ORR (3). Lenvatinib is a multitarget tyrosine kinase inhibitor analogous to sorafenib, with unique high selectivity for FGF receptor 1-4 (FGFR1-4) (3). Of the receptors, FGFR4 is considered a potent target of lenvatinib in the treatment of HCC (44), providing a mechanistic rationale for that lenvatinib resulted in a statistically significant improvement in ORR compared with sorafenib in REFLECT trial. Likewise, Yang et al. have reported TACE plus lenvatinib was significantly superior to TACE plus sorafenib with respect to OS, PFS, and ORR (45). In addition, lenvatinib was also found to target FGFR4 to enhance antitumor response of ICI in HCC (40). Consistently, our data showed that patients treated with ICI+lenvatinib displayed superior PFS versus ICI+sorafenib ( Supplementary Table S2 ). Importantly, we further revealed that in the patients treated with ICI+lenvatinib, the addition of TACE can remarkably improve PFS in comparison with those treated with ICI+sorafenib, suggesting the synergistic effect between TACE and ICI+lenvatinib.

In the ICI+TKI group, the point of treatment initiation was defined as the initiation of systemic therapy, while in the ICI+TKI+TACE group, it was defined as the initiation of TACE or systemic therapy, whichever comes first. Notably, treatment sequences in landmark trials vary. In LEAP-012 trial, TACE was initiated 2-4 weeks after systemic therapy, while in EMERALD-1 study, durvalumab with or without bevacizumab was administered 7 days post-TACE (36, 37). Moreover, current guidelines also lack consensus on optimal sequencing of TACE and systemic therapy. Due to the nature of real-world studies, the point of treatment initiation was inconsistent between the two groups in our study. Nonetheless, the interval between TACE and systemic therapy was within an acceptable range, and patients in triple therapy group could be considered to have received concurrent TACE and ICI+TKI therapy, and the impact of this inconsistency on the results was largely limited.

In clinical trials, post-progression therapy or treatment crossover after progression increases the fragility of overall survival as an endpoint (46). The COSMIC-312 study (42) only observed an improved PFS in HCC patients treated with cabozantinib plus atezolizumab versus sorafenib but not OS. Similarly, our data also didn’t detect significant interactions between preLRT or TKIs-combined with treatment via interaction analysis of OS, which was probably due to that almost 3/4 progressed patients in both triple therapy and doublet therapy groups had received subsequent post-progression therapies.

This study has several limitations. First, this is a single center retrospective study and almost all patients had an etiology of HBV infection, which limits the generalizability of our findings. Second, the portion of patients with ECOG PS 1 or Child-pugh score B was significantly higher in ICI+TKI doublet therapy group than that in triple therapy group before PS matching ( Supplementary Table S1 ), indicating that patients with relatively good performance or preserved liver functions tended to receive ICI+TKI+TACE triple therapy. This bias is probably because, currently, as a potent means of LRT, TACE is preferred to be combined with ICI plus TKI therapy for uHCC when patient conditions permit in the real world. Although PS matching balanced the differences of baseline characteristics between the two groups, our findings still should be confirmed in prospective randomized studies. Third, unlike the LEAP-012 trial (19), in which TACE was limited to 2 treatments per tumor, repeated TACE was allowed in this study, which was also in accordance with the situation in the real world.

In conclusion, compared with ICI+TKI doublet therapy, ICI+TKI+TACE triple therapy showed minimal difference in the efficacy for uHCC. However, patients with no preLRT or treated with lenvatinib can significantly benefit from ICI+TKI+TACE triple therapy in terms of PFS, providing a rationale for conducting prospective studies to assess the efficacy of adding TACE to ICI plus TKI combination therapy in selected patients with uHCC.