Sage Journals: Discover world-class research

Abstract

BACKGROUND:

Inflammation-based prognostic scores have been increasingly used for prognosis prediction in malignant tumors. However, no existing study has comprehensively evaluated these scores in combined hepatocellular-cholangiocarcinoma (cHCC-CCA).

OBJECTIVE:

This study aimed to identify a robust inflammation-based prognostic predictor for cHCC-CCA.

METHODS:

We retrospectively analyzed 220 patients pathologically confirmed as Allen type C cHCC-CCA. The univariate and multivariate analyses were used to explore the associations between clinical variables and prognosis of cHCC-CCA. The propensity score-matching (PSM) was performed to reduce the effects of potential cofounders and selection bias. Finally, the predictive values of different inflammation-based indices were compared by using time-dependent receiver operating characteristic (ROC) curves.

RESULTS:

The systemic immune-inflammation index (SII) and aspartate aminotransferase to platelet ratio index (APRI) were identified as independent prognostic predictors in multivariate analysis. After PSM, the survival differences were still significant between SII-high group and SII-low group ( $P=$ 0.016 for RFS and $P=$ 0.001 for OS). Further ROC analysis showed that the SII harbored the largest 1-, 3- and 5-year area under the curves (AUC) values as compared with other scores.

CONCLUSIONS:

The SII may serve as a preferable predictor of both recurrence-free survival (RFS) and overall survival (OS) in patients with cHCC-CCA.

Keywords

Combined hepatocellular-cholangiocarcinoma systemic immune-inflammation index propensity score-matching overall survival recurrence-free survival

1. Introduction

Combined hepatocellular-cholangiocarcinoma (cHCC-CCA) is a distinct subtype of primary liver cancer (PLC) defined by both hepatocellular and biliary differentiations within the same tumor nodule [1]. cHCC-CC is a rare liver malignancy, only accounting for about 2%–5% of PLCs [2]. A host of evidences have suggested that the demographic and clinical features of cHCC-CCA are intermediate between hepatocellular carcinoma (HCC) and intrahepatic cholangiocarcinoma (iCCA) [3]. Hepatic resection (HR) is the mainstay curative treatment for cHCC-CCA patients [4]. Unfortunately, recent researches showed that cHCC-CCA exhibited significantly higher early recurrence potential and worse prognosis than other primary liver cancers, even after curative resection [5, 6]. Nevertheless, due to the rarity of cHCC-CCA, the prognostic characteristics of this malignancy remain poorly understood. To date, no effective prognostic predictor for cHCC-CCA has been identified.

Inflammation is a hallmark of cancer, which plays a crucial role in tumorigenesis, tumor progression and metastasis [7]. The severity of underlying tumor-associated inflammation can be measured by a number of clinical markers, including peripheral blood counts (neutrophil, lymphocyte and platelet counts) and acute-phase protein (albumin (ALB) and C-reactive protein) [8]. Recently, several inflammation-based prognostic scores have been constructed for prognosis prediction in cancer patients, such as the neutrophil to lymphocyte ratio (NLR) [9], platelet to lymphocyte ratio (PLR) [10], lymphocyte to monocyte ratio (LMR) [11], prognostic nutritional index (PNI) [12], systemic immune-inflammation index (SII) [13] and aspartate aminotransferase to platelet ratio index (APRI) [14]. These scores have showed significant prognostic values in various malignancies including HCC and iCCA [15, 16, 17, 18]. However, no existing study has comprehensively evaluated the prognostic values of inflammation-based indices in cHCC-CCA patients.

In this study, we aimed to evaluate and compare the prognostic values of inflammation-based prognostic scores in a cohort of 220 patients with Allen type C cHCC-CCA following curative resection.

