Comparison of Efficacy between Lenvatinib and Bevacizumab in Combination of Immune Checkpoint Inhibitor and Interventional Triple Therapy in Chinese Advanced Hepatocellular Carcinoma: The CLEAP 2302 Study Open Access

Hepatocellular carcinoma (HCC) is globally recognized as the sixth most prevalent form of cancer and ranks as the third leading cause of cancer-related mortality [1], while it ranks fourth in incidence and second in mortality in China [2]. Most cases of HCC are diagnosed at a relatively advanced stage, making them ineligible for resection. Conversion therapy has been gaining attention in recent years. For unresectable HCC, systemic therapy options like single-agent lenvatinib and sorafenib are considered first-line treatments [3‒5]. However, over time, novel therapeutic alternatives have emerged, such as immune checkpoint inhibitors (ICIs) and ICIs or tyrosine kinase inhibitors (TKIs), such as atezolizumab plus bevacizumab (T+A) and tremelimumab plus durvalumab and nivolumab plus ipilimumab providing patients with advanced HCC a broader array of treatment options [6‒10].

In recent years, systemic therapy for liver cancer has achieved favorable results and has become an important means of conversion therapy. Among these, the efficacy of lenvatinib monotherapy has been proven to be no inferior to sorafenib, with the highest single-agent objective response rate (ORR) being 18.8% [11]. The efficacy rate of the T+A dual-drug regimen is approximately 30% [12] and has been widely used in clinical practice [13]. Although the combination of lenvatinib and pembrolizumab (the “Combo” regimen) showed negative results in the LEAP-002 study [14], there are also phase II studies (such as Keynote 524, ORR = 36%) [15] that have made this regimen a choice for many physicians in clinical practice. Nevertheless, combining local treatments such as radiotherapy, transarterial chemoembolization (TACE), and hepatic arterial infusion chemotherapy (HAIC) with systemic therapy can achieve an even higher conversion rate. For example, triple therapy using the small-molecule TKI lenvatinib and the large-molecule bevacizumab has received support from extensive phase II clinical research data, with an ORR of 56.7%–74.4% [3, 4, 16, 17]. Additionally, phase III clinical trials such as EMERALD-1 [18] and LEAP-012 [19] have already reported their results, and further analyses are ongoing.

Although lenvatinib and bevacizumab have achieved promising results in the treatment of unresectable HCC, the optimal combination regimen remains uncertain, particularly between lenvatinib-based and bevacizumab-based triple combinations. Direct comparisons between these treatments are lacking. Therefore, this study aimed to compare the prognostic impact of lenvatinib-based and bevacizumab-based triple therapy in patients with unresectable HCC. We believe that this comparison is critical in determining the best treatment options for patients with advanced HCC. To achieve this goal, we planned to conduct a retrospective cohort study using medical record data from patients with unresectable HCC who received triple therapy with either lenvatinib or bevacizumab. The primary outcome will be overall survival (OS) and the secondary outcome will be progression-free survival (PFS). Based on data from multiple centers across the country from the CLEAP research group [20‒25], we designed this study. We hope this study can provide valuable information to guide clinical practice and improve the prognosis of patients with unresectable HCC.

The CLEAP database contains data on 735 uHCC patients treated at 21 clinical centers in China between April 2017 and December 2023. Patients with the following features were included: (1) HCC diagnosed by clinical or pathological features according to the American Association for Study of Liver Diseases (AASLD) practice guidelines [26]; (2) belong to Chinese liver cancer (CNLC) stage IIb, IIIa, and IIIb [27], equivalent to Barcelona clinical liver cancer stage B or C; (3) received lenvatinib-based triple therapy as first-line treatment, or bevacizumab-based triple therapy as first-line treatment; and (4) follow-up data were complete and available. Exclusion criteria were as follows: diagnosis of other malignant tumors; use of interventional therapy other than TACE or HAIC; and use of systemic treatment regimens other than PD-1 or PD-L1 antibodies. Eligible patients were divided into the lenvatinib group and the bevacizumab group according to the treatment regimen adopted for analysis. Treatments were administered according to the product package insert and previous studies, based on current efficacy and safety data, prior treatment, and drug availability [28].

