Hepatic arterial infusion or intravenous infusion of adebrelimab, combined with bevacizumab and HAIC of the FOLFOX regimen for advanced unresectable hepatocellular carcinoma (HAIBrave-001 trial): a study protocol of a prospective, double-arm, phase II trial

Introduction

Hepatocellular carcinoma (HCC) is the sixth most common cancer worldwide and the third leading cause of cancer-related death (1). Approximately 906,000 new cases and 830,000 deaths occur each year, with China accounting for about 55% of incident cases (2). As HCC is often insidious in onset, most Chinese patients have advanced stage disease and hence forfeit the chance for radical surgical resection at initial diagnosis. According to the Barcelona Clinic Liver Cancer (BCLC) staging system, systemic therapy is recommended as first-line treatment (3). Currently available first-line targeted agents such as sorafenib and lenvatinib generally yield objective response rates (ORRs) less than 20%. Although programmed cell death ligand 1 (PD-L1) antibody plus bevacizumab (the T+A regimen) has been shown to achieve a median overall survival (OS) of 19.2 months, the ORR remained only 27.3% (4).

Hepatic arterial infusion chemotherapy (HAIC) using the FOLFOX regimen is an emerging effective treatment for advanced stage HCC (5,6). Compared with conventional intravenous chemotherapy, HAIC markedly increases local drug concentrations while reducing systemic toxicity. By catheterizing the hepatic artery, high concentrations of chemotherapeutic agents can be continuously delivered directly into the tumor-feeding artery, thereby increasing intratumoral exposure while reducing distribution into the peripheral circulation and minimizing systemic adverse reactions (7). In 2022, a phase III trial reported by Lyu et al. conducted at Sun Yat-sen University Cancer Center showed that, compared with sorafenib, HAIC-FOLFOX conferred greater survival benefit in advanced HCC, with an ORR of 40–50% (8). However, HAIC alone is still insufficient to achieve satisfactory long-term outcomes.

To overcome this bottleneck, the combination of locoregional interventional therapy, targeted therapy, and immunotherapy has become a trend in current research (9). In 2022, Lai et al. reported the data on HAIC combined with lenvatinib and toripalimab for advanced HCC; among 36 enrolled patients, the response rate reached 66.7% according to Modified Response Evaluation Criteria in Solid Tumors (mRECIST) (10). In 2023, Zhang et al. reported data on HCC patients receiving HAIC combined with camrelizumab and apatinib, with a response rate of 88.6% according to mRECIST (11). Although most of these prior studies were single-center with relatively small sample sizes, these encouraging findings suggest that HAIC plus targeted therapy and immunotherapy is a highly promising direction.

Based on this rationale, we hypothesize that HAIC combined with PD-L1 antibody plus bevacizumab may also exert synergistic antitumor activity in advanced unresectable HCC. Therefore, the HAIBrave-001 study was initiated to explore the efficacy and safety of HAIC-FOLFOX combined with adebrelimab (a PD-L1 antibody) and bevacizumab in advanced unresectable HCC. In clinical practice, immune checkpoint inhibitors (ICIs) are administered intravenously. In recent years, however, accumulating evidence has suggested that arterial infusion of ICIs is safe (12). Theoretically, PD-L1 antibodies target PD-L1 antigens on the surface of tumor cells; local infusion may therefore achieve stronger immunotherapeutic effects while reducing systemic adverse reactions. Accordingly, the HAIBrave-001 study was designed with two randomized treatment arms: one arm receives the PD-L1 antibody by arterial infusion and the other by intravenous infusion. This study may answer two key questions: (I) the efficacy and safety of HAIC-FOLFOX combined with intravenous PD-L1 antibody and bevacizumab in advanced HCC; and (II) whether arterial infusion of a PD-L1 antibody is safe, effective, and synergistic with HAIC-FOLFOX and bevacizumab. We present this article in accordance with the SPIRIT reporting checklist (available at https://jgo.amegroups.com/article/view/10.21037/jgo-2026-0607/rc).

