Survival Outcomes in Patients with Advanced Hepatocellular Carcinoma Based on Pretreatment Serum Lactate Dehydrogenase Level and Albumin-Bilirubin Grade: A Single-Center, Retrospective, Non-Randomized Cohort Study

Abstract

Background

Hepatocellular carcinoma (HCC) is a major global contributor to cancer mortality. The primary objective of this study was to compare the prognostic value of pretreatment albumin-bilirubin (ALBI) grade and serum lactate dehydrogenase (LDH) levels in patients with advanced hepatocellular carcinoma receiving first-line triple therapy consisting of immune checkpoint inhibitors (ICI), targeted agents, and radiotherapy.

Methods

This single-center retrospective cohort study enrolled 143 unresectable advanced HCC patients treated with ICI-targeted combined radiotherapy from 2019 to 2023. Patients were stratified by baseline ALBI grade (1 vs 2) and LDH levels (LDH-0: ≤ 250 U/L vs LDH-1: > 250 U/L). Treatment response was evaluated per mRECIST 1.1, and adverse events (AEs) were graded by NCI-CTCAE v4.0. The primary endpoint was overall survival (OS); secondary endpoints included progression-free survival (PFS), objective response rate (ORR), disease control rate (DCR) and safety profiles. Receiver operating characteristic (ROC) curve analysis was performed to evaluate the standalone predictive performance of ALBI grade and LDH status for OS and PFS endpoints.

Results

ALBI-1 and LDH-0 patients obtained significantly higher ORR and DCR than ALBI-2 and LDH-1 patients (P < 0.05 and P < 0.001, respectively). No efficacy gaps existed between favorable-risk subgroups (ALBI-1 vs LDH-0) or poor-risk subgroups (ALBI-2 vs LDH-1). The cohort’s median OS and PFS were 15.7 and 12.8 months. Univariate analysis showed ALBI-1 was associated with lowered risks of death (Hazard ratio (HR) = 0.27, P < 0.001) and progression (HR = 0.38, P < 0.001), although this association was not statistically significant in multivariate Cox models, while LDH-0 reduced mortality (HR = 0.33, P < 0.001) and progression risks (HR = 0.53, P = 0.007). ALBI-1 and LDH-0 had comparable survival, and LDH-1 patients showed numerically better survival than ALBI-2 patients. ROC curves showed weak predictive ability for single ALBI or LDH (all areas under the curves near 0.5). ALBI-2 patients suffered more grade 1–4 AEs, whereas LDH subgroups showed no consistent toxicity differences.

Conclusions

Pretreatment ALBI grade and LDH showed comparable prognostic stratification performance for advanced HCC receiving triple therapy, while ALBI grade better predicts treatment-related toxicities. Single markers had limited predictive efficacy, and combined ALBI-LDH scoring may improve clinical risk stratification, although its added clinical value requires prospective external validation.

Graphical Abstract

Introduction

Primary liver cancer imposes a substantial disease burden in China, with hepatocellular carcinoma (HCC) accounting for 75% − 85% of all cases.¹ Approximately 70% − 80% of patients are diagnosed at advanced disease stages.² Accurate prognostic stratification guides individualized therapeutic decision-making and quality-of-life optimization. The Barcelona Clinic Liver Cancer (BCLC) staging system serves as the global standard for HCC risk classification. Among untreated HCC patients, median overall survival (OS) differs drastically across BCLC stages: 41.3 months for stage 0/A, 22.1 months for stage B, 8.9 months for stage C, and 2.6 months for stage D, with the proportion of patients surviving less than 12 months reaching 28.5%, 52.1%, 84.3%, and 96.9%, respectively.³

Hepatic functional reserve strongly modulates clinical outcomes in advanced HCC. The Child-Pugh scoring system and albumin-bilirubin (ALBI) grade are two widely adopted liver function assessment tools.⁴ In sorafenib-treated HCC cohorts, median OS was markedly longer in Child-Pugh A patients vs Child-Pugh B patients (11.1 vs 3.9 months, P < 0.001); stratified by ALBI grade, median OS declined sequentially across grades 1 to 3 (12.5, 8.4, and 3.8 months, P < 0.001).⁵ For first-line atezolizumab plus bevacizumab combination therapy, median OS was 21.6, 6.4, and 2.5 months, and median progression-free survival (PFS) was 8.9, 4.8, and 2.5 months for Child-Pugh A, B, and C patients. When stratified by ALBI grade, median OS reached 34.9, 11.7, and 4.7 months, and median PFS reached 11.0, 6.4, and 3.7 months for grades 1, 2, and 3, respectively.⁶

The ALBI grade outperforms Child-Pugh score for survival prediction, largely owing to subjective Child-Pugh variables (ascites and hepatic encephalopathy) that compromise scoring reproducibility.⁷ As an objective, simplified quantitative index, ALBI enables reliable risk stratification across advanced HCC populations.⁸

Serum lactate dehydrogenase (LDH) levels correlate tightly with HCC clinical prognosis. LDH is a core enzyme mediating anaerobic glycolysis; its overexpression drives lactate accumulation within the tumor microenvironment, further promoting malignant proliferation, angiogenesis, and local immune suppression.^9,10 In resectable HCC cohorts, low pretreatment LDH predicts superior 5-year OS (73.7% vs 43.1%) and recurrence-free survival (RFS) (54.0% vs 20.4%).¹¹ Following liver transplantation, elevated LDH (≥ 213.5 U/L) independently predicts shortened RFS.¹² Post-transarterial chemoembolization (TACE), serum LDH and tumor PD-L1 expression increase transiently on days 3–7 and normalize by day 30, indicating a positive correlation between LDH and intratumoral immune evasion.¹³

ALBI grade (an objective liver function biomarker) and serum LDH (a surrogate marker of tumor burden and immune suppression) are both well-validated prognostic indicators for HCC.^3,4 Nevertheless, head-to-head comparative analyses evaluating their predictive performance in patients receiving ICI-targeted combined radiotherapy remain scarce.

Recently, triple-modality regimens integrating immune checkpoint inhibitors, molecular-targeted agents, and radiotherapy have emerged as a promising treatment paradigm for unresectable advanced HCC. Multiple retrospective real-world cohorts have demonstrated encouraging anti-tumor efficacy and acceptable tolerability of such combined strategies; however, considerable inter-patient heterogeneity in survival outcomes and treatment-related toxicity persists. Reliable baseline prognostic biomarkers to support risk-stratified patient selection for this intensive combined approach remain under-investigated.^14,15

This study systematically compared the prognostic efficacy of pretreatment ALBI grade and serum LDH for therapeutic response, long-term survival, and treatment-related toxicity in advanced HCC patients treated with ICI-targeted combined radiotherapy. We hypothesized that serum LDH would demonstrate non-inferior survival predictive value relative to ALBI grade, with potential complementary stratification benefits when combined.

