Drug-eluting bead TACE improves the efficacy of HAIC combined with molecular-targeted therapy and anti-PD-1 immunotherapy in large hepatocellular carcinoma with inferior vena cava and/or right atrium tumor thrombus: a multicenter retrospective cohort study

Abstract

Background

To evaluate the safety and efficacy of combining hepatic arterial infusion chemotherapy (HAIC), molecular-targeted therapy (MTT), and programmed cell death 1 inhibitors (PD-1i) (HAIC-MTT-PD-1i) with or without drug-eluting bead transarterial chemoembolization (dTACE) (dTACE-HAIC-MTT-PD-1i) in large hepatocellular carcinoma (HCC) with tumor thrombus (TT) in the inferior vena cava (IVC) and/or right atrium (RA) (HCC-IVC/RA TT).

Methods

We conducted a multicenter study of patients with large HCC-IVC/RA TT undergoing HAIC-MTT-PD-1i (HMP group) or dTACE-HAIC-MTT-PD-1i (THMP group) between January 01, 2021 and December 31, 2024. Local tumor responses, survival outcomes, and treatment-related adverse events (TRAEs) were analyzed between the two groups.

Results

A total of 113 patients were enrolled, with 62 in the HMP group and 51 in the THMP group. The rates of overall objective response (76.5% vs. 50.0%, P = 0.004) and overall disease control (90.2% vs. 69.4%, P = 0.007) were significantly higher in the THMP group compared to the HMP group. The THMP group achieved a significantly superior median progression-free survival of 11.5 months compared to 8.0 months for the HMP group (P = 0.0050) and a median overall survival of 20.0 months versus 17.0 months (P = 0.0038).

TRAEs of any grade and grade 3–4 were similar between the two groups (P > 0.05). No grade 5 TRAEs or treatment-related death were observed in either group.

Conclusions

dTACE-HAIC-MTT-PD-1i significantly improved efficacy compared to HAIC-MTT-PD-1i, while maintaining an acceptable safety profile for patients with large HCC-IVC/RA TT. Further RCTs are required to validate the findings.

Introduction

Hepatocellular carcinoma (HCC) is one of the most common malignant tumors, which ranks as the sixth most prevalent cancer and the second leading cause of cancer-related mortality worldwide [1]. Given the high malignancy nature of HCC, approximately 14.3% to 39.6% of patients are diagnosed with portal vein tumor thrombus (PVTT) at the time of initial diagnosis [2]. Although the incidence of tumor extension beyond hepatic vein (HV) to inferior vena cava (IVC) and/or right atrium (RA) is less frequent than PVTT, IVC tumor thrombus (IVC TT) and/or RA tumor thrombus (RA TT) represent a significant risk factor for fatal pulmonary embolism and right heart failure, profoundly impacting the prognosis of patients with HCC [3, 4].

According to the available guidelines, HCC with macro-vascular invasion is classified as advanced-stage, systemic therapy is recommended as the standard treatment option, including immune checkpoint inhibitor (ICI) plus molecular-targeted therapy (MTT), dual ICIs, or monotherapy of MTT or ICI [5–7]. However, the clinical benefit of systemic therapy remains to be highly heterogeneous and unsatisfactory. Moreover, the phase 3 trials of IMbrave 150 [8], HIMALAYA [9], CheckMate 9WD [10], and CARES-310 [11], referenced by several international guidelines, provide limited efficacy data concerning HCC with IVC TT and/or RA TT (HCC-IVC/RA TT). Therefore, there is a lack of standardized treatment protocols for the treatment of HCC-IVC/RA TT.

Currently, due to the grim outcomes of a single treatment in advanced HCC, some combination strategies have been explored for HCC-IVC/RA TT including surgery plus radiotherapy, radiotherapy plus chemotherapy, transarterial chemoembolization (TACE) plus MTT/ICI, and hepatic arterial infusion chemotherapy (HAIC) plus MTT/ICI [12–14].

HAIC is a prominent therapeutic modality for advanced HCC [15, 16]. Compared to systemic chemotherapy, HAIC offers several advantages including targeted delivery of chemotherapeutics into the hepatic artery, elevated concentrations within the tumor microenvironment, prolonged interaction between the chemotherapeutics and tumor cells, and limited systemic toxicity of chemotherapy [15–17]. Randomized controlled trials (RCTs) have assessed the feasibility of oxaliplatin, fluorouracil, and leucovorin (FOLFOX)-based HAIC, demonstrating notable survival benefits for advanced HCC [15, 16]. Zhong et al. reported that combing MTT and ICI with HAIC was associated with superior progression-free survival (PFS) (9.5 vs. 4.4 months) and overall survival (OS) (18.0 vs. 7.5 months) over MTT-ICI in patients with HCC-IVC/RA TT [18]. Similarly, Lou et al. reported that the combination treatment of HAIC-lenvatinib-programmed cell death 1 inhibitors (PD-1i) showed a marked improvement in PFS (13.8 vs. 5.1 months) and OS (22.2 vs. 14.4 months) over the lenvatinib-PD-1i regimen for HCC-IVC/RA TT [19].

Drug-eluting bead TACE (dTACE) has demonstrated superior local tumor responses compared to conventional TACE (cTACE) in the management of unresectable HCC [5, 6]. In contrast to cTACE, which employs a Lipiodol-based emulsion, dTACE utilizes drug-eluting beads that facilitate sustained, high-concentration drug release and achieve more uniform embolization [5, 6]. This approach is theoretically beneficial for tumors exhibiting potential arteriovenous shunting, a phenomenon frequently observed in large HCC with IVC/RA TT, as it may enhance the retention of chemotherapeutic agents within the tumor microenvironment.

Previous studies define large HCC as tumor lesions with a maximum diameter exceeding 5 cm, which correlates with a significantly poorer prognosis [20–22]. Furthermore, recent research indicates that a tumor diameter greater than 7 cm further adversely impacts the prognosis of HCC [21, 22]. Nevertheless, few studies concern the safety and efficacy of the combination treatment of dTACE with HAIC, MTT, and PD-1i (dTACE-HAIC-MTT-PD-1i) in treating large HCC-IVC/RA TT. To address this gap, we postulated that dTACE could potentially improve the efficacy of HAIC plus MTT and PD-1i (HAIC-MTT-PD-1i), and developed this multicenter retrospective study to compare dTACE-HAIC-MTT-PD-1i with HAIC-MTT-PD-1i for large HCC-IVC/RA TT. This study aims to propose a promising combined therapeutic approach for large HCC-IVC/RA TT.