Methods

Trial design

This study is designed as a multicenter, open-label, randomized double-arm prospective phase II trial. A total of 76 patients will be randomized to receive hepatic arterial infusion FOLFOX chemotherapy combined with adebrelimab (arterial administration) plus bevacizumab (Arm 1), or hepatic arterial infusion FOLFOX chemotherapy combined with adebrelimab (intravenous administration) plus bevacizumab (Arm 2), at a 1:1 ratio. Figure 1 shows the flowchart of the trial. A list of all participating centers is in online Table S1. The patient recruitment began in December 2024 and will span to December 2026. The 2-year follow-up is designed for all patients, and the final study report will be prepared within 6 months. Therefore, this study is scheduled to end in July 2029.

Figure 1 Study design. ^†, assessed by the investigator based on RECIST v1.1 criteria. ^‡, treatment until intolerable toxicity or disease progression; if the investigator determines the patient can still derive clinical benefit, they may continue study drug treatment. BCLC, Barcelona Clinic Liver Cancer; D1, day 1; D3, day 3; ECOG PS, Eastern Cooperative Oncology Group Performance Status; HAIC-FOLFOX, hepatic arterial infusion chemotherapy-FOLFOX; IV, intravenous; ORR, objective response rate; OS, overall survival; PFS, progression-free survival.

Objectives of the study

This study aims to determine the response rate of hepatic arterial infusion FOLFOX chemotherapy combined with adebrelimab (hepatic arterial or intravenous administration) and bevacizumab in advanced HCC. Additional aims of this study include comprehensively evaluating the response rate and safety profile of arterial infusion of a PD-L1 antibody as a parallel exploration to the intravenous route and to investigate changes in the number and function of immune cells in the liver cancer immune microenvironment after FOLFOX chemotherapy combined with adebrelimab (hepatic arterial or intravenous administration) and bevacizumab.

Endpoints of the study

The primary endpoint is ORR, defined as the proportion of enrolled patients in each group for whom the best overall response during the study is complete response (CR) or partial response (PR) according to RECIST v1.1 criteria; the ORR will be assessed by independent central reviewers. Participants with a first assessment of PR/CR will be required to undergo confirmation 4 weeks later or at the next scheduled assessment. If suspected clinical progression occurs, physical examination and imaging confirmation should be performed immediately rather than waiting for the next scheduled imaging examination.

The secondary endpoints include the following: (I) OS, defined as time from enrollment to death from any cause; (II) progression-free survival (PFS), defined as time from initiation of this regimen to first disease progression or death due to disease; (III) ORR, disease control rate (DCR) and duration of response (DOR) according to mRECIST to assess viable tumor response; (IV) safety, defined as incidence and severity of adverse events (AEs) and serious adverse events (SAEs) according to National Cancer Institute-Common Terminology Criteria for Adverse Events (NCI-CTCAE) v5.0, together with vital signs and laboratory abnormalities; (V) quality of life (QoL), defined as changes in QoL before and after treatment assessed using a QoL scale. The exploratory endpoints include changes in the tumor immune microenvironment (TIME) before initiation and after two cycles of HAIC-FOLFOX combined with adebrelimab (arterial/intravenous) plus bevacizumab in both arms; changes in peripheral blood lymphocyte subsets and key cytokines during combination therapy in both arms.

Study procedures

Patient selection/screening

This study includes patients aged ≥18 years who meet the diagnostic criteria for BCLC stage C (2022 edition) HCC. All patients must have clinically- or pathologically-confirmed HCC and must not have received prior systemic therapy. The investigators will explain the concept of the trial to the patients and obtain written informed consent from all participating patients. Detailed inclusion and exclusion criteria are shown in Table 1.

Randomization

The central randomization system of Sun Yat-sen University Cancer Center will randomly assign eligible participants to Arm 1 or Arm 2, stratified by baseline portal vein tumor thrombosis (PVTT) classification (Vp0–Vp2 vs. Vp3). Stratified randomization will be performed using a computerized central randomization system.