Use of both treatments followed the same inclusion criteria as registered trials. The current study was approved by the Ethics Committee of each center and adhered to clinical practice guidelines, the Declaration of Helsinki, and local laws. All patients provided informed consent to receive treatment for HCC and to have their medical records used for research purposes.

Patients were treated with lenvatinib/bevacizumab in combination with PD-1/PD-L1 antibodies and interventional therapies, including TACE and HAIC, as first-line treatment. The PD-1 antibodies used in this study included pembrolizumab, nivolumab, sintilimab, tislelizumab, toripalimab, and camrelizumab, while the PD-L1 antibodies included durvalumab and atezolizumab. Notably, 32.1% of patients initially underwent interventional therapy, while the remaining patients received a combination of the triple therapy regimen concurrently.

TACE is a minimally invasive procedure in which chemotherapy agents are directly delivered to the tumor through the hepatic artery. This is followed by embolization to block the tumor’s blood supply, thereby enhancing the local effectiveness of the drugs. The standard TACE regimen typically includes epirubicin and lipid-based platinum compounds. For cisplatin, the recommended dose is 50 mg/m² of body surface area, administered at a rate of 4–10 mL/min. TACE treatments are generally performed every 3–6 weeks, with the frequency adjusted based on the patient’s clinical condition and response.

In contrast, HAIC involves the direct infusion of chemotherapy agents into the hepatic artery, achieving high local drug concentrations while minimizing systemic exposure. In this study, the HAIC regimen primarily consisted of the FOLFOX protocol, which includes oxaliplatin, leucovorin, and fluorouracil. Recommended doses for HAIC therapy are 85 mg/m² for oxaliplatin, 400 mg/m² for cisplatin, and 2,400 mg/m² for fluorouracil, administered every 3–4 weeks. As with TACE, the frequency and dosage of HAIC treatments were tailored to the patient’s clinical response and condition.

Treatment with modality continued until disease progression, drug toxicity intolerance, or eligibility for therapeutic liver resection. In clinical practice in China, TACE is predominantly used for patients with CNLC stage IIa–IIb (equivalent to BCLC stage B) HCC, particularly those with multinodular tumors without vascular invasion. Conversely, HAIC is more commonly employed for patients with CNLC stage IIIa–IIIb (BCLC stage C) HCC, especially those with large tumors or portal vein invasion. Additionally, all patients with HBV infection received oral antiviral drugs, such as entecavir or tenofovir, as part of their cancer treatment.

Tumor evaluation was conducted every 6–8 weeks using CT or MRI for intrahepatic tumors and upper abdominal metastases, chest CT for lung metastases, and blood tests to assess liver function and tumor markers. OS was measured from the initiation of treatment until death. PFS was measured from the start of treatment until disease progression or death from any cause. Safety was evaluated based on the occurrence of severe (grade ≥3) treatment-related adverse events (TRAEs).

Categorical variables are summarized as counts (percentages), and continuous variables are summarized as means ± standard deviation. Differences in baseline characteristics between groups were estimated using Fisher’s exact test or Kruskal-Wallis test. The Kaplan-Meier method was used for survival analysis, and the log-rank test was used to compare OS and PFS curves. Survival outcomes were also analyzed by Cox regression, using univariable and multivariable models that included relevant variables. Subgroup analysis was conducted based on age (≤65 or >65 years), gender, HBV infection status, BCLC stage, CNLC stage, ALT level (<40 U/L or ≥40 U/L), platelet level (<100 × 10⁹ or ≥100 × 10⁹/L), and AFP level, with logarithms taken based on commonly used clinical standard values in China (log value <1.40 or ≥1.40).

In order to eliminate differences in potential confounding variables, logistic regression was used to calculate the propensity score (PS), and 1:3 matching was performed on the nearest individuals based on the PS, with the caliper value set to 0.05. Propensity score matching (PSM) was conducted using the “Matchit” package in R software, with covariates including HBV infection status, ECOG PS, Child-Pugh liver function class, and other variables. Sensitivity analysis using inverse probability of treatment weighting was performed on the same variables using the “RISCA” package in R software. All statistical analyses were performed using R software version 4.3.1 (http://www.rproject.org/). p value <0.05 was considered statistically significant.