Methods and materials

Study design and patients

This study complied with the ethical standards outlined by the Declaration of Helsinki, and obtained ethical approval (SYSKY-2025-1049-01) from the Sun Yat-sen Memorial Hospital, Sun Yat-sen University. The medical records of patients were retrospectively reviewed, and the requirement for written informed consent for this retrospective analysis was waived by the ethics committee due to the retrospective nature of the study, and the use of de-identified data. The American Association for the Study of Liver Diseases and the European Association for the Study of the Liver criteria were used for the diagnosis of HCC. The contrast-enhanced computed tomography (CT) / magnetic resonance imaging (MRI) signs were employed to diagnose HCC-IVC/RA TT, including HCC with the tumor tissue extension from HV to IVC/RA and presence of blood flow deficiency within IVC/RA. Based on the correlation between the anatomical location of IVC TT and prognosis, the classification of IVC TT was defined into three categories: (1) subdiaphragmatic type (type I): IVC TT located below the diaphragm; (2) supradiaphragmatic type (type II): IVC TT located above the diaphragm but outside the atrium; and (3) RA type (type III): IVC TT extended into the RA [23, 24]. Between January 01, 2021 and December 31, 2024, patients with large HCC-IVC/RA TT undergoing HAIC-MTT-PD-1i (HMP group) or dTACE-HAIC-MTT-PD-1i (THMP group) were enrolled from four tertiary hospitals retrospectively.

The inclusion criteria for eligible participants were: (a) aged 18–80 years, (b) Eastern Cooperative Oncology Group (ECOG) performance status score < 2, (c) maximum tumor diameter ≥ 7 cm, (d) IVC/RA TT diagnosed by contrast-enhanced CT / MRI before study treatment, (e) unresectable HCC (uHCC) confirmed by a specialized multidisciplinary team, and (f) liver function of Child-Pugh class A or B. The exclusion criteria were: (a) with other malignancies, (b) prior anti-tumor therapy, (c) history of drug allergy, (d) autoimmune diseases, (e) digestive hemorrhage of any grade, and (f) insufficient renal, bone marrow, or coagulation function. The patient flowchart is depicted in Fig. 1.

Fig. 1.

Flowchart of the enrolled patients. Abbreviations: HCC, hepatocellular carcinoma; TACE, transarterial chemoembolization; HAIC, hepatic arterial infusion chemotherapy; MTT, molecular-targeted therapy; PD-1i, programmed cell death 1 inhibitors; dTACE, drug-eluting bead transarterial chemoembolization; IVC, inferior vena cava; RA, right atrium; TT, tumor thrombus; cm, centimeter

Treatment protocol

In this study, all TACE and HAIC procedures were performed by experienced interventional radiologists with over 10 years’ experience using digital subtraction angiography (DSA). In the THMP group, sequential dTACE-HAIC treatments were conducted. Whereas, in the HMP group, HAIC was the only interventional treatment, which was associated with the same regimen as that in the THMP group.

In dTACE procedure, following normative hepatic artery angiography and super-selective catheterization, 100–300 μm drug-eluting beads (DEBs) (CalliSpheres Beads, Hengrui Medical, China) loaded with epirubicin (EPI) (Hisun Pharmaceutical, China) were used for chemoembolization. The dose of DEBs-EPI mixture was established as 45 mg EPI / ml beads, which was injected at a flow rate of 1 ml per minute. If necessary, 300–500 μm blank microspheres (8Spheres, Hengrui Medical, China) were administrated to achieve embolization endpoint. Ideally, the embolization endpoint was defined as complete stasis in feeding vessels and/or disappearance of tumor stains, typically achieved within three dTACE sessions. After the completion of dTACE, subsequent cycles of HAIC were administered independently.

In HAIC procedure, a micro-catheter was positioned in the primary tumor-feeding artery based on the indication of hepatic artery angiography. Subsequently, a chemotherapy regimen comprising oxaliplatin, fluorouracil, and leucovorin (oxaliplatin, 85 mg/m², 2 h; leucovorin, 400 mg/m², 2 h; fluorouracil, 400 mg/m² bolus, and 2400 mg/m², 46 h) was infused. HAIC sessions were scheduled every three weeks, up to eight cycles, even if dTACE was discontinued. Each HAIC session involved femoral artery puncture, catheterization, infusion of chemotherapeutics, and catheter removal.

All eligible patients received MTT within 3 to 5 days post dTACE-HAIC or HAIC procedures, including lenvatinib (12 mg/day for patients weighing > 60 kg, 8 mg/day for patients weighing ≤ 60 kg), apatinib (250 mg/day), donafenib (200 mg, twice daily), and bevacizumab (5 mg/kg, once every 3 weeks intravenously). Subsequently, a PD-1 inhibitor (comprising a 200 mg dose of pembrolizumab, camrelizumab, sintilimab, or tislelizumab; or a 240 mg dose of toripalimab) was administered once every 3 weeks intravenously.

Adjustments or discontinuation of systemic treatments occurred in cases of intolerable adverse events or a modification of the treatment regimen due to uncontrollable progressive disease (PD).

Following completion of the study treatments, TACE was administered based on the review of medical records, after completion of the study treatments, additional TACE was performed at the discretion of the treating physicians when clinically indicated (e.g., disease progression), until resistance developed. Furthermore, alternative “second-line” treatments were contemplated, encompassing local modalities such as ablation techniques, ¹²⁵I seed implantation, or radiotherapy, as well as systemic interventions like chemotherapy, other MTT, or ICI.

Efficacy and safety evaluation

Local tumor responses, PFS, OS, and treatment-related adverse events (TRAEs) were compared between the THMP and HMP groups. The duration from the initiation of the study treatments to the final follow-up or death was defined as OS, and the duration from the first treatment cycle to the initial tumor progression, patient death, or final follow-up was defined as PFS. Radiologic responses of overall tumor, intrahepatic tumor, and IVC/RA TT were evaluated using the modified Response Evaluation Criteria in Solid Tumors (mRECIST).

Objective response (OR) comprised complete response (CR) and partial response (PR), while disease control (DC) encompassed CR, PR, and stable disease (SD). Tumor responses were assessed according to the mRECIST every 4–6 weeks during the follow-up period. The best overall response achieved at any time during the study period was reported. TRAEs were evaluated according to the Common Terminology Criteria for Adverse Events (CTCAE) version 5.0.

Follow-up

A regular follow-up of every four to six weeks was performed for the patients. Imaging examinations (CT and/or MRI) and laboratory tests (complete blood count, liver function, and tumor markers) were conducted to evaluate tumor responses during the follow-up period.