2. Materials and methods

2.1 Patients selection

A total of 307 patients with cHCC-CCA who underwent hepatic resection at liver cancer institute, Zhongshan Hospital between 1993 and 2015 were retrospectively screened. Given that the 2019 World Health Organization (WHO) classification simplifies the subcategorization of cHCC-CCA and defines cHCC-CCA as the subtype with unequivocal components of both HCC and CC within the same mass [19], we only enrolled patients with Allen type C cHCC-CCA in this research [20]. The detailed inclusion criteria were as follows: 1) received liver resection and pathologically confirmed as Allen type C cHCC-CCA; 2) Eastern Cooperative Oncology Group (ECOG) performance-status score of 0 or 1; 3) Child-Pugh grade A or B; 4) with complete preoperative laboratory results and survival data. The exclusion criteria were as follows: 1) received previous anti-tumor therapy before surgery; 2) concomitance with other malignancies in addition to cHCC-CCA; 3) presence of serious medical comorbidities such as heart or renal dysfunction. Finally, 220 consecutive cHCC-CCA patients were enrolled for further analyses.

2.2 Data collection

Clinical variables were retrospectively retrieved from the electronic medical record, including patient characteristics (age, sex, etiology, liver cirrhosis, diabetes mellitus, ascites, Child-Pugh grade and TNM stage), hepatitis B virus (HBV) status (HBV-DNA load and HbeAg positivity), tumor-related factors (tumor size, tumor number, tumor capsule, microvascular/macrovascular invasion, hilar lymph node involvement and tumor differentiation), laboratory results (ALB (g/L), total bilirubin (TB, umol/L), alanine aminotransferase (ALT, U/L), aspartate aminotransferase (AST, U/L), alkaline phosphatase (AKP, U/L), $\gamma$ -glutamyl transpeptidase ( $\gamma$ -GT, U/L), Alfa-Fetoprotein (AFP, ng/mL), carcinoembryonic antigen (CEA, ug/L) and carbohydrate antigen 19-9 (CA19-9, ku/L)) and treatment details (blood loss, blood transfusion, portal vascular occlusion and postoperative adjuvant transarterial chemoembolization (TACE)). The lymph node metastasis and vascular invasion were indicated by preoperative radiological examination and further confirmed by pathology.

2.3 Inflammation-based prognostic scores

All the blood samples were acquired within 5 days before surgery for measurement of ALB, AST, neutrophil, lymphocyte, monocyte and platelet counts. The inflammation-based prognostic scores including the SII, PNI, NLR, PLR, LMR and APRI were calculated and described in Supplementary Table 1. The optimal threshold value for each score was determined by performing receiver operating characteristic (ROC) analysis regarding overall survival (OS). The value with the maximum Youden index was chosen as the optimal cut-off value(Supplementary Fig. 1).

2.4 Treatment and follow-up

The standard technique for hepatic resection has been described in previous studies [21]. The adjuvant TACE was performed under the standard protocol of our department as described by Ren and Yin et al. [22, 23]. Briefly, a 5F or 4F catheter (Cordis ${}^{\copyright}$ , Baltimore, USA) was introduced into the abdominal aorta under fluoroscopy guidance. Angiography was performed to show the real hepatic artery anatomy and the patency of portal vein. Anticancer drugs (750–1000 mg of 5-fluorouracil, 100–150 mg of oxaliplatin or 60 mg of cisplatin) were infused into the common hepatic artery. Super-selective embolization was performed by using a microcatheter (Terumo ${}^{\copyright}$ , Tokyo, Japan) if needed. An emulsion of 2–5 mL lipiodol with 30 mg epirubicin or 10–20 mg mitomycin C was slowly injected into the tumor feeding arteries. No patients received gelatin or Drug-eluting-beads (DEB) embolization in postoperative adjuvant TACE. All patients were followed up every 2–3 months within the first year after surgery, and every 6 months thereafter until death or dropout. Abdominal ultrasound, liver function tests, tumor marker tests (AFP, CEA and CA19-9) were examined every 3 months. Abdominal computerized tomography (CT) or magnetic resonance imaging (MRI) scanning was performed every 6 months for screening tumor recurrence. The recurrence-free survival (RFS) was defined as the interval between the treatment date and recurrence. The OS was defined as the interval between the treatment date and the death date or last follow-up date.