This analysis included 371 consecutive patients who received lenvatinib or bevacizumab-based triple therapy between April 1, 2017, and December 31, 2023. The patients’ disposition is shown in Figure 1. In the overall population (Table 1), the mean age was 55.3 years (standard deviation, 10.7), and the majority were male (85%). Most patients had hepatitis B virus (HBV) infection (HBsAg, HBcAb, or HBeAb seropositivity) as the cause of uHCC (85%), and most had ECOG PS 0 (69%) and Child-Pugh class A liver function (87%). Based on CNLC criteria, 24%, 37%, and 39% of patients were diagnosed as stage IIb, stage IIIa, and stage IIIb, respectively.

Although most baseline variables were comparable between the two groups, significantly more patients were infected with HBV (88% vs. 79%; p = 0.021) and had ECOG PS 1 (40% vs. 4.4%; p < 0.001) in the lenvatinib group compared with the bevacizumab group. There were also significantly more patients with Child-Pugh class A liver function (90% vs. 78%; p = 0.001) in the lenvatinib group.

As of December 2023, a total of 102 (27.5%) patients had died in the overall population. The median OS for all patients was 32.3 months (95% CI, 27.9–36.0 months), and the median PFS was 17.7 months (95% CI, 13.9–22.1 months).

During the follow-up period, 62 patients (24.0%) died in the lenvatinib group and 40 patients (35.4%) died in the bevacizumab group. Compared with the bevacizumab group, the median OS in the lenvatinib group was significantly longer (36.0 months vs. 27.9 months, HR, 0.536; 95% CI, 0.344–0.835; p = 0.0016; Fig. 2a). The OS rates in the lenvatinib group at 12, 24, and 36 months were 85.8%, 62.5%, and 46.5%, respectively. Additionally, the PFS in the lenvatinib group was also significantly longer than that in the bevacizumab group (20.0 months vs. 12.1 months, HR, 0.649; 95% CI, 0.457–0.922; p = 0.0078; Fig. 2b). The PFS rates at 12, 24, and 36 months in the lenvatinib group were 64.2%, 44.0%, and 31.1%, respectively.

Multivariate analysis showed that the lenvatinib group had improved OS (HR, 0.51; 95% CI, 0.31–0.84; p = 0.009). In addition, aspartate aminotransferase (AST) ≥40 U/L indicated a poor prognosis (HR, 2.03; 95% CI, 1.13–3.63; p = 0.017) (Table 2).

PSM processing was performed on variables such as HBV infection, ECOG PS, and Child-Pugh liver function classification. To retain as many enrolled cases as possible, the matching ratio was set to 1:3. After matching, all baseline variables between the two treatment groups were comparable (all p > 0.05) (Table 1). Notably, after PSM adjustment, the OS in the lenvatinib group was significantly prolonged, with OS rates at 12, 24, and 36 months of 86.6%, 64.5%, and 55.3%, respectively, compared to 65.5%, 55.4%, and 49.9% in the bevacizumab group (HR, 0.524; 95% CI, 0.305–0.900; p = 0.01; Fig. 2c). However, after PSM adjustment, there was no significant difference in PFS between the two groups. The 12-, 24-, and 36-month PFS rates in the lenvatinib group were 59.4%, 41.4%, and 29.5%, respectively, while those in the bevacizumab group were 50.5%, 35.3%, and 23.5%, respectively (HR, 0.808; 95% CI, 0.535–1.222; p = 0.33; Fig. 2d).

After inverse probability of treatment weighting, the OS in the lenvatinib group remained significantly longer, with OS rates at 12, 24, and 36 months of 86.5%, 63.3%, and 54.5%, respectively, compared to 68.1%, 57.2%, and 51.9% in the bevacizumab group (HR, 0.549; 95% CI, 0.331–0.908; p = 0.01; Fig. 2e). The 12-, 24-, and 36-month PFS rates in the lenvatinib group were 59.2%, 40.4%, and 34.3%, respectively, while the PFS rates in the bevacizumab group were 52.3%, 38.3%, and 29.7% (Fig. 2f).

The analysis of OS and PFS among different subgroups revealed several noteworthy trends:

Conversion surgery (Fig. 3a): OS was significantly longer in the lenvatinib group compared to the bevacizumab group for patients undergoing conversion surgery (HR, 0.469; 95% CI, 0.256–0.859; p = 0.012). This highlights the potential survival benefit of conversion surgery in the lenvatinib cohort.

