Abstract
Purpose
The purpose of this study was to investigate the predictive capability of neutrophil‐to‐apolipoprotein A1 ratio (NAR) for predicting overall survival (OS) among patients with hepatocellular carcinoma (HCC) receiving transarterial chemoembolization (TACE).
Patients and Methods
We investigated the clinical features of 554 patients with HCC receiving TACE and assessed NAR's predictive value for OS with 222 patients (the discovery cohort) and 332 patients (the validation cohort). The association of NAR with circulation lectin‐type oxidized low‐density lipoprotein receptor‐1–positive (LOX‐1+) polymorphonuclear myeloid‐derived suppressor cells (PMN‐MDSCs) was illustrated.
Results
Multivariate Cox regression revealed that lymphocyte count; Tumor, Node, Metastasis (TNM) stage; and NAR were independent prognostic factors in the discovery cohort. The validation cohort confirmed the independent prognostic value of TNM stage and NAR. Patients with low NAR (<2.7) displayed significantly increased OS in the discovery cohort (59.8 months vs. 21 months), the validation group (38.0 months vs. 23.6 months), and the total cohort (44.1 months vs. 22.0 months). A Cox proportional hazards model was used to combine Cancer of the Liver Italian Program (CLIP) score with discretized NAR. C‐index illustrated that NAR‐integrated CLIP score was the best model compared with NAR and CLIP score. Furthermore, NAR‐CLIP presented superior predictive capacity for 10‐, 20‐, 30‐, 40‐, 50‐, and 60‐month survival compared with CLIP score by survival receiver‐operator characteristic analysis in the discovery cohort, validation cohort, and total cohort. NAR was significantly associated with LOX‐1+ PMN‐MDSCs by linear regression.
Conclusion
This study identified NAR as an independent predictor for OS among patients with HCC receiving TACE. NAR reflected circulation LOX‐1+ PMN‐MDSC level.
Implications for Practice
The present study identified neutrophil‐to‐apolipoprotein A1 ratio (NAR) as an independent predictor for overall survival among patients with hepatocellular carcinoma receiving transarterial chemoembolization. NAR reflected circulation level of lectin‐type oxidized low‐density lipoprotein receptor‐1–positive polymorphonuclear myeloid‐derived suppressor cells.
Keywords: Hepatocellular carcinoma, Transarterial chemoembolization, Overall survival, Neutrophil‐to‐apolipoprotein A1 ratio, Polymorphonuclear myeloid‐derived suppressor cells
Short abstract
This article reports on the predictive capability of neutrophil‐to‐apolipoprotein A1 ratio for predicting overall survival among patients with hepatocellular carcinoma receiving transarterial chemoembolization.
Introduction
The treatment of advanced hepatocellular carcinoma (HCC) has become increasingly complicated with multiple local, regional, and systematic treatments displaying improved efficacy [1]. Transarterial chemoembolization (TACE) was the fundamental therapy for locally advanced HCC and presented certain efficacy for patients with metastasis outside liver [2]. However, liver transplantation [3], radiofrequency ablation, percutaneous ethanol injection, and radiotherapy [4] also presented certain efficacy for advanced HCC [1]. Meanwhile, systematic treatment using lenvatinib seemed to present superior prognosis for than TACE in selected patients [5, 6]. Successful initial treatment is critically important for patients with HCC. Thus, in order to maximize the efficacy of multidisciplinary teamwork (MDT) in advanced HCC, patient selection has become increasingly important for various treatments, especially TACE.
Lipid metabolism has been found to be involved in dysfunction of myeloid cells. Hypochlorite‐oxidized low‐density lipoprotein possesses a substantial proinflammatory potential by modulating polymorphonuclear neutrophils [7]. Native high‐density lipoprotein (HDL) rapidly and efficiently protects resting neutrophils from respiratory burst activation by hypochlorite‐oxidized low‐density lipoprotein (LDL) [8]. Apolipoprotein A‐1 (apoA‐I), the major protein form of high‐density lipoprotein, has anti‐inflammatory, antiapoptotic, and antioxidant properties and inhibits the formation of tumor vessels [9]. Some studies have implied that apoA‐I might pose an obstacle to tumorigenesis and progression in HCC [8, 9, 10]. Furthermore, previous studies indicated that apoA‐I was an independent prognostic factor for patients with HCC [11, 12, 13, 14]. Moreover, apoA‐I suppresses of neutrophil activation, migration, and adhesion. In malignant conditions, myeloid cells are unable to differentiate effectively into mature neutrophils, are poorly phagocytic, and acquire immune‐suppressive potential. Dysfunction of lipid metabolism has been identified as key mechanism in the generation of polymorphonuclear myeloid‐derived suppressor cells (PMN‐MDSCs). Lectin‐type oxidized LDL receptor‐1 (LOX‐1) was reported to be the specific marker for human PMN‐MDSCs. The prognostic value of neutrophils‐containing indexes and PMN‐MDSCs has been reported for HCC. Thus, lipid metabolism, especially apoA‐I, and neutrophil counts might reflect tumor immunologic status. However, the clinical significance of the combination of serum apoA‐I and neutrophil counts levels have not been investigated in patients with HCC receiving TACE. Thus, we hypothesized that the combination of apoA‐I and neutrophils may be a potent prognostic indicator among patients with HCC receiving TACE.
Clinical blood indexes have also been found to predict the survival of patients with HCC, such as tumor markers (including α‐fetoprotein [AFP], etc.), systemic inflammation (including lymphocyte‐to‐monocyte ratio, platelet‐to‐lymphocyte ratio, etc.), and lipid metabolism (including apolipoprotein A, total cholesterol, etc.) [15, 16, 17, 18, 19, 20, 21]. However, few of them have achieved worldwide acknowledgment. The latent reason might lie in the uninvestigated mechanism of those prognostic factors [22]. In this study, we investigated its prognostic value of neutrophil‐to‐apolipoprotein A ratio (NAR) among patients with HCC receiving TACE in two independent cohorts and its latent immunological mechanism.