Materials and Methods

Study Design

This single-center retrospective cohort study analyzed per-protocol patient data. The trial protocol, data collection scheme and written informed consent documents were all reviewed and approved by the Ethics Committee and Institutional Review Board of the Affiliated Hospital of Southwest Medical University (approval ID: 16/10,2024 / KY2024389). As only de-identified retrospective medical records lacking identifiable personal data were analyzed, written informed consent from each patient was waived by the Ethics Committee following the Declaration of Helsinki. Eligible participants presented with unresectable BCLC stage B/C advanced HCC and received first-line ICI plus targeted therapy combined with radiotherapy. An independent imaging review committee assessed treatment response per modified Response Evaluation Criteria in Solid Tumors (mRECIST). All adverse events (AEs) were graded per National Cancer Institute Common Terminology Criteria for Adverse Events, version 4.0 (NCI-CTCAE v4.0), while radiation-induced liver injury was classified using Radiation Therapy Oncology Group/European Organization for Research and Treatment of Cancer (RTOG / EORTC) grading standards.¹⁶ OS was defined as the primary endpoint; secondary endpoints included PFS, objective response rate (ORR), and disease control rate (DCR). Patient follow-up was censored on March 30, 2025. The core study objective was to compare the prognostic value of pretreatment ALBI grade and serum LDH levels for survival endpoints in advanced HCC patients receiving first-line ICI-targeted combined radiotherapy. This research strictly complied with the ethical principles outlined in the 1975 Declaration of Helsinki and adhered to the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) reporting guidelines. A schematic flowchart illustrating patient screening and grouping is presented in Figure 1.

Figure 1.

Study flowchart.

Abbreviations: HCC, hepatocellular carcinoma; CT, computed tomography; MRI, magnetic resonance imaging; ALBI, albumin-bilirubin grade; LDH, lactate dehydrogenase; CR, complete response; PR, partial response; ORR, objective response rate; DCR, disease control rate; OS, overall survival; PFS, progression-free survival.

Diagnostic, Inclusion and Exclusion Criteria

HCC diagnosis was established in accordance with international clinical guidelines,^17,18 integrating typical contrast-enhanced computed tomography (CT) and/or magnetic resonance imaging (MRI) imaging features and serum alpha-fetoprotein (AFP) levels ≥ 400 ng/mL. Patients failing to meet radiologic and serologic diagnostic thresholds underwent histopathological confirmation for definitive HCC diagnosis.

Core inclusion criteria: (1) Treatment-naïve unresectable advanced HCC; (2) Laboratory parameters meeting ICI administration safety thresholds; (3) Baseline ALBI grade restricted to 1 or 2; (4) Eastern Cooperative Oncology Group (ECOG) performance status 0–1; (5) No prior hepatic radiotherapy history or decompensated cirrhosis.

Core exclusion criteria: (1) Concurrent autoimmune disorders; (2) Female patients during pregnancy or lactation; (3) Active hepatic encephalopathy at baseline; (4) Uncontrolled severe hypertension or prior cerebral infarction; (5) Moderate-to-severe hypothyroidism. Detailed descriptions of eligibility criteria have been published in our previous report.¹⁹

All patients underwent contrast-enhanced abdominal CT / MRI at baseline, with serial laboratory testing (liver/kidney function, routine blood work, coagulation function, thyroid function, hepatitis B virus DNA (HBV-DNA), and AFP) conducted pre-treatment and every 3 months throughout follow-up.

Treatment Regimens

All therapeutic regimens were finalized via multidisciplinary tumor board discussion, consisting of ICIs, molecular targeted agents, and sequential radiotherapy. ICI agents and dosing schedules: camrelizumab (3 mg/kg, Q3W. Hengrui, China), sintilimab (200 mg, Q3W. Innovent, China), tislelizumab (200 mg, Q3W. Beigene, China), pembrolizumab (200 mg, Q3W. Merck, US). Targeted agents and dosing schedules: lenvatinib (8 mg oral daily. Eisai Co, Ltd, Japan), apatinib (250 mg oral daily. Hengrui, China), anlotinib (8 mg oral days 1–14, 21-day cycle. Chia Tai Tianqing, China), bevacizumab (15 mg/kg Q3W. Innovent, China). Patients received ICIs for a minimum of four treatment cycles, or until disease progression or intolerable toxicity, with a maximum treatment duration limited to 2 years. Given the heterogeneity of ICI agents, targeted agents, and radiotherapy schedules, treatment-related heterogeneity was acknowledged as a potential confounder and was addressed in sensitivity analyses.

The detailed procedures of radiotherapy positioning and scanning requirements have been fully described in our previous study.¹⁹ All patients were immobilized with a thermoplastic body mask, and abdominal compression was applied to minimize liver movement while maintaining normal breathing. Notably, the positioning simulation scan for target volume delineation must cover both the arterial and venous phases, and target volume delineation should be performed based on the venous phase images.

Target volumes were delineated strictly in accordance with relevant clinical guidelines.²⁰ The gross tumor volume (GTV) encompassed the primary tumor lesion, metastatic lymph nodes, and tumor thrombus. Clinical target volume (CTV) was generated by expanding 5 mm outward from GTV, and the planning target volume (PTV) was further expanded by 5 mm based on CTV. Intrahepatic lesions were defined as GTV for patients with intrahepatic metastasis; for patients with exclusive extrahepatic metastases, the dominant metastatic lesion was contoured as GTV. Organs at risk (OARs) included the normal liver parenchyma, entire stomach, small intestine, bilateral kidneys, as well as the spinal cord, heart and lungs within 5 cm outside PTV (α/β = 3). Radiotherapy was administered on days 2–3 of the first ICI-targeted treatment cycle.

Intensity-modulated radiotherapy was delivered. The prescribed radiation doses were 15.0–48.0 Gy for GTV and 13.0–42.0 Gy for CTV, administered in 5–16 fractions (tumor α/β ratio = 10). All treatment plans satisfied the following dose constraints. For GTV and CTV, the dose covering 95% of the target volume (D_95%) was required to be ≥ 14.2–46.0 Gy and ≥ 12.5–40.0 Gy, respectively. The mean dose to the normal liver parenchyma was limited to < 8.0–20.0 Gy, and the mean dose to bilateral normal kidneys was < 10.0–20.0 Gy. The maximum dose to the stomach was constrained to < 6.0–20.0 Gy. The maximum dose to the small intestine and spinal cord was limited to 10.0–18.0 Gy.