Body mass index (BMI) (Fig. 3b–e): In the lenvatinib group, obese patients demonstrated the best OS outcomes, while underweight patients had the poorest prognosis (p = 0.02, Fig. 3b). In contrast, no significant OS differences were observed among BMI subgroups in the bevacizumab group (p = 0.26, Fig. 3c). Regarding PFS, a trend toward improved outcomes was observed for obese patients in the lenvatinib group, though not statistically significant (p = 0.082, Fig. 3d). No PFS differences were observed among BMI subgroups in the bevacizumab group (p = 0.32, Fig. 3e).

Hypertension (Fig. 3f, g): For OS, patients without hypertension in the bevacizumab group exhibited a rapid mid-term decline, stabilizing later, with no significant differences observed overall (p = 0.09, Fig. 3f). For PFS, no significant differences were detected between hypertensive and nonhypertensive subgroups in either treatment group (p = 0.77, Fig. 3g).

Diabetes mellitus (Fig. 3h, i): In both the lenvatinib and bevacizumab groups, diabetic patients showed a trend toward better OS compared to nondiabetic patients, though this was not statistically significant (p = 0.062, Fig. 3h). Similarly, a trend toward improved PFS was noted in diabetic patients in both groups (p = 0.19, Fig. 3i).

Hyperlipidemia (Fig. 3j, k): No significant differences were observed in OS (p = 0.093, Fig. 3j) or PFS (p = 0.64, Fig. 3k) between patients with and without hyperlipidemia in either treatment group.

Alcohol consumption (Fig. 3l, m): In the lenvatinib group, patients without a history of alcohol consumption exhibited significantly better OS compared to other subgroups (p = 0.0066, Fig. 3l). However, no significant differences in PFS were observed based on alcohol consumption in either group (p = 0.36, Fig. 3m).

Number of interventional treatments (Fig. 3n, o): In the lenvatinib group, patients undergoing 4–6 interventional treatments tended to achieve higher and more stable OS rates, though not statistically significant (p = 0.064, Fig. 3n). In the bevacizumab group, no significant OS differences were observed across subgroups with varying numbers of interventional treatments (p = 0.79, Fig. 3o).

These findings provide insights into how patient background factors and treatment modalities influence survival outcomes, underscoring the importance of personalized approaches in managing HCC.

In subgroup analysis, patients aged ≤65 years significantly benefited from lenvatinib-based triple therapy compared with bevacizumab-based triple therapy (OS HR, 0.48; 95% CI, 0.27–0.85; p = 0.012), whereas no difference was observed in patients older than 65 years (OS HR, 1.02; 95% CI, 0.36–2.93; p = 0.968; Fig. 4). Male patients significantly benefited from lenvatinib combination therapy (HR, 0.48; 95% CI, 0.28–0.82; p = 0.008), whereas no difference was observed in female patients (HR, 1.70; 95% CI, 0.31–9.34). Patients without HBV infection also significantly benefited from lenvatinib combination therapy (HR, 0.18; 95% CI, 0.04–0.80; p = 0.025), whereas no difference was observed in patients with previous HBV infection (HR, 0.63; 95% CI, 0.36–1.11; p = 0.108). Interestingly, patients with extrahepatic metastasis (BCLC stage C, CNLC stage IIIb) (HR, 0.51; 95% CI, 0.30–0.89; p = 0.017) or vascular tumor thrombus (BCLC stage C, CNLC stage IIIa) (HR, 0.39; 95% CI, 0.17–0.87; p = 0.021) showed a significant survival benefit with lenvatinib combination therapy. In addition, patients with ALT <40 U/L (HR, 0.45; 95% CI, 0.23–0.88; p = 0.019), platelets ≥100 × 10⁹/L (HR, 0.51; 95% CI, 0.29–0.91; p = 0.022), and log 10 AFP ≥1.40 ng/mL (HR, 0.55; 95% CI, 0.31–0.98; p = 0.044) also showed a significant survival benefit with lenvatinib combination therapy (Fig. 4).

