TACE, lenvatinib, and PD-1 inhibitors in unresectable advanced hepatocellular carcinoma: an exploratory retrospective cohort study Authors: Hongyi Zhao, Tingting Li, Yu Zhang, Yang Song, Miaomiao Li Published: 2026 ABSTRACT BackgroundFor patients with unresectable advanced hepatocellular carcinoma (HCC), therapeutic options are limited, and transarterial chemoembolization (TACE) alone often yields suboptimal outcomes. This study aimed to evaluate clinical outcomes and recorded safety events associated with combining TACE, lenvatinib, and programmed cell death protein 1 (PD-1) inhibitors for HCC.MethodsA retrospective cohort study was conducted involving 209 patients with unresectable advanced HCC treated from June 2016 to December 2022. Patients were divided into the TACE + len group (n = 96) and the TACE + len+PD-1 group (n = 113). Tumor response, recorded adverse events, treatment exposure, OS, and PFS were analyzed. Propensity score-based IPTW was performed using baseline demographic and clinical variables to reduce measured baseline imbalance.ResultsAfter IPTW adjustment, baseline covariates achieved acceptable balance. Median OS was 16.49 months in the TACE + len+PD-1 group and 13.37 months in the TACE + len group; IPTW-weighted Cox regression showed an association between triple therapy and longer OS (HR = 0.591, 95% CI 0.417–0.836, P = 0.003). Median PFS was 10.02 months and 3.42 months, respectively, and the IPTW-weighted HR for PFS was 0.638 (95% CI 0.479–0.848, P = 0.002). Grade ≥ 3 adverse events occurred in 18.58% and 14.58% of patients, respectively (P = 0.440).ConclusionIn this exploratory retrospective cohort, adding PD-1 inhibitors to TACE plus lenvatinib was associated with improved tumor response and longer OS/PFS after IPTW adjustment, without a significant increase in recorded Grade ≥ 3 adverse events. These findings require prospective validation. Introduction Hepatocellular carcinoma (HCC) represents a formidable global health challenge, ranking as the third leading cause of cancer-related mortality worldwide [1–3]. Despite surveillance programs and curative-intent therapies for early-stage disease, a significant proportion of patients experience recurrence, often presenting at an advanced or unresectable stage [4]. The prognosis for unresectable advanced HCC remains particularly dismal, with limited therapeutic options and a pressing need for more effective strategies [4, 5]. Transarterial chemoembolization (TACE) has long been established as a cornerstone locoregional therapy for intermediate-stage HCC, offering palliative control by inducing tumor ischemia and delivering localized chemotherapy [4]. However, its efficacy in the advanced setting is frequently suboptimal due to factors such as tumor heterogeneity, compromised liver function, vascular invasion, and the development of treatment resistance [4, 5]. Systemic therapy forms the backbone of treatment for advanced HCC unsuitable for locoregional approaches. The advent of molecular targeted agents, particularly multi-kinase inhibitors (MKIs) like lenvatinib which targets vascular endothelial growth factor receptors (VEGFRs), platelet-derived growth factor receptor (PDGFR), and fibroblast growth factor receptors (FGFRs), represented a significant advancement [6–8]. Lenvatinib, demonstrating non-inferiority to sorafenib in the landmark REFLECT phase III trial (median OS: 13.6 vs. 12.3 months, HR 0.92; PFS: 7.3 vs. 3.6 months; ORR: 24% vs. 9%) [9], offers anti-angiogenic effects crucial for countering HCC’s hypervascular nature. Nevertheless, real-world studies indicate that lenvatinib monotherapy or its combination with TACE provides median OS of only 11–14 months [4, 6], with transient durability and limited survival gains, underscoring persistent unmet needs. The emergence of immune checkpoint inhibitors (ICIs), pioneered by Nobel-recognized discoveries in T-cell regulation (specifically the identification of CTLA-4 and PD-1 as negative regulators of T-cell activation) [10], has transformed oncology since the first PD-1 inhibitor (nivolumab) gained FDA approval for advanced melanoma in December 2014 [11]. In HCC, pivotal trials like the phase I/II CheckMate 040 trial [12], demonstrated single-agent anti-PD-1 efficacy in sorafenib-experienced patients, with an objective response rate (ORR) of 15% in dose-escalation cohorts and 20% in dose-expansion cohorts (by RECIST 1.1), and a median overall survival (OS) of 15.6 months (95% CI: 11–20) in the overall population, extending to 28 months in Child-Pugh A patients [12]. ICIs harness the host immune system to recognize and eliminate tumor cells, demonstrating encouraging single-agent activity (e.g., nivolumab monotherapy achieving 15–20% ORR in sorafenib-experienced HCC) [12], and subsequently forming the basis of landmark combination regimens like atezolizumab plus bevacizumab, which significantly improved median OS to 19.2 months vs. 13.4 months with sorafenib in the IMbrave150 trial [13]. This success highlights the potential of immunotherapy within the HCC treatment paradigm [14, 15]. Importantly, a compelling biological rationale exists for combining TACE, anti-angiogenic agents, and immunotherapy. TACE-induced tumor necrosis can release tumor-associated antigens, potentially priming an adaptive immune response – an effect often hampered by the immunosuppressive tumor microenvironment (TME) characteristic of HCC [16–18]. Anti-angiogenic agents like lenvatinib may modulate this TME by normalizing aberrant vasculature, reducing immunosuppressive cell populations, and enhancing T-cell infiltration, thereby potentially overcoming resistance and synergizing with the immune-activating effects of PD-1 inhibitors [16, 19]. While the triple combination of TACE, lenvatinib, and PD-1 inhibitors is increasingly being explored clinically based on this biological rationale, robust comparative data evaluating its efficacy and safety against established dual-modality approaches (TACE plus lenvatinib) in the specific context of unresectable advanced HCC remain scarce. Real-world evidence comparing these strategies is crucial to inform clinical decision-making, as this patient population often presents with unique challenges, including prior treatment exposure and potentially more aggressive