Treatment

Each treatment cycle is defined as 3 weeks (21 days). The treatment regimen during each cycle is shown below in Table 2. In both arms, participants undergo digital subtraction angiography (DSA)-guided hepatic arterial catheterization followed by the first cycle of HAIC-FOLFOX. This consists of oxaliplatin 85 mg/m² and leucovorin (calcium folinate) 400 mg/m² administered by arterial infusion over 1.5 hours on Day 1, followed by 5-FU 2,500 mg/m² by continuous arterial infusion over 46 hours. Of note, the pharmacologically active L-isomer, calcium levofolinate (200 mg/m²), may be used as an equivalent substitute for leucovorin based on institutional availability. Subsequent HAIC-FOLFOX cycles are administered after completion of the previous cycle using the same procedure, with chemotherapy doses adjusted as needed based on toxicity.

For Arm 1, immediately after completion of HAIC-FOLFOX, adebrelimab 1,200 mg will be given by arterial infusion on Day 3, once every 21 days (Q3W), in combination with bevacizumab 15 mg/kg by intravenous infusion on the same day, also Q3W. For Arm 2, adebrelimab 1,200 mg and bevacizumab 15 mg/kg will both be administered by intravenous infusion on Day 3, Q3W. In both arms, adebrelimab and bevacizumab must be given on the same day, and the first dose will be given on Day 3 of the first cycle.

Administration of adebrelimab more than 3 days after the intended date is defined as delayed dosing, and subsequent dosing intervals are calculated from the actual prior dosing date. To maintain consistency in dose adjustments, any dose reduction related to AEs requires a prior dose interruption, whereas adjustments unrelated to AEs may be made without interruption.

The hepatic arterial catheterization procedure is performed as follows: the patient is placed in the supine position, and the right femoral artery is punctured using the Seldinger technique. After successful sheath placement, a 5-F catheter is inserted into the common hepatic artery for angiography. Based on tumor size, location, blood supply, and liver function, the catheter is selectively placed into the tumor-feeding artery or the left/right hepatic artery. If the catheter cannot be maneuvered successfully, a 2.9-F microcatheter is used for superselective catheterization. The interventional physician positions the catheter appropriately under angiographic guidance, and repeat angiography is performed to confirm that the catheter tip is in the desired position. The catheter is then carefully secured, and the puncture site is covered with a dressing. If the tumor receives blood supply from both the celiac trunk and the superior mesenteric artery, or from other feeding arteries, the non-dominant feeding vessels may be embolized, and the catheter is placed in the dominant feeding artery; if the microcatheter tip cannot be stabilized, coil embolization of the gastroduodenal artery may be necessary.

Each HAIC-FOLFOX cycle lasts 3 weeks, comprising 3 days of drug infusion followed by 18 days of rest. If a participant cannot tolerate the planned schedule during the study because of performance status, liver function, or other reasons, the interval between treatments may be appropriately extended. Patients will receive up to six cycles of combination treatment, followed by maintenance with intravenous PD-L1 antibody plus bevacizumab every 3 weeks. All patients without disease progression after six cycles are eligible to enter the maintenance phase, which is generally recommended for up to one year, unless disease progression or unacceptable toxicity occurs. If non-progressive disease is sustained after one year and continued treatment is considered necessary, maintenance therapy may be continued at the investigator’s discretion. The use of investigational drugs, anticancer agents, immunosuppressants, or immunotherapies that might interfere with the treatment protocol will also be prohibited throughout the study period.

Treatment will continue until the first occurrence of intolerable toxicity, withdrawal of informed consent, conversion to curative surgical resection, or other protocol-defined discontinuation criteria. Disease progression per RECIST v1.1 does not automatically mandate treatment discontinuation; patients who progress radiologically may continue treatment if the investigator judges that the patient remains in clinical benefit and tolerates treatment adequately, with ECOG performance status ≤2, no significant clinical signs or symptoms suggesting tumor progression, no rapid progression or involvement of critical organs, and documented approval from the study team. If the investigator determines that the patient no longer derives clinical benefit, treatment will be discontinued. Patients receiving treatment beyond progression must be re-consented with clear explanation of risks and alternatives, and will continue scheduled evaluations; if clinical deterioration occurs without objective progression, it will be reported as “global deterioration”, with efforts to obtain imaging confirmation.