Tumor responses were assessed by a multidisciplinary team comprising radiologists and oncologists with over a decade of experience in interventional radiology and oncology. The OS was calculated as the time from the initiation of the study treatments to the last follow-up for the patients who were lost to the follow-up. Follow-up was no longer continued after December 31, 2025.

Statistical analysis

Data analysis and figure generation were conducted using SAS software version 9.4 (SAS Institute) and GraphPad Prism version 9.5 (GraphPad Inc.). Continuous variables were assessed using the Mann-Whitney U test, while categorical data were analyzed with Pearson’s χ2 or Fisher’s exact tests. The Kaplan-Meier method, along with the log-rank test, was utilized to evaluate PFS and OS. Prognostic factors affecting PFS and OS were examined through univariate and multivariate analyses employing the Cox proportional hazards regression model, with a backward stepwise selection method applied to identify independent predictors.

All analyses were two-tailed, and a significance level of P < 0.05 was established.

Results

Patient characteristics

A total of 113 eligible patients were enrolled, comprising 51 patients who underwent dTACE-HAIC-MTT-PD-1i (THMP group) and 62 patients who received HAIC-MTT-PD-1i (HMP group). In the THMP group, 43 patients (84.3%) were male, whereas in the HMP group, 49 patients (79.1%) were male (P = 0.473). In the THMP group, 16 patients (31.4%) presented with the largest tumor size exceeding 10 cm, while 13 patients (21.0%) in the HMP group exhibited similar tumor sizes (P = 0.208). Tumor counts exceeding three were observed in 19 patients (37.3%) in the THMP group and in 17 patients (27.4%) in the HMP group (P = 0.264). Within the THMP group, Type I, II, and III IVC/RATT were identified in 47.1%, 29.4%, and 23.5% of patients, respectively. In contrast, the HMP group displayed these types in 48.4%, 35.5%, and 16.1% of patients, respectively (P = 0.573). The baseline characteristics of the two groups are comparable (P > 0.05), as summarized in Table 1.

Table 1.

Baseline characteristics

Characteristics	All patients (n = 113)	dTACE-HAIC-MTT-PD-1i (THMP group, n = 51)		HAIC-MTT-PD-1i (HMP group, n = 62)		p value
Age (years)						0.202
< 60	82 (72.6)		34 (66.7)		48 (77.4)
≥ 60	31 (27.4)		17 (33.3)		14 (22.6)
Gender						0.473
Male	92 (81.4)		43 (84.3)		49 (79.1)
Female	21 (18.6)		8 (15.7)		13 (20.9)
Hepatitis						0.214
Absent	15 (13.3)		9 (17.6)		6 (9.7)
Present	98 (86.7)		42 (82.4)		56 (90.3)
ECOG PS						0.248
0 or 1	85 (75.2)		41 (80.4)		44 (71.0)
2	28 (24.8)		10 (19.6)		18 (29.0)
Child-Pugh class						0.739
A	69 (61.1)		32 (62.7)		37 (59.7)
B	44 (38.9)		19 (37.3)		25 (40.3)
ALBI Grade						0.976
1	53 (46.9)		24 (41.2)		29 (51.6)
2	60 (53.1)		27 (58.8)		33 (48.4)
Tumor size (cm)						0.208
7–10	84 (74.3)		35 (68.6)		49 (79.0)
> 10	29 (25.7)		16 (31.4)		13 (21.0)
Number of tumors						0.264
1–3	77 (68.1)		32 (62.7)		45 (72.6)
> 3	36 (31.9)		19 (37.3)		17 (27.4)
AFP (ng/mL)						0.739
< 400	44 (38.9)		19 (37.3)		25 (40.3)
≥ 400	69 (61.1)		32 (62.7)		37 (59.7)
PIVKA-II (mAU/mL)						0.137
< 1000	18 (15.9)		11 (21.6)		7 (11.3)
≥ 1000	95 (84.1)		40 (78.4)		55 (88.7)
IVC/RATT						0.573
Type I	52 (46.0)		24 (47.1)		30 (48.4)
Type II	39 (34.5)		15 (29.4)		22 (35.5)
Type III	22 (19.5)		12 (23.5)		10 (16.1)
PVTT						0.114
Absent	71 (62.8)		28 (54.9)		43 (69.4)
Present	42 (37.2)		23 (45.1)		19 (30.6)
Extrahepatic spread						0.086
Absent	78 (69.1)		31 (60.8)		47 (75.8)
Present	35 (30.9)		20 (39.2)		15 (24.2)

Data are number of patients, with percentage in parentheses

Categorical variables were compared using the Chi-square test or Fisher’s exact test

Abbreviations: dTACE Drug-eluting bead transarterial chemoembolization, HAIC Hepatic arterial infusion chemotherapy, MTT Molecular-targeted therapy, PD-1i Programmed cell death 1 inhibitors, THMP dTACE-HAIC-MTT-PD-1i, HMP HAIC-MTT-PD-1i, IQR Interquartile range, SD Standard deviation, ECOG Eastern Cooperative Oncology Group, PS Performance score, ALBI Albumin-bilirubin, cm Centimeter, AFP Alpha-fetoprotein, PIVKA-II Protein induced by vitamin K absence or antagonist-II, IVC/RATT Inferior vena cava tumor thrombus and/or right atrium tumor thrombus, PVTT Portal vein tumor thrombus

Treatment compliance

All dTACE and HAIC procedures were successfully conducted. Patients in the THMP group underwent a median of two dTACE treatments (range: 1–4), four HAIC cycles (range: 2–8), and six PD-1i cycles (range: 3–54). In contrast, patients in the HMP group received a median of five HAIC cycles (range: 2–8) and eight PD-1i cycles (range: 2–40). HAIC delay and discontinuation affected 29.4% and 22.6%, and 7.8% and 6.4% of patients in the THMP and HMP groups, respectively. Among patients in the THMP group, 23.5% experienced interruptions in MTT administration, 15.7% had their doses reduced, and 9.8% discontinued the drug entirely. Similarly, among those in the HMP group, 25.8% experienced interruptions, 12.9% required dose reductions, and 8.1% discontinued MTT. Furthermore, 5.9% and 9.6% of patients discontinued PD-1 inhibitor treatment due to TRAEs in the THMP and HMP groups, respectively. Regarding the “second-line” treatments, TACE (88.2% vs. 93.5%), switched MTT (68.6% vs. 79.0%), switched ICIs (49.0% vs. 61.3%), thermal ablation (33.3% vs. 24.2%), and radiotherapy (11.8% vs. 17.7%) were the most often performed protocols following dTACE-HAIC-MTT-PD-1i and HAIC-MTT-PD-1i.