Materials and Methods
Patients
During the period between June 2009 and September 2016, 554 patients with HCC receiving TACE as initial therapy were included into the present study, with 222 from the Third Affiliated Hospital of Sun Yat‐sen University, 191 from the Lin‐Nan Hospital of Sun Yat‐sen University, 94 from the Third People's Hospital of Shenzhen, and 47 from the Third Affiliated Hospital of Guangzhou Medical University. The diagnosis of HCC was confirmed by pathology or the American Association for the Study of Liver Diseases radiological criteria by either computed tomography or magnetic resonance imaging. All patients were also screened for serum HIV antibody, hepatitis B surface antigen, hepatitis C virus antibody, hepatitis D virus (HDV) antigen, and HDV antibody. Patients who were positive for HIV, presented acute infections (including pneumonia, urinary tract infection, etc.), were pregnant, received systematic corticosteroids or immunosuppressive agents, or had fever were excluded from this study. This study was approved by the Clinical Ethics Review Board of the Third Affiliated Hospital of Sun Yat‐sen University, the Lin‐Nan Hospital of Sun Yat‐sen University, the Third Affiliated Hospital of Guangzhou Medical University, and the Third People's Hospital of Shenzhen. Written informed consent was obtained from all the patients at the time of admission.
Data Collection
The clinical baseline data and biomarkers included in our study were measured in each included patient before TACE within 1 week using laboratory devices in our hospital. Glucose, aspartate transaminase (AST), alanine aminotransferase (ALT), gamma‐glutamyl transpeptidase (GGT), albumin (ALB), triglyceride (TG), HDL, LDL, apoA‐I, and apolipoprotein B (APOB) were included in a biochemical test performed using a Hitachi Automatic Analyzer 7600–020 (Tokyo, Japan), and AFP in serum tumor marker tests was measured using a Roche Elecsys 2010 Chemistry Analyzer (Basel, Switzerland). White blood cells, neutrophils, lymphocytes, monocytes, hemoglobin, and platelets were measured by routine blood examination (XE‐5000TM Automated Hematology System, Sysmex, Japan). The normal ranges of white blood cell (WBC), neutrophil, lymphocyte, monocyte, hemoglobin （HGB, platelets, glucose, total cholesterol (TC), TG, HDL, LDL, apoA‐I, APOB, ALT, AST, ALB, GGT, total bilirubin, blood urea nitrogen, and AFP levels in blood were 3.5–9.5 × 109/L, 0.4–0.75 × 109/L, 0.2–0.5 × 109/L, 0.03–0.1 × 109/L, 115–150 g/L, 100–350 × 109/L. 3.9–6.19 mmol/L, 3.1–5.7 mmol/L, 0.34–1.92 mmol/L, 1.0–1.6 g/L, 0.6–1.1 g/L, 3–35 U/L, 13–35 U/L, 36–51 g/L, 7–45 U/L, 4.0–23.9 μmol/L, 2.4–8.2 mmol/L, and 0–8.1 ng/mL, respectively. Level of NAR was defined as absolute neutrophil count divided by the level of apolipoprotein A as specific value; it did not have a standard normal range. Cirrhosis was diagnosed by a radiologist. The level of LOX‐1–positive (LOX‐1+) PMN‐MDSCs available closest to the time of TACE was collected and used for analysis. Patients were followed until the last follow‐up or death.
Flow Cytometric Analysis
LOX‐1+ PMN‐MDSC level was tested using the blood sample for baseline blood cell counts. Blood sample were analyzed within 6 hours after sampling and underwent red blood cell lysis before flow cytometric analysis. The following anti‐human antibodies were purchased from eBioscience (San Diego, CA): CD15‐eFluor450, LOX‐1–Allophycocyanin (APC), and their corresponding isotype controls. The cell phenotypes were analyzed by flow cytometry on a flow cytometer FACSAria II flow cytometer (BD Biosciences, San Jose, CA), and data were analyzed with FlowJo version 10.0.7 (FlowJo, Ashland, OR).
Statistical Analysis
Overall survival (OS) was the main endpoint of this study, which was defined as the time between the first diagnosis and death or the last follow‐up. Maximally selected rank statistics, provided in the survminer package of R software, were used to determine the optimal cutoff point for continuous variables. Multivariate analysis using a stepwise Cox proportional hazards model was used to test for the independent significance of baseline characteristics and explanatory variables. The performance of relevant parameters was assessed using the Kaplan‐Meier method, and differences in survival between groups were compared using the log‐rank test. Hazard ratios (HRs) and 95% confidence intervals were pooled to measure the effects of relevant parameters on prognosis. An HR greater than 1 indicated a worse prognosis in patients with a relevant parameter, whereas an HR less than 1 indicated a better prognosis. Survival receiver‐operator characteristic (ROC) curves of relevant parameters, to predict 1‐year OS for patients with HCC, were plotted using the survivalROC package in R software. We calculated the area under the curves (AUC) of the 1‐year OS to evaluate the prognostic and predictive accuracy. The timeROC package showed the dynamic change of AUC value on predicting 10‐, 20‐, 30‐, 40‐, 50‐, and 60‐month survival. The criterion for statistical significance was set at a α of .05, and all p values were based on two‐sided tests. All statistical analyses were performed with R statistical software version 3.6.0 and IBM SPSS Statistics software version 22.
Results
Clinicopathological Characteristics
We identified 554 patients. Of these patients, 222 from the Third Affiliated Hospital of Sun Yat‐sen University were included in the discovery cohort and 332 from other three hospitals composed the validation cohort. For the discovery and validation cohorts, median OS was 19.0 and 17.0 months, respectively. For the discovery and the validation cohorts, median time to follow‐up was 43.3 and 44.8 months, respectively. Baseline clinicopathologic characteristics are summarized in Table 1.
