Prognostic Significance of Complete Blood Counts-Derived Inflammatory Markers in Compensated Hepatocellular Carcinoma

Abstract

Background

Inflammation-based indices derived from complete blood counts (CBC) have emerged as potential prognostic markers in hepatocellular carcinoma (HCC), yet their comparative value in compensated patients remains unclear. This study aimed to evaluate the prognostic significance of seven CBC-derived inflammatory markers in patients with compensated HCC.

Methods

This retrospective cohort study included 2802 patients with compensated HCC diagnosed between 2012 and 2025 at the First Affiliated Hospital of Anhui Medical University. Seven inflammatory indices (NLR, dNLR, MLR, NMLR, SIRI, SII, and AISI) were calculated from baseline blood tests. LASSO Cox regression with 10-fold cross-validation was used to select prognostically relevant indices. Restricted cubic spline (RCS) models were used to determine optimal cutoff values. Overall survival (OS) was analyzed using Kaplan-Meier curves and multivariable Cox regression models.

Results

MLR, NMLR, and SIRI were identified as significant prognostic markers. RCS-derived cutoffs were 0.27 for MLR, 2.62 for NMLR, and 0.84 for SIRI. Kaplan–Meier analysis showed that high levels of all three indices were significantly associated with worse OS (log-rank P < 0.001). In adjusted Cox models, high MLR (HR = 1.972, 95% CI: 1.704–2.283), NMLR (HR = 1.822, 95% CI: 1.576–2.107), and SIRI (HR = 1.736, 95% CI: 1.503–2.006) remained independently associated with poorer survival. These associations remained robust in sensitivity analyses using the complete-case dataset.

Conclusion

These readily available and cost-effective markers (MLR, NMLR, and SIRI) may serve as useful prognostic indicators in patients with compensated HCC.

Introduction

Hepatocellular carcinoma (HCC) is the most prevalent type of primary liver cancer and remains a leading cause of cancer-related mortality worldwide.^1–3 Despite marked heterogeneity in clinical characteristics among patients with HCC, a substantial proportion present with compensated liver function, representing a key subgroup eligible for potentially curative or life-prolonging therapies.⁴ Despite recent advances in surgical techniques, locoregional interventions, and systemic treatments such as tyrosine kinase inhibitors and immune checkpoint inhibitors, the long-term prognosis of these patients remains unsatisfactory.^5,6 In particular, many are still diagnosed at intermediate or advanced stages, precluding the possibility of curative treatment. These challenges underscore the urgent need for simple, reliable, accessible, and cost-effective prognostic biomarkers to guide treatment decision-making and improve clinical outcomes in compensated HCC.

Chronic inflammation plays a central role in the initiation and progression of HCC, particularly in patients with chronic liver diseases such as viral hepatitis, alcohol-related liver injury, and non-alcoholic fatty liver disease.^7,8 Chronic inflammation contributes to HCC development not only by sustaining a protumorigenic microenvironment but also by activating key oncogenic signaling pathways.⁹ Accumulating evidence indicates that HCC progression is driven by dysregulation of receptor tyrosine kinase signaling and downstream pathways such as MAPK, PI3K/AKT/mTOR, and JAK/STAT, which are involved in tumor cell growth, survival, angiogenesis, metastasis, and treatment resistance. VEGF-associated angiogenic signaling also plays a central role in shaping the biological behavior of HCC. Recent studies on targeted therapeutic formulations and molecular interaction-based strategies further support the relevance of these signaling pathways in HCC biology.^10,11 Therefore, inflammation-based hematologic markers may reflect, at least in part, the systemic consequences of these tumor-promoting signaling and inflammatory processes.

Accordingly, systemic inflammation-based biomarkers measurable through routine blood tests have gained increasing attention as prognostic indicators in various malignancies. Inflammation-based hematologic indices, such as the neutrophil-to-lymphocyte ratio (NLR), platelet-to-lymphocyte ratio (PLR), and systemic immune-inflammation index (SII), have been widely investigated as prognostic markers in various cancers, including HCC.^12–14 In recent years, newer composite indices such as the systemic inflammation response index (SIRI) have also been proposed, with emerging evidence linking high SIRI levels to inferior overall survival and progression-free survival in patients undergoing curative or palliative treatments.^15–17 However, direct comparisons of these inflammation-based markers in well-defined HCC subpopulations, such as compensated patients, remain limited.