Tumor response

The mRECIST criteria were utilized to assess tumor responses. Comparative analysis of overall tumor, intrahepatic tumor, and IVC/RA TT between the two groups indicated that the rates of CR, OR, DC were 9.8%, 11.8%, and 35.3%, 76.5%, 78.5%, and 78.4%, and 90.2%, 92.2%, and 88.3%, respectively, in the THMP group. In contrast, the HMP group exhibited rates of 4.8%, 6.5%, and 19.4%, 50.0%, 50.0%, and 48.4%, and 69.4%, 74.2%, and 67.8%, respectively (overall tumor: P_OR = 0.004, P_DC = 0.007; intrahepatic tumor: P_OR = 0.002, P_DC = 0.013; IVC/RATT: P_OR = 0.001, P_DC = 0.010) (Table 2; Fig. 2). These findings indicate a significantly superior tumor response in the THMP group compared to the HMP group. Waterfall plots depicting tumor responses revealed that more than half of the patients achieved objective responses (Fig. 2). Representative example with imaging response of a patient with large HCC-IVC/RA TT after dTACE-HAIC-MTT-PD-1i was depicted in Fig. 3.

Table 2.

Comparison of tumor response

Response				dTACE-HAIC-MTT-PD-1i(THMP group, n = 51)	HAIC-MTT-PD-1i(HMP group, n = 62)		p value
Overall tumor
mRECIST
CR		5 (9.8)			3 (4.8)
PR		34 (66.7)			28 (45.2)
SD		7 (13.7)			12 (19.4)
PD		5 (9.8)			19 (30.6)
ORR		39 (76.5)			31 (50.0)	0.004
DCR		46 (90.2)			43 (69.4)	0.007
Intrahepatic tumor
CR		6 (11.8)			4 (6.5)
PR		34 (66.7)			27 (43.5)
SD		7 (13.7)			15 (24.2)
PD		4 (7.8)			16 (25.8)
ORR		40 (78.5)			31 (50.0)	0.002
DCR		47 (92.2)			46 (74.2)	0.013
IVC/RATT
CR		18 (35.3)			12 (19.4)
PR		22 (43.2)			18 (29.0)
SD		5 (9.8)			12 (19.4)
PD		6 (11.7)			20 (32.2)
ORR		40 (78.4)			30 (48.4)	0.001
DCR		45 (88.3)			42 (67.8)	0.010

Data are number of patients, with percentage in parentheses

Categorical variables were compared using the Chi-square test or Fisher’s exact test

Abbreviations: mRECIST Modified Response Evaluation Criteria in Solid Tumors, dTACE Drug-eluting bead transarterial chemoembolization, HAIC Hepatic arterial infusion chemotherapy, MTT Molecular-targeted therapy, PD-1i Programmed cell death 1 inhibitors, THMP dTACE-HAIC-MTT-PD-1i, HMP HAIC-MTT-PD-1i, IVC/RATT Inferior vena cava tumor thrombus and/or right atrium tumor thrombus, CR Complete response, PR Partial response, SD Stable disease, PD Progressive disease, ORR Objective response rate, DCR Disease control rate

Fig. 2.

Tumor responses assessment by mRECIST in all patients (n = 113) undergoing THMP and HMP in terms of overall tumor (A&D), intrahepatic tumor (B&E), and IVC/RA TT (C&F). A, B, & C: Percent change from baseline in target lesion; D, E, & F: Histogram depicting the tumor responses. Abbreviations: mRECIST, modify Response Evaluation Criteria in Solid Tumors; THMP, dTACE-HAIC-MTT-PD-1i; HMP, HAIC-MTT-PD-1i; dTACE, drug-eluting bead transarterial chemoembolization; HAIC, hepatic arterial infusion chemotherapy; MTT, molecular-targeted therapy; PD-1i, programmed cell death 1 inhibitors. IVC, inferior vena cava; RA, right atrium; TT, tumor thrombus; CR, complete response; PR, partial response; SD, stable disease; PD, progressive disease

Fig. 3.

Representative images of a patient with large HCC-IVC/RA TT after dTACE-HAIC-MTT-PD-1i. A, B, C, G, H, & I: Baseline status of the images in terms of DSA (A) and CT (B: white arrow indicates IVC TT; C: red arrow indicates RA TT; G-I: indicate intrahepatic tumor lesion); D, E, F, J, K, & L: Post-treatment images show an efficacy of CR by mRECIST in terms of DSA (D) and MRI (E&F: indicate a CR efficacy of IVC&RA TT; J-L: indicate a CR efficacy of intrahepatic tumor lesion). Abbreviations: HCC, hepatocellular carcinoma; IVC, inferior vena cava; RA, right atrium; TT, tumor thrombus; dTACE, drug-eluting bead transarterial chemoembolization; HAIC, hepatic arterial infusion chemotherapy; MTT, molecular-targeted therapy; PD-1i, programmed cell death 1 inhibitors; DSA, digital subtraction angiography; CT, computed tomography; IVC, inferior vena cava; RA, right atrium; TT, tumor thrombus; CR, complete response; mRECIST, modify Response Evaluation Criteria in Solid Tumors; MRI, magnetic resonance imaging

Survival

The median follow-up duration was 17.6 (range: 6.5–41.0) months for the THMP group and 16.0 (range: 6.5–30.5) months for the HMP group, with no participants lost to follow-up.

At the conclusion of the follow-up period, 93 patients had died. Comparative analyses revealed that the THMP group achieved a significantly superior median PFS (mPFS) of 11.5 months compared to 8.0 months for the HMP group (hazard ratio [HR]: 0.59 [95% confidence interval (CI): 0.40–0.88], P = 0.0050) and a median OS (mOS) of 20.0 months versus 17.0 months (HR: 0.57 [95% CI: 0.38–0.86], P = 0.0038) (Fig. 4). A marked improvement in both PFS and OS was observed, indicating potential benefits of THMP over HMP. Further analyses demonstrated that patients with objective responses (CR/PR) in overall tumor, IVC/RATT, and intrahepatic tumor exhibited significantly better mOS (22.0 vs. 15.8 months, HR: 0.44 [95% CI: 0.26–0.74], P < 0.0001; 22.0 vs. 12.0 months, HR: 0.21 [95% CI: 0.10–0.43], P < 0.0001; 22.0 vs. 14.0 months, HR: 0.31 [95% CI: 0.17–0.57], P < 0.001) compared to non-responders (SD/PD) (Fig. 5). Subsequent analyses indicated consistent trends of improved PFS and OS with dTACE-HAIC-MTT-PD-1i compared to HAIC-MTT-PD-1i across all subgroups (Fig. 6).