tumor biology. Therefore, a critical knowledge gap exists regarding whether the addition of PD-1 inhibitors to the TACE-lenvatinib backbone translates into tangible clinical benefits for this challenging cohort, warranting rigorous investigation. This study aims to address this gap by evaluating the clinical association of combining TACE, lenvatinib, and PD-1 inhibitors in patients with unresectable advanced HCC. The findings may provide real-world evidence for further evaluation of this combined treatment strategy. Materials and methods Participants This retrospective cohort study included 209 patients with unresectable advanced hepatocellular carcinoma who were treated with TACE combined with lenvatinib or TACE combined with lenvatinib and PD-1 inhibitors at our hospital from June 2016 to December 2022. Patients were divided into two groups based on the treatment they received: the TACE combined with lenvatinib group (hereinafter referred to as the TACE + len group, n = 96) and the TACE combined with lenvatinib and PD-1 inhibitors group (hereinafter referred to as the TACE + len+PD-1 group, n = 113). Demographic information of the patients was collected through the case system. The inclusion criteria are as follows: (1) age 18–80 years; (2) histologically/cytologically or clinically (meeting AASLD/EASL imaging diagnostic criteria) confirmed hepatocellular carcinoma (HCC) [20]; (3) Barcelona Clinic Liver Cancer stage (BCLC) C, or locally advanced (Stage B) unsuitable for curative treatment and not suitable/refused surgical resection or ablation therapy; (4) according to mRECIST criteria, at least one measurable intrahepatic lesion; (5) liver function Child-Pugh grade A or B (≤ 7 points); (6) Eastern Cooperative Oncology Group (ECOG) Performance Status: 0–2 points; (7) no prior immunotherapy or anti-angiogenic treatment. Exclusion criteria for patients are as follows: (1) mainly extrahepatic metastasis or combined with main portal vein tumor thrombus involving contralateral branches or inferior vena cava tumor thrombus; (2) those with central nervous system metastases or other primary malignant tumors in other sites/organs; (3) any contraindications to the drugs used in this study; (4) those with severe comorbidities, including heart, brain, kidney, lung diseases; (5) recent history of gastrointestinal bleeding; (6) previously received other local or systemic treatments (radiofrequency ablation, iodine-125 seed implantation, etc.); (7) pregnant or lactating women; (8) incomplete clinical data. As this retrospective study uses de-identified patient data, it will not cause any potential harm to the patient, thus informed consent is waived. This waiver and the study have been approved by the Ethics Review Committee of our hospital, meeting relevant regulatory and ethical standards. Treatment methods TACE treatment TACE was performed for intrahepatic tumor control and palliation. TACE was performed by the same professional radiologist, and all patients were evaluated for liver function before undergoing TACE. The procedure for TACE is as follows: after routine disinfection and draping, 2% lidocaine injection was used for local anesthesia in the inguinal area. Using the Seldinger technique, femoral artery puncture was performed and a 5 F vascular sheath was inserted. A 5 F-RH catheter was then introduced through the sheath, and celiac artery angiography was performed under digital subtraction angiography (DSA) to determine the size, location, feeding arteries, and distribution of blood vessels of the tumor. The microcatheter was superselected into the tumor-feeding artery, and an appropriate amount of iodized oil chemotherapy drug emulsion (chemotherapy drugs such as oxaliplatin, raltitrexed, epirubicin) was slowly infused through the microcatheter, followed by embolization with gelatin sponge particles until blood flow ceased. After treatment, the catheter was removed, and local pressure dressing was applied, with the patient lying flat for 24 h to prevent bleeding and hematoma formation at the puncture site. Lenvatinib treatment Patients in both groups started continuous oral administration of the molecular targeted drug lenvatinib one week after TACE treatment. The dosage was calculated based on body weight: for patients weighing less than 60 kg, the oral dose of lenvatinib was 8 mg/day; for those weighing 60 kg or more, the oral dose was 12 mg/day. Dose reductions or temporary interruptions were allowed when clinically indicated to manage treatment-emergent adverse events, particularly Grade ≥ 3 events, according to drug instructions and institutional practice, until the adverse events were alleviated or resolved. PD-1 inhibitor treatment Patients receiving triple therapy underwent PD-1 inhibitor treatment every three weeks, with the first administration generally initiated after TACE according to institutional practice. Specific PD-1 inhibitors, dosages, and their distributions are shown in Table 1. The median number of PD-1 inhibitor cycles was 13 (IQR, 4–18; range, 2–35). Because this study was not powered to compare individual PD-1 agents, PD-1 inhibitors were analyzed as a treatment class. Table 1PD-1 inhibitors used in the triple-therapy regimen and their distributionPD-1 InhibitorDosageTrade NameApproval NumberNo. of patients, n (%)Pembrolizumab200 mgKeytrudaNMPA Approval No. SJ2018001912 (10.6%)Sintilimab200 mgTyvytNMPA Approval No. S2018001632 (28.3%)Toripalimab240 mgTuoyiNMPA Approval No. S2018001518 (15.9%)Camrelizumab200 mgAiRuiKaNMPA Approval No. S2019002729 (25.7%)Tislelizumab200 mgBaiZeAnNMPA Approval No. S2019004522 (19.5%)The median number of PD-1 inhibitor cycles was 13 (IQR, 4–18; range, 2–35) Table 1PD-1 inhibitors used in the triple-therapy regimen and their distributionPD-1 InhibitorDosageTrade NameApproval NumberNo. of patients, n (%)Pembrolizumab200 mgKeytrudaNMPA Approval No. SJ2018001912 (10.6%)Sintilimab200 mgTyvytNMPA Approval No. S2018001632 (28.3%)Toripalimab240 mgTuoyiNMPA Approval No. S2018001518 (15.9%)Camrelizumab200 mgAiRuiKaNMPA Approval No. S2019002729 (25.7%)Tislelizumab200 mgBaiZeAnNMPA Approval No. S2019004522 (19.5%)The median number of PD-1 inhibitor cycles was 13 (IQR, 4–18; range, 2–35) PD-1 inhibitors used in the triple-therapy regimen and their distribution