To date, few studies have systematically compared multiple inflammation-based indices within the same clinical cohort, limiting the ability to assess their relative prognostic significance. Furthermore, the predictive value of newer indices, such as neutrophil-plus-monocyte-to-lymphocyte ratio and aggregate index of systemic inflammation, remains unexplored in HCC. Therefore, this study aimed to systematically compare seven CBC-derived inflammatory indices and determine their prognostic relevance in a large cohort of patients with compensated HCC.

Materials and Methods

Study Design and Participants

This retrospective cohort study included patients diagnosed with HCC at the First Affiliated Hospital of Anhui Medical University from August 2012 to May 2025. Eligible patients were identified through the hospital’s electronic medical record system. The inclusion criteria were as follows: (1) age ≥ 18 years at the time of HCC diagnosis; (2) diagnosis of HCC based on imaging criteria consistent with the EASL/AASLD guidelines (typical radiological features on multiphasic CT or MRI) or histopathological confirmation; (3) compensated liver function at baseline, defined as no history of ascites, hepatic encephalopathy, or variceal bleeding, and classified as Child-Pugh A; and (4) availability of the baseline blood cell counts required to calculate all seven CBC-derived inflammatory markers. Patients were excluded if they met any of the following criteria: (1) decompensated liver disease at baseline; (2) diagnosis of any other malignancy; (3) incomplete baseline data for CBC-derived inflammatory markers or unavailable survival follow-up information; or (4) lost to follow-up immediately after diagnosis. This study was conducted in accordance with the Declaration of Helsinki and was approved by the Institutional Review Board of the First Affiliated Hospital of Anhui Medical University. This study was reported according to the RECORD guidelines.

Definition of Inflammatory Indices

Laboratory data were obtained from the first available blood test at baseline, including neutrophil, lymphocyte, monocyte, platelet, and total white blood cell (WBC) counts, to calculate CBC-derived inflammatory indices. These included: Neutrophil-to-lymphocyte ratio (NLR): neutrophil count divided by lymphocyte count; Derived NLR (dNLR): neutrophil count divided by (WBC) count minus neutrophil count; Monocyte-to-lymphocyte ratio (MLR): monocyte count divided by lymphocyte count; Neutrophil-plus-monocyte-to-lymphocyte ratio (NMLR) (neutrophil count + monocyte count) divided by lymphocyte count; Systemic immune-inflammation index (SII): (platelet count × neutrophil count) divided by lymphocyte count; Systemic inflammatory response index (SIRI): (neutrophil count × monocyte count) divided by lymphocyte count; Aggregate index of systemic inflammation (AISI): (neutrophil count × monocyte count × platelet count) divided by lymphocyte count.

Covariates

Demographic and clinical information was extracted from electronic medical records, including age, sex, etiology of liver disease (eg, HBV, HCV, alcohol), tumor characteristics (size, number, BCLC stage), and treatments (antiviral therapy, TACE, resection, etc).

Outcome Definition

All patients had compensated hepatocellular carcinoma at baseline. The primary outcome of this study was overall survival (OS). OS was calculated from the date of initial diagnosis to death from any cause or the date of the last documented follow-up, whichever came first. Survival status was ascertained through review of death certificates or, when necessary, telephone contact with patients or their family members. Follow-up was administratively censored on November 30, 2025.

Statistical Analysis

Continuous variables were expressed as medians with interquartile ranges (IQRs), while categorical variables were presented as counts and percentages. Baseline blood cell counts required for calculation of the seven CBC-derived inflammatory markers were complete in all included patients. Missingness was present in several covariates, and these variables were imputed using multiple imputation by chained equations (MICE), with 20 imputed datasets generated. The primary exposure variables and survival outcome were not imputed.