2.5 Statistical analysis

Continuous variables were presented as median with interquartile range (IQR), whereas categorical variables were summarized as counts with percentages. Binary variables were compared by using either the chi-square test or Fisher’s exact test, whereas ordinal categorical variables were compared by using the Kruskal-Wallis test. Survival curves were estimated using the Kaplan-Meier method and compared through log-rank test. The univariate analysis was employed to identify significant factors associated with RFS and OS. Then, all variables that had a $P$ value $<$ 0.1 in univariate analysis were fitted into a multivariate Cox proportional hazard model to determine independent prognostic indicators. To reduce the effects of potential cofounders and selection bias, the propensity score-matching (PSM) analysis (1:1 ratio, with nearest-neighbor matching or caliper width of 0.1 of the standard deviation of the logit) was performed. Covariates entered into the propensity model included age, gender, etiology, HBV-DNA load, HbeAg positivity, ascites, liver cirrhosis, diabetes mellitus, Child-Pugh grade, TNM stage, tumor size, tumor number, tumor capsule, vascular invasion, hilar lymph node involvement, tumor differentiation, blood loss, blood transfusion, portal vascular occlusion, postoperative adjuvant TACE, ALB, ALT, AKP, $\gamma$ -GT, AFP, CEA and CA19-9. Finally, the discriminative values of the 6 inflammation-based prognostic scores were compared by using the time-dependent receiver operating characteristic curves.

All statistical analyses were performed using the Stata version 15.1 (StataCorp, College station, TX). Differences were considered statistically significant when two-tailed $P$ values $<$ 0.05.

2.6 Ethics approval statement

This study was approved by Medical Ethics Committee of Zhongshan Hospital affiliated to Fudan University (No: B2019-169) and complied with the standards of the Declaration of Helsinki. Informed consent was obtained from each patient for using their data for research.

3. Results

3.1 Patient characteristics

A total of 220 patients with Allen type C cHCC-CCA were enrolled in this study. The entire cohort consists of 162 (73.6%) males and 58 (26.4%) females, with a median age of 54 (IQR: 47, 61) years. Among these patients, 168 (76.4%) were positive for HBV surface antigen (HBsAg), 43 (19.6%) were positive for HBV e antigen (HBeAg), and 2 (0.9%) were positive for hepatitis C virus (HCV) antibody. Liver cirrhosis was observed in 126 (57.3%) patients and a number of 41 (18.6%) patients exhibited Child-Pugh grade B. With regard to tumor characteristics, the median tumor size was 4.7 (IQR: 3, 7.5) cm and 63 (28.6%) patients presented with multiple tumors. Microvascular invasion, macrovascular invasion and hilar lymph node involvement were detected in 40 (18.2%) patients, 16 (7.3%) patients and 21 (9.6%) patients, respectively. The numbers of patients with AJCC TNM-8 stage I, II, III and IV were 107 (48.6%), 44 (20.0%), 38 (17.3%) and 31 (14.1%), respectively. Other clinical characteristics including the 6 inflammation-based scores (SII, PNI, NLR, PLR, LMR and APRI) were presented in Table 1.

Table 1
Clinical characteristics of enrolled patients ( $N=$ 220)