Fig. 4.

Kaplan-Meier survival curves for estimating probability of PFS (A) and OS (B) of the patients with HCC-IVC/RA TT receiving THMP and HMP. Abbreviations: PFS, progression-free survival; OS, overall survival; HR, hazard ratio; CI, confidence interval; HCC, hepatocellular carcinoma; IVC, inferior vena cava; RA, right atrium; TT, tumor thrombus; THMP, dTACE-HAIC-MTT-PD-1i; HMP, HAIC-MTT-PD-1i; dTACE, drug-eluting bead transarterial chemoembolization; HAIC, hepatic arterial infusion chemotherapy; MTT, molecular-targeted therapy; PD-1i, programmed cell death 1 inhibitors

Fig. 5.

Kaplan-Meier survival curves for estimating OS probability of the entire cohort patients with local tumor efficacy of “Response” and “Non-Response” in terms of overall tumor (A), IVC/RA TT (B), and intrahepatic tumor (C). Abbreviations: IVC, inferior vena cava; RA, right atrium; TT, tumor thrombus; HR, hazard ratio; CI, confidence interval; R, Response; N, Non-Response

Fig. 6.

Subgroup analyses of PFS and OS. Abbreviations: PFS, progression-free survival; OS, overall survival; HR, hazard ratio; 95%CI, 95% confidence interval; ECOG, Eastern Cooperative Oncology Group; PS, performance status; ALBI, albumin-bilirubin; cm, centimeter; AFP, alpha-fetoprotein; PIVKA-II, protein induced by vitamin K absence or antagonist-II; IVC, inferior vena cava; RA, right atrium; TT, tumor thrombus; PVTT, portal vein tumor thrombus; THMP, dTACE-HAIC-MTT-PD-1i; HMP, HAIC-MTT-PD-1i; dTACE, drug-eluting bead transarterial chemoembolization; HAIC, hepatic arterial infusion chemotherapy; MTT, molecular-targeted therapy; PD-1i, programmed cell death 1 inhibitors

Prognostic factors of survival

The univariate analyses identified several significant prognostic factors for OS, including ECOG score greater than 1 (P < 0.001), Child-Pugh class beyond A (P < 0.001), advanced ALBI grade (P < 0.001), tumor size larger than 10 cm (P < 0.001), tumor number exceeding 3 (P < 0.001), alpha-fetoprotein (AFP) level ≥ 400 ng/mL (P < 0.001), different IVC/RATT types (P < 0.001), absence of PVTT (P < 0.001), extrahepatic tumor lesion (P < 0.001), and dTACE treatment (P < 0.001) (Table 3). Subsequent multivariate analyses confirmed ECOG score (P = 0.002), ALBI grade (P = 0.001), tumor size (P < 0.001), tumor number (P < 0.001), AFP level (P = 0.002), IVC/RA TT (P < 0.001), PVTT (P = 0.001), extrahepatic spread (P = 0.002), and dTACE treatment (P < 0.001) as independent predictors of OS (Table 3).

Table 3.

Prognostic factors of progression-free survival and overall survival

Factors	Progression-free survival				Overall survival
	Univariate analysis		Multivariate analysis		Univariate analysis		Multivariate analysis
	HR (95%CI)	p value	HR (95%CI)	p value	HR (95%CI)	p value	HR (95%CI)	p value
Age (years)
≥60 vs. < 60	1.09 (0.47–2.18)	0.615	—	—	2.11 (0.95–3.72)	0.166	—	—
Gender
Male vs. Female	1.48 (0.79–2.61)	0.224	—	—	1.92 (0.87–2.70)	0.195	—	—
Hepatitis
Present vs. Absent	1.07 (0.63–2.06)	0.782	—	—	1.78 (0.59–3.91)	0.319	—	—
ECOG score
2 vs. 0/1	1.15 (0.61–2.49)	0.315	—	—	2.96 (0.81–6.38)	< 0.001	1.99 (0.77–5.30)	0.002
Child-Pugh class
B vs. A	1.45 (0.91–2.55)	0.231	—	—	2.75 (1.51–5.44)	< 0.001	1.12 (0.45–3.20)	0.069
ALBI grade
2 vs. 1	1.69 (0.72–3.07)	0.357	—	—	4.64 (2.05–7.27)	< 0.001	3.15 (1.46–6.01)	0.001
Tumor size (cm)
> 10 vs. 7–10	4.37 (2.05–7.91)	< 0.001	3.99(1.87–5.19)	< 0.001	5.15 (1.98–8.67)	< 0.001	2.79 (0.92–6.11)	< 0.001
Tumor number
> 3 vs. 1–3	3.06 (1.87–7.93)	< 0.001	2.21 (0.98–4.75)	0.005	4.14 (1.84–8.59)	< 0.001	3.13 (1.55–6.43)	< 0.001
AFP (ng/mL)
≥ 400 vs. <400	1.01 (0.23–1.74)	0.718	—	—	3.96 (1.55–9.81)	< 0.001	2.78 (1.97–6.03)	0.002
IVC/RATT
Type III vs. I/ II	2.81 (1.56–6.43)	< 0.001	2.11 (1.27–4.96)	0.001	3.45 (2.03–8.17)	< 0.001	3.39 (1.48–8.66)	< 0.001
PVTT
Present vs. Absent	3.41 (1.42–8.77)	< 0.001	2.97(1.55–5.01)	< 0.001	5.95(2.16–11.67)	< 0.001	3.12 (1.98–6.44)	0.001
Extrahepatic spread
Present vs. Absent	4.51 (1.35–8.41)	< 0.001	2.07 (0.92–5.38)	0.002	4.22(1.96–10.01)	< 0.001	2.15 (0.75–5.63)	0.002
dTACE treatment
Without vs. With	5.32 (2.33–9.87)	< 0.001	3.10(1.72–7.53)	< 0.001	3.60(1.12–8.07)	< 0.001	2.97 (0.98–5.99)	< 0.001

The Cox proportional hazards regression model was used for multivariable analysis

For progression-free survival, the prognostic factors of “Tumor size”, “Tumor number”, “IVC/RATT”, “PVTT”, “Extrahepatic spread”, and “dTACE treatment” were entered into the final multivariable analysis. For overall survival, the prognostic factors of “ECOG score”, “Child-Pugh class”, “ALBI grade”, “Tumor size”, “Tumor number”, “AFP “, “IVC/RATT”, “PVTT”, “Extrahepatic spread”, and “dTACE treatment” were entered into the final multivariable analysis