Spearman correlation analysis was performed to explore the relationships among inflammation-based indices. Least absolute shrinkage and selection operator (LASSO) Cox regression was conducted using the cv.glmnet function with family = “cox”, alpha = 1, and 10-fold cross-validation, and the optimal penalty parameter was selected according to the lambda.1se criterion. Restricted cubic spline (RCS) models were used to determine optimal cutoff values for the selected markers (three knots). Survival analysis was conducted using the Kaplan–Meier method, and differences between groups were compared using the Log rank test. Multivariable Cox proportional hazards regression models were constructed to estimate hazard ratios (HRs) and 95% confidence intervals (CIs). The proportional hazards assumption was formally assessed using Schoenfeld residuals and the results showed no significant violation of the proportional hazards assumption in the Cox models. Because several CBC-derived inflammatory indices were mathematically related and highly correlated, they may capture overlapping inflammatory information. Therefore, MLR, NMLR, and SIRI were evaluated in separate Cox regression models rather than being entered simultaneously into a single model. To assess the robustness of the findings, sensitivity analyses were conducted using the complete-case dataset, in which MLR, NMLR, and SIRI were modeled as continuous variables in Cox regression analyses.

All statistical analyses were performed using R software (version 4.4.1, http://www.R-project.org/). A two-sided P-value < 0.05 was considered statistically significant.

Results

Baseline Characteristics and Overall Survival

A total of 2802 HCC patients were included in this study. The median age was 59.00 years [IQR: 52.00–68.00], with 51.4% of patients were younger than 60 years. The majority of patients were male (83.5%). Detailed baseline characteristics of the study population were presented in Table 1. The median follow-up time was 45.7 months [44.8–49.3 months], and the 5-year OS rate was 67.6%.

Table 1.

Baseline Characteristics of HCC Patients

Variables	Overall (n=2802)
Age (median [IQR])	59.00 [52.00, 68.00]
Age (%)
<60	1441 (51.4)
≥60	1361 (48.6)
Gender (%)
Male	2341 (83.5)
Female	461 (16.5)
NLR (median [IQR])	2.36 [1.62, 3.42]
dNLR (median [IQR])	0.87 [0.82, 0.90]
MLR (median [IQR])	0.27 [0.20, 0.36]
NMLR (median [IQR])	2.63 [1.84, 3.80]
SIRI (median [IQR])	0.84 [0.51, 1.45]
SII (median [IQR])	338.34 [197.84, 579.34]
AISI (median [IQR])	123.02 [60.51, 253.34]
ALB (median [IQR])	41.10 [37.40, 44.40]
CR (median [IQR])	67.60 [58.00, 77.00]
TBil (median [IQR])	15.81 [11.90, 21.09]
INR (median [IQR])	1.03 [0.96, 1.10]
Alkaline Phosphatase (median [IQR])	101.00 [77.20, 139.00]
LDH (median [IQR])	199.00 [169.00, 245.00]
WBC (median [IQR])	5.26 [4.08, 6.67]
NEU (median [IQR])	3.24 [2.32, 4.50]
LYC (median [IQR])	1.38 [1.03, 1.79]
MONO (median [IQR])	0.37 [0.28, 0.49]
RBC (median [IQR])	4.41 [4.01, 4.78]
HB (median [IQR])	136.00 [124.00, 148.00]
PLT (median [IQR])	149.00 [104.00, 199.00]
AFP (median [IQR])	39.76 [4.41, 893.01]
CEA (median [IQR])	2.12 [1.32, 3.30]

Note: Continuous variables are presented as median and interquartile range (IQR), and categorical variables are presented as number (percentage).

Abbreviations: NLR, neutrophil-to-lymphocyte ratio; dNLR, derived neutrophil-to-lymphocyte ratio; MLR, monocyte-to-lymphocyte ratio; NMLR, neutrophil plus monocyte-to-lymphocyte ratio; SIRI, systemic inflammation response index; SII, systemic immune-inflammation index; AISI, aggregate index of systemic inflammation; ALB, albumin; CR, creatinine; TBil, total bilirubin; INR, international normalized ratio; LDH, lactate dehydrogenase; WBC, white blood cell count; NEU, neutrophil count; LYC, lymphocyte count; MONO, monocyte count; RBC, red blood cell count; HB, hemoglobin; PLT, platelet count; AFP, alpha-fetoprotein; CEA, carcinoembryonic antigen.