Characteristics	Median (IQR) or NO. of patients (%)
Age (years)	54 (47, 61)
Sex (male/female)	162/58 (73.6/26.4)
Etiology (HBV/HCV/HBV $+$ HCV/Other)	168/2/1/49 (76.4/0.9/0.5/22.3)
HBeAg (negative/positive)	177/43 (80.5/19.6)
HBV-DNA load ( $\leqslant$ 100/ $>$ 100 IU/mL)	160/60 (72.7/27.3)
Liver cirrhosis (absence/presence)	94/126 (42.7/57.3)
Diabetes mellitus (absence/presence)	207/13 (94.1/5.9)
Ascites (absence/presence)	200/20 (90.9/9.1)
Child-Pugh grade (A/B)	179/41 (81.4/18.6)
AJCC TNM-8 stage (I/II/III/IV)	107/44/38/31 (48.6/20.0/17.3/14.1)
Tumor size (cm)	4.7 (3, 7.5)
Tumor number (single/multiple)	157/63 (71.4/28.6)
Tumor capsule (complete/incomplete)	79/141 (35.9/64.1)
Microvascular invasion (absence/presence)	180/40 (81.8/18.2)
Macrovascular invasion (absence/presence)	204/16 (92.7/7.3)
Lymph node involvement (absence/presence)	199/21 (90.5/9.6)
Tumor differentiation (I, II/III, IV)	160/60 (72.7/27.3)
Blood loss (mL)	200 (100, 400)
Blood transfusion (no/yes)	196/24 (89.1/10.9)
Portal vascular occlusion (no/yes)	109/111 (49.6/50.5)
Adjuvant TACE (no/yes)	176/44 (80.0/20.0)
ALB (g/L)	41 (38, 44)
TB (umol/L)	12.5 (9.2, 15.8)
ALT (U/L)	30 (21, 49.5)
AST (U/L)	33.5 (26, 49.5)
AKP (U/L)	90.5 (72, 115)
$\gamma$ -GT (U/L)	61.5 (36, 107.5)
AFP (ng/mL)	30 (5.3, 298.3)
CEA (ug/L)	2.4 (1.3, 3.7)
CA19-9 (ku/L)	21.2 (10.2, 45.1)
Neutrophil (10 ${}^{9}$ /L)	3.4 (2.4, 5.3)
Lymphocyte (10 ${}^{9}$ /L)	1.5 (1.1, 1.9)
Monocyte (10 ${}^{9}$ /L)	0.4 (0.3, 0.5)
Platelet (10 ${}^{9}$ /L)	144 (102, 195)
SII (0/1)	135/85 (61.4/38.6)
PNI (0/1)	12/208 (5.5/94.6)
NLR (0/1)	147/73 (66.8/33.2)
PLR (0/1)	104/116 (47.3/52.7)
LMR (0/1)	34/186 (15.5/84.6)
APRI (0/1)	80/140 (36.4/63.6)

Abbreviations: IQR: interquartile range; HBV: hepatitis B virus; HCV: hepatitis C virus; HBeAg: hepatitis B virus e antigen; AJCC TNM-8: the 8 ${}^{\text{th}}$ edition of American Joint Committee on Cancer TNM staging; TACE: transarterial chemoembolization; ALB: albumin; TB, total bilirubin; ALT: alanine aminotransferase; AST: aspartate aminotransferase; AKP: alkaline phosphatase; $\gamma$ -GT: $\gamma$ -glutamyl transpeptidase; AFP: Alfa-Fetoprotein; CEA: carcinoembryonic antigen; CA19-9: carbohydrate antigen 19-9; SII: systemic immune-inflammation index; PNI: prognostic nutritional index; NLR: neutrophil to lymphocyte ratio; PLR: platelet to lymphocyte ratio; LMR: lymphocyte to monocyte ratio; APRI: aspartate aminotransferase to platelet ratio index.

Table 2

Univariate and multivariate analyses of prognostic factors in patients with combined hepatocellular-cholangiocarcinoma ( $N=$ 220)