The median number of PD-1 inhibitor cycles was 13 (IQR, 4–18; range, 2–35) Data collection Baseline Data Collection: Demographic and clinical characteristics including age, gender, body mass index (BMI), Eastern Cooperative Oncology Group Performance Status (ECOG PS), Barcelona Clinic Liver Cancer stage (BCLC), Child-Pugh grade, tumor size, tumor number, presence of vascular invasion, presence of extrahepatic metastasis, underlying etiology (e.g., hepatitis B/C, alcohol), and number of transarterial chemoembolization (TACE) sessions were retrieved from electronic medical records.Laboratory Parameters: Peripheral blood samples were collected at baseline (pre-treatment) and 1-month post-treatment initiation. The following tests were performed:① Hematological parameters: Complete blood counts including white blood cell count (WBC), absolute neutrophil count (ANC), hemoglobin (Hb), and platelet count (PLT) were measured using an automated hematology analyzer (Sysmex XN-9000, Sysmex Corporation, Japan). Neutrophil-to-lymphocyte ratio (NLR) and platelet-to-lymphocyte ratio (PLR) were calculated from differential counts. ②Coagulation function: Prothrombin time (PT), international normalized ratio (INR), and activated partial thromboplastin time (APTT) were determined using a fully automated coagulation analyzer (Sysmex CS-5100, Sysmex Corporation, Japan) with standardized reagents.③ Liver and kidney function: Serum alanine aminotransferase (ALT), aspartate aminotransferase (AST), total bilirubin (TBIL), albumin (ALB), creatinine (Cr), and blood urea nitrogen (BUN) levels were quantified using enzymatic/colorimetric methods on an automated biochemistry analyzer (Roche Cobas® 8000, Roche Diagnostics, Switzerland).④ Tumor markers: Serum alpha-fetoprotein (AFP), carcinoembryonic antigen (CEA), and carbohydrate antigen 19-9 (CA19-9) concentrations were measured via chemiluminescent immunoassays (Roche Cobas® 8000, Roche Diagnostics, Switzerland) using manufacturer-calibrated reagents.Short-term Clinical Efficacy Evaluation: Tumor response was assessed 3 months post-treatment initiation using contrast-enhanced CT/MRI, evaluated using mRECIST-based criteria focused on viable arterial-enhancing lesions [21]. The specific evaluation criteria are as follows: Complete Response (CR) refers to the disappearance of any arterial enhancement within all target lesions. Partial Response (PR) is defined as a decrease of at least 30% in the sum of diameters of target lesions during the arterial enhancement phase. Stable Disease (SD) indicates that the reduction in the sum of diameters of target lesions does not meet the criteria for PR or the increase does not meet the criteria for PD. Progressive Disease (PD) refers to an increase of at least 20% in the sum of diameters of viable enhancing target lesions or the appearance of new intrahepatic or extrahepatic lesions. Objective response rate (ORR: CR + PR) and disease control rate (DCR: CR + PR + SD) were calculated.Documentation of Adverse Events: Treatment-emergent adverse events recorded in the medical records were summarized as binary incidence variables, with CTCAE version 5.0 referenced for adverse event identification [22]. The relevant adverse reactions observed in this study include hand-foot syndrome, diarrhea, abdominal pain, decreased appetite, nausea, rash, hypothyroidism, hyperbilirubinemia, and proteinuria. Baseline Data Collection: Demographic and clinical characteristics including age, gender, body mass index (BMI), Eastern Cooperative Oncology Group Performance Status (ECOG PS), Barcelona Clinic Liver Cancer stage (BCLC), Child-Pugh grade, tumor size, tumor number, presence of vascular invasion, presence of extrahepatic metastasis, underlying etiology (e.g., hepatitis B/C, alcohol), and number of transarterial chemoembolization (TACE) sessions were retrieved from electronic medical records. Baseline Data Collection: Demographic and clinical characteristics including age, gender, body mass index (BMI), Eastern Cooperative Oncology Group Performance Status (ECOG PS), Barcelona Clinic Liver Cancer stage (BCLC), Child-Pugh grade, tumor size, tumor number, presence of vascular invasion, presence of extrahepatic metastasis, underlying etiology (e.g., hepatitis B/C, alcohol), and number of transarterial chemoembolization (TACE) sessions were retrieved from electronic medical records. Laboratory Parameters: Peripheral blood samples were collected at baseline (pre-treatment) and 1-month post-treatment initiation. The following tests were performed:① Hematological parameters: Complete blood counts including white blood cell count (WBC), absolute neutrophil count (ANC), hemoglobin (Hb), and platelet count (PLT) were measured using an automated hematology analyzer (Sysmex XN-9000, Sysmex Corporation, Japan). Neutrophil-to-lymphocyte ratio (NLR) and platelet-to-lymphocyte ratio (PLR) were calculated from differential counts. ②Coagulation function: Prothrombin time (PT), international normalized ratio (INR), and activated partial thromboplastin time (APTT) were determined using a fully automated coagulation analyzer (Sysmex CS-5100, Sysmex Corporation, Japan) with standardized reagents.③ Liver and kidney function: Serum alanine aminotransferase (ALT), aspartate aminotransferase (AST), total bilirubin (TBIL), albumin (ALB), creatinine (Cr), and blood urea nitrogen (BUN) levels were quantified using enzymatic/colorimetric methods on an automated biochemistry analyzer (Roche Cobas® 8000, Roche Diagnostics, Switzerland).④ Tumor markers: Serum alpha-fetoprotein (AFP), carcinoembryonic antigen (CEA), and carbohydrate antigen 19-9 (CA19-9) concentrations were measured via chemiluminescent immunoassays (Roche Cobas® 8000, Roche Diagnostics, Switzerland) using manufacturer-calibrated reagents. Laboratory Parameters: Peripheral blood samples were collected at baseline (pre-treatment) and 1-month post-treatment initiation. The following tests were performed:① Hematological parameters: Complete blood counts including white blood cell count (WBC), absolute neutrophil count (ANC), hemoglobin (Hb), and platelet count (PLT) were measured using an automated hematology analyzer (Sysmex XN-9000, Sysmex Corporation, Japan). Neutrophil-to-lymphocyte ratio (NLR) and platelet-to-lymphocyte ratio (PLR) were calculated from differential counts. ②Coagulation function: Prothrombin time (PT), international normalized ratio (INR), and activated partial thromboplastin time (APTT) were determined using a fully automated coagulation analyzer (Sysmex CS-5100, Sysmex Corporation, Japan) with standardized reagents.