Data collection procedure

Data collection

All data shall be recorded promptly and truthfully in the case report form (CRF). After completeness and accuracy are checked by supervisors assigned by the project leader, the data will be entered into the database by designated personnel and regularly verified. Source documents include medical records, examination reports, imaging results, laboratory data, AE records, surgical records, and other materials that must correspond to and be traceable to the CRF data. Source documents that constitute medical records shall be retained in accordance with hospital regulations, but for no less than five years after trial completion.

Follow-up and assessments

Imaging examinations (computed tomography or magnetic resonance imaging) will be conducted every 2 cycles. Tumor response will be assessed by the investigator as per RECIST 1.1 and mRECIST. RECIST v1.1 will serve as the primary criteria for statistical evaluation and defining primary/secondary survival endpoints. mRECIST will be used exclusively to evaluate secondary efficacy endpoints accounting for tumor necrosis. In cases of discrepancy between the two criteria (e.g., disease progression per RECIST v1.1 but not per mRECIST), the formal statistical date of progression will be strictly based on RECIST v1.1. However, investigators will utilize parallel mRECIST evaluations to help determine whether a patient is experiencing tumor necrosis/pseudo-progression and may qualify for treatment beyond progression per protocol guidelines. For patients who discontinue the study treatment due to reasons other than progression, subsequent imaging examinations will be performed every three months until disease progression, death, or initiation of other anticancer therapies. Survival status will be followed every two months until loss to follow-up, death, or study termination. AEs will be evaluated and recorded before each treatment cycle and within 30 days after the last treatment and will be graded as per CTCAE 5.0. Patients with documented disease progression during the follow-up period will receive standard-of-care treatment as determined by their treating physicians, including systemic therapy in accordance with national guidelines.

Translational research

Participants may choose to participate in the translational research programme as stated in the informed consent (Figure 2). For those who sign the informed consent, blood samples are obtained before initiation of treatment and after every two cycles of HAIC-FOLFOX. In patients who undergo the initial liver biopsy at diagnosis, a repeat biopsy is recommended after two treatment cycles to obtain tissue specimens. Analyses include immunohistochemistry (PD-1, PD-L1, CTLA-4, CD8, CD4, etc.) of the tissue samples, and single-cell sequencing of both blood and tumor tissue samples. Dynamic changes in peripheral blood lymphocyte subsets and key cytokines are also evaluated. The primary interests of this translational research are to investigate the changes in the tissue immune microenvironment after HAIC-FOLFOX combined with adebrelimab (arterial or intravenous) plus bevacizumab in the two arms, to characterize the corresponding changes in peripheral blood immune profiles, and to explore whether the expression of PD-L1/PD-1 in pretreatment liver cancer tissue correlates with the efficacy of the combination regimen.

Figure 2 The diagram of the timeline shows the HAIBrave treatment regimen and sample collection. After patients sign informed consent, specimen collection is performed. ¹, Blood samples are collected before initiation and after every two cycles of HAIBrave therapy. ², Biopsy samples of tumor tissue are collected before initiation of HAIBrave therapy. For patients who undergo an initial liver biopsy at diagnosis, a repeat biopsy is recommended after two treatment cycles to obtain tissue specimens. Analyses include immunohistochemistry (PD-1, PD-L1, CTLA-4, CD8, CD4, etc.) of tissue samples, and single-cell sequencing of both blood and tumor tissue samples. CTLA-4, cytotoxic T-lymphocyte-associated protein 4; HAIC-FOLFOX, hepatic arterial infusion chemotherapy-FOLFOX; IV, intravenous; PD, progressive disease; PD-1, programmed cell death protein 1; PD-L1, programmed cell death ligand 1.