Correlation Among Inflammatory Indices

NLR showed a strong positive correlation with NMLR (r = 0.998, P < 0.001). Significant correlations were also observed between NMLR and SII (r = 0.792), SIRI and AISI (r = 0.855), and SII and AISI (r = 0.864), all with P < 0.001. A weak negative correlation was found between MLR and dNLR (r = −0.044, P < 0.05, Figure 1).

Figure 1.

Correlation matrix of seven systemic inflammatory indices, including NLR, dNLR, MLR, NMLR, SIRI, SII, and AISI. The lower triangle shows pairwise scatterplots with marginal density distributions; the upper triangle displays Spearman correlation coefficients with significance levels. ***P < 0.001, *P < 0.05.

Abbreviations: NLR, neutrophil-to-lymphocyte ratio; dNLR, derived neutrophil-to-lymphocyte ratio; MLR, monocyte-to-lymphocyte ratio; NMLR, neutrophil plus monocyte-to-lymphocyte ratio; SIRI, systemic inflammation response index; SII, systemic immune-inflammation index; AISI, aggregate index of systemic inflammation.

Variable Selection Using LASSO Regression

LASSO Cox regression was conducted to identify prognostic inflammatory indices. The optimal penalization parameter was determined based on 10-fold cross-validation, as shown in Figure 2B. At this lambda value, three variables were selected, including MLR, NMLR, and SIRI (Figure 2A).

Figure 2.

LASSO regression for variable selection. (A) Coefficient profiles of candidate variables as a function of the regularization parameter (log(λ)). (B) Ten-fold cross-validation for tuning parameter selection. The two dotted vertical lines indicate lambda.min, the value of λ yielding the minimum mean cross-validated partial likelihood deviance, and lambda.1se, the largest value of λ within one standard error of the minimum. Variables for the final model were selected based on lambda.1se.

Prognostic Impact of MLR, NMLR, and SIRI

Based on RCS analysis, optimal cutoff values for MLR, NMLR, and SIRI were identified as 0.27, 2.62, and 0.84, respectively. Patients were stratified into high and low groups according to these thresholds (Table S1). Kaplan–Meier survival analysis demonstrated that individuals with high levels of MLR, NMLR, or SIRI exhibited worse OS compared to those with lower levels (log-rank P < 0.001 for all comparisons, Figure 3). In multivariable Cox regression models adjusted for age and sex, elevated MLR (≥0.27), NMLR (≥2.62), and SIRI (≥0.84) remained independently associated with poorer OS (Table 2).

Figure 3.

Kaplan-Meier survival curves for overall survival in patients with HCC stratified by RCS-derived cutoff values of (A) MLR (cutoff = 0.27), (B) NMLR (cutoff = 2.62), and (C) SIRI (cutoff = 0.84).

Abbreviations: HCC, hepatocellular carcinoma; RCS, restricted cubic spline; MLR, monocyte-to-lymphocyte ratio; NMLR, neutrophil plus monocyte-to-lymphocyte ratio; SIRI, systemic inflammation response index.

Table 2.

Cox Regression Analysis of MLR, NMLR, and SIRI for OS in HCC

Variables	HR (95% CI)	P
MLR
<0.27	Ref
≥0.27	1.972 (1.704–2.283)	<0.001
NMLR
<2.62	Ref
≥2.62	1.822 (1.576–2.107)	<0.001
SIRI
<0.84	Ref
≥0.84	1.736 (1.503–2.006)	<0.001

Note: Hazard ratios (HRs) and 95% confidence intervals (CIs) were presented for each dichotomized variable based on spline-derived cutoff values.

Abbreviations: HR, hazard ratio; CI, confidence interval; MLR, monocyte-to-lymphocyte ratio; NMLR, neutrophil plus monocyte-to-lymphocyte ratio; SIRI, systemic inflammation response index.

Sensitivity Analysis

Sensitivity analyses were performed using the complete-case dataset, in which MLR, NMLR, and SIRI were modeled as continuous variables. All three markers remained significantly associated with poorer OS, with effect sizes and statistical significance generally consistent with the primary analyses based on the RCS-derived cutoff values (Figure S1).

Discussion

In this study, seven CBC-derived inflammatory markers were systematically evaluated in a large cohort of patients with compensated HCC. Among them, MLR, NMLR, and SIRI emerged as independent predictors of overall survival. These findings support the prognostic importance of systemic inflammation in compensated HCC and suggest that simple, inexpensive, and routinely available hematologic indices may provide useful prognostic information in this population.