③ Liver and kidney function: Serum alanine aminotransferase (ALT), aspartate aminotransferase (AST), total bilirubin (TBIL), albumin (ALB), creatinine (Cr), and blood urea nitrogen (BUN) levels were quantified using enzymatic/colorimetric methods on an automated biochemistry analyzer (Roche Cobas® 8000, Roche Diagnostics, Switzerland).④ Tumor markers: Serum alpha-fetoprotein (AFP), carcinoembryonic antigen (CEA), and carbohydrate antigen 19-9 (CA19-9) concentrations were measured via chemiluminescent immunoassays (Roche Cobas® 8000, Roche Diagnostics, Switzerland) using manufacturer-calibrated reagents. Short-term Clinical Efficacy Evaluation: Tumor response was assessed 3 months post-treatment initiation using contrast-enhanced CT/MRI, evaluated using mRECIST-based criteria focused on viable arterial-enhancing lesions [21]. The specific evaluation criteria are as follows: Complete Response (CR) refers to the disappearance of any arterial enhancement within all target lesions. Partial Response (PR) is defined as a decrease of at least 30% in the sum of diameters of target lesions during the arterial enhancement phase. Stable Disease (SD) indicates that the reduction in the sum of diameters of target lesions does not meet the criteria for PR or the increase does not meet the criteria for PD. Progressive Disease (PD) refers to an increase of at least 20% in the sum of diameters of viable enhancing target lesions or the appearance of new intrahepatic or extrahepatic lesions. Objective response rate (ORR: CR + PR) and disease control rate (DCR: CR + PR + SD) were calculated. Short-term Clinical Efficacy Evaluation: Tumor response was assessed 3 months post-treatment initiation using contrast-enhanced CT/MRI, evaluated using mRECIST-based criteria focused on viable arterial-enhancing lesions [21]. The specific evaluation criteria are as follows: Complete Response (CR) refers to the disappearance of any arterial enhancement within all target lesions. Partial Response (PR) is defined as a decrease of at least 30% in the sum of diameters of target lesions during the arterial enhancement phase. Stable Disease (SD) indicates that the reduction in the sum of diameters of target lesions does not meet the criteria for PR or the increase does not meet the criteria for PD. Progressive Disease (PD) refers to an increase of at least 20% in the sum of diameters of viable enhancing target lesions or the appearance of new intrahepatic or extrahepatic lesions. Objective response rate (ORR: CR + PR) and disease control rate (DCR: CR + PR + SD) were calculated. Documentation of Adverse Events: Treatment-emergent adverse events recorded in the medical records were summarized as binary incidence variables, with CTCAE version 5.0 referenced for adverse event identification [22]. The relevant adverse reactions observed in this study include hand-foot syndrome, diarrhea, abdominal pain, decreased appetite, nausea, rash, hypothyroidism, hyperbilirubinemia, and proteinuria. Documentation of Adverse Events: Treatment-emergent adverse events recorded in the medical records were summarized as binary incidence variables, with CTCAE version 5.0 referenced for adverse event identification [22]. The relevant adverse reactions observed in this study include hand-foot syndrome, diarrhea, abdominal pain, decreased appetite, nausea, rash, hypothyroidism, hyperbilirubinemia, and proteinuria. Endpoint definitions OS and PFS durations were summarized as observed follow-up outcomes. OS was defined as the recorded duration from treatment initiation to death or the last available follow-up record. PFS was defined as the recorded duration from treatment initiation to radiological progression, death, or the last available follow-up record. Follow up Follow-up evaluations began 4 to 10 weeks after the completion of the last treatment and were subsequently conducted at regular intervals of 1 to 3 months. Follow-up data were collected through outpatient or inpatient systems, records from external hospitals, and scheduled follow-up visits to monitor the patients’ survival status and disease condition. The follow-up cutoff date was December 2024. Statistical analysis All statistical analyses were performed using SPSS (version 22.0; IBM Corp., Armonk, NY, USA) and R software (version 4.3.1; R Foundation). Continuous variables with normal distribution were expressed as mean ± standard deviation (SD) and compared between the two treatment groups using independent samples t-tests. Non-normally distributed continuous data were analyzed with the Mann-Whitney U test. Categorical variables were presented as frequencies (%) and compared using Pearson’s χ² test or Fisher’s exact test for small cell counts. Survival outcomes were analyzed using Kaplan–Meier methods and Cox proportional hazards models. OS and PFS were reported as median survival times, and between-group differences were assessed using the log-rank test. Hazard ratios (HRs) with 95% confidence intervals (CIs) were estimated using Cox regression models. Propensity scores for receiving TACE + len+PD-1 were estimated using baseline demographic and clinical variables listed in Table 2, including age, sex, BMI, education level, employment status, marital status, residence, alcohol history, smoking history, ECOG PS, BCLC stage, Child-Pugh grade, tumor size, tumor number, vascular invasion, extrahepatic metastasis, TACE session category, and etiology. Stabilized inverse probability of treatment weights were then generated. Covariate balance before and after weighting was assessed using standardized mean differences (SMDs), with an absolute SMD < 0.10 indicating acceptable balance. The primary comparative analyses, including tumor response, OS/PFS, and Cox regression, were performed in the IPTW-weighted cohort. Unweighted results were retained as descriptive or supplementary analyses. Table 2Baseline demographic and clinical characteristics before and after IPTW adjustmentParametersTACE + len group (n = 96)TACE + len+PD-1 group (n = 113)P valueUnweighted SMDWeighted SMDBaseline DemographicAge (years)58.73 ± 8.3457.92 ± 9.160.5060.0930.036Gender (Male/Female)74 (77.08%) /22 (22.92%)89 (78.76%) /24 (21.24%)0.7700.0400.021BMI (kg/m²)22.68 ± 3.1723.12 ± 2.940.2970.1450.026Education Level [n (%)]0.6920.1170.022 - Primary or below38(39.58%)42(37.17%) - Secondary School32(33.33%)44(38.94%) - College or above26(27.08%)27(23.89%)Employment Status [n (%)]0.7250.0490.001 - Employed41(42.71%)51(45.13%) - Unemployed55(57.29%)62(54.87%)Marital Status [n (%)]0.6210.1330.060 - Married82(85.42%)93(82.30%) - Single8(8.33%)14(12.39%) - Divorced6(6.25%)6(5.31%)Current Residence [n (%)]0.8530.0260.004 - Urban54(56.25%)65(57.52%) - Rural42(43.75%)48(42.48%)History of Alcohol (Yes/No)43 (44.79%) /53 (55.21%)48 (42.48%) /65 (57.52%)0.7370.0470.010History of Smoking (Yes/No)51 (53.12%) /45 (46.88%)63 (55.75%) /50 (44.25%)0.7040.0530.011Clinical Disease CharacteristicsECOG PS Score [n (%)]0.6250.0680.016 − 0–1 points72(75.00%)88(77.88%) − 2 points24(25.00%)25(22.12%)BCLC Staging [n (%)]0.6560.0620.012 - B32(33.33%)41(36.28%) - C64(66.67%)72(63.72%)Child-Pugh grade [n (%)]0.7700.0400.010 - Class A74(77.08%)89(78.76%) - Class B22(22.92%)24(21.24%)Tumor Size [n (%)]0.8840.0200.001 - ≤3 cm28(29.17%)34(30.09%) - >3 cm68(70.83%)79(69.91%)Tumor Number [n (%)]0.7700.0410.013 - ≤341(42.71%)46(40.71%) - >355(57.29%)67(59.29%)Vascular Invasion [n (%)]0.7000.0530.006 - Present63(65.63%)77(68.14%) - Absent33(34.37%)36(31.86%)Extrahepatic Metastasis [n (%)]0.8690.0230.010 - Present38(39.58%)46(40.71%) - Absent58(60.42%)67(59.29%)TACE Sessions [n (%)]0.3750.1230.027 - ≤361(63.54%)65(57.52%) - >335(36.46%)48(42.48%)Etiology [n (%)]0.9260.0840.028 - Hepatitis B71(73.96%)87(76.99%) - Hepatitis C9(9.38%)8(7.08%) - Alcohol11(11.46%)13(11.50%) - Others5(5.21%)5(4.42%)P values were calculated using independent-samples t tests, χ² tests, or Fisher’s exact tests, as appropriate. SMD standardized mean difference, IPTW inverse probability of treatment weighting. For categorical variables with more than two levels, the reported SMD represents the maximum absolute SMD across category levels. Covariate balance was considered acceptable when the absolute SMD was < 0.10 Table 2Baseline demographic and clinical characteristics before and after IPTW adjustmentParametersTACE + len group (n = 96)TACE + len+PD-1 group (n = 113)P valueUnweighted SMDWeighted SMDBaseline DemographicAge (years)58.73 ± 8.3457.92 ± 9.160.5060.0930.036Gender (Male/Female)74 (77.08%) /22 (22.92%)89 (78.76%) /24 (21.24%)0.7700.0400.021BMI (kg/m²)22.68 ± 3.1723.12 ± 2.940.2970.1450.026Education Level [n (%)]0.6920.1170.022 - Primary or below38(39.58%)42(37.17%) - Secondary School32(33.33%)44(38.94%) - College or above26(27.08%)27(23.89%)Employment Status [n (%)]0.7250.0490.001 - Employed41(42.71%)51(45.13%) - Unemployed55(57.29%)62(54.87%)Marital Status [n (%)]0.6210.1330.060 - Married82(85.42%)93(82.30%) - Single8(8.33%)14(12.39%) - Divorced6(6.25%)6(5.31%)Current Residence [n (%)]0.8530.0260.004 - Urban54(56.25%)65(57.52%) - Rural42(43.75%)48(42.48%)History of Alcohol (Yes/No)43 (44.79%) /53 (55.21%)48 (42.48%) /65 (57.52%)0.7370.0470.010History of Smoking (Yes/No)51 (53.12%) /45 (46.88%)63 (55.75%) /50 (44.25%)0.7040.0530.011Clinical Disease CharacteristicsECOG PS Score [n (%)]0.6250.0680.016 − 0–1 points72(75.00%)88(77.88%) − 2 points24(25.00%)25(22.12%)BCLC Staging [n (%)]0.6560.0620.012 - B32(33.33%)41(36.28%) - C64(66.67%)72(63.72%)Child-Pugh grade [n (%)]0.7700.0400.010 - Class A74(77.08%)89(78.76%) - Class B22(22.92%)24(21.24%)Tumor Size [n (%)]0.8840.0200.001 - ≤3 cm28(29.17%)34(30.09%) - >3 cm68(70.83%)79(69.91%)Tumor Number [n (%)]0.7700.0410.013 - ≤341(42.71%)46(40.71%) - >355(57.29%)67(59.29%)Vascular Invasion [n (%)]0.7000.0530.006 - Present63(65.63%)77(68.14%) - Absent33(34.37%)36(31.86%)Extrahepatic Metastasis [n (%)]0.8690.0230.010 - Present38(39.58%)46(40.71%) - Absent58(60.42%)67(59.29%)TACE Sessions [n (%)]0.3750.1230.027 - ≤361(63.54%)65(57.52%) - >335(36.46%)48(42.48%)Etiology [n (%)]0.9260.0840.028 - Hepatitis B71(73.96%)87(76.99%) - Hepatitis C9(9.38%)8(7.08%) - Alcohol11(11.46%)13(11.50%) - Others5(5.21%)5(4.42%)P values were calculated using independent-samples t tests, χ² tests, or Fisher’s exact tests, as appropriate. SMD standardized mean difference, IPTW inverse probability of treatment weighting. For categorical variables with more than two levels, the reported SMD represents the maximum absolute SMD across category levels. Covariate balance was considered acceptable when the absolute SMD was < 0.10 Baseline demographic and clinical characteristics before and after IPTW adjustment 74 (77.08%) / 22 (22.92%) 89 (78.76%) / 24 (21.24%) 43 (44.79%) / 53 (55.21%) 48 (42.48%) / 65 (57.52%) 51 (53.12%) / 45 (46.88%) 63 (55.75%) / 50 (44.25%) P values were calculated using independent-samples t tests, χ² tests, or Fisher’s exact tests, as appropriate. SMD standardized mean difference, IPTW inverse probability of treatment weighting. For categorical variables with more than two levels, the reported SMD represents the maximum absolute SMD across category levels. Covariate balance was considered acceptable when the absolute SMD was < 0.10 Results Baseline demographic and clinical characteristics Before weighting, baseline demographic and clinical variables were generally comparable between the two groups. However, because P values alone are insufficient to demonstrate covariate balance in a non-randomized cohort, IPTW was applied and covariate balance was assessed using SMDs. After IPTW adjustment, all baseline variables achieved acceptable balance, with absolute weighted SMDs below 0.10. The distribution of TACE sessions was also comparable between groups (≤ 3 vs. >3 sessions, P = 0.375; Table 2). Comparison of laboratory parameters before treatment and 1 month after treatment Post-treatment hematological changes revealed significant reductions in systemic inflammatory markers in the TACE + len+PD-1 group. At 1-month follow-up, neutrophil-to-lymphocyte ratio (NLR) and platelet-to-lymphocyte ratio (PLR) were significantly lower in the TACE + len+PD-1 group compared to TACE + len (NLR: 2.46 ± 0.93 vs. 2.87 ± 1.06, P = 0.004; PLR: 112.83 ± 32.46 vs. 124.36 ± 37.21, P = 0.018). No intergroup differences were observed in WBC, ANC, Hb, PLT, or coagulation parameters (PT, INR, APTT; all P > 0.05; Table 3). Table 3Comparison of laboratory parameters before treatment and 1 month after treatmentParametersTimeTACE + len group (n = 96)TACE + len+PD-1 group (n = 113)tP valueHematological and Coagulation ParametersWBC (×10⁹/L)Before6.38 ± 1.736.42 ± 1.680.1550.8771 month4.87 ± 1.264.92 ± 1.330.3010.763ANC (×10⁹/L)Before3.84 ± 1.123.91 ± 1.080.4640.6431 month2.73 ± 0.942.68 ± 0.870.4230.672Hb (g/L)Before123.42 ± 16.38125.73 ± 17.210.9920.3221 month112.86 ± 14.73114.27 ± 15.620.6680.505PLT (×10⁹/L)Before146.83 ± 48.72142.67 ± 51.360.5970.5511 month118.24 ± 42.16113.89 ± 39.870.7650.445NLRBefore3.18 ± 1.243.22 ± 1.310.2490.8041 month2.87 ± 1.062.46 ± 0.932.9500.004PLRBefore136.74 ± 42.63132.89 ± 38.720.6840.4951 month124.36 ± 37.21112.83 ± 32.462.3910.018PT (s)Before13.28 ± 1.4613.34 ± 1.520.2870.7741 month13.87 ± 1.6313.92 ± 1.710.2030.840INRBefore1.12 ± 0.181.14 ± 0.160.6510.5151 month1.18 ± 0.211.19 ± 0.230.3050.761APTT (s)Before36.42 ± 4.8335.87 ± 4.670.8440.3991 month38.16 ± 5.2437.92 ± 5.130.3270.744Hepatic and Renal Function ParametersALT (U/L)Before48.73 ± 22.1646.87 ± 20.940.6250.5331 month62.34 ± 28.7373.86 ± 31.422.7460.007AST (U/L)Before56.28 ± 24.3754.16 ± 23.840.6360.5261 month78.43 ± 32.1692.67 ± 36.282.9780.003TBIL (µmol/L)Before21.34 ± 8.7322.16 ± 9.240.6580.5111 month28.67 ± 11.3432.84 ± 12.732.4830.014ALB (g/L)Before36.84 ± 4.2737.12 ± 4.530.4610.6451 month34.26 ± 4.1833.78 ± 4.620.7790.437Cr (µmol/L)Before76.34 ± 18.4274.89 ± 17.630.5800.5621 month82.16 ± 21.3784.73 ± 23.160.8270.409BUN (mmol/L)Before5.82 ± 1.735.76 ± 1.680.2550.7991 month6.34 ± 2.166.28 ± 2.070.1770.860Tumor Marker TrajectoriesAFP (ng/mL)Before634.83 ± 287.46646.12 ± 292.370.2800.7801 month398.34 ± 104.67354.37 ± 113.282.8950.004CEA (U/mL)Before8.73 ± 3.248.92 ± 3.170.0280.9781 month7.86 ± 2.936.24 ± 2.164.501< 0.001CA199 (U/mL)Before42.16 ± 18.7343.28 ± 19.240.4240.6721 month38.42 ± 16.2831.73 ± 14.863.1040.002WBC White Blood Cell, ANC Absolute Neutrophil Count, Hb Hemoglobin, PLT Platelet, NLR Neutrophil-to-Lymphocyte Ratio, PLR Platelet-to-Lymphocyte Ratio, PT Prothrombin Time, INR International Normalized Ratio, APTT Activated Partial Thromboplastin Time, ALT Alanine Aminotransferase, AST Aspartate Aminotransferase, TBIL Total Bilirubin, ALB Albumin, Cr Creatinine, BUN Blood Urea Nitrogen, AFP Alpha-fetoprotein, CEA Carcinoembryonic Antigen, CA19-9 Carbohydrate Antigen 19 − 9 Table 3Comparison of laboratory parameters before treatment and 1 month after treatmentParametersTimeTACE + len group (n = 96)TACE + len+PD-1 group (n = 113)tP valueHematological and Coagulation ParametersWBC (×10⁹/L)Before6.38 ± 1.736.42 ± 1.680.1550.8771 month4.87 ± 1.264.92 ± 1.330.3010.763ANC (×10⁹/L)Before3.84 ± 1.123.91 ± 1.080.4640.6431 month2.73 ± 0.942.68 ± 0.870.4230.672Hb (g/L)Before123.42 ± 16.38125.73 ± 17.210.9920.3221 month112.86 ± 14.73114.27 ± 15.620.6680.505PLT (×10⁹/L)Before146.83 ± 48.72142.67 ± 51.360.5970.5511 month118.24 ± 42.16113.89 ± 39.870.7650.445NLRBefore3.18 ± 1.243.22 ± 1.310.2490.8041 month2.87 ± 1.062.46 ± 0.932.9500.004PLRBefore136.74 ± 42.63132.89 ± 38.720.6840.4951 month124.36 ± 37.21112.83 ± 32.462.3910.018PT (s)Before13.28 ± 1.4613.34 ± 1.520.2870.7741 month13.87 ± 1.6313.92 ± 1.710.2030.840INRBefore1.12 ± 0.181.14 ± 0.160.6510.5151 month1.18 ± 0.211.19 ± 0.230.3050.761APTT (s)Before36.42 ± 4.8335.87 ± 4.670.8440.3991 month38.16 ± 5.2437.92 ± 5.130.3270.744Hepatic and Renal Function ParametersALT (U/L)Before48.73 ± 22.1646.87 ± 20.940.6250.5331 month62.34 ± 28.7373.86 ± 31.422.7460.007AST (U/L)Before56.28 ± 24.3754.16 ± 23.840.6360.5261 month78.43 ± 32.1692.67 ± 36.282.9780.003TBIL (µmol/L)Before21.34 ± 8.7322.16 ± 9.240.6580.5111 month28.67 ± 11.3432.84 ± 12.732.4830.014ALB (g/L)Before36.84 ± 4.2737.12 ± 4.530.4610.6451 month34.26 ± 4.1833.78 ± 4.620.7790.437Cr (µmol/L)Before76.34 ± 18.4274.89 ± 17.630.5800.5621 month82.16 ± 21.3784.73 ± 23.160.8270.409BUN (mmol/L)Before5.82 ± 1.735.76 ± 1.680.2550.7991 month6.34 ± 2.166.28 ± 2.070.1770.860Tumor Marker TrajectoriesAFP (ng/mL)Before634.83 ± 287.46646.12 ± 292.370.2800.7801 month398.34 ± 104.67354.37 ± 113.282.8950.004CEA (U/mL)Before8.73 ± 3.248.92 ± 3.170.0280.9781 month7.86 ± 2.936.24 ± 2.164.501< 0.001CA199 (U/mL)Before42.16 ± 18.7343.28 ± 19.240.4240.6721 month38.42 ± 16.2831.73 ± 14.863.1040.002WBC White Blood Cell, ANC Absolute Neutrophil Count, Hb Hemoglobin, PLT Platelet, NLR Neutrophil-to-Lymphocyte Ratio, PLR Platelet-to-Lymphocyte Ratio, PT Prothrombin Time, INR International Normalized Ratio, APTT Activated Partial Thromboplastin Time, ALT Alanine Aminotransferase, AST Aspartate Aminotransferase, TBIL Total Bilirubin, ALB Albumin, Cr Creatinine, BUN Blood Urea Nitrogen, AFP Alpha-fetoprotein, CEA Carcinoembryonic Antigen, CA19-9 Carbohydrate Antigen 19 − 9 Comparison of laboratory parameters before treatment and 1 month after treatment WBC White Blood Cell, ANC Absolute Neutrophil Count, Hb Hemoglobin, PLT Platelet, NLR Neutrophil-to-Lymphocyte Ratio, PLR Platelet-to-Lymphocyte Ratio, PT Prothrombin Time, INR International Normalized Ratio, APTT Activated Partial Thromboplastin Time, ALT Alanine Aminotransferase, AST Aspartate Aminotransferase, TBIL Total Bilirubin, ALB Albumin, Cr Creatinine, BUN Blood Urea Nitrogen, AFP Alpha-fetoprotein, CEA Carcinoembryonic Antigen, CA19-9 Carbohydrate Antigen 19 − 9 Liver function parameters exhibited divergent trends post-treatment. The TACE + len+PD-1 group showed significantly higher ALT (73.86 ± 31.42 vs. 62.34 ± 28.73 U/L, P = 0.007), AST (92.67 ± 36.28 vs. 78.43 ± 32.16 U/L, P = 0.003), and TBIL (32.84 ± 12.73 vs. 28.67 ± 11.34 µmol/L, P = 0.014) at 1 month compared to TACE + len. Renal function markers (Cr, BUN) and albumin remained comparable between groups (all P > 0.05; Table 3). Reductions in tumor-marker levels were observed in the triple-therapy group. At 1-month follow-up, the TACE + len+PD-1 group demonstrated lower AFP (354.37 ± 113.28 vs. 398.34 ± 104.67 ng/mL, P = 0.004), CEA (6.24 ± 2.16 vs. 7.86 ± 2.93 U/mL, P < 0.001), and CA19-9 (31.73 ± 14.86 vs. 38.42 ± 16.28 U/mL, P = 0.002) compared to TACE + len (Table 3). Comparison of short-term treatment response