ALBI and LDH Stratification

ALBI grades and LDH levels were calculated using pretreatment laboratory results. The validated ALBI scoring formula was applied: ALBI = 0.66 × log₁₀ [total bilirubin (μmol/L)] − 0.085 × serum albumin (g/L) ¹¹. Two independent grouping strategies were adopted: single-factor stratification and combined dual-biomarker stratification. Single-factor grouping: ALBI subgroups: ALBI-1 (ALBI ≤ −2.60) and ALBI-2 (−2.60 < ALBI ≤ −1.39); LDH subgroups: LDH-0 (LDH ≤ 250 U/L) and LDH-1 (LDH > 250 U/L). Although this 250 U/L cut-off was derived from published literature,²¹ it coincidentally matched the upper reference limit of LDH in our institutional laboratory. Combined dual-biomarker grouping: four mutually exclusive subgroups integrating ALBI grade and LDH status: ALBI-1+LDH-0, ALBI-1+LDH-1, ALBI-2+LDH-0, ALBI-2+LDH-1.

Efficacy, Safety and Follow-Up Definitions

Treatment response was assessed at the 6-month landmark timepoint via serial CT / MRI scans per mRECIST criteria: Complete response (CR): full disappearance of all target lesions; Partial response (PR): ≥ 30% reduction in the sum of longest target lesion diameters; Stable disease (SD): < 30% reduction or < 20% increase in target lesion diameter sum; Progressive disease (PD): ≥ 20% increase in target lesion diameter sum or new lesion emergence. ORR was defined as the proportion of patients achieving CR and PR; DCR encompassed CR, PR, and SD. Notably, PD status was determined purely by radiological findings, independent of the time-point relative to the 6-month landmark. Death events without radiological progression were coded as death events in survival analysis but not classified as PD.

All emergent AEs were prospectively recorded, graded per NCI-CTCAE v4.0, and attributed to study combination therapy by treating investigators. Follow-up was performed every 3 months via outpatient clinic visits or telephone contact. PFS was calculated from the first day of combination therapy to radiologically confirmed disease progression. OS was defined as the interval from treatment initiation to all-cause death, loss to follow-up, or the pre-specified data cut-off date.

Statistical Analysis

Sample size estimation was performed using PASS 15 software. A two-tailed statistical test was pre-specified with α = 0.01 and statistical power (1−β) = 0.95. The anticipated AUC for prognostic biomarker prediction was set to 0.7, with a projected 10% patient dropout rate. Clinically meaningful intergroup difference thresholds were predefined as follows: ORR / DCR intergroup discrepancy < 10%, OS difference < 4.5 months, and PFS difference < 3.0 months. Calculations yielded a minimal required sample size of 138 patients; a total of 143 eligible participants were enrolled to satisfy the pre-specified statistical power requirements.

All statistical analyses were conducted in SPSS 26.0. Two-tailed P values < 0.05 were deemed statistically significant. Categorical variables were summarized as patient counts and corresponding percentages. Intergroup comparisons of categorical data utilized the chi-square test; Fisher’s exact test was substituted for contingency tables with insufficient expected cell frequencies. Kaplan–Meier survival curves were generated to estimate median OS and PFS, with Log-rank (Mantel-Cox) tests applied to compare survival distributions across subgroups stratified by ALBI grade, LDH status, and integrated ALBI-LDH classification.

Sequential univariate and multivariate Cox proportional hazards regression models were constructed to identify independent prognostic correlates of OS and PFS. All variables exhibiting statistical significance in univariate analysis were incorporated into the multivariate regression model to screen for independent prognostic predictors. Hazard ratios (HR) with corresponding 95% confidence intervals (CI) were computed to quantify relative risks of all-cause mortality and tumor progression. To account for treatment heterogeneity, sensitivity multivariable Cox regression models additionally adjusted for ICI category and targeted therapy category were fitted to evaluate the robustness of the primary findings. Multicollinearity among covariates was assessed using variance inflation factors (VIF); a value < 5 indicated no serious multicollinearity.

Receiver operating characteristic (ROC) curve analysis was performed to evaluate the standalone predictive performance of pretreatment ALBI grade and LDH status for OS and PFS endpoints. The area under the curve (AUC), Youden index, sensitivity, specificity, and positive likelihood ratio (+LR) were calculated to quantify each biomarker’s discriminative capacity. Treatment response endpoints (CR, PR, SD, ORR, DCR) and the incidence of all-grade treatment-related AEs were statistically compared across single-factor subgroups and the four integrated ALBI-LDH combined subgroups.

Results

Baseline Patient Characteristics

A total of 152 consecutive patients with advanced HCC receiving first-line ICI-targeted combined radiotherapy between January 2019 and December 2023 were initially screened. Nine patients were excluded due to incomplete follow-up, failure to complete four ICI cycles, or missing core baseline clinical data. The final analytical cohort comprised 143 patients (114 males, 29 females; median age 54.9 years, range 27–76 years).

Table 1 summarizes the baseline clinical characteristics of the 143 enrolled patients, stratified by ALBI grade (ALBI-1, n = 82; ALBI-2, n = 61) and serum LDH level (LDH-0, n = 71; LDH-1, n = 72). No significant intergroup differences were observed in median age, age distribution, gender, ECOG performance status, BCLC stage, drinking history, HBV-DNA positivity, lymphatic metastasis, ascites or portal hypertension (all P > 0.05). Compared with ALBI-1 patients, the ALBI-2 subgroup presented marginally higher proportions of elevated LDH, liver cirrhosis, intrahepatic metastasis, portal vein invasion and portal vein tumor thrombus (PVTT) (P values ranged from 0.065 to 0.238). For LDH subgroups, the LDH-1 group tended to have more patients with ALBI grade 2, AFP ≥ 400 ng/mL, liver cirrhosis, intrahepatic metastasis and PVTT relative to the LDH-0 group (all P > 0.05).

Table 1.