Table 1.
Clinical characteristics of the study patients in discovery and validation cohort
| Characteristic | Discovery cohort (n = 222) | Validation cohort (n = 332) | p value |
|---|---|---|---|
| Age (years) | 51 (22–78) | 52.3 (12–86) | .065 |
| Sex | .675 | ||
| Male | 203 (91.4) | 297 (90.0) | |
| Female | 19 (8.6) | 35 (10.0) | |
| White blood cells (×109/L) | 5.48 (1.95–15.3) | 5.41 (1.54–26.84) | .839 |
| Neutrophil (×109/L) | 3.16 (0.51–10.9) | 3.35 (0.79–23.95) | .266 |
| Lymphocyte (×109/L) | 1.41 (0.46–4.74) | 1.34 (0.14–8.87) | .225 |
| Monocyte (×109/L) | 0.41 (0.08–11) | 0.42 (0–1.65) | .048 |
| HGB (g/L) | 135 (60–177) | 132 (66–192) | .683 |
| Platelets (×109/L) | 154 (23–466) | 160 (3–682) | .926 |
| Glucose (mmol/L) | 4.92 (2.16–22.09) | 5.13 (0.91–25.81) | .05 |
| TC (mmol/L) | 4.39 (1.8–14.51) | 4.33 (1.51–10.12) | .127 |
| TG (mmol/L) | 0.95 (0.18–7.66) | 0.96 (0.14–11.8) | .827 |
| HDL (mmol/L) | 1.04 (0.11–2.16) | 1.07 (0.14–2.76) | .241 |
| LDL (mmol/L) | 2.82 (1.01–12.68) | 2.81 (0.21–9.44) | .055 |
| APOA (g/L) | 1.23 (0.23–2.03) | 1.24 (0.14–2.13) | .423 |
| APOB (g/L) | 0.82 (0.27–2.8) | 0.79 (0.16–2.41) | .3 |
| LPa (mg/L) | 71 (0–833.8) | 61 (0–799) | .129 |
| AFP (ng/mL) | 172.4 (1–33,104) | 106.7 (0–41,726) | .899 |
| ALT (U/L) | 43 (5–502) | 45 (9–387) | .571 |
| AST (U/L) | 48 (14–842) | 55 (9–931) | .154 |
| ALB (g/L) | 38.5 (24.6–50.5) | 38.5 (20.4–54.1) | .624 |
| GGT (U/L) | 103 (11–878) | 111 (11–1,536) | .267 |
| TB (μmol/L) | 16.5 (1.8–197.2) | 17.3 (2.1–93.9) | .653 |
| BUN (mmol/L) | 4.61 (2.13–22.14) | 4.91 (2.07–13.39) | .476 |
| NAR | 2.70 (0.45–23.6) | 2.66 (0.56–69.3) | .356 |
| TNM 7th edition stage | .237 | ||
| I | 22 (10.2) | 27 (8.3) | |
| II | 47 (21.9) | 60 (18.4) | |
| III | 85 (39.5) | 158 (48.5) | |
| IV | 61 (28.4) | 81 (24.8) | |
| Child‐Pugh score | .736 | ||
| 5 | 77 (23.2) | 128 (38.6) | |
| 6 | 59 (26.6) | 89 (26.8) | |
| 7 | 57 (25.7) | 77 (23.2) | |
| 8 | 25 (11.3) | 28 (8.4) | |
| 9 | 2 (0.9) | 5 (1.5) | |
| 10 | 2 (0.9) | 5 (1.5) | |
| CLIP score | .266 | ||
| 0 | 25 (11.3) | 48 (14.5) | |
| 1 | 51 (23.0) | 87 (26.2) | |
| 2 | 62 (27.9) | 72 (21.7) | |
| 3 | 46 (20.7) | 55 (16.6) | |
| 4 | 29 (13.1) | 57 (17.2) | |
| 5 | 9 (4.1) | 13 (3.9) |
Data are presented as median (range) or numbers (percentage). Continuous variables were compared using Mann‐Whitney U test and categorical variables using Pearson's chi‐square or Fisher's exact probability test. A value of p < .05 is considered significant.
Abbreviations: AFP, α‐fetoprotein; ALB, albumin; ALT, alanine aminotransferase; APOA, apolipoprotein a; APOB, apolipoprotein b; AST, aspartate transaminase; BUN, blood urea nitrogen; CLIP, Cancer of the Liver Italian Program; GGT, gamma‐glutamyl transpeptidase; HDL, high‐density lipoprotein; HGB, hemoglobin; LDL, low‐density lipoprotein; LPa, lipoprotein a; NAR neutrophil‐to‐apolipoprotein a ratio; TB, total bilirubin; TC, total cholesterol; TG, triglyceride; TNM, Tumor, Node, Metastasis.
Identification NAR as the Predictors for OS
Based on the discovery cohort, univariate analysis by Cox regression revealed that WBC, neutrophil count, lymphocyte count, TC, LCL cholesterol, apoA‐I, T stage, distant metastasis, Tumor, Node, Metastasis (TNM) stage, glutamic oxaloacetic transaminase (AST), GGT, Child‐Pugh score, Cancer of the Liver Italian Program (CLIP) score, and NAR were associated with unfavorable OS. In order to avoid multicollinearity in the multivariate analysis, WBC, neutrophil count, apoA‐I, T stage, and distant metastasis were not included in further analysis. The interaction between NAR and CLIP score or TNM stage was not significant (supplemental online Table 1), which justified its inclusion into multivariate analysis. Validation cohort was used to confirm the results from discovery cohort. Univariate analysis, including lymphocyte count, TC, LDL, TNM stage, GGT, AST, and NAR, revealed that lymphocyte count and NAR were the prognostic factors for patients with HCC receiving TACE. Multivariate Cox regression found that TNM stage and NAR were the independent predictors for OS among patients with HCC receiving TACE in the validation group (Table 2). Above all, NAR was identified as an independent prognostic factor for patients with HCC receiving TACE.