Abbreviations: HR Hazard ratio, 95%CI 95% Confidence interval, ECOG Eastern Cooperative Oncology Group, ALBI Albumin-bilirubin, cm Centimeter, AFP Alpha-fetoprotein, IVC/RATT Inferior vena cava tumor thrombus and/or right atrium tumor thrombus, PVTT Portal vein tumor thrombus, dTACE Drug-eluting bead transarterial chemoembolization

Furthermore, univariate analyses revealed significant associations between PFS and tumor size > 10 cm (P < 0.001), tumor number > 3 (P < 0.001), different IVC/RA TT types (P < 0.001), presence of PVTT (P < 0.001), extrahepatic tumor spread (P < 0.001), and absence of dTACE treatment (P < 0.001) (Table 3). Subsequent multivariate analyses confirmed that tumor size (P < 0.001), number of tumors (P = 0.005), IVC/RA TT (P = 0.001), PVTT (P < 0.001), extrahepatic tumor lesion (P = 0.002), and dTACE treatment (P < 0.001) (Table 3).

Safety

All enrolled patients in both the THMP and HMP groups experienced TRAEs of any grade. In the THMP group, 28 (54.9%) patients experienced grade 3–4 TRAEs, which primarily included elevated aminotransferase (27.5%), hypertension (15.7%), abdominal pain (9.8%), elevated bilirubin (7.8%), hand-foot syndrome (5.9%), gastrointestinal hemorrhage (5.9%), and immune-related adverse events (5.9%). In the HMP group, 25 (40.3%) patients experienced grade 3–4 TRAEs, with the most common being elevated aminotransferase (14.5%), hypertension (11.3%), immune-related adverse events (9.6%), hand-foot syndrome (4.8%), elevated bilirubin (3.3%), and gastrointestinal hemorrhage (3.2%). All TRAEs were well-tolerated and managed, and TRAEs of any grade and grade 3–4 were similar between the two groups (P > 0.05). Notably, no grade 5 TRAEs or treatment-related death were observed in either group. Detailed information on all TRAEs is presented in Table 4.

Table 4.

Treatment-related adverse events

Adverse event	Any grade			Grade 3–4
	dTACE-HAIC-MTT-PD-1i (n = 51)	HAIC-MTT-PD-1i (n = 62)	p value	dTACE-HAIC-MTT-PD-1i (n = 51)	HAIC-MTT-PD-1i (n = 62)	p value
Total	51 (100)	62 (100)	1.000	28 (54.9)	25 (40.3)	0.122
Decreased appetite	38 (74.4)	39 (62.9)	0.188	0 (0)	0 (0)	—
Nausea	46 (90.2)	52 (83.9)	0.324	0 (0)	0 (0)	—
Fatigue	41 (80.4)	46 (74.2)	0.436	0 (0)	0 (0)	—
Vomiting	31 (60.8)	32 (51.6)	0.329	0 (0)	0 (0)	/
Abdominal pain	47 (92.2)	40 (64.5)	< 0.001	5 (9.8)	0 (0)	0.012
Fever	33 (64.7)	32 (51.6)	0.082	0 (0)	0 (0)	—
Leukopenia/thrombopenia	43 (84.3)	49 (79.0)	0.473	0 (0)	0 (0)	—
Hand-foot syndrome	11 (21.6)	12 (19.4)	0.771	3 (5.9)	3 (4.8)	0.806
Proteinuria	5 (9.8)	7 (11.3)	0.799	0 (0)	0 (0)	—
Rash	4 (7.8)	3 (4.8)	0.510	0 (0)	0 (0)	—
Diarrhea	9 (17.6)	7 (11.3)	0.335	2 (3.9)	1 (1.6)	0.447
Hypertension	30 (58.8)	29 (46.8)	0.202	8 (15.7)	7 (11.3)	0.493
Elevated bilirubin	17 (33.3)	12 (19.4)	0.090	4 (7.8)	2 (3.3)	0.276
Elevated ALT/AST	42 (82.4)	40 (64.5)	0.034	14 (27.5)	9 (14.5)	0.089
Decreased albumin	18 (35.3)	17 (27.4)	0.368	0 (0)	0 (0)	—
Ascites	14 (27.5)	11 (17.7)	0.216	0 (0)	0 (0)	—
Gastrointestinal hemorrhage	3 (5.9)	2 (3.2)	0.494	3 (5.9)	2 (3.2)	0.494
Liver abscess	2 (3.9)	0 (0)	0.116	2 (3.9)	0 (0)	0.116
Immune-related adverse events	11 (21.6)	15 (24.2)	0.741	3 (5.9)	6 (9.6)	0.458
Immune-related hepatitis	1 (1.9)	2 (3.2)	0.677	1 (1.9)	2 (3.2)	0.677
Immune-related pneumonitis	1 (1.9)	1 (1.6)	0.889	1 (1.9)	1 (1.6)	0.889
Immune-related dermatitis	2 (3.9)	1 (1.6)	0.447	0 (0)	0 (0)	—
Immune-related myocarditis	2 (3.9)	3 (4.8)	0.813	0 (0)	0 (0)	—
Immune-related hypothyroidism	5 (9.8)	8 (12.9)	0.609	1 (1.9)	3 (4.8)	0.410

Data are number of patients, with percentage in parentheses

Abbreviations: dTACE Drug-eluting bead transarterial chemoembolization, HAIC Hepatic arterial infusion chemotherapy, MTT Molecular-targeted therapy, PD-1i Programmed cell death 1 inhibitors, ALT Alanine aminotransferase, AST Aspartate aminotransferase

Discussion

HCC is associated with a high risk of vascular invasion. Although the formation of a tumor thrombus (TT) in the IVC or RA is infrequent, it is classified as an advanced stage due to vascular invasion and is associated with an extremely poor prognosis. Currently, there is a lack of global guidelines or consensus regarding the treatment of HCC with IVC/RA TT. In addition, RCTs performing systemic treatment for advanced HCC have excluded the cases with IVC/RA TT [8, 11, 25]. Consequently, in clinical practice, we adopt treatment strategies designed for advanced HCC to manage cases with IVC/RA TT. The combination of MTT and ICIs is recommended as first-line treatments for advanced HCC according to several international guidelines, such as the Barcelona Clinic Liver Cancer (BCLC) guideline [5] and the National Comprehensive Cancer Network (NCCN) guideline [7]. Furthermore, combination therapy involving HAIC, MTT, and PD-1i has shown significantly superior efficacy compared to monotherapy with HAIC or MTT-PD-1i in multiple studies [18, 26, 27].