Statistical analysis

Sample size estimation

The primary endpoint of this study is the ORR in Arm 1 and Arm 2 (based on RECIST v1.1). Participants who sign informed consent will be randomly assigned to either Arm 1 or Arm 2, and each Arm will independently use Simon’s optimal two-stage design (13). Although evaluated independently, both arms share identical statistical parameters because they aim to surpass the same historical benchmark and achieve the same target efficacy. Based on previous studies, the unacceptable response rate (null hypothesis, P₀) is set at 27.3% for both arms, reflecting the established ORR of standard first-line atezolizumab plus bevacizumab. The expected response rate (alternative hypothesis, P₁) of the new treatment is hypothesized to be approximately 50% for both arms, representing the unified clinical threshold required to justify further phase III evaluation. With α=0.05 (one-sided) and β=0.20, the optimal two-stage design is 13/21. Specifically, for each arm, in the first stage, 13 patients will be treated; if the number of responders is ≤4, the trial will be terminated. Otherwise, the second stage will continue until a total of 34 patients have been enrolled. If the total number of responders across both stages exceeds 13, the new treatment will be considered effective; if it is ≤13, the new regimen will be considered ineffective. Considering a 10% dropout rate, 38 participants are planned for each arm, with a total sample size of 76.

Statistical methods

No formal statistical comparison between the two arms is planned, as the sample size is not powered for a head-to-head comparison. This design allows independent evaluation of each arm against the historical control while reducing selection and temporal biases through concurrent randomization; descriptive comparisons and exploratory analyses between arms will be performed for hypothesis generation for future phase III trials.

All data in this study will be summarized using statistical measures appropriate to the data type: continuous data will be described by mean, standard deviation, median, minimum, and maximum; categorical data will be summarized using frequency and proportion. Parametric statistical methods will be considered first; if the distribution deviates substantially from assumptions required for hypothesis testing, nonparametric methods will be used. Time-to-event data will be analyzed using the Kaplan-Meier method to estimate median survival time, and survival curves and 95% confidence intervals (CIs) for the median may be presented when necessary. In addition, three analysis sets will be used in this study: (I) Full Analysis Set (FAS): all randomized participants who receive at least one study treatment; (II) Per-Protocol Set (PPS): participants without major protocol deviations who complete at least two treatment cycles and complete the primary efficacy assessment; and (III) Safety Set (SS): all participants who receive at least one study treatment and have safety records.

Data management and monitoring

Data management and monitoring will be conducted in accordance with the protocol, good clinical practice (GCP) requirements, and the data management plan to ensure data authenticity, completeness, accuracy, and traceability. It is independent and without competing interests. The principal investigator (PI) takes responsibility for trial design and performance. The PIs will make the final trial database available.

Ethics and dissemination

The study will be conducted in accordance with the Declaration of Helsinki and its subsequent amendments, the International Council for Harmonisation Good Clinical Practice (ICH-GCP, 2016 edition), and China’s Good Clinical Practice for Pharmaceutical Products (2020 edition). It was formally approved by the Medical Ethics Committee of Sun Yat-sen University Cancer Center (approval No. SL-B2024-588-02) and passed annual continuing review on 12 November 2025 (approval No. B2024-588-Y01). All participating hospitals were informed of and agreed with this study. The protocol of this study was registered at ClinicalTrials.gov (identification number: NCT06737913). Participants who experience harm from trial participation will receive compensation from the insurance company. The trial results will be presented at scientific and professional conferences and submitted for publication in peer-reviewed journals.

Discussion

The HAIBrave-001 study aims to explore the efficacy and safety of hepatic arterial infusion (HAI) or intravenous infusion (IV) of the PD-L1 inhibitor adebrelimab, combined with bevacizumab and HAIC-FOLFOX, as first-line treatment for advanced unresectable HCC. This trial seeks to answer a clinically important scientific question: given that HAIC-FOLFOX has already established an arterial route of administration, will its combination with HAI/IV of PD-L1 antibody and IV of bevacizumab improve the survival outcome of HCC patients?