Table 2.
Significant univariate and multivariate Cox regression analyses of factors associated with overall survival in discovery and validation cohorts
| Variable | Univariate analysis | Multivariate analysis | ||||
|---|---|---|---|---|---|---|
| Discovery cohort | Discovery cohort | Validation cohort | ||||
| HR (95% CI) | p value | HR (95% CI) | p value | HR (95% CI) | p value | |
| Sex | 0.92 (0.48–1.76) | .803 | — | — | — | — |
| Age | 1.00 (0.98–1.02) | .923 | — | — | — | — |
| White blood cells | 1.14 (1.06–1.22) | <.001 | — | — | — | — |
| Neutrophil | 1.14 (1.05–1.24) | .002 | — | — | — | — |
| Lymphocyte | 1.43 (1.07–1.93) | .018 | 1.62 (1.19–2.20) | .002 | 0.89 (0.70–1.15) | .387 |
| Monocyte | 1.12 (0.96–1.30) | .150 | — | — | — | — |
| HGB | 1.00 (1.00–1.01) | .545 | — | — | — | — |
| Platelets | 1.00 (1.00–1.00) | .183 | — | — | — | — |
| Glucose | 0.98 (0.90–1.08) | .711 | — | — | — | — |
| TC | 1.15 (1.02–1.30) | .021 | 1.07 (0.75–1.53) | .715 | 0.91 (0.70–1.18) | .474 |
| TG | 1.09 (0.86–1.40) | .470 | — | — | — | — |
| HDL | 0.60 (0.36–1.01) | .055 | — | — | — | — |
| LDL | 1.17 (1.03–1.33) | .011 | 0.93 (0.63–1.38) | .735 | 1.08 (0.80–1.47) | .596 |
| APOA | 0.45 (0.26–0.79) | .005 | — | — | — | — |
| APOB | 1.45 (0.91–2.30) | .114 | — | — | — | — |
| LPa | 1.00 (1.00–1.00) | .720 | — | — | — | — |
| AFP | 1.00 (1.00–1.00) | .957 | — | — | — | — |
| ALT | 1.00 (1.00–1.00) | .113 | — | — | — | — |
| AST | 1.01 (1.00–1.01) | <.001 | 1.01 (1.00–1.01) | <.001 | 1.00 (1.00–1.00) | .666 |
| ALB | 1.01 (0.97–1.05) | .542 | — | — | — | — |
| GGT | 1.00 (1.00–1.00) | .041 | 1.00 (1.00–1.00) | .365 | 1.00 (1.00–1.00) | .513 |
| TB | 0.99 (0.99–1.00) | .287 | — | — | — | — |
| BUN | 1.09 (0.98–1.22) | .111 | — | — | — | — |
| T stage | 1.57 (1.29–1.91) | <.001 | — | — | — | — |
| N stage | 1.12 (0.72–1.74) | .612 | — | — | — | — |
| M stage | 2.76 (1.43–5.31) | .003 | — | — | — | — |
| TNM stage | 1.41 (1.15–1.73) | <.001 | 1.30 (0.99–1.69) | .056 | 1.27 (1.02–1.58) | .030 |
| Child‐Pugh score | 0.83 (0.70–0.99) | .039 | — | — | — | — |
| CLIP score | 1.23 (1.07–1.41) | .004 | 1.03 (0.85–1.24) | .772 | 1.06 (0.93–1.21) | .363 |
| NAR | 1.07 (1.03–1.12) | .002 | 1.05 (1.00–1.11) | .044 | 1.02 (1.01–1.04) | .007 |
Abbreviations: —, not applicable; AFP, α‐fetoprotein; ALB, albumin; ALT, alanine aminotransferase; APOA, apolipoprotein a; APOB, apolipoprotein b; AST, aspartate transaminase; BUN, blood urea nitrogen; CI, confidence interval; CLIP, Cancer of the Liver Italian Program; GGT, gamma‐glutamyl transpeptidase; HDL, high‐density lipoprotein; HGB, hemoglobin; HR, hazard ratio; LDL, low‐density lipoprotein; LPa, lipoprotein a; NAR neutrophil‐to‐apolipoprotein a ratio; TB, total bilirubin; TC, total cholesterol; TG, triglyceride; TNM, Tumor, Node, Metastasis.
The discretization of NAR, which was a continuous variable, was performed by maximally selected rank statistics (maxstat) to determine the optimal cutoff point for OS among discovery cohort. The cutoff point for NAR was 2.7 (Fig. 1A). Kaplan‐Meier curve illustrated that patients with low NAR displayed significantly increased OS in the discovery cohort, with doubled median OS compared with patients with NAR (59.8 months vs. 21 months; Fig. 1B). Similarly, patients with low NAR in the validation group also displayed improved OS (38.0 months vs. 23.6 months; Fig. 1C). Above all, Kaplan‐Meier analysis revealed that NAR influenced OS for the total cohort (44.1 months vs. 22.0 months; Fig. 1D).
Figure 1.
Survival analysis of patients with hepatocellular carcinoma (HCC) with different NAR level. (A): The discretization of NAR by maximally selected rank statistics to determine the optimal cutoff point for overall survival among discovery cohort. (B–D): Kaplan‐Meier survival curves of overall survival for the subgroups with different levels of NAR in the discovery cohort (B), validation cohort (C), and total cohort (D) of patients with HCC receiving transarterial chemoembolization. Abbreviations: CI, confidence interval; HR, hazard ratio; NAR, neutrophil‐to‐apolipoprotein A1 ratio.
