Organ-specific response to atezolizumab + bevacizumab combination therapy for unresectable hepatocellular carcinoma with extrahepatic metastases: a retrospective cohort study

Introduction

Hepatocellular carcinoma (HCC) is the third leading cause of cancer-related deaths worldwide (1). Sorafenib, a tyrosine kinase inhibitor, has been approved for use in unresectable HCC in Japan in 2009 after the SHARP and Asia-Pacific studies demonstrated a significant overall survival benefit compared with placebo (2,3). Lenvatinib, another tyrosine kinase inhibitor, subsequently proved statistically non-inferior to sorafenib regarding overall survival in the REFLECT trial and has been approved in Japan since 2018 (4). The combination of atezolizumab, an immune checkpoint inhibitor (ICI), and bevacizumab, a vascular endothelial growth factor (VEGF) inhibitor, showed a significant overall survival advantage over sorafenib in the IMbrave150 trial, and has been first approved in Japan since 2020 (5). In addition, the combination of tremelimumab plus durvalumab, both ICIs, showed a significant overall survival advantage over sorafenib in the HIMALAYA trial, and has been approved in Japan since 2023 (6). The combination of nivolumab plus ipilimumab, both ICIs, showed a significant overall survival advantage over sorafenib in the Check Mate-9DW trial (7). For these reasons, the opportunity to provide systemic drug therapy to patients with HCC is increasing rapidly. The 2023 edition of Clinical Practice Guidelines for HCC: The Japan Society of Hepatology (JSH) recommends the combination of atezolizumab and bevacizumab (Atezo + Bev) and tremelimumab plus durvalumab as first-line therapy in Child-Pugh class A cases of unresectable HCC with distant metastases, good performance status, and good liver reserve (8,9). Investigating organ-specific response (OSR) is particularly crucial in HCC because the liver possesses a uniquely immunosuppressive microenvironment, characterized by abundant myeloid-derived suppressor cells, regulatory T cells, and continuous exposure to gut-derived antigens. This tolerogenic milieu often impairs systemic anti-tumor immunity and can lead to discordant therapeutic efficacies between intrahepatic lesions and extrahepatic metastases. Therefore, optimizing treatment regimens based on organ-specific susceptibility represents a critical, yet highly controversial, frontier in clinical oncology.

However, current clinical practice guidelines lack definitive consensus on OSR, primarily due to evaluation gaps in prior literature. Most existing studies on multi-organ efficacy have relied strictly on the conventional Response Evaluation Criteria in Solid Tumors (RECIST), which only provide a dichotomous assessment (responder vs. non-responder) at specific cross-sectional timepoints (10-13). This approach fundamentally overlooks key therapeutic dimensions: how deeply a tumor responds (depth of response, DpR) and how long that response is sustained (duration of response, DoR) in an organ-specific manner. For instance, although Cheon et al. reported a seemingly consistent antitumor effect of Atezo + Bev across intrahepatic and metastatic sites, their analysis was constrained by a brief median observation period of 10.1 months and did not capture the longitudinal dynamics of DpR or DoR (14).

Resolving these intricate clinical questions ideally warrants large-scale, highly powered cohorts. Nevertheless, mass-scale retrospective or registry data often suffer from heterogeneous scanning intervals, missing granular multi-lesion measurements, and brief follow-up periods that obscure long-term DoR. To bridge these precise knowledge blanks, we conducted an exploratory study utilizing a meticulously curated, albeit small-sample, retrospective dataset. The methodology of this study is justified by its highly uniform treatment protocol, rigorous longitudinal multi-organ lesion tracking, and a substantially extended observation period. By shifting the analytical focus from a simple binary response to granular, multi-dimensional metrics—specifically organ-specific DpR and DoR—this cohort is well-positioned to yield high-fidelity, hypothesis-generating insights that larger, less granular datasets frequently fail to capture. We present this article in accordance with the STROBE reporting checklist (available at https://jgo.amegroups.com/article/view/10.21037/jgo-2026-0282/rc).