In the present study, high MLR was associated with inferior OS in patients with compensated HCC. The spline-based modeling further supported a nonlinear relationship between MLR and mortality risk and identified 0.27 as the optimal prognostic threshold. Notably, when all seven CBC-derived indices were evaluated concurrently, MLR showed the strongest effect size among individual ratios, and remained significant even after comprehensive adjustment. Previous investigations have demonstrated that elevated MLR predicts poor prognosis in heterogeneous HCC populations, particularly in advanced-stage disease or after treatments such as TACE or immunotherapy.^18,19 Wang et al reported that elevated pre‑treatment MLR was associated with increased recurrence and reduced disease‑free survival in patients with resectable HCC.¹⁹ However, the magnitude and consistency of this association have varied across studies, possibly owing to differences in cohort size, disease stage composition, treatment heterogeneity, liver functional reserve, follow-up duration, and population characteristics. In contrast, our study specifically focused on compensated patients with Child-Pugh A status and included more than 2800 participants, representing one of the largest cohorts to date for evaluating MLR. An elevated MLR reflects an imbalance between increased peripheral monocytes and reduced lymphocyte-mediated anti-tumor immunity.^20,21 In HCC, circulating monocytes can be recruited into tumor tissue and differentiate into tumor-associated macrophages, which promote angiogenesis, extracellular matrix remodeling, and immune suppression within the tumor microenvironment.^22,23 Conversely, a lower lymphocyte count may indicate weakened adaptive immune surveillance and reduced cytotoxic antitumor activity.²⁴ Therefore, a high MLR may represent a systemic signature of a more permissive tumor microenvironment, in which myeloid inflammation predominates over effective immune control, ultimately contributing to adverse survival outcomes.²⁵

In our study, an elevated NMLR was independently associated with worse OS among patients with compensated HCC, even after adjustment for known clinical confounders. This finding suggests that NMLR captures distinct inflammatory dynamics relevant to cancer progression. Compared to traditional markers such as NLR or PLR, NMLR may offer additional prognostic value by incorporating neutrophil elevation alongside monocyte-mediated inflammation. Although research specifically investigating NMLR in HCC is currently limited, evidence from other clinical contexts supports the potential utility of this composite inflammatory marker. Although these studies do not involve cancer populations, they establish that NMLR captures systemic immune alterations relevant to adverse outcomes across disease states. Peripheral blood ratios that incorporate both neutrophil and monocyte counts relative to lymphocytes therefore provide a more comprehensive reflection of the balance between pro‑inflammatory innate responses and adaptive antitumor immunity than single ratios such as NLR or MLR alone. Evidence from a broad cancer literature establishes that elevated NLR and low LMR are associated with poorer prognosis across multiple tumor types, highlighting the clinical relevance of leukocyte dynamics in disease progression.^26,27

In our cohort, a higher SIRI was independently associated with poorer OS in compensated HCC patients after adjustment for established clinical confounders. This finding implies that SIRI captures aspects of the host inflammatory milieu that are relevant to cancer progression and survival beyond traditional clinical predictors. Subsequent studies across diverse malignancies have consistently reported that elevated SIRI is associated with worse prognosis. For example, a recent retrospective analysis in colorectal cancer showed that high preoperative SIRI was significantly correlated with both reduced overall and recurrence‑free survival.²⁸ In advanced non‑small cell lung cancer treated with immune checkpoint inhibitors, elevated SIRI predicted shorter progression‑free and overall survival, underscoring its relevance in the immunotherapy era.²⁹ Meta‑analytic evidence further supports that high SIRI is a consistent predictor of poor survival across multiple cancer types, including hepatobiliary, gastric, and urologic malignancies.³⁰ By incorporating both neutrophil and monocyte elevation relative to lymphocytes, SIRI provides a more comprehensive representation of the host immune status than traditional CBC-derived ratios, which may explain its stronger association in our study. These alterations may also partly reflect an immunosuppressive tumor microenvironment in HCC, where myeloid-driven inflammation and weakened lymphocyte-mediated surveillance interact with cytokine signaling pathways such as IL-6/JAK/STAT3 and VEGF-related signaling to facilitate tumor progression.³¹

The present findings suggest that MLR, NMLR, and SIRI may provide additional prognostic information in compensated HCC. In combination with established clinicopathologic parameters, these hematologic indices may contribute to a more comprehensive assessment of prognosis. However, their clinical utility should be further evaluated in independent cohorts.