Predictive Value of NAR on Patients with Early‐ and Late‐Stage HCC
The CLIP system is most acknowledged prognostic system for HCC; its independent prognostic value was confirmed by the present study. Thus, in order to specify the clinical usage of NAR, patients with HCC in the total group were divided into early and late stage by CLIP score and TNM stage. Kaplan‐Meier analysis was used to determine the prognostic value of NAR in different stages of HCC. As a result, patients with different NAR presented significantly variant survival among patients with early‐stage disease (TNM stage I–II and CLIP score 0–2; Fig. 2A and C). NAR has prognostic value among patients with TNM stages III–IV instead of patients with CLIP score 3–5 (Fig. 2B and D). Thus, NAR might be more clinically useful among patients with early stages of HCC.
Figure 2.
Kaplan‐Meier survival curves of overall survival for the subgroups with different levels of neutrophil‐to‐apolipoprotein A1 ratio (NAR) in patients with hepatocellular carcinoma with Tumor, Node, Metastasis (TNM) stage I–II (A), TNM stage III–IV (B), Cancer of the Liver Italian Program (CLIP) score 0–2 (C), and CLIP score 3–5 (D) from the total cohort. Abbreviations: CI, confidence interval; HR, hazard ratio; NAR, neutrophil‐to‐apolipoprotein A1 ratio.
NAR Integration Improved the Prognostic Value of CLIP Score
To investigate the clinical usage of NAR, the distribution of NAR among different CLIP scores was analyzed by violin plot. It seems that NAR was lower in patients with lower CLIP score; however, the different between different CLIP scores was of no clinical importance (supplemental online Fig. 1). Cox proportional hazards model was used to combine CLIP score with discretized NAR. The discretization of NAR‐CLIP score was performed by the maxstat statistic to determine the optimal cutoff point for OS among discovery cohort. The cutoff point for NAR was 0.61. (Fig. 3A). Kaplan‐Meier curves were used to illustrate the prognosis of patients with different NAR‐CLIP scores. Patients with high and low NAR‐adjusted CLIP score presented significantly different OS in the discovery cohort (65.0 months vs. 20.5 months), validation cohort (40.4 months vs. 17.1 months), and total cohort (44.7 months vs. 18.2 months; Fig. 3B–D). C‐index illustrated that NAR‐CLIP is the best model compared with NAR and CLIP score system, with NAR presenting similar results (supplemental online Table 2).
Figure 3.
Survival analysis of patients with hepatocellular carcinoma (HCC) with different NAR‐adjusted CLIP score. (A): The discretization of NAR‐CLIP by maximally selected rank statistics to determine the optimal cutoff point for overall survival among discovery cohort. (B–D): Kaplan‐Meier survival curves of overall survival for the subgroups with different levels of NAR‐CLIP score in the discovery cohort (B), validation cohort (C), and total cohort (D) of patients with HCC receiving transarterial chemoembolization.
Abbreviations: CI, confidence interval; CLIP, Cancer of the Liver Italian Program; HR, hazard ratio; NAR, neutrophil‐to‐apolipoprotein A1 ratio.
NAR Integration Improved the Accuracy of CLIP Score in Predicting Specific Time Point Survival Rate
To further support the result, ROC analysis was conducted to illustrate the predictive value of NAR and NAR‐adjusted CLIP score for 1‐year OS in the total cohort. NAR presented similar AUC compared with CLIP score, with NAR‐CLIP displaying highest AUC (Fig. 4 A, C, E). Then, the predictive value on 10‐, 20‐, 30‐, 40‐, 50‐, and 60‐month survival was compared using ROC analysis between NAR, CLIP score, and NAR‐CLIP. NAR‐CLIP and NAR presented superior predictive capacity compared with CLIP score. In the validation cohort and total cohort, NAR‐CLIP displayed the best result. (Fig. 4 B, D, F).
Figure 4.
Time survival receiver‐operator characteristic (ROC) curves for NAR, CLIP score, and NAR‐CLIP to predict 1‐year overall survival among the discovery cohort (A), validation cohort (C), and total cohort (E). The predictive value of NAR, CLIP score, and NAR‐CLIP on 10‐, 20‐, 30‐, 40‐, 50‐, and 60‐month survival compared using ROC analysis of the discovery cohort (B), validation cohort (D), and total cohort (F). Abbreviations: AUC, area under the curve; CLIP, Cancer of the Liver Italian Program; NAR, neutrophil‐to‐apolipoprotein A1 ratio.
NAR Associated with LOX‐1+ PMN‐MDSCs
PMN‐MDSCs are critical immunosuppressive cells in inflammation and malignancies, which suppress immune cell activation and induce immune tolerance. Notably, lipometabolism was reported to a be key regulator of myeloid‐derived suppressor cells, with LOX‐1 reported to be the specific marker for human PMN‐MDSCs. According to our previous study, LOX‐1+ PMN‐MDSCs in circulation suppressed immune response and promoted HCC development [23]. In the present study, the association of circulating frequencies of LOX‐1+ PMN‐MDSCs (LOX‐1+ CD15+) (Fig. 5A) in peripheral blood mononuclear cells with NAR were tested by linear regression, which showed that NAR was significantly associated with LOX‐1+ PMN‐MDSCs (Fig. 5B).
Figure 5.
Association of LOX‐1+ polymorphonuclear myeloid‐derived suppressor cells (PMN‐MDSCs) with neutrophil‐to‐apolipoprotein A1 ratio (NAR). (A): Gating strategy for LOX‐1+ PMN‐MDSCs. (B): Linear regression of LOX‐1+ PMN‐MDSCs with NAR. Abbreviations: CI, confidence interval; FSC‐A, forward scatter‐area; LOX‐1+, lectin‐type oxidized low‐density lipoprotein receptor‐1–positive; SSC‐A, side scatter‐area.