Baseline Characteristics of 143 Patients (n = 143)

Characteristics	ALBI-1 (n=82)	ALBI-2 (n=61)	LDH-0 (n=71)	LDH-1 (n=72)	P (ALBI)	P (LDH)
Median age, years (range)	53.7 (27–76)	56.3 (31–76)	54.8 (28–76)	54.8 (27–76)	0.120	0.910
Age, n (%)
<50 years	32 (39.0)	16 (26.2)	23 (32.4)	22 (30.6)	0.152	0.854
≥50 years	50 (61.0)	45 (73.8)	48 (67.6)	50 (69.4)
Gender, n (%)
Male	63 (76.8)	51 (83.6)	55 (77.5)	59 (81.9)	0.402	0.539
Female	19 (23.2)	10 (16.4)	16 (22.5)	13 (18.1)
ECOG score, n (%)
0	61 (74.4)	48 (78.7)	52 (73.2)	54 (75.0)	0.692	0.850
1	21 (25.6)	13 (21.3)	19 (26.8)	18 (25.0)
BCLC stage, n (%)
B	47 (57.3)	39 (63.9)	39 (54.9)	37 (51.4)	0.491	0.738
C	35 (42.7)	22 (36.1)	32 (45.1)	35 (48.6)
AFP ≥400 ng/mL, n (%)	26 (31.7)	29 (47.5)	19 (26.8)	30 (41.7)	0.059	0.078
Drinking history, n (%)	27 (32.9)	25 (41.0)	23 (32.4)	29 (40.3)	0.381	0.386
HBV-DNA positive, n (%)	50 (61.0)	43 (70.5)	47 (66.2)	46 (63.9)	0.289	0.861
Liver cirrhosis, n (%)	32 (39.0)	30 (49.2)	25 (35.2)	36 (50.0)	0.238	0.091
Portal hypertension, n (%)	36 (43.9)	35 (57.4)	27 (38.0)	34 (47.2)	0.130	0.312
Intrahepatic metastasis, n (%)	51 (62.2)	43 (70.5)	38 (53.5)	49 (68.1)	0.374	0.088
Lymphatic metastasis, n (%)	45 (54.9)	38 (62.3)	40 (56.3)	43 (59.7)	0.397	0.736
Ascites, n (%)	24 (29.3)	24 (39.3)	17 (23.9)	21 (29.2)	0.216	0.571
Portal vein invasion, n (%)	34 (41.5)	35 (57.4)	30 (42.3)	39 (54.2)	0.065	0.182
PVTT, n (%)	29 (35.4)	31 (50.8)	30 (42.3)	34 (47.2)	0.086	0.615

Abbreviations: ECOG, Eastern Cooperative Oncology Group; BCLC, Barcelona Clinical Liver Cancer staging; AFP, alpha fetoprotein; HBV, hepatitis B virus; PVTT, portal vein tumor thrombus; ALBI, albumin-bilirubin grade; LDH, lactate dehydrogenase.

Treatment Efficacy

The ALBI-1 subgroup achieved superior therapeutic response vs ALBI-2 patients, with higher ORR (62.2% vs 37.7%) and DCR (92.7% vs 65.6%). Similarly, LDH-0 patients exhibited improved treatment outcomes relative to LDH-1 patients (ORR: 70.4% vs 33.3%; DCR: 95.8% vs 66.7%). Complete tumor regression was exclusively observed in ALBI-1 and LDH-0 patients; no CR events were recorded in ALBI-2 or LDH-1 subgroups.

Dramatic therapeutic heterogeneity was observed across the four integrated ALBI-LDH subgroups. The ALBI-1+LDH-0 subgroup achieved the optimal treatment response profile, with the highest ORR (74.1%), DCR (96.3%), and CR rate (11.1%). In contrast, the ALBI-2+LDH-1 subgroup demonstrated the poorest anti-tumor efficacy, with the lowest ORR (29.5%) and DCR (54.5%). Intermediate therapeutic outcomes were observed in ALBI-1+LDH-1 (ORR =3 9.3%, DCR = 85.7%) and ALBI-2+LDH-0 (ORR = 58.8%, DCR = 94.1%) subgroups. Collectively, preserved hepatic function combined with lower LDH levels correlated with favorable treatment response, while impaired liver function and elevated LDH predicted limited therapeutic benefit in advanced HCC (Table 2).

Table 2.

Treatment Response by ALBI and LDH Subgroups (n = 143)

Group	CR, n (%)	PR, n (%)	SD, n (%)	PD, n (%)	ORR, %	DCR, %
ALBI-1 (n=82)	6 (7.3)	45 (54.9)	25 (30.5)	6 (7.3)	62.2	92.7
ALBI-2 (n=61)	0 (0.0)	23 (37.7)	17 (27.9)	21 (34.4)	37.7	65.6
LDH-0 (n=71)	6 (8.5)	44 (62.0)	18 (25.4)	3 (4.2)	70.4	95.8
LDH-1 (n=72)	0 (0.0)	24 (33.3)	24 (33.3)	24 (33.3)	33.3	66.7
ALBI-1+LDH-0 (n=54)	6 (11.1)	34 (63.0)	12 (22.2)	2 (3.7)	74.1	96.3
ALBI-1+LDH-1 (n=28)	0 (0.0)	11 (39.3)	13 (46.4)	4 (14.3)	39.3	85.7
ALBI-2+LDH-0 (n=17)	0 (0.0)	10 (58.8)	6 (35.3)	1 (5.9)	58.8	94.1
ALBI-2+LDH-1 (n=44)	0 (0.0)	13 (29.5)	11 (25.0)	20 (45.5)	29.5	54.5

Abbreviations: ALBI, albumin-bilirubin grade; LDH, lactate dehydrogenase; CR, complete response; PR, partial response; SD, stable disease; PD, progressive disease; ORR, objective response rate; DCR, disease control rate.

Survival Analyses

Median OS and PFS for the entire cohort reached 15.7 and 12.8 months, respectively. Head-to-head prognostic comparison confirmed significantly prolonged OS and PFS in ALBI-1 and LDH-0 patients relative to ALBI-2 and LDH-1 counterparts. ALBI-1 vs ALBI-2: median OS 28.7 vs 11.4 months (χ² = 40.520, P < 0.001); median PFS 24.6 vs 9.1 months (χ² = 18.420, P < 0.001). LDH-0 vs LDH-1: median OS 24.8 vs 13.2 months (χ² = 25.530, P < 0.001); median PFS 23.7 vs 11.9 months (χ² = 7.127, P = 0.007).

In univariate analyses, risk regression demonstrated that compared with patients with ALBI-2 grade, ALBI grade 1 reduced all-cause mortality risk by 73.0% (HR = 0.27, 95% CI: 0.16–0.45) and disease progression risk by 62.0% (HR = 0.38, 95% CI: 0.22–0.65) (HRs were calculated with ALBI-2 / LDH-1 as the reference group). Compared with patients with LDH-1 status, lower LDH levels decreased mortality risk by 67.0% (HR = 0.33, 95% CI: 0.21–0.54) and progression risk by 47.0% (HR = 0.53, 95% CI: 0.32–0.89). These univariate associations support the prognostic relevance of both biomarkers in this cohort; however, after multivariable adjustment, ALBI grade did not retain statistical significance for either OS or PFS, whereas elevated LDH remained independently associated with OS but not with PFS (Table 3).

Table 3.