Variable	Recurrence-free survival			Overall survival
	Univariate analysis	Multivariate analysis		Univariate analysis	Multivariate analysis
	$P$ value	HR (95%CI)	$P$ value	$P$ value	HR (95%CI)	$P$ value
Age, years ( $>$ 65)	0.920			0.022 ${}^{*}$
Sex (male)	0.006 ${}^{*}$	1.78 (1.06–2.99)	0.029 ${}^{*}$	0.920
HBV infection (presence)	0.598			0.780
HCV infection (presence)	0.907			0.110
HBeAg positivity	0.001 ${}^{*}$	2.30 (1.45–3.65)	$<$ 0.001 ${}^{*}$	0.219
HBV-DNA load, IU/mL ( $>$ 100)	0.747			0.760
Ascites (presence)	0.749			0.008 ${}^{*}$
Liver cirrhosis (presence)	0.026 ${}^{*}$			0.003 ${}^{*}$
Diabetes mellitus (presence)	0.534			0.475
Child-Pugh grade (A vs. B)	0.197			0.002 ${}^{*}$
AJCC TNM-8 stage (I/II vs. III/IV)	0.001 ${}^{*}$			$<$ 0.001 ${}^{*}$
Tumor size, cm ( $>$ 5)	0.003 ${}^{*}$			$<$ 0.001 ${}^{*}$	1.64 (1.08–2.49)	0.019 ${}^{*}$
Tumor number (multiple)	0.024 ${}^{*}$			0.059
Tumor capsule (incomplete)	0.021 ${}^{*}$	1.58 (1.02–2.43)	0.038 ${}^{*}$	0.015 ${}^{*}$	1.66 (1.04–2.65)	0.033 ${}^{*}$
Microvascular invasion (presence)	0.596			0.328
Macrovascular invasion (presence)	$<$ 0.001 ${}^{*}$	2.15 (1.09–4.27)	0.028 ${}^{*}$	$<$ 0.001 ${}^{*}$	2.10 (1.10–4.01)	0.024 ${}^{*}$
Lymph node involvement (presence)	0.041 ${}^{*}$			$<$ 0.001 ${}^{*}$	2.59 (1.41–4.75)	0.002 ${}^{*}$
Tumor differentiation (1/2 vs. 3/4)	0.316			0.595
Blood loss, mL ( $\geqslant$ 1000)	0.086			0.030 ${}^{*}$
Blood transfusion (yes)	0.690			0.014 ${}^{*}$
Portal vascular occlusion (yes)	0.313			0.922
Adjuvant TACE (yes)	$<$ 0.001 ${}^{*}$	2.73 (1.81–4.10)	$<$ 0.001 ${}^{*}$	0.482
ALB, g/L ( $\leqslant$ 35)	0.662			$<$ 0.001 ${}^{*}$	6.73 (3.34–13.55)	$<$ 0.001 ${}^{*}$
TB, umol/L ( $\geqslant$ 34)	0.851			0.252
ALT, U/L ( $\geqslant$ 40)	0.300			0.060
AST, U/L ( $\geqslant$ 35)	0.651			0.014 ${}^{*}$
AKP, U/L ( $\geqslant$ 135)	0.024 ${}^{*}$			$<$ 0.001 ${}^{*}$
$\gamma$ -GT, U/L ( $>$ 50)	0.068			$<$ 0.001 ${}^{*}$
AFP, ng/mL ( $\geqslant$ 400)	0.237			0.207
CEA, ug/L ( $\geqslant$ 5)	0.755			0.001 ${}^{*}$	2.73 (1.67–4.45)	$<$ 0.001 ${}^{*}$
CA19-9, ku/L ( $\geqslant$ 37)	0.724			0.008 ${}^{*}$
SII (0 vs. 1)	$<$ 0.001 ${}^{*}$	2.10 (1.39–3.17)	$<$ 0.001 ${}^{*}$	$<$ 0.001 ${}^{*}$	2.73 (1.76–4.23)	$<$ 0.001 ${}^{*}$
PNI (0 vs. 1)	0.742			0.576
NLR (0 vs. 1)	$<$ 0.001 ${}^{*}$			$<$ 0.001 ${}^{*}$
PLR (0 vs. 1)	0.477			0.009 ${}^{*}$
LMR (0 vs. 1)	0.380			0.450
APRI	0.021 ${}^{*}$	1.71 (1.10–2.66)	0.018 ${}^{*}$	0.002 ${}^{*}$	2.26 (1.40–3.67)	0.001 ${}^{*}$

Abbreviations: HR: hazard ratio; CI: confidence interval; HBV: hepatitis B virus; HCV: hepatitis C virus; HBeAg: hepatitis B virus e antigen; AJCC TNM-8: the 8 ${}^{\text{th}}$ edition of American Joint Committee on Cancer TNM staging; TACE: transarterial chemoembolization; ALB: albumin; TB, total bilirubin; ALT: alanine aminotransferase; AST: aspartate aminotransferase; AKP: alkaline phosphatase; $\gamma$ -GT: $\gamma$ -glutamyl transpeptidase; AFP: Alfa-Fetoprotein; CEA: carcinoembryonic antigen; CA19-9: carbohydrate antigen 19-9; SII: systemic immune-inflammation index; PNI: prognostic nutritional index; NLR: neutrophil to lymphocyte ratio; PLR: platelet to lymphocyte ratio; LMR: lymphocyte to monocyte ratio; APRI: aspartate aminotransferase to platelet ratio index. ${}^{*}$ Statistically significant.

Figure 1.