Methods

Study design and ethics

This was a single-center, retrospective cohort study designed to evaluate the organ-specific therapeutic efficacy of Atezo + Bev in patients with unresectable HCC presenting with distant metastases. The study was performed at Kanazawa University Hospital in strict compliance with the guidelines of the Declaration of Helsinki and its subsequent amendments and the Ethical Guidelines for Medical and Health Research Involving Human Subjects in Japan. The study protocol was approved by the Institutional Review Board of Kanazawa University Hospital (No. 2017-012). The requirement for written informed consent was waived due to the retrospective, non-interventional nature of the study, utilizing completely anonymized clinical data. Instead, an opt-out approach was implemented: detailed information regarding the study objectives and data usage was disclosed via public notices on the hospital website, providing eligible patients with the opportunity to decline participation.

Patient selection and diagnostic criteria

We screened all consecutive patients who received Atezo + Bev for unresectable HCC at our institution between October 2020 and November 2024. The definitive diagnosis of HCC was established according to the clinical practice guidelines of the JSH and the American Association for the Study of Liver Diseases (AASLD). This was achieved pathologically via tissue biopsy or radiologically using contrast-enhanced computed tomography (CECT) or gadoxetic acid-enhanced magnetic resonance imaging (Gd-EOB-DTPA MRI), demonstrating hyperattenuation during the arterial phase with washout in the portal venous or delayed phases. The specific inclusion criteria were as follows:

• Age ≥20 years at the initiation of Atezo + Bev therapy.
• Radiologically confirmed distant (extrahepatic) metastasis prior to the initiation of systemic therapy.
• At least one measurable intrahepatic or extrahepatic lesion in accordance with the RECIST version 1.1. (15).

The exclusion criteria were as follows:

• No evaluable target or non-target lesions per RECIST 1.1.
• Absence of any post-baseline radiological treatment response assessment.
• Concurrent administration of radiation therapy, or initiation of systemic therapy within 6 months of completing radiation therapy, to avoid confounding treatment responses.

Sample size justification

Because this was an exploratory, retrospective cohort study using a highly specialized and sequentially treated patient population within a defined historical window (October 2020 to November 2024), a formal a priori sample size calculation was not performed. Instead, a strict total-eligible-population sampling approach was utilized. To maximize statistical power and minimize selection bias, all consecutive patients meeting the aforementioned eligibility criteria during the 4-year study period were included in the final analysis.

Data extraction and quality control

A structured, standardized electronic data collection form was developed prior to data extraction to ensure consistency across all variables. To minimize information and measurement biases, clinical data were independently extracted from the hospital’s electronic medical record system by two investigators. Discrepancies were resolved through consensus or by consulting a senior hepatologist. To guarantee the integrity of tumor response evaluations, all computed tomography (CT) images were independently reviewed by an experienced hepatologist who was completely blinded to the patients’ clinical outcomes and laboratory data.

Baseline characteristics and follow-up procedures

Comprehensive baseline clinical characteristics were extracted within 14 days prior to the first dose of Atezo + Bev. These factors included: age, sex, performance status (ECOG-PS), etiology of background liver disease (hepatitis B, hepatitis C, alcoholic steatohepatitis, or others), prior systemic/locoregional treatments, treatment line (first-line vs. second- or later-line), Child-Pugh classification, modified albumin-bilirubin (mALBI) grade (16), presence of intrahepatic lesions, specific sites of distant metastasis, and serum alpha-fetoprotein (AFP) levels. Following the initiation of Atezo + Bev, patients were followed up longitudinally. In accordance with our institutional protocol for HCC management, clinical visits and laboratory assessments (including liver function tests and tumor markers) were performed every 3 to 4 weeks. Radiological assessments using high-resolution CECT or MRI were performed at regular intervals of every 6 to 8 weeks until November 2025, or until disease progression, death, or loss to follow-up occurred.