Factors Affecting OS and PFS Analyzed by Univariate and Multivariate Cox Regression (n=143)

Characteristic	Univariate analysis of OS		Multivariate analysis of OS		Univariate analysis of PFS		Multivariate analysis of PFS
	HR (95%CI)	p-values	HR (95%CI)	p-values	HR (95%CI)	p-values	HR (95%CI)	p-values
BCLC stage	1.78 (1.12–2.82)	0.015	2.16 (1.26–3.73)	0.005	2.00 (1.23–3.26)	0.010	2.03 (1.21–3.39)	0.010
ALBI grade	0.27 (0.16–0.45)	<0.001	1.83 (0.95–3.53)	0.070	0.38 (0.22–0.65)	<0.001	1.64 (0.82–3.28)	0.160
Child-Pugh score	3.88 (2.37–6.34)	<0.001	2.77 (1.49–5.11)	0.001	2.45 (1.44–4.14)	<0.001	1.83 (0.95–3.55)	0.070
AFP	2.34 (1.47–3.71)	<0.001	1.04 (0.62–1.75)	0.870	2.41 (1.49–3.92)	<0.001	1.45 (0.84–2.49)	0.180
Liver cirrhosis	1.67 (1.06–2.64)	0.020	1.20 (0.73–1.98)	0.470	-	-	-	-
Intrahepatic metastasis	2.92 (1.67–5.11)	<0.001	2.44 (1.35–4.43)	0.003	-	-	-	-
Ascites	1.75 (1.07–2.87)	0.020	0.79 (0.45–1.41)	0.430	-	-	-	-
PVTT	1.98 (1.25–3.14)	0.004	2.47 (1.27–4.79)	0.010	-	-	-	-
Portal vein invaded	1.66 (1.05–2.62)	0.030	0.81 (0.42–1.58)	0.540	1.91 (1.18–3.10)	0.010	1.56 (0.91–2.67)	0.110
LDH level	3.24 (2.01–5.24)	<0.001	3.20 (1.53–6.69)	0.002	2.00 (1.22–3.27)	0.006	1.33 (0.14–12.75)	0.804
TBIL	-	-	-	-	1.01 (1.00–1.02)	0.049	1.02 (0.99–1.06)	0.233

Abbreviations: BCLC, Barcelona Clinical Liver Cancer staging; ALBI, albumin-bilirubin grade; AFP, alpha fetoprotein; PVTT, portal vein tumor thrombus; LDH, lactate dehydrogenase; TBIL, total bilirubin; OS, overall survival; PFS, progression-free survival; HR, Hazard Ratio; CI, confidence interval.

In univariate analyses, cross-comparison of favorable and poor-risk subgroups confirmed equivalent predictive performance between the two biomarkers. No significant differences in median OS or PFS were detected between ALBI-1 and LDH-0 patients (OS: 28.7 vs 24.8 months, χ² = 0.022, P = 0.881; PFS: 24.6 vs 23.7 months, χ² = 0.148, P = 0.700). Similarly, ALBI-2 and LDH-1 patients displayed comparable survival endpoints (OS: 11.4 vs 13.8 months, χ² = 1.168, P = 0.280; PFS: 9.1 vs 11.9 months, χ² = 0.980, P = 0.322). Overall, ALBI grade and serum LDH levels delivered highly equivalent survival predictive capacity (Figure 2).

Figure 2.

Kaplan-meier OS and PFS curves stratified by baseline ALBI grade and LDH level. (a-d) Survival comparisons between ALBI-1 vs ALBI-2 and LDH-0 vs LDH-1. (e-h) Direct pairwise comparisons of favorable subgroups (ALBI-1 vs LDH-0) and poor subgroups (ALBI-2 vs LDH-1). Median survival times, log-rank χ² and P values are shown for all comparisons. No significant survival differences were detected between ALBI-1 and LDH-0, or between ALBI-2 and LDH-1.

Abbreviations: ALBI, albumin-bilirubin grade; LDH, lactate dehydrogenase; OS, overall survival; PFS, progression-free survival.

Integrated ALBI-LDH subgroup analysis revealed distinct survival gradients across all four cohorts. The ALBI-1+LDH-0 subgroup achieved the longest median OS (39.0 months), significantly exceeding all other subgroups: ALBI-1+LDH-0 vs ALBI-1+LDH-1: OS 39.0 vs 18.3 months (χ² = 7.958, P = 0.005); PFS comparable (23.7 vs 17.6 months, χ² = 0.669, P = 0.414). ALBI-1+LDH-0 vs ALBI-2+LDH-0: OS 39.0 vs 14.4 months (χ² = 19.120, P < 0.001); PFS 23.7 vs 9.7 months (χ² = 5.267, P = 0.022). ALBI-1+LDH-0 vs ALBI-2+LDH-1: OS 39.0 vs 11.4 months (χ² = 47.720, P < 0.001); PFS 23.7 vs 7.8 months (χ² = 22.150, P < 0.001).

Survival endpoints were comparable between ALBI-1+LDH-1 and ALBI-2+LDH-0 subgroups (OS: 18.3 vs 14.4 months, χ² = 2.093, P = 0.148; PFS: 17.6 vs 9.7 months, χ² = 1.634, P = 0.250). Nevertheless, ALBI-1+LDH-1 patients demonstrated superior OS and PFS vs ALBI-2+LDH-1 patients (OS: 18.3 vs 11.4 months, χ² = 8.557, P = 0.003; PFS: 17.6 vs 7.8 months, χ² = 7.278, P = 0.007). The two ALBI grade 2 subgroups exhibited similar survival outcomes (OS: 14.4 vs 11.4 months, χ² = 2.168, P = 0.141; PFS: 9.7 vs 7.8 months, χ² = 1.634, P = 0.201) (Figure 3).

Figure 3.

Kaplan-meier survival analyses of OS and PFS in HCC patients stratified by ALBI grade and LDH level. (a and b) OS and PFS comparison among ALBI-1+LDH-0 and ALBI-1+LDH-1 (median OS: 39.0 vs 18.3 months, χ² = 7.958, P = 0.005; median PFS: 23.7 vs 17.6 months, χ² = 0.669, P = 0.414). (c and d) OS and PFS comparison among ALBI-1+LDH-0 and ALBI-2+LDH-0 (median OS: 39.0 vs 14.4 months, χ² = 19.120, P < 0.001; median PFS: 23.7 vs 9.7 months, χ² = 5.267, P = 0.022). (e and f) OS and PFS comparison among ALBI-1+LDH-0 and ALBI-2+LDH-1 (median OS: 39.0 vs 11.4 months, χ² = 47.720, P < 0.001; median PFS: 23.7 vs 7.8 months, χ² = 22.150, P < 0.001). (g and h) OS and PFS comparison among ALBI-1+LDH-1 vs ALBI-2+LDH-0 (median OS: 18.3 vs 14.4 months, χ² = 2.093, P = 0.148; median PFS: 17.6 vs 9.7 months, χ² = 1.634, P = 0.250). (i and j) OS and PFS comparison among ALBI-1+LDH-1 and ALBI-2+LDH-1 (median OS: 18.3 vs 11.4 months, χ² = 8.557, P = 0.003; median PFS: 17.6 vs 7.8 months, χ² = 7.278, P = 0.007). (k and l) OS and PFS comparison among ALBI-2+LDH-0 and ALBI-2+LDH-1 (median OS: 14.4 vs 11.4 months, χ² = 2.168, P = 0.141; median PFS: 9.7 vs 7.8 months, χ² = 1.634, P = 0.201).