However, due to the limitations caused by drug resistance and repeated cycles of MTT-PD-1i and/ or HAIC, the prognosis for HCC remains to be improved.

In this study, all enrolled patients exhibited a high burden of HCC, characterized by a large intrahepatic tumor (≥ 7 cm) in 100% of cases, IVC/RA TT in 100%, PVTT in 37.2%, and extrahepatic spread in 30.9%. These characteristics underscore the challenges and necessity of combination therapy. Consequently, we administered d-TACE-HAIC-MTT-PD-1i and HAIC-MTT-PD-1i for the treatment of large HCC-IVC/RA TT. Our results demonstrated that the tumor response rates for overall tumor, intrahepatic tumor, and IVC/RA TT following d-TACE-HAIC-MTT-PD-1i were significantly higher than those observed after HAIC-MTT-PD-1i. Survival analysis indicated that d-TACE-HAIC-MTT-PD-1i yielded improved PFS and OS compared to HAIC-MTT-PD-1i. Furthermore, patients who achieved a tumor response of CR or PR exhibited markedly superior OS compared to those with SD or PD, suggesting that the combination therapy of d-TACE-HAIC-MTT-PD-1i may improve the prognosis of HCC by enhancing local tumor responses.

Systemic therapy is recommended as the preferred first-line treatment for advanced HCC according to Western guidelines [5–7]; however, the treatment remains inadequate due to its limited effectiveness. In contrast, Eastern guidelines [28–30] favor a combination of systemic therapy and local treatments. RCTs have confirmed the safety and efficacy of HAIC for advanced HCC [15, 16]. Moreover, several studies have demonstrated that HAIC combined with systemic therapy, such as MTT or MTT-ICIs, exhibits superior efficacy compared to HAIC or systemic therapy alone in advanced HCC [18, 26, 27]. Regarding the management of HCC-IVC/RA TT, Lou et al. reported that the combination of HAIC and lenvatinib with PD-1i was associated with significantly improved mOS compared to lenvatinib with PD-1i alone (22.2 vs. 14.4 months) [19]. Similarly, Zhong et al. compared the efficacy of HAIC-MTT-ICIs to that of MTT-ICIs or HAIC in treating HCC-IVC/RA TT, revealing that HAIC-MTT-ICIs achieved superior mOS compared to both MTT-ICIs (18.0 vs. 7.5 months) and HAIC (18.5 vs. 7.1 months) [18]. In this study, we administered HAIC-MTT-PD-1i for large HCC-IVC/RA TT, achieving a mOS of 17.0 months, which aligns with previous findings.

In East Asia, on basis of systemic therapy, TACE has been widely performed for advanced HCC to improve local tumor response, even for those cases with macrovascular invasion or extrahepatic spread [31–33]. Several studies have investigated the combination of TACE with MTT-PD-1i in the treatment of advanced HCC, demonstrating significantly improved efficacy compared to either TACE or MTT-PD-1i alone [34–36]. Chung et al. reported that TACE monotherapy for advanced HCC with IVC/RA TT resulted in a mOS of 4.2 to 10.9 months, which is considered an unsatisfactory outcome [37]. In contrast, Ning et al. found that TACE combined with lenvatinib and sintilimab for HCC-IVC/RA TT achieved mPFS and mOS of 13.0 months and 17.3 months, respectively [38]. Additionally, Huang et al. indicated that TACE in combination with lenvatinib and camrelizumab resulted in an mOS of 8.92 months for HCC patients with IVC/RA TT [39]. Collectively, these studies affirm the safety and efficacy of the TACE-MTT-PD-1i combination for advanced HCC with IVC/RA TT.

HAIC-MTT-PD-1i demonstrates reliable efficacy in treating HCC-IVC/RA TT [18, 19].

However, a subset of highly heterogeneous HCCs may develop resistance to this triple combination strategy. Moreover, in real-world practice, the HAIC procedure is typically limited to six to eight sessions during a single treatment period to mitigate the risk of serious TRAEs [40–42]. This limitation may lead to inadequate tumor response in patients with a high burden of HCC. Previous studies indicate that TACE-HAIC combined with MTT-PD-1i offers superior efficacy compared to HAIC-MTT-PD-1i for unresectable HCC [43, 44]. Super-selective TACE has shown efficacy in enhancing local tumor response by inducing devascularization in high-burden HCC, particularly in cases exhibiting potential drug resistance to HAIC. Previous studies have confirmed that dTACE is associated with superior local tumor response over conventional (Lipiodol)-TACE [5, 6]. Furthermore, dTACE can effectively embolize arteriovenous shunting, ensuring the delivery of an adequate dose of chemotherapeutics to the HCC microenvironment following HAIC. Consequently, we implemented dTACE in addition to HAIC-MTT-PD-1i for large HCC-IVC/RA TT to address these limitations. Our findings revealed that dTACE-HAIC-MTT-PD-1i significantly improved both mPFS (11.5 vs. 8.0 months) and mOS (20.0 vs. 17.0 months) compared to HAIC-MTT-PD-1i. Nonetheless, few studies have reported on the safety and efficacy of dTACE-HAIC-MTT-PD-1i for large HCC-IVC/RA TT. Furthermore, to our knowledge, no published studies have specifically evaluated the combination of cTACE, HAIC, MTT, and PD-1i in this context. The existing literature primarily comprises studies utilizing standard TACE without specifying the type of beads, which may encompass both cTACE and dTACE. Our study offers initial evidence pertinent to the dTACE platform.

In addition to standard treatments for advanced HCC, alternative therapies for HCC-IVC/RA TT have been reported, primarily encompassing surgery and radiotherapy [39, 45–47]. Lee et al. summarized the efficacy of surgical intervention for HCC-IVC/RA TT patients exhibiting optimal liver function, adequate future liver remnant volume, and intrahepatic tumors, reporting a mOS ranging from 13.8 to 19.0 months [45]. However, only a limited percentage of patients are deemed suitable for surgery in clinical practice. Furthermore, surgical options may be constrained by the complexity of the techniques involved and the elevated risk of postoperative complications. In contrast, HAIC-MTT-PD-1i, as a minimally invasive treatment, demonstrates significantly superior safety and non-inferior efficacy, with a mOS ranging from 8.92 to 22.2 months for patients with HCC-IVC/RA TT. Our findings indicate that the addition of dTACE to HAIC-MTT-PD-1i significantly enhances local tumor response, PFS, and OS in patients with high-burden HCC-IVC/RA TT. Consequently, for patients who are not candidates for surgery, systemic therapy or a combination of systemic therapy and local treatment remains the most viable option. In terms of radiotherapy, Huang et al. reported that the combination of radiotherapy with MTT and ICIs resulted in a mOS of 16.98 months for 60 patients with HCC-IVC/RA TT in a retrospective study [39].