Efficacy and prognostic outcome measurements

Tumor response was assessed for each organ system (liver, lung, lymph nodes, peritoneal dissemination, bone, and adrenal glands) according to RECIST version 1.1. The primary prognostic and efficacy outcomes were defined as follows (17):

• Best overall response (BOR): the highest tumor response recorded from the start of treatment until disease progression.
• Objective response rate (ORR): the percentage of patients achieving a complete response (CR) or partial response (PR) as their BOR.
• Disease control rate (DCR): the percentage of patients achieving CR, PR, or stable disease (SD) maintained for at least 6 weeks.
• Organ-specific response rate (OSRR): the percentage of patients achieving CR or PR in target lesions within a specific organ system, considering a maximum of 5 target lesions per organ.
• DpR: the maximum percentage change (reduction) in the sum of diameters of target lesions compared to the baseline measurement.
• DoR: the time from the initial documentation of CR or PR to the first radiologically confirmed disease progression (PD) per RECIST 1.1 or death from any cause.
• Time to progression (TTP): the time from the initiation of Atezo + Bev therapy to the first documentation of objective disease progression. Patients who died without objective progression were censored at the date of death.
• Progression-free survival (PFS): the time from treatment initiation to radiologically defined PD or death from any cause.
• Overall survival (OS): the time from treatment initiation to death from any cause.

For patients who did not reach the designated endpoints (progression or death) by the data cut-off date (November 30, 2025), data were treated as censored cases at the date of the last reliable radiological or clinical follow-up.

Statistical analysis

EZR version 1.37 was used for the statistical analyses (18). Continuous variables were presented as medians with interquartile ranges (IQR) and compared across multiple organ subgroups using the Kruskal-Wallis test. Categorical variables, including OSRRs and disease control rates, were expressed as numbers with percentages and compared using Fisher’s exact test. For time-to-event outcomes (organ-specific DoR and TTP), survival curves were estimated using the Kaplan-Meier method, and inter-group differences were evaluated via the log-rank test. Two-sided P values less than 0.05 were considered statistically significant.

Crucially, due to the restricted sample size of certain organ-specific subgroups (e.g., bone or adrenal metastases) and the retrospective nature of this cohort, multivariate adjustments (e.g., Cox proportional hazards or logistic regression models) to control for key baseline confounding factors—such as viral etiology, baseline tumor burden, and liver function classification (mALBI grade/Child-Pugh class)—were methodologically unfeasible.

With this mathematical reality, all inter-organ comparisons in this study were strictly exploratory. To mitigate the unacceptably high risk of type II statistical errors (false-negative results) inherent in underpowered, fragmented subgroups, any lack of statistically significant differences (P≤0.05) between organ sites was not interpreted as definitive evidence of uniform therapeutic efficacy. Instead, such findings were treated strictly as descriptive and hypothesis-generating, highlighting the need for larger, highly powered validation cohorts.

Results

Patient disposition and baseline characteristics

Among the 159 patients initially screened, 117 were excluded for the following reasons: 97 patients lacked distant metastasis at the start of therapy, 4 had no follow-up imaging for response evaluation, 12 did not have measurable lesions according to RECIST version 1.1, and 4 were excluded due to concurrent or recent (within 6 months) radiation therapy. These exclusions ensured that the final cohort of 42 patients had evaluable systemic disease for organ-specific analysis (Figure 1). The baseline characteristics of the 42 included patients are shown in Table 1. There were no missing data for the primary variables, including treatment response and survival outcomes, for the 42 included patients.

Figure 1 Consort diagram. Among 159 patients initially screened, 117 were excluded for the following reasons: 97 patients lacked distant metastasis at the start of therapy, 4 had no follow-up imaging for response evaluation, 12 did not have measurable lesions according to RECIST version 1.1, and 4 were excluded due to concurrent or recent (within 6 months) radiation therapy. These exclusions ensured that the final cohort of 42 patients had evaluable systemic disease for organ-specific analysis. Atezo, atezolizumab; Bev, bevacizumab; RECIST, Response Evaluation Criteria in Solid Tumors.

Overall survival and treatment efficacy

The median follow-up period, as calculated by the reverse Kaplan-Meier method, was 36.3 months. Kaplan–Meier curves for OS and PFS are shown in Figure 2. The median survival was 17.5 months, and the median duration of drug treatment was 5.5 months.

Figure 2 OS and PFS for all patients. (A) Median OS was 17.5 months; 6-month survival was 81%, and 1-year survival rate was 61%. (B) The median PFS was 5.5 months; 6-month PFS was 48%, and 1-year PFS was 21%. OS, overall survival; PFS, progression-free survival.