The central hypothesis of this study is based on two interrelated theoretical foundations. First, HAIC-FOLFOX may remodel the TIME by inducing immunogenic cell death (ICD) (14). Oxaliplatin-based chemotherapy can promote the release of damage-associated molecular patterns and tumor antigens from tumor cells, thereby activating dendritic cells and initiating tumor-specific T-cell responses (15). This process has a distinct time window—within several days after chemotherapy, the activation of antigen-presenting cells and T-cell priming reaches a peak. In this study, adebrelimab is scheduled immediately after completion of HAIC-FOLFOX infusion (Day 3 of each cycle) in order to precisely capture this immune activation window and maximize synergy between ICD and checkpoint blockade. Second, the TIME exhibits marked spatial heterogeneity (16). The peritumoral region usually has richer blood supply and greater immune cell infiltration, whereas the tumor core often has poorer perfusion and more immune-suppressive cells, making it a stronghold of resistance to immunotherapy. By hepatic arterial infusion, high concentrations of PD-L1 antibody may penetrate the tumor core more effectively, overcome uneven drug distribution caused by impaired perfusion, and achieve more comprehensive immune activation. The translational component of this study, using paired tumor biopsy samples analyzed by multiplex immunofluorescence and single-cell sequencing, will directly compare differences between the HAI and IV groups in immune cell infiltration, PD-L1 occupancy, and functional status in the tumor core, providing direct evidence for this hypothesis.

In recent years, triple therapy combining HAIC, targeted therapy, and immunotherapy has achieved a series of advances in advanced HCC, and multiple studies have shown that compared with dual therapy, triple regimens can yield higher tumor response rates. However, in most of these studies, ICIs were administered intravenously, and the sequencing of HAIC and ICIs varied. The unique contribution of HAIBrave-001 lies in its prospective randomized two-arm design, which goes one step further within the framework of triple therapy to address optimization of the administration route. Once triple therapy has shown preliminary superiority, can efficacy be pushed to a new level by switching ICIs from intravenous infusion to hepatic arterial infusion? In this regard, the positioning of the present study is conceptually similar to the NACI study by Li et al. in locally advanced cervical cancer, which optimized neoadjuvant immunochemotherapy through chemotherapy–immunotherapy sequencing, whereas the present study refines triple therapy through route-of-administration optimization (17).

Several design choices in this trial are based on deeper scientific considerations. First, chemotherapy-first sequencing is rooted in the time dynamics of ICD—chemotherapy-induced immune remodeling takes time, and preceding chemotherapy may create an immune-permissive microenvironment for subsequent ICIs. This concept has been widely recognized in neoadjuvant immunotherapy for lung cancer, breast cancer, and other malignancies. Second, the decision to switch to intravenous maintenance after six cycles of combination therapy is based on the following considerations: (I) prior studies suggest that six cycles of HAIC are sufficient to achieve maximal tumor response, and additional cycles offer limited benefit while increasing cumulative toxicity; (II) HAI requires repeated arterial catheterization, and six cycles strikes a balance between clinical feasibility and patient tolerability; and (III) a uniform intravenous maintenance phase may reduce confounding from subsequent treatment differences in survival analyses. This study adopts 1:1 randomization, with ORR as the primary endpoint, to provide high-level phase II evidence for the independent evaluation of both HAI and IV administration routes. Although this trial is not formally powered for a direct comparison between the HAI and IV routes, the parallel randomized design offers a unique opportunity to explore potential differences in efficacy and safety between the two administration routes. These exploratory analyses may inform the design and sample size calculation of future phase III trials should a promising signal emerge (6).