Discussion
Patient selection has become increasingly critical for treatment of advanced HCC, not only because successful initial therapy determines the long‐term prognosis of patients with cancer, but also because various treatments have promising efficacy for advanced HCC [1, 4, 24, 25]. TACE is still the fundamental remedy for advanced HCC [1], but it is challenged by radiotherapy, liver transplantation, and lenvatinib [3, 4, 5, 6]. Thus, patient selection for TACE becomes imperative for the MDT groups making decisions on advanced HCC. In the present study based on multiple centers, we found that NAR was an independent prognostic factor for patients with advanced HCC receiving TACE. NAR presented prognostic capacity similar to that of CLIP score, which was confirmed by multivariate Cox regression and survival ROC analysis. Furthermore, NAR‐adjusted CLIP score presented higher prognostic capacity on OS and survival on specific time points. Because NAR is much simpler than CLIP score, NAR will be more practical in daily practice.
ApoA‐I acts as an obstacle to cancer development by multiple mechanisms. Serum apoA‐I is associated with the diagnosis of many types of cancer and the appearance of metastases. A meta‐analysis revealed that apoA‐I levels also predicted the prognosis of patients with various cancers treated with surgery, chemotherapy, radiotherapy, or immunotherapy [26]. mRNA and protein levels of apoA‐I were reduced in HCC [27, 28]. ApoA‐I affects the proliferative, survival, and migratory behavior of various carcinoma cells, largely through cell‐autonomous mechanisms [29] Moreover, the immune regulator role apoA‐I contributes to its antitumor capacity. It contributes to the anti‐infection immunity of multiple pathogens [29]. Clinically, reduced serum apoA‐I levels in patients with sepsis are associated with poor prognosis [30]. ApoA‐I also regulates neutrophil development. Abca1‐ and Abcg1‐deficient mice display marked leukocytosis and a transplantable myeloproliferative disease, which can be suppressed by transgenic overexpression of apoA‐I [31]. ApoA‐I, via ABCA1, reduces the abundance of lipid rafts and leads to downregulation of neutrophil activation, migration, and adhesion [32]. Thus, the balance of serum apoA‐I and neutrophil count reflects the immunological status of patients with cancer, which endows it prognostic value for patients with HCC receiving TACE.
Prognostic factors similar to NAR have been identified for HCC. A prognostic score system was built with apoA‐I and C‐reactive protein for patients with HCC [13]. ApoA‐I was indicated as a novel serum biomarker for predicting the prognosis of hepatocellular carcinoma after curative resection [14]. Furthermore, mRNA expressions of apolipoprotein A and C family genes predicted hepatitis B virus–related HCC survival [11]. Neutrophil‐to‐lymphocyte ratio [33] and neutrophil count [22] were investigated in specific group of patients with HCC. However, previous studies seldom investigated the immune mechanism of these prognostic factors, except for the association of neutrophil count and PMN‐MDSC level [22]. In the present study, we found that NAR mediates the prognosis of patients with HCC receiving TACE by reflecting LOX‐1+ PMN‐MDSC level.
PMN‐MDSCs are critical immunosuppressive cells in inflammation and malignancies, suppressing immune cell activation and inducing immune tolerance [34, 35]. PMN‐MDSC level was shown to be of clinical use for patients with HCC in studies by us and other groups [22, 23, 36, 37]. Higher level of LOX‐1+ PMN‐MDSCs predicted decreased survival [23]. However, the methodology of testing PMN‐MDSCs has not been standardized until now, and quality control was not parallel between studies [23, 38, 39]. Thus, myeloid‐derived suppressor cell–related clinical parameters were more practical in daily usage. Lipometabolism was reported to be a key regulator of PMN‐MDSCs [40], and PMN‐MDSCs morphologically resemble and are quantitatively related to neutrophils [22]. Thus, clinical parameters combining lipometabolism and neutrophils count should directly reflect the PMN‐MDSC level. LOX‐1+ PMN‐MDSCs, as novel human‐specific PMN‐MDSCs [41], were confirmed to be prognostic for HCC [23]. In the present study, we found that NAR directly relates to PMN‐MDSCs, which explains the mechanism of the prognostic value of NAR.
The present study did not include progression‐free survival because of lack of data for a critical proportion of patients. Treatment choice and response of the later lines might influence OS. Thus, the current study included larger number of patients from multiple centers to achieve reliable results. Secondly, we did not dynamically monitor LOX‐1+ PMN‐MDSCs and NAR for each patient, which might provide more accurate prognostic value and detailed mechanism analysis. Also, the underlying mechanism of apoA‐I and LOX‐1+ PMN‐MDSCs requires further investigation.
Conclusion
The present study identified NAR as an independent predictor for OS among patients with HCC receiving TACE. NAR reflects circulation LOX‐1+ PMN‐MDSC level.
Author Contributions
Conception/design: Jie Chen, Yong‐Jian Chen, Nan Jiang, Jian‐Liang Xu, Zi‐Ming Liang, Ming‐Jun Bai, Yan‐Fang Xing, Xiang‐Yuan Wu, Xing Li
Provision of study material or patients: Jie Chen, Yong‐Jian Chen, Nan Jiang, Jian‐Liang Xu, Zi‐Ming Liang, Ming‐Jun Bai, Yan‐Fang Xing, Xiang‐Yuan Wu, Xing Li
Collection and/or assembly of data: Jie Chen, Yong‐Jian Chen, Nan Jiang, Jian‐Liang Xu, Zi‐Ming Liang, Ming‐Jun Bai, Yan‐Fang Xing, Xiang‐Yuan Wu, Xing Li
Data analysis and interpretation: Jie Chen, Yong‐Jian Chen, Nan Jiang, Jian‐Liang Xu, Zi‐Ming Liang, Ming‐Jun Bai, Yan‐Fang Xing, Xiang‐Yuan Wu, Xing Li
Manuscript writing: Jie Chen, Yong‐Jian Chen, Nan Jiang, Xiang‐Yuan Wu, Xing Li.