The BOR and the OSRR are summarized in Table 2. Regarding the BOR, a complete response (CR) was observed in 0 of 42 patients (0%), a partial response (PR) in 13 (31%), stable disease (SD) in 15 (36%), and progressive disease (PD) in 14 (33%). The overall disease control rate (DCR) and ORR were 67% and 31%, respectively. The OSRR was 18 of 27 patients (30%) for liver, 2 of 18 patients (11%) for lung, 9 of 18 patients (50%) for LN, and 4 of 7 patients (57%) for peritoneal dissemination. Among the 14 patients with PD at the initial assessment, 20 specific target lesions demonstrated an increase in size. Among the 28 patients who achieved a BOR of CR, PR, or SD, 26 experienced disease progression prior to the data cut-off date. A waterfall plot illustrating the maximum percentage change in each target lesion is presented in Figure 3.

Figure 3 Waterfall plot of the evaluation of the best overall response for each organ with target lesions. The figure shows the best overall response in target lesions from baseline for all organs (n=42), liver (n=27), lung (n=18), lymph node (n=18), and peritoneal dissemination (n=7). There were no significant differences in the OSRR between the organs. The gray lines at −30% and 20% indicate partial response and progression, respectively, in accordance with the RECIST criteria. OSRR, organ-specific response rate; RECIST, Response Evaluation Criteria in Solid Tumors.

Exploratory organ-specific tumor response

The OSRR was 30% (18/27) for the liver, 11% (2/18) for the lung, 50% (9/18) for lymph nodes (LN), and 57% (4/7) for peritoneal dissemination.

The median DpR for each organ was as follows: 1% (range, −92% to 100%) for the liver and 7% (range, −100% to 193%) for the lung, -28% (range, −79% to 104%) for LN, and -32% (range, −72% to 155%) for peritoneal dissemination, 9% (range, 0% to 17%) for bone (n=2), and 8% (range, −27% to 17%) for the adrenal glands (n=3). Although a Kruskal-Wallis test revealed no statistically significant differences in DpR among these organs (P=0.31) (Table 3), this exploratory comparison was severely underpowered due to the restricted sample size and fragmentation across subgroups, carrying a high risk of a false-negative result.

Kaplan-Meier curves for organ-specific DoR and TTP are shown in Figure 4. The median DoR was 16.5 months for the liver [95% CI: 4–not reached (NR)], NR for the lung (95% CI: NR–NR), NR for LN (95% CI: 6–NR), and 15.9 months for peritoneal dissemination (95% CI: 14–NR), indicating no significant differences (P=0.37). Although the log-rank test showed no statistically significant differences among these evaluable organs (P=0.37), the statistical power was critically limited by the small sample size, and these values should be interpreted strictly as descriptive.

Figure 4 Organ specific duration of response and time to progression for all patients. (A) A median duration of response: 16.5 months for liver (95% CI: 4–not reached), not reached for lung (95% CI: not reached–not reached), not reached for lymph nodes (95% CI: 6–not reached), and 15.9 month for peritoneum dissemination (95% CI: 14–not reached), indicating no significant differences (P=0.37). (B) A median time to progression: 9.7 months for liver (95% CI: 1–11%), 5.5 months for lung (95% CI: 1–14%), 21.3 months for lymph nodes (95% CI: 6–not reached), and 19.8 months for peritoneum dissemination (95% CI: 1–not reached), indicating no significant differences (P=0.09). CI, confidence interval.

The median TTP was 9.7 months for the liver (95% CI: 1–11), 5.5 months for the lung (95% CI: 1–14), 21.3 months for LN (95% CI: 6–NR), and 19.8 months for peritoneal dissemination (95% CI: 1–NR). No statistically significant differences were observed (P=0.09), though a definitive conclusion of uniform efficacy is precluded by the restricted cohort size. Furthermore, within individual patients, we observed no patterns of diametrically opposed treatment effects across different organs (e.g., simultaneous PR in one organ and PD in another within the same individual).

At the time of data cut-off, disease progression was documented across various sites. Of these progressed patients, 23 patients exhibited an increase in target lesions (liver: 10, lung: 10, LN: 3, peritoneal dissemination: 2, adrenal metastasis: 2, bone metastasis: 1), 12 patients had progression in non-target lesions (liver: 7, lung: 1, LN: 4, peritoneal dissemination: 2), and 6 patients developed entirely new lesions (liver: 3, lung: 1, LN: 1, adrenal metastasis: 1).