The swimmer plot for OS and the last dose of treatment are shown in Figure 5. As of the last follow-up, 62 patients (40.8%) in the lenvatinib group and 34 patients (41.5%) in the bevacizumab group were still receiving treatment. Among those who discontinued the triple therapy, in the lenvatinib group, 63.3% of patients switched to regorafenib, while 26.7% switched to the T+A regimen, and the remaining patients were lost to follow-up. In the bevacizumab group, 54.2% of patients switched to regorafenib, 25.0% switched to donafenib, 14.6% switched to lenvatinib, and the remaining patients were lost to follow-up. The reasons for discontinuation included successful conversion to resectable disease, adverse events, drug resistance or tumor progression, lack of response, or loss to follow-up.

Regarding the best overall response, 6 patients (4.8%) in the lenvatinib group achieved a complete response (CR), and 47 patients (37.9%) achieved a partial response. In the bevacizumab group, 1 patient (1.9%) achieved CR, and 18 patients (33.3%) achieved partial response (Figure 6). The ORRs were 42.7% in the lenvatinib group and 35.2% in the bevacizumab group. Additionally, the disease control rates (DCRs) were 88.7% for lenvatinib and 96.3% for bevacizumab.

The conversion rate to surgery was evaluated in both groups during the study. The rates were 31.0% in the lenvatinib group and 35.4% in the bevacizumab group (p = 0.406), with no statistically significant difference observed. These results suggest that both therapies have a similar potential to enable conversion to surgery, despite differences in other efficacy metrics such as OS and PFS.

The safety profile of lenvatinib and bevacizumab was assessed through the incidence of TRAEs (Table 3). The overall incidence of any grade TRAEs was 86% in the lenvatinib group and 91% in the bevacizumab group, with no statistically significant difference observed between the two (p = 0.22). For grade 3 and 4 TRAEs, the incidence was 34% in the lenvatinib group and 38% in the bevacizumab group, also showing no significant difference (p = 0.67).

The most frequent TRAEs included elevated gamma-glutamyl transferase, elevated AST, and decreased hemoglobin (HGB). Elevated gamma-glutamyl transferase was observed in 52% of the lenvatinib group and 43% of the bevacizumab group (p = 0.40), with grade 3 and 4 incidences of 9.2% and 9.8%, respectively (p = 1.00). Elevated AST was reported in 52% of the lenvatinib group and 55% of the bevacizumab group (p = 0.68), with grade 3 and 4 incidences of 6.6% and 7.3%, respectively (p = 0.79). Decreased HGB was noted in 45% of the lenvatinib group and 44% of the bevacizumab group (p = 0.89), with grade 3 and 4 incidences of 6.6% and 4.9%, respectively (p = 0.78).

Proteinuria was observed in 15.8% of the lenvatinib group and 9.8% of the bevacizumab group (p = 0.24), with grade 3 and 4 incidences of 5.9% and 3.7%, respectively (p = 0.55). Gastrointestinal (GI) bleeding was reported in 14% of the lenvatinib group and 23% of the bevacizumab group (p = 0.10), with grade 3 and 4 incidences of 7.9% and 15%, respectively (p = 0.12). Notably, the bevacizumab group demonstrated a higher incidence of decreased platelet counts (PLT) compared to the lenvatinib group (39.0% vs. 29.6% for any grade, and 12.2% vs. 5.9% for grades 3 and 4), which may contribute to the higher incidence of GI bleeding in this cohort. This observation suggests a potential link between thrombocytopenia and GI hemorrhagic events (Fig. 7).