Several limitations warrant consideration. First, the open-label design is unavoidable due to the procedural nature of HAI, which may introduce bias in efficacy and safety assessments. To mitigate this, the primary endpoint (ORR) is evaluated by centrally trained investigators using RECIST v1.1 criteria, with an independent blinded radiology review planned as a sensitivity analysis; AEs are graded per NCI-CTCAE v5.0 and reviewed by independent monitors. Second, the sample size (76 patients, Simon two-stage design) is adequate for the primary comparison (ORR improvement from 27.3% to 50%) but limits detailed subgroup analyses; these will be addressed in future phase III studies. Third, the HAI procedure requires specialized expertise, and although the participating centers are experienced high-volume institutions, broader generalizability remains to be confirmed. Finally, paired tumor biopsies are optional and may introduce selection bias in translational analyses; this is mitigated by detailed documentation of refusal reasons, baseline comparisons, and the incorporation of high-compliance peripheral blood sampling for dynamic immune monitoring.

In summary, the HAIBrave-001 study is grounded in the current academic frontier where triple therapy has become a major focus. By placing optimization of the administration route of ICIs at the center of its scientific question, and by combining a rigorous randomized design, a rational treatment schedule, and in-depth translational research, it is expected to open a new avenue of exploration in the treatment of advanced HCC. Regardless of the eventual outcome, the study will provide valuable experience and insights for refinement of triple therapy, development of individualized treatment strategies, and design of larger future clinical studies.

Acknowledgments

We thank Ms. Meizhen Zhu and Mr. Xueting Shen for giving their best love to their son, Dr. Lujun Shen and his family when he pursues the excellence in medicine. We also thank Ms. Juan Nie for her unwavering support to her husband, Dr. Lujun Shen, over the past 14 years.

Footnote

Reporting Checklist: The authors have completed the SPIRIT reporting checklist. Available at https://jgo.amegroups.com/article/view/10.21037/jgo-2026-0607/rc

Peer Review File: Available at https://jgo.amegroups.com/article/view/10.21037/jgo-2026-0607/prf

Funding: This study was funded by Comprehensive Basic Medical Research Public Welfare Project (No. 5101-X-026-QLY001).

Conflicts of Interest: All authors have completed the ICMJE uniform disclosure form (available at https://jgo.amegroups.com/article/view/10.21037/jgo-2026-0607/coif). L.S. reports funding support from the Comprehensive Basic Medical Research Public Welfare Project (Project No. 5101-X-026-QLY001). X.F. is from Jiangsu Hengrui Pharmaceuticals Co., Ltd. The other authors have no conflicts of interest to declare.

Ethical Statement: The authors are accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. The study will be conducted in accordance with the Declaration of Helsinki and its subsequent amendments, the International Council for Harmonisation Good Clinical Practice (ICH-GCP, 2016 edition), and China’s Good Clinical Practice for Pharmaceutical Products (2020 edition). It was formally approved by the Medical Ethics Committee of Sun Yat-sen University Cancer Center (approval No. SL-B2024-588-02). All participating hospitals were informed of and agreed with this study. The investigators will explain the concept of the trial to the patients and obtain written informed consent from all participating patients.

Open Access Statement: This is an Open Access article distributed in accordance with the Creative Commons Attribution-NonCommercial-NoDerivs 4.0 International License (CC BY-NC-ND 4.0), which permits the non-commercial replication and distribution of the article with the strict proviso that no changes or edits are made and the original work is properly cited (including links to both the formal publication through the relevant DOI and the license). See: https://creativecommons.org/licenses/by-nc-nd/4.0/.