Discussion

This study evaluated the organ-specific therapeutic dynamics of Atezo + Bev in patients with advanced HCC presenting with distant metastases. Our exploratory analysis demonstrated objective responses across diverse metastatic sites, with an OSRR of 30% for the liver, 11% for the lungs, 50% for the lymph nodes, and 57% for peritoneal dissemination. A previous retrospective study by Cheon et al. (14) reported an OSRR of 28% for the liver, 29% for the lung, 42% for LNs, and 21% for other sites, suggesting a generally balanced distribution of efficacy. Crucially, our study extended these findings by investigating organ-specific DpR, DoR, and TTP in greater depth, utilizing a substantially longer median follow-up period of 36.3 months.

However, in interpreting our findings—specifically the lack of statistically significant differences in median DpR (P=0.31), DoR (P=0.37), and TTP (P=0.09) across organs—we must strictly avoid the statistical fallacy of interpreting a lack of significance as definitive proof of uniform or consistent therapeutic efficacy. Given our restricted sample size and the fragmentation of patients into small organ-specific subgroups, these comparisons were heavily underpowered and carry an inherently high risk of type II statistical errors (false negatives). For instance, heavily restricted subgroups, such as bone (n=2) or adrenal metastases (n=3), preclude any robust statistical inference. Thus, rather than serving as a definitive validation of multi-organ equivalence, our data provide detailed, high-fidelity descriptive benchmarks that are hypothesis-generating in nature.

Despite these statistical constraints, the biological trends observed in this cohort merit clinical consideration, particularly regarding the intrahepatic response. The liver is widely recognized as possessing a uniquely immunosuppressive microenvironment (19). This tolerogenic milieu is maintained by continuous antigen exposure via the portal vein and a high density of immunosuppressive cell types, including regulatory T cells (Tregs), tumor-associated macrophages (TAMs), and myeloid-derived suppressor cells (MDSCs). Consequently, ICI monotherapy often yields suboptimal outcomes within the liver; for example, a study of ICI monotherapy in advanced HCC noted that intrahepatic lesions achieved the lowest OSRR (10.1%) compared to extrahepatic sites (9). Similarly, in advanced esophageal cancer, ICI efficacy was significantly impaired in intrahepatic lesions (20).

The combining of atezolizumab with bevacizumab represents a mechanistic strategy to overcome this hepatic immune resistance. VEGF-A, secreted abundantly by HCC tumors, acts as a potent driver of systemic and regional immunosuppression by promoting the recruitment and activation of Tregs, TAMs, and MDSCs, while inhibiting dendritic cell maturation (21,22). By neutralizing VEGF-A, bevacizumab not only promotes tumor vessel normalization to enhance T-cell infiltration, but also reprograms the tolerogenic liver microenvironment into an immune-supportive state. This synergistic mechanism likely explains why, contrary to ICI monotherapy, the intrahepatic OSRR (30%) and median TTP (9.7 months) in our Atezo + Bev cohort did not appear inferior to extrahepatic responses. Given that intrahepatic progression constitutes the primary cause of mortality in HCC patients (23), utilizing a regimen capable of counteracting the liver’s unique immunosuppressive defenses is of paramount clinical importance, even in the presence of extrahepatic disease.

This study has several critical methodological limitations that must be transparently acknowledged. First, its retrospective, single-center design may have introduced selection bias, potentially shifting the clinical outcomes toward an overestimation of efficacy, as our cohort was managed in a specialized tertiary care center. Second, the small sample size (n=42) and the severe fragmentation into small organ subgroups critically restricted our statistical power. Consequently, we were methodologically unable to perform multivariate regression analyses to control for key baseline confounding factors—such as viral etiology, baseline tumor burden, or liver function classification (mALBI/Child-Pugh)—which are well-established determinants of treatment response. Third, calculating the “true” organ-specific DoR presents an inherent methodological challenge in a retrospective setting; patients were censored at the time of overall disease progression based on the sum of all target lesions per RECIST 1.1, which may obscure the isolated, long-term clinical trajectory of an individual organ site.