This study benefits from a relatively large and well-characterized cohort of patients with compensated HCC, which reduces confounding related to advanced liver dysfunction. By focusing on Child-Pugh A individuals without prior decompensation, the study specifically evaluated the prognostic relevance of systemic inflammation in a clinically important and relatively understudied population. In addition, the use of multiple imputation for incomplete covariates, LASSO-based variable selection, and restricted cubic spline analysis strengthened the analytical framework. Clinically, MLR, NMLR, and SIRI may provide complementary prognostic information in patients with compensated HCC. Because these indices are inexpensive and routinely available, they may help support risk stratification and closer follow-up in patients with less favorable inflammatory profiles. However, several limitations should be acknowledged. First, as a retrospective single-center study, the findings remain susceptible to selection bias, information bias, and limited generalizability. Second, although multivariable adjustment was performed, residual confounding from unmeasured or incompletely measured factors cannot be excluded. Third, all inflammatory markers were assessed only at baseline, and their longitudinal changes during follow-up were not available. Fourth, the cutoff values for MLR, NMLR, and SIRI were derived internally from the present dataset, which may introduce a degree of overfitting and limit direct applicability to other populations. Finally, no external validation cohort was available, and therefore the generalizability of these findings requires confirmation in independent prospective studies.

Conclusion

In conclusion, this study demonstrates that MLR, NMLR, and SIRI were independent prognostic indicators of overall survival in patients with compensated HCC. These simple and routinely available indices reflect the systemic inflammatory state and may serve as complementary prognostic indicators alongside existing clinical parameters in patients with compensated HCC. Further validation in independent cohorts is needed before clinical application.

Funding Statement

The authors declared that financial support was not received for this work and/or its publication.

Abbreviations

AISI, Aggregate index of systemic inflammation; CBC, Complete blood counts; CIs, Confidence intervals; dNLR, Derived NLR; HCC, Hepatocellular carcinoma; HRs, Hazard ratios; IQRs, Interquartile ranges; LASSO, Least absolute shrinkage and selection operator; MICE, Multiple imputation by chained equations; MLR, Monocyte-to-lymphocyte ratio; NLR, Neutrophil-to-lymphocyte ratio; NMLR, Neutrophil-plus-monocyte-to-lymphocyte ratio; OS, Overall survival; PLR, Platelet-to-lymphocyte ratio; RCS, Restricted cubic spline; SII, Systemic immune-inflammation index; SIRI, Systemic inflammation response index; WBC, White blood cell.

Data Sharing Statement

The datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request.

Ethics Approval and Consent to Participate

This study was approved by the Ethics Committee of the First Affiliated Hospital of Anhui Medical University (Approval number: P2021-12-30). The Institutional Review Board waived the requirement for informed consent because this was a low-risk retrospective medical record review study using existing clinical data, with no direct patient contact and no impact on patient care. All data were anonymized or de-identified before analysis and were kept strictly confidential. The study was conducted in accordance with the Declaration of Helsinki.

Author Contributions

Ning Yang: Conceptualization, Data curation, Formal analysis, Visualization, Writing – original draft, Writing – review and editing. Lei Zhang: Formal analysis, Visualization, Writing – original draft. Shiyu Liu: Data curation, Writing – review and editing. Zhentao Li: Validation, Writing – review and editing. Xin Yang: Conceptualization, Writing – review and editing. Chuannan Wu: Data curation, Validation, Writing – review and editing. Tao Li: Conceptualization, Writing – review and editing. Hu Chen: Methodology, Resources, Writing – review and editing. Yongheng Wen: Data curation, Project administration, Supervision, Writing – review and editing. Yun Zheng: Conceptualization, Writing – review and editing. Guangxia Chen: Conceptualization, Funding acquisition, Writing – review and editing. All authors 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 research was carried out without any financial or commercial ties that might be seen as having a conflict of interest, the authors disclose.