Abbreviations: ALBI, albumin–bilirubin grade; LDH, lactate dehydrogenase; HCC, hepatocellular carcinoma; OS, overall survival; PFS, progression-free survival.

Collectively, integrated ALBI-LDH classification enabled refined prognostic stratification; patients with ALBI grade 1 and normal LDH represented the optimal risk group with the most favorable long-term survival.

Cox Regression and ROC Curve Analyses

Univariate Cox regression identified BCLC stage, ALBI grade, Child-Pugh score, AFP level, liver cirrhosis, intrahepatic metastasis, clinical ascites, PVTT, portal vein invasion, and serum LDH as significant predictors of OS. Significant PFS prognostic correlates included BCLC stage, ALBI grade, Child-Pugh score, AFP, portal vein invasion, LDH, and total bilirubin (TBIL).

Multivariate Cox regression further confirmed BCLC stage, Child-Pugh score, intrahepatic metastasis, PVTT, and elevated LDH as independent adverse risk factors for shortened OS. No clinical variables retained independent predictive value for PFS in the multivariate model (Table 3). Specifically, ALBI grade was not statistically significant in either multivariable model (OS: HR = 1.83, 95% CI: 0.95–3.53, P = 0.070; PFS: HR = 1.64, 95% CI: 0.82–3.28, P = 0.160), and LDH was not independently associated with PFS (P = 0.804).

ROC curve analysis was performed to evaluate the prognostic performance of the ALBI score, LDH level, and their combined model, with the AUC, sensitivity, specificity, Youden’s index, and +LR calculated for each.

For OS, the combined ALBI-LDH model yielded the highest predictive accuracy, with an AUC of 0.670 (95% CI: 0.574–0.766), which was numerically higher than that of ALBI alone (AUC = 0.547, 95% CI: 0.447–0.647) and LDH alone (AUC = 0.511, 95% CI: 0.411–0.611). At the optimal cut-off value, the combined model achieved a sensitivity of 80.00% (95% CI: 0.670–0.888), a specificity of 47.31% (95% CI: 0.375–0.574), a Youden’s index of 0.273, and a +LR of 1.518, whereas ALBI alone showed a sensitivity and specificity of 57.37% and 52.43%, respectively, and LDH alone of 49.83% and 52.27%, respectively.

For PFS, the combined model likewise demonstrated the best performance, with an AUC of 0.672 (95% CI: 0.583–0.761), outperforming ALBI alone (AUC = 0.521, 95% CI: 0.426–0.616) and LDH alone (AUC = 0.518, 95% CI: 0.423–0.613). The combined model had a sensitivity of 59.09% (95% CI: 0.471–0.701), a specificity of 71.43% (95% CI: 0.605–0.803), a Youden’s index of 0.305, and a +LR of 2.068.

In conclusion, the combined ALBI–LDH model provided superior prognostic discrimination for both OS and PFS compared with either marker alone, suggesting that their combination may serve as an effective tool for prognostic assessment in this patient population (Figure 4).

Figure 4.

ROC curves and diagnostic performance of ALBI, LDH, and the combined ALBI-LDH model for predicting OS and PFS. (a) ROC curves for OS prediction; (b) ROC curves for PFS prediction; (c) Sensitivity, specificity, AUC, youden index, and positive likelihood ratio of each model at the optimal cutoff value.

Abbreviations: ROC, receiver operating characteristic; ALBI, albumin-bilirubin grade; LDH, lactate dehydrogenase; OS, overall survival; PFS, progression-free survival; AUC, area under the curve; +LR, positive likelihood ratio; CI, confidence interval.

Treatment-Related Adverse Events

The majority of treatment-emergent AEs were grade 1–2, with a low incidence of grade 3–4 severe toxicities. Stratified by ALBI grade, ALBI-2 patients demonstrated higher rates of grade 1–2 fatigue, cutaneous pruritus/rash, thrombocytopenia, leukopenia, and dyspepsia relative to ALBI-1 patients. Numerically elevated rates of grade 3–4 ALT / AST elevation, hematologic toxicity, and gastrointestinal adverse reactions were also observed in the ALBI-2 subgroup.

Stratification by LDH status revealed no consistent intergroup disparities in overall grade 1–2 or grade 3–4 AE incidence; minor isolated differences were limited to hypothyroidism and nausea/vomiting, without systematic disparities in severe toxicity distribution.

Subgroup comparison demonstrated that safety disparities were far more pronounced across ALBI grades than LDH subgroups. Patients with impaired baseline hepatic reserve (ALBI-2) were more susceptible to both mild and severe treatment-related toxicities, whereas baseline LDH status delivered minimal predictive value for AE incidence and severity. No fatal or unexpected severe immune-related adverse events were documented within the cohort (Table 4).

Table 4.

Adverse Events Stratified by ALBI Grade and LDH Level (n = 143)