Despite these severe constraints, our study provides valuable real-world clinical insights. Unlike strictly controlled clinical trials, our cohort reflects the unselected, heterogeneous patient backgrounds encountered in daily practice within the Japanese population. Furthermore, the involvement of an independent, blinded hepatologist blinded to all clinical data for all radiological assessments ensured high objectivity and minimized measurement bias. In conclusion, while our small cohort cannot definitively settle the controversy surrounding organ-specific efficacy due to limited statistical power and unadjusted confounders, it successfully delivers granular, longitudinal descriptive data regarding DpR and DoR dynamics. These findings establish a key hypothesis-generating framework that justifies the initiation of large-scale, highly powered, prospective multi-center registries to safely resolve these complex clinical questions.

Conclusions

In conclusion, this exploratory study demonstrated that the combination therapy of Atezo + BEV elicited objective tumor responses across various intrahepatic and extrahepatic metastatic sites in patients with advanced HCC. Although no statistically significant differences were observed in organ-specific OSRR, DpR, DoR, and TTP, these comparative findings must be interpreted with caution as strictly hypothesis-generating rather than definitive, due to the limited statistical power and lack of multivariate control inherent in our small-sample cohort. Nevertheless, the descriptive trends suggesting a preserved therapeutic effect within the uniquely immunosuppressive liver microenvironment provide a clinical rationale for using this regimen. Future large-scale, highly powered multi-center prospective studies are warranted to validate these organ-specific efficacy profiles.

Acknowledgments

We thank Jane Charbonneau, DVM, from Edanz (https://jp.edanz.com/ac) for editing a draft of this manuscript. During the revision of this manuscript, the authors utilized Gemini (Google LLC, Mountain View, CA, USA) as an AI-assisted technology to improve the linguistic quality, grammatical structure, and overall clarity of the revised text. The tool was used solely for English language editing and text refinement, and not for data generation, analysis, or medical interpretation. Following the AI-assisted editing, the authors thoroughly reviewed, verified, and manually edited the content to ensure scientific accuracy. The authors maintain full accountability and responsibility for the integrity and final content of this manuscript.

Footnote

Reporting Checklist: The authors have completed the STROBE reporting checklist. Available at https://jgo.amegroups.com/article/view/10.21037/jgo-2026-0282/rc

Data Sharing Statement: Available at https://jgo.amegroups.com/article/view/10.21037/jgo-2026-0282/dss

Peer Review File: Available at https://jgo.amegroups.com/article/view/10.21037/jgo-2026-0282/prf

Funding: None.

Conflicts of Interest: All authors have completed the ICMJE uniform disclosure form (available at https://jgo.amegroups.com/article/view/10.21037/jgo-2026-0282/coif). The 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. This study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments. The study protocol was reviewed and approved by the Institutional Review Board of Kanazawa University Hospital (No. 2017-012). Written informed consent was obtained from all participants prior to their inclusion in the study.