The treatment of liver cancer is progressing rapidly. Although dual therapy is the mainstream, triple therapy combining systemic therapy with locoregional therapy has gained attention due to its high ORR rates [3, 4]. The EMERALD-1 study [29] found that combining interventional therapy with bevacizumab and durvalumab shows a statistically significant improvement in PFS (median PFS 15.0 vs. 8.2 months; HR = 0.77; CI: 0.61–0.98; p = 0.032). The LEAP-012 study [19] demonstrated a significant improvement in PFS with TACE plus lenvatinib and pembrolizumab compared to TACE plus placebo (median PFS 14.6 vs. 10.0 months; HR = 0.66; 95% CI: 0.51–0.84; p = 0.0002). Additionally, several phase III studies combining TACE/HAIC with targeted immunotherapy have been conducted, including TALEN-TACE (NCT04712643), TACE-3 (NCT04268888), and CheckMate-74W (NCT04340193) [30]. This includes previous retrospective studies by CLEAP [23], which also demonstrated the advantages of triple therapy over systemic therapy without TACE/HAIC. Targeted therapy is mainly divided into two categories: large-molecule bevacizumab and small-molecule TKIs. Bevacizumab has shown the best outcomes when combined with atezolizumab [13, 28] or sintilimab [31]. Small-molecule multitarget drugs include lenvatinib, sorafenib, donafenib, apatinib, and second-line regorafenib and cabozantinib. The large-molecule ramucirumab, which targets VEGFR2, is rarely used. Among small molecule multitarget drugs, lenvatinib has the highest ORR of 18.8% [32]. Given the positive results from the LEAP-012 study [19], which demonstrated superior efficacy of TACE combined with lenvatinib and pembrolizumab compared to TACE alone in patients with HCC, the combination of lenvatinib and immune checkpoint inhibitors (ICI) continues to be favored by many clinical experts. This supports the potential of lenvatinib plus ICB to enhance the therapeutic outcomes of TACE, further validated by encouraging results from several phase II studies [3, 4, 17, 33, 34]. Currently, a phase III RCT combining lenvatinib and toripalimab is also in progress. Therefore, comparing the large-molecule bevacizumab with the small-molecule lenvatinib in the context of triple therapy is both important and interesting.

This study is the first real-world, multicenter comparison of the efficacy and safety of triple therapy between lenvatinib and bevacizumab, both combined with PD-1 antibodies and interventional therapy, as first-line treatment in patients with uHCC. It reveals the advantages of lenvatinib combined with immunotherapy and interventional therapy in improving survival, particularly in patients who are not infected with HBV and have HCC with macrovascular invasion, normal alanine transaminase levels, normal platelet count, and low AFP levels. These findings are consistent with previous clinical observations and theories [6, 35‒43]. Notably, our study found that patients without HBV infection and those with macrovascular invasion tend to benefit more from lenvatinib-based triple therapy, resonating with the findings of Camma et al. [44]. Additionally, our study highlights the survival differences in patients with extrahepatic metastasis and tumor thrombus, which aligns with discussions on treatment strategies for metastatic HCC in previous studies [45]. Our findings suggest that multiple factors need to be considered when selecting appropriate patients for treatment, including but not limited to viral status, tumor characteristics, liver inflammation, and AFP levels [8, 27, 41, 42, 46]. This underscores the importance of personalized treatment in uHCC.

The current study revealed a higher ORR in lenvatinib-based triple therapy compared to bevacizumab-based triple therapy. Several factors may contribute to this difference. First, as reported in the literature, the ORR is 18.8% for lenvatinib monotherapy [11], 21.0% for sintilimab combined with bevacizumab biosimilar IBI305 [31], and 30% for T+A [47], while the phase II Keynote-524 study [15] and others have shown a 36% ORR. The higher proportion of sintilimab plus IBI305 in the bevacizumab group may explain part of this outcome. Second, lenvatinib targets VEGFRs, PDGFRs, Kit, FGFRs, and other pathways, and its multitarget mechanism of action makes it more effective in treatment [7, 26, 35, 42, 48], potentially leading to a higher ORR and faster clinical response. Third, lenvatinib has also been reported to modulate antitumoral immunity [49, 50], which may contribute to its higher ORR.

Our study found that OS in the lenvatinib group was significantly better than in the bevacizumab group, which may be attributed to several factors, including a higher depth of response in the lenvatinib group. Literature has indicated a correlation between depth of response and OS [51]. This may also explain the higher ORR in the lenvatinib group. Second, bevacizumab treatment may be associated with adverse events such as bleeding [13], which could negatively impact survival. Also, the similar PFS but higher percentage of death and shorter OS in the bevacizumab may suggest a higher portion of nonprogressive death. However, this hypothesis requires further evidence. Third, for patients undergoing conversion surgery, the withdrawal period for lenvatinib is relatively short (1–2 weeks), whereas bevacizumab requires 4–6 weeks. The shorter discontinuation time may have reduced the risk of tumor progression and spreading due to discontinuation in the lenvatinib group, thereby positively impacting OS. Additionally, available data provide further clues. In the LEAP-002 study [14], the median OS of the lenvatinib combination was 21.2 months, and the OS of the Chinese subgroup was 32.3 months. The OS data from the T+A regimen, which was 24.0 months in the Asian subgroup [13], also support the long-term survival advantage of lenvatinib in triple therapy. Additionally, our study observed that patients with diabetes mellitus and obesity in the lenvatinib group exhibited better OS. This finding aligns with emerging evidence that metabolic factors may positively influence treatment outcomes in advanced HCC. Specifically, the improved OS in these subgroups could be related to altered drug metabolism and enhanced efficacy of lenvatinib, potentially mediated through its impact on the tumor microenvironment. This hypothesis is supported by recent studies, demonstrating that metabolic alterations, such as glucose restriction or inhibition of glucose transporters, can significantly enhance the anticancer effects of targeted therapies in HCC models [52]. Therefore, our findings underscore the need to further explore the role of metabolic status in optimizing personalized treatment strategies for patients with unresectable HCC.