Adverse events	ALBI-1 group (n=82)		ALBI-2 group (n=61)		LDH-0 group (n=71)		LDH-1 group (n=72)
	Grade 1–2(%)	Grade 3–4(%)	Grade 1–2(%)	Grade 3–4(%)	Grade 1–2(%)	Grade 3–4(%)	Grade 1–2(%)	Grade 3–4(%)
Hypothyroidism	31 (37.8)*	-	23 (37.7)&	-	22 (31.0)	-	32 (44.4)	-
ALT elevation	55 (67.1)*	2 (2.4)#	34 (55.7)&	3 (4.9) **	45 (63.4)	2 (2.8)	44 (61.1)	3 (4.2)
AST elevation	53 (64.6)*	3 (3.7)#	39 (63.9)&	5 (8.2) **	49 (69.0)	2 (2.8)	43 (59.7)	6 (8.3)
Hypoalbuminemia	64 (78.0)*	3 (3.7)#	42 (68.9)&	3 (4.9) **	52 (73.2)	3 (4.2)	54 (75.0)	3 (4.2)
Fatigue	51 (62.2)*	-	45 (73.8)&	-	47 (66.2)	-	49 (68.1)	-
Nausea/vomiting	35 (42.7)*	-	27 (44.3)&	-	36 (50.7)	-	26 (36.1)	-
Pruritus/rash	58 (70.7)*	2 (2.4)#	54 (88.5)&	2 (3.3) **	51 (71.8)	3 (4.2)	61 (84.7)	1 (1.4)
Thrombocytopenia	54 (65.9) *	-	51 (83.6)&	-	54 (76.1)	-	51 (70.8)	-
Hyperbilirubinemia	21 (25.6)*	-	17 (27.9)&	-	19 (26.8)	-	19 (26.4)	-
Leukopenia	20 (24.4)*	5 (6.1)#	21 (34.4)&	3 (4.9) **	18 (25.4)	5 (7.0)	23 (31.9)	3 (4.2)
Anemia	18 (22.0)*	3 (3.7)#	15 (24.6)&	2 (3.3) **	14 (19.7)	3 (4.2)	19 (26.4)	2 (2.8)
Diarrhea	9 (11.0)*	5 (6.1)#	9 (14.8)&	1 (1.6)**	8 (11.3)	3 (4.2)	10 (13.9)	3 (4.2)
Hypertension	13 (15.9)*	-	7 (11.5)&	-	10 (14.1)	-	10 (13.9)	-
Dyspepsia	2 (2.4)*	-	6 (9.8)&	-	4 (5.6)	-	4 (5.6)	-
Myocarditis	1 (1.2)*	-	-	-	1 (1.4)	-	-	-

Notes: * Comparison of grade 1–2 adverse events between the ALBI-1 group and the LDH-0 group (P>0.05); # Comparison of grade 3–4 adverse events between the ALBI-1 group and the LDH-0 group (P>0.05); and Comparison of grade 1–2 adverse events between the ALBI-2 group and the LDH-1 group (P>0.05); ** Comparison of grade 3–4 adverse events between the ALBI-2 group and the LDH-1 group (P>0.05).

Abbreviations: ALT, alanine transaminase; AST, aspartate transaminase; ALBI, albumin-bilirubin grade; LDH, lactate dehydrogenase.

Discussion

Of note, the patient cohort analyzed in the present study is entirely distinct from those of our previous publications,^19,22 with no overlap in patient population, study period, or dataset. The ALBI grade is a validated objective quantitative metric for evaluating hepatic functional reserve in HCC patients,^4,5,7 and additionally predicts ICI therapeutic efficacy in this population.²³ Our data demonstrated that ALBI-2 patients presented higher baseline rates of PVTT, liver cirrhosis, and clinical ascites compared with ALBI-1 counterparts. In terms of anti-tumor response, ALBI-1 patients achieved superior ORR and DCR, alongside markedly prolonged OS and PFS. Univariate regression confirmed ALBI grade as a strong prognostic correlate of survival endpoints, although this association was not retained as statistically significant in multivariate Cox models after adjustment for other clinical covariates. ROC curve analysis indicated limited standalone predictive accuracy of ALBI grade for OS and PFS, a finding likely attributable to the substantial clinical heterogeneity of advanced HCC. Mechanistically, ALBI grade reflects combined tumor burden and hepatic functional reserve; higher ALBI grades signify poorer liver function and heavier tumor load. For patients receiving radiotherapy combined with ICIs and targeted agents, pre-existing hepatic dysfunction is further exacerbated by triple therapy toxicity, ultimately attenuating anti-tumor response and compromising long-term survival. Furthermore, the heterogeneity of ICI, targeted, and radiotherapy regimens across patients may have influenced the observed associations; although treatment type was adjusted for in sensitivity analyses, residual confounding related to radiotherapy parameters and other unmeasured factors cannot be fully excluded.

Serum LDH levels are tightly linked to HCC patient survival.^11,12 In this cohort, patients with lower LDH levels achieved superior therapeutic response and longer OS/PFS vs those with elevated LDH. Multivariate Cox regression validated serum LDH as an independent adverse prognostic biomarker for OS. Consistent with ALBI grade, ROC analysis revealed limited standalone predictive performance of LDH for survival endpoints, again driven by heterogeneous disease biology across advanced HCC cases. Based on previously published mechanistic literature, serum LDH is regarded as a surrogate marker of tumor microenvironment metabolic reprogramming. According to pre-clinical and translational studies, elevated LDH may drive lactate accumulation via glycolysis-related gene upregulation, remodel the immunosuppressive tumor microenvironment, promote tumor proliferation and angiogenesis, and abrogate endogenous anti-tumor immune activity.^24,25 We emphasize that our present retrospective clinical cohort lacks molecular or immunological assays to experimentally verify these mechanistic links. Increased circulating LDH reflects enhanced tumor glycolysis, which typically coincides with elevated invasive and metastatic potential. Of note, LDH was not independently associated with PFS in the multivariable model, and no clinical variable retained independent predictive value for PFS.

Head-to-head comparison of ALBI grade and LDH revealed equivalent therapeutic response and survival predictive capacity across favorable-risk (ALBI-1 vs LDH-0) and poor-risk (ALBI-2 vs LDH-1) subgroups. This congruent prognostic performance likely stems from the intrinsic biological correlation between hepatic function and tumor immune microenvironment. Patients with preserved liver function harbor abundant intratumoral anti-tumor lymphocytes, which suppress malignant proliferation and angiogenesis and amplify the anti-tumor activity of radiotherapy plus ICI-targeted therapy. In contrast, patients with impaired hepatic reserve exhibit depleted anti-tumor immune cell populations, accelerating tumor progression, stimulating angiogenesis, and establishing a systemic immunosuppressive state that blunts combination therapeutic efficacy.

HCC clinical outcomes are modulated by multiple interacting factors, including hepatic functional reserve,^4,5 tumor microenvironment characteristics,^11,12 tumor staging,³ and PVTT status.²² Single biomarkers carry insufficient predictive resolution, while integrated multi-factor models better capture heterogeneous clinical outcomes in HCC patients.²⁶ In our ROC analyses, the combined ALBI-LDH model achieved higher AUC values for OS and PFS compared with either biomarker alone, indicating improved discriminative power relative to single-marker prediction within our cohort. Importantly, we did not directly benchmark this combined signature against other established HCC prognostic systems such as Child-Pugh or BCLC staging; whether ALBI-LDH outperforms those conventional indices remains unproven in the present study and warrants further comparative investigation. Nevertheless, the absolute AUC values remained modest, and these findings should be regarded as exploratory pending external validation. Patients in the ALBI-1+LDH-0 subgroup achieved superior therapeutic response and longer OS/PFS relative to single-biomarker subgroups and all other dual-biomarker cohorts. This finding indicates that patients with preserved liver function and non-immunosuppressive tumor microenvironment (ALBI-1+LDH-0) represent optimal candidates for triplet combination therapy. Remaining dual-biomarker subgroups are characterized by either impaired hepatic reserve or glycolytically active immunosuppressive tumor microenvironments, both of which undermine therapeutic efficacy.