References

1. Toh MR, Wong EYT, Wong SH, et al. Global Epidemiology and Genetics of Hepatocellular Carcinoma. Gastroenterology 2023;164:766-82. [Crossref] [PubMed]
2. Shan S, Jia J. The clinical management of hepatocellular carcinoma in China: Progress and challenges. Clin Mol Hepatol 2023;29:339-41. [Crossref] [PubMed]
3. Reig M, Sanduzzi-Zamparelli M, Forner A, et al. BCLC strategy for prognosis prediction and treatment recommendations: The 2026 update. J Hepatol 2026;84:631-54. [Crossref] [PubMed]
4. Finn RS, Qin S, Ikeda M, et al. Atezolizumab plus Bevacizumab in Unresectable Hepatocellular Carcinoma. N Engl J Med 2020;382:1894-905. [Crossref] [PubMed]
5. Lyu N, Lin Y, Kong Y, et al. FOXAI: a phase II trial evaluating the efficacy and safety of hepatic arterial infusion of oxaliplatin plus fluorouracil/leucovorin for advanced hepatocellular carcinoma. Gut 2018;67:395-6. [Crossref] [PubMed]
6. Shen L, Jiang Y, Lu L, et al. Dynamic prognostication and treatment planning for hepatocellular carcinoma: A machine learning-enhanced survival study using multi-centric data. Innov Med 2025;3:100125. [Crossref]
7. Zhou X, Zhang B, Hu J, et al. Igniting cold tumors of intrahepatic cholangiocarcinoma: An insight into immune evasion and tumor immune microenvironment. Innov Med 2024;2:100052. [Crossref]
8. Lyu N, Wang X, Li JB, et al. Arterial Chemotherapy of Oxaliplatin Plus Fluorouracil Versus Sorafenib in Advanced Hepatocellular Carcinoma: A Biomolecular Exploratory, Randomized, Phase III Trial (FOHAIC-1). J Clin Oncol 2022;40:468-80. [Crossref] [PubMed]
9. Wang Y, Zhang N. Sandwich-Model Perioperative Therapy Ushers in a New Era of Chronic Disease Management in Hepatocellular Carcinoma. Innov Med 2026;4:100213. [Crossref]
10. Lai Z, He M, Bu X, et al. Lenvatinib, toripalimab plus hepatic arterial infusion chemotherapy in patients with high-risk advanced hepatocellular carcinoma: A biomolecular exploratory, phase II trial. Eur J Cancer 2022;174:68-77. [Crossref] [PubMed]
11. Zhang TQ, Geng ZJ, Zuo MX, et al. Camrelizumab (a PD-1 inhibitor) plus apatinib (an VEGFR-2 inhibitor) and hepatic artery infusion chemotherapy for hepatocellular carcinoma in Barcelona Clinic Liver Cancer stage C (TRIPLET): a phase II study. Signal Transduct Target Ther 2023;8:413. [Crossref] [PubMed]
12. Shen L, Cao F, Liu Y, et al. Hepatic artery infusion of FOLFOX chemotherapy plus camrelizumab combined with sorafenib for advanced hepatocellular carcinoma in Barcelona Clinic Liver Cancer stage C (Double-IA-001): a phase II trial. BMC Med 2025;23:275. [Crossref] [PubMed]
13. Simon R. Optimal two-stage designs for phase II clinical trials. Control Clin Trials 1989;10:1-10. [Crossref] [PubMed]
14. Xing R, Mei J, Zuo Z, et al. Enhanced formation of tertiary lymphoid structures shapes the anti-tumor microenvironment in hepatocellular carcinoma after FOLFOX-HAIC therapy. Cell Rep Med 2025;6:102298. [Crossref] [PubMed]
15. Liu W, Zhang F, Quan B, et al. NLRP3/IL-1β induced myeloid-derived suppressor cells recruitment and PD-L1 upregulation promotes oxaliplatin resistance of hepatocellular carcinoma. MedComm (2020) 2023;4:e447.
16. Qiu X, Zhou T, Li S, et al. Spatial single-cell protein landscape reveals vimentin(high) macrophages as immune-suppressive in the microenvironment of hepatocellular carcinoma. Nat Cancer 2024;5:1557-78. [Crossref] [PubMed]
17. Li K, Chen J, Hu Y, et al. Neoadjuvant chemotherapy plus camrelizumab for locally advanced cervical cancer (NACI study): a multicentre, single-arm, phase 2 trial. Lancet Oncol 2024;25:76-85. [Crossref] [PubMed]

(English Language Editor: J. Jones)