Despite the ORR and OS advantage in the lenvatinib group, the DCR was relatively lower in the lenvatinib group, which may account for the nonsignificant difference in PFS between the two treatment groups. The median PFS for the LEAP002 study [14] was 8.2 months, compared to 4.6 months in the ORIENT-32 study [28] and 6.8 months in the IMBRAVE-150 study [13]. Since PFS was similar in both groups, these data may explain why lenvatinib and bevacizumab performed similarly in triple therapy. This study also revealed that the PD rate was higher for lenvatinib than for bevacizumab, indicating that bevacizumab may be more stable in controlling disease progression [53].

The difference between PFS and OS can be attributed to several factors. First, we speculate that the duration of response for bevacizumab may be relatively short [47], meaning that lenvatinib may produce a more rapid tumor shrinkage effect once it takes effect [54], thus acceleratting conversion to surgical resection, with a relatively short period of drug discontinuation. Additionally, the incidence of bevacizumab-related adverse reactions, such as GI bleeding [55], is relatively higher, which may lead to a scenario where bevacizumab achieves a better DCR but has a higher proportion of deaths.

However, retrospective studies are at risk of selection bias, and our results, although statistically significant, need to be validated by more rigorous prospective studies and randomized controlled trials [56, 57]. Moreover, our study did not directly compare lenvatinib and bevacizumab triple therapy in terms of drug resistance, toxicity, and quality of life, which are important considerations in practical clinical decisions [58, 59].

In conclusion, our study provides real-world evidence for the OS benefits of lenvatinib-based triple therapy over bevacizumab-based therapy in a subset of uHCC patients. Moreover, individual patient characteristics, including viral status and biomarkers, should be considered when selecting the most appropriate treatment. Further research is essential to refine patient stratification and explore biomarkers for predicting treatment response [60, 61].

This study involving human participants was reviewed and approved by Zhongshan Hospital, Fudan University (Approval No. B2022-195R). The patient/participants provided their written informed consent to participate in this study.

The authors have no conflicts of interest to declare.

This research was supported by funds as follows: (1) Tianjin Medical University Cancer Hospital, Precision Treatment Technology Construction Project for Cancer Surgery, ZLWKJZZL14; (2) Zhongguancun Precision Medicine Foundation, Medical and Health Public Welfare – Cancer Medical Research Special Project, ZGC-YXKY-ZL004 & ZGC-YJWC-ZLYXKY-9. (3) Tianjin Key Medical Discipline(Specialty) Construction Project (TJYXZDXK-011A).

Z.P., J.C., and W.Z. designed experiments and drafted the manuscript. D.L., L.F., X.Z., X.Z., Y.P., J.L., C.H., R.J., S.S., X.Z., H.L., X.Y., K.H., X.S., D.W., Y.Z., J.Z., B.X., S.G., T.L., S.Z., L.Z., H.Z., YZ, TW, MK, X.L., T.P., K.W., L.X., H.L., T.S., and H.S. collected the cases. W.Z. approved the final version. The China Liver Cancer Study Group Yonug Investigators (CLEAP) provided the platform for data maintenance. All authors contributed to the article and approved the submitted version.

Zhaolong Pan, Dongming Liu, Junbo Cao, and Xiaodong Zhu contributed equally to this work.

The original contributions presented in the study are included in the article. Further inquiries can be directed to the corresponding authors.