Final approval of manuscript: Jie Chen, Yong‐Jian Chen, Nan Jiang, Jian‐Liang Xu, Zi‐Ming Liang, Ming‐Jun Bai, Yan‐Fang Xing, Zhuo Liu, Xiang‐Yuan Wu, Xing Li
Disclosures
The authors indicated no financial relationships.
Supporting information
See http://www.TheOncologist.com for supplemental material available online.
Supplementary figure 1 Violin plot for the distribution of NAR among different CLIP scores.
Supplemental Table 1 Interaction between NAR and TNM or CLIP.
Supplementary Table 2: C‐index in three cohort
Acknowledgments
This study was supported by Guangzhou Science and Technology Project (201904010461), National Natural Science Foundation of China (81972677, 81871999), Natural Science Foundation of Guangdong (2019A1515012198, 2019A1515011187 and 2017A030313537), Special Funds for Fundamental Research Fund of Sun Yat‐sen University (19ykpy17), Shenzhen science and technology project (JCYJ20190809165813331), and Guangdong High‐Level Personnel of Special Support Program Outstanding Young Scholar in Science and Technology Innovation (2019TQ05Y266).
Disclosures of potential conflicts of interest may be found at the end of this article.
No part of this article may be reproduced, stored, or transmitted in any form or for any means without the prior permission in writing from the copyright holder. For information on purchasing reprints contact commercialreprints@wiley.com. For permission information contact permissions@wiley.com.
Contributor Information
Xiang‐Yuan Wu, Email: wuxiangy@mail.sysu.edu.cn.
Xing Li, Email: lixing9@mail.sysu.edu.cn.
References
- 1. Heimbach JK, Kulik LM, Finn RS et al. AASLD guidelines for the treatment of hepatocellular carcinoma. Hepatology 2018;67:358–380. [DOI] [PubMed] [Google Scholar]
- 2. Bettinger D, Spode R, Glaser N et al. Survival benefit of transarterial chemoembolization in patients with metastatic hepatocellular carcinoma: A single center experience. BMC Gastroenterol 2017;17:98. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3. Sinha J, Mehta N, Dodge JL et al. Are there upper limits in tumor burden for down‐staging of hepatocellular carcinoma to liver transplant? Analysis of the all‐comers protocol. Hepatology 2019;70:1185–1196. [DOI] [PubMed] [Google Scholar]
- 4. Hara K, Takeda A, Tsurugai Y et al. Radiotherapy for hepatocellular carcinoma results in comparable survival to radiofrequency ablation: A propensity score analysis. Hepatology 2019;69:2533–2545. [DOI] [PubMed] [Google Scholar]
- 5. Kudo M, Ueshima K, Chan S et al. Lenvatinib as an initial treatment in patients with intermediate‐stage hepatocellular carcinoma beyond up‐to‐seven criteria and Child‐Pugh A liver function: A proof‐of‐concept study. Cancers (Basel) 2019;11:1084. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6. Llovet J, Shepard KV, Finn RS et al. A phase Ib trial of lenvatinib (LEN) plus pembrolizumab (PEMBRO) in unresectable hepatocellular carcinoma (uHCC): Updated results. Ann Oncol 2019;30(suppl 5):747Pa. [Google Scholar]
- 7. Steinberg D. Low density lipoprotein oxidation and its pathobiological significance. J Biol Chem 1997;272:20963–20966. [DOI] [PubMed] [Google Scholar]
- 8. Kopprasch S, Pietzsch J, Graessler J. The protective effects of HDL and its constituents against neutrophil respiratory burst activation by hypochlorite‐oxidized LDL. Mol Cell Biochem 2004;258:121–127. [DOI] [PubMed] [Google Scholar]
- 9. Sirniö P, Väyrynen JP, Klintrup K et al. Decreased serum apolipoprotein A1 levels are associated with poor survival and systemic inflammatory response in colorectal cancer. Sci Rep 2017;7:5374. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10. Tuft Stavnes H, Nymoen DA, Hetland Falkenthal TE et al. ApoA1 mRNA expression in ovarian serous carcinoma effusions is a marker of longer survival. Am J Clin Pathol 2014;142:51–57. [DOI] [PubMed] [Google Scholar]
- 11. Wang X, Gong Y, Deng T et al. Diagnostic and prognostic significance of mRNA expressions of apolipoprotein A and C family genes in hepatitis B virus‐related hepatocellular carcinoma. J Cell Biochem 2019;120:18246–18265. [DOI] [PubMed] [Google Scholar]
- 12. Zhang Y, Yang X. Prognostic significance of pretreatment apolipoprotein A‐I as a noninvasive biomarker in cancer survivors: A meta‐analysis. Dis Markers 2018;2018:1034037. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13. Mao M, Wang X, Sheng H et al. A novel score based on serum apolipoprotein A‐1 and C‐reactive protein is a prognostic biomarker in hepatocellular carcinoma patients. BMC Cancer 2018;18:1178. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14. Ma XL, Gao XH, Gong ZJ et al. Apolipoprotein A1: A novel serum biomarker for predicting the prognosis of hepatocellular carcinoma after curative resection. Oncotarget 2016;7:70654–70668. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15. Chen ZH, Zhang XP, Cai XR et al. The predictive value of albumin‐to‐alkaline phosphatase ratio for overall survival of hepatocellular carcinoma patients treated with trans‐catheter arterial chemoembolization therapy. J Cancer 2018;9:3467–3478. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16. Cai XR, Chen ZH, Liu MM et al. Modified CLIP score with the albumin‐bilirubin grade retains prognostic value in HBV‐related hepatocellular carcinoma patients treated with trans‐catheter arterial chemoembolization therapy. J Cancer 2018;9:2380–2388. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17. Chen ZH, Hong YF, Chen X et al. Comparison of five staging systems in predicting the survival rate of patients with hepatocellular carcinoma undergoing trans‐arterial chemoembolization therapy. Oncol Lett 2018;15:855–862. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18. Loosen SH, Schulze‐Hagen M, Bruners P et al. Sarcopenia is a negative prognostic factor in patients undergoing transarterial chemoembolization (TACE) for hepatic malignancies. Cancers (Basel) 2019;11:1503. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19. Shen Y, Wang H, Chen X et al. Prognostic significance of lymphocyte‐to‐monocyte ratio and platelet‐to‐lymphocyte ratio in patients with hepatocellular carcinoma undergoing transcatheter arterial chemoembolization and radiofrequency ablation. Onco Targets Ther 2019;12:7129–7137. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20. Du QC, Hu CL, Wang YY et al. Comparison of the prognostic value of platelet‐based prognostic models in patients with malignant hepatic tumors after TACE therapy. Acta Cir Bras 2019;34:e201900710. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21. Wan S, Kuo N, Kryczek I et al. Myeloid cells in hepatocellular carcinoma. Hepatology 2015;62:1304–1312. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22. Li X, Xing YF, Lei AH et al. Neutrophil count is associated with myeloid derived suppressor cell level and presents prognostic value of for hepatocellular carcinoma patients. Oncotarget 2017;8:24380–24388. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23. Nan J, Xing YF, Hu B et al. Endoplasmic reticulum stress induced LOX‐1(+) CD15(+) polymorphonuclear myeloid‐derived suppressor cells in hepatocellular carcinoma. Immunology 2018;154:144–155. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24. Mehta N, Dodge JL, Roberts JP et al. Alpha‐fetoprotein decrease from > 1,000 to < 500 ng/ml in patients with hepatocellular carcinoma leads to improved posttransplant outcomes. Hepatology 2019;69:1193–1205. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25. Nishikawa H, Arimoto A, Wakasa T et al. Effect of transcatheter arterial chemoembolization prior to surgical resection for hepatocellular carcinoma. Int J Oncol 2013;42:151–160. [DOI] [PubMed] [Google Scholar]
- 26. Wu J, Zhang C, Zhang G et al. Association between pretreatment serum apolipoprotein A1 and prognosis of solid tumors in Chinese population: A systematic review and meta‐analysis. Cell Physiol Biochem 2018;51:575–588. [DOI] [PubMed] [Google Scholar]
- 27. Jiang Y, Sun A, Zhao Y et al. Proteomics identifies new therapeutic targets of early‐stage hepatocellular carcinoma. Nature 2019;567:257–261. [DOI] [PubMed] [Google Scholar]
- 28. Ai J, Tan Y, Ying W et al. Proteome analysis of hepatocellular carcinoma by laser capture microdissection. Proteomics 2006;6:538–546. [DOI] [PubMed] [Google Scholar]
- 29. Georgila K, Vyrla D, Drakos E. Apolipoprotein A‐I (ApoA‐I), immunity, inflammation and cancer. Cancers (Basel) 2019;11:1097. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30. Morin EE, Guo L, Schwendeman A et al. HDL in sepsis ‐ risk factor and therapeutic approach. Front Pharmacol 2015;6:244. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31. Yvan‐Charvet L, Pagler T, Gautier EL et al. ATP‐binding cassette transporters and HDL suppress hematopoietic stem cell proliferation. Science 2010;328:1689–1693. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32. Murphy AJ, Woollard KJ, Suhartoyo A et al. Neutrophil activation is attenuated by high‐density lipoprotein and apolipoprotein A‐I in in vitro and in vivo models of inflammation. Arterioscler Thromb Vasc Biol 2011;31:1333–1341. [DOI] [PubMed] [Google Scholar]
- 33. Li X, Chen ZH, Ma XK et al. Neutrophil‐to‐lymphocyte ratio acts as a prognostic factor for patients with advanced hepatocellular carcinoma. Tumour Biol 2014;35:11057–11063. [DOI] [PubMed] [Google Scholar]
- 34. Talmadge JE, Gabrilovich DI. History of myeloid‐derived suppressor cells. Nat Rev Cancer 2013;13:739–752. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35. Zhou J, Nefedova Y, Lei A et al. Neutrophils and PMN‐MDSC: Their biological role and interaction with stromal cells. Semin Immunol 2018;35:19–28. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36. Lu LC, Chang CJ, Hsu CH. Targeting myeloid‐derived suppressor cells in the treatment of hepatocellular carcinoma: Current state and future perspectives. J Hepatocell Carcinoma 2019;6:71–84. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37. Li YM, Liu ZY, Wang JC et al. Receptor‐interacting protein kinase 3 deficiency recruits myeloid‐derived suppressor cells to hepatocellular carcinoma through the chemokine (C‐X‐C motif) ligand 1‐chemokine (C‐X‐C motif) receptor 2 axis. Hepatology 2019;70:1564–1581. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 38. Grutzner E, Stirner R, Arenz L et al. Kinetics of human myeloid‐derived suppressor cells after blood draw. J Transl Med 2016;14:2. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 39. Bronte V, Brandau S, Chen SH et al. Recommendations for myeloid‐derived suppressor cell nomenclature and characterization standards. Nat Commun 2016;7:12150. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 40. Veglia F, Tyurin VA, Blasi M et al. Fatty acid transport protein 2 reprograms neutrophils in cancer. Nature 2019;569:73–78. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 41. Condamine T, Dominguez GA, Youn JI et al. Lectin‐type oxidized LDL receptor‐1 distinguishes population of human polymorphonuclear myeloid‐derived suppressor cells in cancer patients. Sci Immunol 2016;1:aaf8943. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
See http://www.TheOncologist.com for supplemental material available online.
Supplementary figure 1 Violin plot for the distribution of NAR among different CLIP scores.
Supplemental Table 1 Interaction between NAR and TNM or CLIP.
Supplementary Table 2: C‐index in three cohort