at 3 months Tumor response was evaluated according to mRECIST-based criteria focused on viable arterial-enhancing lesions. The CR/PR/SD/PD distributions were 2/18/23/53 in the TACE + len group and 3/36/29/45 in the TACE + len+PD-1 group. In the unweighted analysis, ORR was higher in the TACE + len+PD-1 group than in the TACE + len group (39/113, 34.51% vs. 20/96, 20.83%), and DCR was also higher (68/113, 60.18% vs. 43/96, 44.79%). In the IPTW-weighted analysis, DCR remained higher in the TACE + len+PD-1 group than in the TACE + len group (58.91% vs. 45.11%, P = 0.047), whereas the ORR difference was attenuated and did not reach conventional statistical significance (32.92% vs. 21.40%, P = 0.065; Table 4). Table 4Short-term tumor response at 3 months according to mRECIST-based criteria before and after IPTW adjustmentParametersTACE + len group unweightedTACE + len+PD-1 group unweightedUnweighted P valueTACE + len IPTW-weightedTACE + len+PD-1 IPTW-weightedWeighted P valueORR20/96(20.83%)39/113 (34.51%)0.02921.40%32.92%0.065DCR43/96 (44.79%)68/113 (60.18%)0.02645.11%58.91%0.047- CR2/96 (2.08%)3/113 (2.65%)- PR18/96 (18.75%)36/113 (31.86%)- SD23/96 (23.96%)29/113 (25.66%)- PD53/96 (55.21%)45/113 (39.82%)ORR objective response rate, DCR disease control rate, CR Complete Response, PR Partial Response, SD Stable Disease, PD Progressive Disease Table 4Short-term tumor response at 3 months according to mRECIST-based criteria before and after IPTW adjustmentParametersTACE + len group unweightedTACE + len+PD-1 group unweightedUnweighted P valueTACE + len IPTW-weightedTACE + len+PD-1 IPTW-weightedWeighted P valueORR20/96(20.83%)39/113 (34.51%)0.02921.40%32.92%0.065DCR43/96 (44.79%)68/113 (60.18%)0.02645.11%58.91%0.047- CR2/96 (2.08%)3/113 (2.65%)- PR18/96 (18.75%)36/113 (31.86%)- SD23/96 (23.96%)29/113 (25.66%)- PD53/96 (55.21%)45/113 (39.82%)ORR objective response rate, DCR disease control rate, CR Complete Response, PR Partial Response, SD Stable Disease, PD Progressive Disease Short-term tumor response at 3 months according to mRECIST-based criteria before and after IPTW adjustment ORR objective response rate, DCR disease control rate, CR Complete Response, PR Partial Response, SD Stable Disease, PD Progressive Disease Comparison of treatment-related adverse events during follow-up The recorded incidence of selected adverse events was not significantly different between groups (Table 5). Grade ≥ 3 adverse events occurred in 21 patients (18.58%) in the TACE + len+PD-1 group and 14 patients (14.58%) in the TACE + len group, with no significant difference between groups (P = 0.440). Lenvatinib dose reduction occurred in 33 patients (29.2%) and 25 patients (26.0%), respectively (P = 0.611); temporary interruption occurred in 33 patients (29.2%) and 31 patients (32.3%), respectively (P = 0.629); and permanent discontinuation occurred in 5 patients (4.4%) and 3 patients (3.1%), respectively (P = 0.729; Table 5). Table 5Recorded adverse events and treatment exposure during follow-upParametersTACE + len group (n = 96)TACE + len+PD-1 group (n = 113)P valueHand-foot syndrome20 (20.83%)32 (28.32%)0.212Diarrhea15 (15.63%)25 (22.12%)0.234Abdominal pain10 (10.42%)18 (15.93%)0.244Decreased appetite25 (26.04%)29 (25.66%)0.950Nausea22 (22.92%)28 (24.78%)0.753Rash8 (8.33%)15 (13.27%)0.255Hypothyroidism5 (5.21%)12 (10.62%)0.154Hyperbilirubinemia12 (12.50%)21 (18.58%)0.229Proteinuria7 (7.29%)10 (8.85%)0.681Any Grade ≥ 3 adverse event14 (14.58%)21 (18.58%)0.440Lenvatinib dose reduction25 (26.0%)33 (29.2%)0.611Temporary interruption of lenvatinib31 (32.3%)33 (29.2%)0.629Permanent discontinuation of lenvatinib3 (3.1%)5 (4.4%)0.729 Table 5Recorded adverse events and treatment exposure during follow-upParametersTACE + len group (n = 96)TACE + len+PD-1 group (n = 113)P valueHand-foot syndrome20 (20.83%)32 (28.32%)0.212Diarrhea15 (15.63%)25 (22.12%)0.234Abdominal pain10 (10.42%)18 (15.93%)0.244Decreased appetite25 (26.04%)29 (25.66%)0.950Nausea22 (22.92%)28 (24.78%)0.753Rash8 (8.33%)15 (13.27%)0.255Hypothyroidism5 (5.21%)12 (10.62%)0.154Hyperbilirubinemia12 (12.50%)21 (18.58%)0.229Proteinuria7 (7.29%)10 (8.85%)0.681Any Grade ≥ 3 adverse event14 (14.58%)21 (18.58%)0.440Lenvatinib dose reduction25 (26.0%)33 (29.2%)0.611Temporary interruption of lenvatinib31 (32.3%)33 (29.2%)0.629Permanent discontinuation of lenvatinib3 (3.1%)5 (4.4%)0.729 Recorded adverse events and treatment exposure during follow-up Survival outcomes analysis Survival outcomes were analyzed using Kaplan–Meier methods and IPTW-weighted Cox regression. Median OS was 16.49 months in the TACE + len+PD-1 group and 13.37 months in the TACE + len group (log-rank P = 0.0014; Fig. 1a). In the IPTW-weighted Cox model, triple therapy was associated with longer OS (HR = 0.591, 95% CI 0.417–0.836, P = 0.003; Table 6). Median PFS was 10.02 months in the TACE + len+PD-1 group and 3.42 months in the TACE + len group (log-rank P = 0.0011; Fig. 1b). The IPTW-weighted Cox model also showed an association between triple therapy and longer PFS (HR = 0.638, 95% CI 0.479–0.848, P = 0.002; Table 6). Fig. 1IPTW-weighted Kaplan–Meier curves for OS and PFS. a: OS; b: PFS. OS, overall survival; PFS, progression-free survival; IPTW, inverse probability of treatment weighting IPTW-weighted Kaplan–Meier curves for OS and PFS. a: OS; b: PFS. OS, overall survival; PFS, progression-free survival; IPTW, inverse probability of treatment weighting Table 6IPTW-weighted Cox regression analyses for OS and PFSOutcomeComparisonIPTW-weighted HR95% CIP valueOSTACE + len+PD-1 vs. TACE + len0.591(0.417–0.836)0.003PFSTACE + len+PD-1 vs. TACE + len0.638(0.479–0.848)0.002HR hazard ratio, CI confidence interval, OS overall survival, PFS progression-free survival, IPTW inverse probability of treatment weighting. HRs were estimated using IPTW-weighted Cox proportional hazards models. The TACE + len group was used as the reference group Table 6IPTW-weighted Cox regression analyses for OS and PFSOutcomeComparisonIPTW-weighted HR95% CIP valueOSTACE + len+PD-1 vs. TACE + len0.591(0.417–0.836)0.003PFSTACE + len+PD-1 vs. TACE + len0.638(0.479–0.848)0.002HR hazard ratio, CI confidence interval, OS overall survival, PFS progression-free survival, IPTW inverse probability of treatment weighting. HRs were estimated using