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. Bray F, Laversanne M, Sung H, et al. Global cancer statistics 2022: GLOBOCAN estimates of incidence and mortality worldwide for 36 cancers in 185 countries. CA Cancer J Clin 2024;74:229-63. [Crossref] [PubMed]
2. Llovet JM, Ricci S, Mazzaferro V, et al. Sorafenib in advanced hepatocellular carcinoma. N Engl J Med 2008;359:378-90. [Crossref] [PubMed]
3. Cheng AL, Kang YK, Chen Z, et al. Efficacy and safety of sorafenib in patients in the Asia-Pacific region with advanced hepatocellular carcinoma: a phase III randomised, double-blind, placebo-controlled trial. Lancet Oncol 2009;10:25-34. [Crossref] [PubMed]
4. Kudo M, Finn RS, Qin S, et al. Lenvatinib versus sorafenib in first-line treatment of patients with unresectable hepatocellular carcinoma: a randomised phase 3 non-inferiority trial. Lancet 2018;391:1163-73. [Crossref] [PubMed]
5. 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]
6. Abou-Alfa GK, Lau G, Kudo M, et al. Tremelimumab plus Durvalumab in Unresectable Hepatocellular Carcinoma. NEJM Evid 2022;1:EVIDoa2100070.
7. Galle PR, Decaens T, Kudo M, et al. Nivolumab (NIVO) plus ipilimumab (IPI) vs lenvatinib (LEN) or sorafenib (SOR) as first-line treatment for unresectable hepatocellular carcinoma (uHCC): First results from CheckMate 9DW. J Clin Oncol 2024;42:LBA4008.
8. Kiyoshi H, Nobuyuki T, Tatsuya Y, et al. Clinical practice guidelines for hepatocellular carcinoma. 2021 version (5th JSH-HCC Guidelines). Hepatol Res 2023;10:383-90. Available online.
9. Hasegawa K, Takemura N, Yamashita T, et al. The modifications of CQ39. 2021 version (5th JSH-HCC Guidelines). Japan Society of Hepatology. Available online: https://www.jsh.or.jp/lib/files/medical/guidelines/jsh_guidlines/medical/guideline_jp_2021_cq39_algorithm.pdf
10. Kim HS, Kim CG, Hong JY, et al. The presence and size of intrahepatic tumors determine the therapeutic efficacy of nivolumab in advanced hepatocellular carcinoma. Ther Adv Med Oncol 2022;14:17588359221113266. [Crossref] [PubMed]
11. Lu LC, Hsu C, Shao YY, et al. Differential Organ-Specific Tumor Response to Immune Checkpoint Inhibitors in Hepatocellular Carcinoma. Liver Cancer 2019;8:480-90. [Crossref] [PubMed]
12. Ao H, Xin Z, Jian Z. Liquid biopsy to identify biomarkers for immunotherapy in hepatocellular carcinoma. Biomark Res 2021;9:91. [Crossref] [PubMed]
13. Jung SK, Kim Y, Kang B, et al. Organ-Specific Responses to Nivolumab Plus Ipilimumab in Advanced Hepatocellular Carcinoma: A Multicenter, Retrospective Study. Cancer Med 2025;14:e70997.
14. Cheon J, Jung S, Kim JS, et al. Organ-specific responses to atezolizumab plus bevacizumab in advanced hepatocellular carcinoma: A multicentre, retrospective study. Liver Int 2024;44:1961-70. [Crossref] [PubMed]
15. Schwartz LH, Litière S, de Vries E, et al. RECIST 1.1-Update and clarification: From the RECIST committee. Eur J Cancer 2016;62:132-7. [Crossref] [PubMed]
16. Hsu WF, Hsu SC, Chen TH, et al. Modified Albumin-Bilirubin Model for Stratifying Survival in Patients with Hepatocellular Carcinoma Receiving Anticancer Therapy. Cancers (Basel) 2022;14:5083. [Crossref] [PubMed]
17. Schmid S, Diem S, Li Q, et al. Organ-specific response to nivolumab in patients with non-small cell lung cancer (NSCLC). Cancer Immunol Immunother 2018;67:1825-32. [Crossref] [PubMed]
18. Kanda Y. Investigation of the freely available easy-to-use software 'EZR' for medical statistics. Bone Marrow Transplant 2013;48:452-8. [Crossref] [PubMed]
19. Tiegs G, Lohse AW. Immune tolerance: what is unique about the liver. J Autoimmun 2010;34:1-6. [Crossref] [PubMed]
20. Guo JC, Lin CY, Lin CC, et al. Response to Immune Checkpoint Inhibitors in Recurrent or Metastatic Esophageal Squamous Cell Carcinoma May Be Affected by Tumor Sites. Oncology 2021;99:652-8. [Crossref] [PubMed]
21. Kudo M. Combination Cancer Immunotherapy in Hepatocellular Carcinoma. Liver Cancer 2018;7:20-7. [Crossref] [PubMed]
22. Voron T, Marcheteau E, Pernot S, et al. Control of the immune response by pro-angiogenic factors. Front Oncol 2014;4:70. [Crossref] [PubMed]
23. Uchino K, Tateishi R, Shiina S, et al. Hepatocellular carcinoma with extrahepatic metastasis: clinical features and prognostic factors. Cancer 2011;117:4475-83. [Crossref] [PubMed]