Additionally, Qian et al. conducted a case report that explored the efficacy of radiotherapy combined with ICI-MTT for HCC-IVC/RA TT, achieving a PFS exceeding six months [48]. Although radiotherapy may represent a promising treatment option for advanced HCC, the limited number of available studies necessitates further RCTs to validate these findings.

In terms of safety and compliance, all TRAEs were well-tolerated and effectively managed following dTACE-HAIC-MTT-PD-1i or HAIC-MTT-PD-1i. Throughout the follow-up period, no unexpected toxicities were identified in either treatment group, aligning with the established TRAEs associated with each combination strategy. Although instances of treatment interruption, dose reduction, and discontinuation were noted, no grade 5 TRAEs or treatment-related mortality occurred. Consistent with previous studies, the most prevalent TRAEs among all study participants included elevated aminotransferase levels, bilirubin, hypertension, hand-foot syndrome, immune-related adverse events, and abdominal pain, all of which were manageable. Importantly, the incidence of TRAEs of any grade and those classified as grade 3–4 did not differ significantly between the two groups.

To the best of our knowledge, this study represents the first investigation of dTACE-HAIC-MTT-PD-1i in the management of large HCC-IVC/RA TT. Several limitations are present in the current study. First, as a retrospective study, it is inherently subject to selection bias despite baseline comparability. We were unable to adjust for unmeasured confounders, including socioeconomic status, treatment adherence, and institutional differences in interventional techniques. Second, although the MTT and PD-1i used were selected according to international guidelines, the variability in MTT and PD-1i may have influenced the consistency of targeted and immunological therapies. However, few studies have provided sufficient evidence regarding the heterogeneity of efficacy among the recommended first-line MTT and PD-1i. Third, while this study confirmed the safety and efficacy of dTACE-HAIC-MTT-PD-1i for large HCC-IVC/RA TT, further validation is necessary across different types of HCC-IVC/RA TT. Fourth, a significant majority of the study participants (86.7%) were diagnosed with hepatitis B-related HCC, highlighting the need for additional research on other etiologies to validate these findings. Finally, the potential bias in subsequent treatments between the two treatment groups, which may affect the overall survival curve, was not fully controlled. Future RCTs should standardize post-progression management.

In conclusion, our data indicated that dTACE-HAIC-MTT-PD-1i significantly improved efficacy compared to HAIC-MTT-PD-1i, while preserving an acceptable safety profile for patients with large HCC-IVC/RA TT. Consequently, the incorporation of dTACE into HAIC-MTT-PD-1i may represent a reasonable and effective treatment option for carefully selected patients in experienced centers; however, confirmatory prospective randomized controlled trials are essential.

Abbreviations

HCC

Hepatocellular carcinoma

PVTT

Portal vein tumor thrombus

HV

Hepatic vein

IVC

Inferior vena cava

RA

Right atrium

TT

Tumor thrombus

IVC TT

IVC tumor thrombus

RA TT

RA tumor thrombus

ICI

Immune checkpoint inhibitor

MTT

Molecular targeted therapy

TACE

Transarterial chemoembolization

HAIC

Hepatic arterial infusion chemotherapy

dTACE

Drug-eluting bead TACE

cTACE

Conventional TACE

THMP

dTACE-HAIC-MTT-PD-1i

HMP

HAIC-MTT-PD-1i

RCT

Randomized controlled trial

PD-1

Programmed cell death 1

PD-1i

PD-1i inhibitors

PFS

Progression-free survival

OS

Overall survival

CT

Computed tomography

MRI

Magnetic resonance imaging

ECOG

Eastern Cooperative Oncology Group

cm

Centimeter

uHCC

Unresectable HCC

BCLC

Barcelona Clinic Liver Cancer

DSA

Digital subtraction angiography

EPI

Epirubicin

TRAEs

Treatment-related adverse events

mRECIST

Modified Response Evaluation Criteria in Solid Tumors

OR

Objective response

CR

Complete response

PR

Partial response

DC

Disease control

SD

Stable disease

PD

Progressive disease

ORR

Objective response rate

DCR

Disease control rate

CTCAE

Common Terminology Criteria for Adverse Events

IQR

Interquartile range

SD

Standard deviation

PS

Performance score

ALBI

Albumin-bilirubin

AFP

Alpha-fetoprotein

PIVKA-II

Protein induced by vitamin K absence or antagonist-II

HR

Hazard ratio

CI

Confidence interval

ALT

Alanine aminotransferase

AST

Aspartate aminotransferase

R

Response

N

Non-Response

Authors' contributions

Jiayan Ni: Conceptualization, Data curation, Formal analysis, Funding acquisition, Methodology, Investigation, Visualization, Writing – original draft, Writing – review and editing. Hongliang Sun: Conceptualization, Data curation, Software, Resources. Chengzhi Li: Data curation, Methodology, Formal analysis. Yulong Liu: Data curation, Methodology, Formal analysis. Gefan Guo: Data curation, Resources. Xiong Zhou: Formal analysis, Methodology. Zhiyong Zhou: Data curation, Formal analysis. Rui Fan: Data curation, Formal analysis. Hao Hu: Data curation, Formal analysis. Zhuting Fang: Conceptualization, Validation, Investigation, Visualization, Supervision, Writing – review and editing. Linfeng Xu: Project administration, Resources, Supervision, Validation, Visualization. Writing – review and editing.

Funding

This study was supported by the National Natural Science Foundation of China (Grant No.81901854), and the Natural Science Foundation of Guangdong Province (Grant No.2025A1515010664).

Data availability

Data generated or analyzed during the study are available from the corresponding author by request.

Declarations

Ethics approval and consent to participate

The study protocol conformed to the ethical guidelines of the 1975 Declaration of Helsinki and the Ethics Committee of the Sun Yat-sen Memorial Hospital, Sun Yat-sen University (Approval No. SYSKY-2025-1049-01). Written informed consent was obtained before treatment. The requirement for written informed consent for this retrospective analysis was waived by the ethics committee due to the retrospective nature of the study.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.