Most treatment-related AEs were mild (grade 1–2), with low rates of severe grade 3–4 toxicity. ALBI-2 patients exhibited higher incidence of both mild and severe AEs vs ALBI-1 patients, while LDH subgroups displayed comparable overall toxicity profiles, consistent with prior published research.²⁷ Intergroup safety disparities were far more prominent across ALBI grades than LDH strata; patients with poor baseline liver function demonstrate heightened vulnerability to combination therapy toxicity, whereas pretreatment LDH status provides minimal predictive insight into AE frequency and severity.

This research bears inherent limitations associated with single-center retrospective observational design. First, the single-center cohort with limited sample size lacks external validation via multicenter prospective cohorts. Notably, two combined-biomarker subgroups (ALBI-2+LDH-0, n = 17; ALBI-1+LDH-1, n = 28) had relatively small sample sizes; survival and efficacy findings from these subsets should be interpreted cautiously as exploratory results. Second, only static baseline ALBI and LDH values were analyzed; dynamic longitudinal changes in these biomarkers during treatment were not evaluated for prognostic impact. In addition, our LDH grouping relied on a literature-derived fixed cut-off rather than laboratory-specific reference limits or cohort-optimized thresholds, which may limit generalisability across centers with different biochemistry assays. Third, due to substantial advanced HCC heterogeneity, we did not conduct stratified analyses integrating tumor stage, AFP, PVTT, and intrahepatic metastasis simultaneously. Fourth, safety analysis only compared crude AE incidence and grading without adjusting for confounding variables including treatment duration, drug dosage intensity, and concurrent hepatoprotective supportive care. Fifth, formal subgroup analyses evaluating differential biomarker performance across distinct ICIs, targeted agents, or radiotherapy dose/volume schedules were not performed; treatment type was instead adjusted for in sensitivity analyses, which cannot fully eliminate residual confounding. Sixth, mechanistic interpretations of ALBI and LDH’s impact on efficacy, survival, and toxicity were derived exclusively from published literature; no translational molecular experiments were conducted to validate specific underlying pathways. Seventh, because patients who did not complete at least four cycles of ICI therapy were excluded, this per-protocol design may introduce selection bias and limit the generalizability of the findings. Future prospective multicenter research will address these limitations and further define the clinical utility of ALBI and LDH in standardized HCC management workflows.

Conclusions

In summary, pretreatment ALBI grade (a hepatic function biomarker) and serum LDH (a tumor glycolysis surrogate) represent two low-cost, readily accessible, clinically practical prognostic biomarkers for advanced HCC patients receiving ICI-targeted combined radiotherapy. The two biomarkers deliver equivalent predictive performance for therapeutic response and long-term survival in univariate analyses, although ALBI grade did not retain statistical significance as an independent prognostic factor in multivariate Cox regression, whereas ALBI grade is superior for forecasting treatment-related toxicities. Within our cohort, combined utilization of ALBI grade and serum LDH achieved better prognostic discrimination than either biomarker alone; direct comparisons against other established HCC prognostic systems were not performed, so superiority relative to conventional risk tools remains to be validated. Mechanistic inferences regarding LDH and the tumor immune microenvironment are derived from external published literature rather than experimental data generated in this study. Combined utilization of ALBI grade and serum LDH may enable refined patient risk stratification and support personalized clinical decision-making, although these findings require confirmation in prospective, externally validated cohorts. For patients presenting with impaired liver function (ALBI-2) or elevated pretreatment LDH, comprehensive assessment of therapeutic tolerance and long-term prognosis should be performed prior to initiating triplet combination therapy. During treatment, clinicians should dynamically monitor therapeutic response and emergent adverse events, with timely regimen adjustments to maximize net clinical benefit for individual HCC patients.

Funding Statement

This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.

Abbreviations

HCC, hepatocellular carcinoma; OS, overall survival; BCLC, Barcelona Clinic Liver Cancer; ALBI, albumin-bilirubin grade; PFS, progression-free survival; ICI, immune checkpoint inhibitor; mRECIST, modified Response Evaluation Criteria in Solid Tumors; AEs, adverse events; NCI-CTCAE v4;0, National Cancer Institute Common Terminology Criteria for Adverse Events, version 4;0; RTOG/EORTC, Radiation Therapy Oncology Group/European Organization for Research and Treatment of Cancer; ORR, objective response rate; DCR, disease control rate; STROBE, Strengthening the Reporting of Observational Studies in Epidemiology; CT, computed tomography; MRI, magnetic resonance imaging; AFP, alpha-fetoprotein; ECOG, Eastern Cooperative Oncology Group; Q3W, every 3 weeks; HBV-DNA, hepatitis B virus DNA; GTV, gross tumor volume; CTV, clinical target volume; PTV, planning target volume; OARs, organs at risk; D95%, the dose covering 95% of the target volume; CR, Complete response; PR, Partial response; SD, Stable disease; PD, Progressive disease; PASS 15, PASS version 15; SPSS 26;0, SPSS version 26;0; HR, Hazard ratio; CI, confidence interval; ROC, Receiver Operating Characteristic; AUC, Area Under Curve; +LR, positive likelihood ratio.

Data Sharing Statement

The data that support the findings of this study are not openly available due to sensitivity reasons and are available from the corresponding author upon reasonable request.

Ethics Approval and Informed Consent

The trial protocol, data collection scheme and written informed consent documents were all reviewed and approved by the Ethics Committee and Institutional Review Board of the Affiliated Hospital of Southwest Medical University (approval ID: 16/10,2024 / KY2024389). As only de-identified retrospective medical records lacking identifiable personal data were analyzed, written informed consent from each patient was waived by the Ethics Committee following the Declaration of Helsinki.

Consent for Publication

The authors affirm that the human research participants provided informed consent for the publication.

Author Contributions

All authors made a significant contribution to the work reported, whether that is in the conception, study design, execution, acquisition of data, analysis and interpretation, or in all these areas; took part in drafting, revising or critically reviewing the article; gave final approval of the version to be published; have agreed on the journal to which the article has been submitted; and agree to be accountable for all aspects of the work.

Disclosure

The authors have no relevant financial or non-financial interests to disclose.