IPTW-weighted Cox proportional hazards models. The TACE + len group was used as the reference group IPTW-weighted Cox regression analyses for OS and PFS HR hazard ratio, CI confidence interval, OS overall survival, PFS progression-free survival, IPTW inverse probability of treatment weighting. HRs were estimated using IPTW-weighted Cox proportional hazards models. The TACE + len group was used as the reference group Discussion Possible biological explanations for the observed association The observed association may be partly explained by multifaceted biological interactions [4]. TACE induces tumor ischemia and necrosis, releasing tumor-associated antigens (TAAs) and damage-associated molecular patterns (DAMPs) [23–25]. This promotes dendritic cell maturation and antigen presentation, as demonstrated by Ren et al. [26], who observed increased CD4+/CD8 + T-cell ratios and PD-1 expression in peripheral monocytes post-TACE. However, the ensuing hypoxic microenvironment often upregulates immunosuppressive factors such as VEGF, PD-L1, and hypoxia-inducible factor-1α (HIF-1α), thereby counteracting T-cell activation [24, 27, 28]. Zhong et al.[29] further elucidated that TACE-induced hypoxia elevates S100A9 via HIF-1α, driving mitochondrial dysfunction and tumor progression—a key resistance mechanism. Lenvatinib mitigates this immunosuppression by normalizing aberrant tumor vasculature. Kimura et al.[30] demonstrated in immunocompetent HCC models that lenvatinib reduces VEGF/FGF-driven immunosuppressive cytokine secretion and enhances intratumoral CD8 + T-cell infiltration, effects not observed with sorafenib. PD-1 inhibitors may further contribute to the observed association by reversing T-cell exhaustion triggered by TACE-induced hypoxia [25, 31]. Preclinical models indicate that antiangiogenic agents enhance intratumoral CD8 + T-cell infiltration, while PD-1 blockade prevents their functional impairment [32–34]. Clinical evidence from a phase 1b study of lenvatinib plus pembrolizumab in patients with unresectable HCC further supports the feasibility of combining anti-angiogenic therapy with PD-1 blockade [35]. This is consistent with the lower AFP, CEA, and CA19-9 levels observed in the triple-therapy group, which may reflect changes in tumor-marker profiles (Table 3). The triple-therapy group exhibited higher post-treatment ALT, AST, and TBIL (Table 3). While this may reflect transient immune-mediated hepatocyte injury or enhanced tumor lysis, this finding indicates that liver function should be carefully monitored during triple therapy. Clinical implications Our findings suggest that the addition of PD-1 inhibitors to TACE plus lenvatinib may warrant further evaluation in selected patients. After IPTW adjustment, triple therapy remained associated with longer OS and PFS (Fig. 1; Table 6). The incidence of recorded Grade ≥ 3 adverse events and lenvatinib dose modification did not differ significantly between groups, suggesting no obvious increase in documented severe toxicity or treatment interruption (Table 5). Study limitations and future research directions Several limitations of this study must be acknowledged. The retrospective design introduces inherent selection bias and unmeasured confounding factors, as treatment allocation was not randomized, despite balanced baseline characteristics. Additionally, the study utilized five different PD-1 inhibitors, each with potential subtle differences in efficacy or toxicity profiles that cannot be fully accounted for (Table 1). The follow-up duration was limited, which restricted the assessment of long-term durability of response and late toxicities. Furthermore, the absence of tissue-based or circulating biomarker analyses limits the ability to provide deeper mechanistic insights or identify predictive biomarkers for response, although NLR/PLR data offer some systemic inflammatory clues. The single-center experience also raises concerns about the generalizability of the findings. To address these limitations, future research should prioritize prospective randomized controlled trials (RCTs) comparing TACE plus lenvatinib plus PD-1 inhibition against current standards such as atezolizumab/bevacizumab specifically in patients with unresectable advanced HCC. Biomarker-driven studies are essential to identify patients most likely to benefit from this intensive combination therapy and to understand resistance mechanisms. Additionally, optimizing the timing, technique (e.g., drug-eluting beads), and frequency of TACE in combination with systemic therapies may help optimize treatment effects and reduce liver injury. Exploring sequential or adaptive treatment strategies based on early response assessments, such as using AFP dynamics or imaging, could further enhance therapeutic outcomes and personalization of care. These steps may help clarify the role of this combination strategy in advanced HCC. The shorter observed OS/PFS duration compared with prospective trials may be related to the real-world nature of the cohort, inclusion of advanced disease features, treatment heterogeneity, differences in TACE indications and treatment exposure, and differences in endpoint reporting. Although IPTW reduced measured baseline imbalance, residual confounding from unmeasured factors cannot be excluded. In addition, five different PD-1 inhibitors were used, and the sample size was insufficient for agent-specific efficacy or safety comparisons. Although treatment exposure variables, including TACE session category, lenvatinib dose modification, and PD-1 cycles, were supplemented, detailed relative dose intensity was not fully assessed. In addition, subsequent anticancer treatment after progression was not predefined as an exposure variable and was not uniformly captured in this retrospective dataset. Because post-progression treatment may influence OS, the OS results should be interpreted cautiously. Conclusion This exploratory retrospective cohort study showed that adding PD-1 inhibitors to TACE plus lenvatinib was associated with improved tumor response and longer OS/PFS after IPTW adjustment in patients with unresectable advanced HCC (Table 4; Fig. 1, and Table 6). No significant increase in recorded Grade ≥ 3 adverse events was observed (Table 5). Prospective validation is required.