Kaplan-Meier analysis of the recurrence-free survival of patients with combined hepatocellular-cholangiocarcinoma. (A) systemic immune-inflammation index (SII); (B) prognostic nutritional index (PNI); (C) neutrophil to lymphocyte ratio (NLR); (D) platelet to lymphocyte ratio (PLR); (E) lymphocyte to monocyte ratio (LMR); (F) aspartate aminotransferase to platelet ratio index (APRI). ${}^{*}$ Statistically significant.

Figure 2.

Kaplan-Meier analysis of the overall survival of patients with combined hepatocellular-cholangiocarcinoma. (A) systemic immune-inflammation index (SII); (B) prognostic nutritional index (PNI); (C) neutrophil to lymphocyte ratio (NLR); (D) platelet to lymphocyte ratio (PLR); (E) lymphocyte to monocyte ratio (LMR); (F) aspartate aminotransferase to platelet ratio index (APRI). ${}^{*}$ Statistically significant.

3.2 Survival analyses

The mean duration of follow-up was 22.1 (Range: 1–118) months. A total of 104 (47.3%) patients died and 110 (50.0%) patients suffered tumor recurrence at the end of the follow-up. The median RFS was 18.2 (95% CI: 13, 22.1) months and the 1-, 3- and 5-year RFS rates were 59.8%, 34.5% and 21.8%, respectively. The median OS was 30 (95% CI: 23, 49) months and the 1-, 3- and 5-year OS rates were 77.8%, 42.8% and 26.3%, respectively.

3.3 Univariable and multivariate analyses

The associations between patient characteristics and clinical outcomes were examined by univariable and multivariate analyses (Table 2 and Figs 1 and 2). With regard to RFS, our univariable analysis indicated that sex, HBeAg positivity, liver cirrhosis, TNM stage, tumor size, tumor number, tumor capsule, macrovascular invasion, lymph node involvement, adjuvant TACE, AKP, SII, NLR and APRI were significant factors. And sex ( $P=$ 0.013), HBeAg positivity ( $P<$ 0.001), tumor capsule ( $P=$ 0.008), macrovascular invasion ( $P=$ 0.016), adjuvant TACE ( $P<$ 0.001), SII ( $P=$ 0.001) and APRI ( $P=$ 0.008) were found to be independent indicators of RFS in multivariate analysis. As for OS, the univariable analysis suggested that age, ascites, liver cirrhosis, Child-Pugh stage, TNM stage, tumor size, tumor capsule, macrovascular invasion, lymph node involvement, blood loss, blood transfusion, ALB, AST, AKP, $\gamma$ -GT, CEA, CA19-9, SII, NLR, PLR and APRI were associated with OS. And the multivariate analysis identified tumor size ( $P=$ 0.019), tumor capsule ( $P=$ 0.033), macrovascular invasion ( $P=$ 0.024), lymph node involvement ( $P=$ 0.002), ALB ( $P<$ 0.001), CEA ( $P<$ 0.001), SII ( $P<$ 0.001) and APRI ( $P=$ 0.001) as independent prognostic factors. Collectively, among all the inflammation-based scores, only the SII and APRI were identified as independent predictor of both RFS and OS in patients with cHCC-CCA.

3.4 Propensity score-matching analysis

Clinical characteristics were compared between groups divided by the SII and APRI (Tables 3 and 4). Before PSM, clinical parameters of SII 0 group and SII 1 group statistically differed in TNM stage ( $P<$ 0.001), tumor size ( $P<$ 0.001), lymph node involvement ( $P=$ 0.002), blood transfusion ( $P=$ 0.006), portal vascular occlusion ( $P=$ 0.012), adjuvant TACE ( $P=$ 0.013), AKP ( $P=$ 0.030), $\gamma$ -GT ( $P=$ 0.005), AFP ( $P=$ 0.022) and CEA ( $P=$ 0.018). For APRI 0 and APRI 1 groups, HBV infection ( $P=$ 0.002), liver cirrhosis ( $P=$ 0.009), Child-Pugh stage ( $P=$ 0.001), tumor number ( $P=$ 0.006), macrovascular invasion ( $P=$ 0.009), blood loss ( $P=$ 0.038), blood transfusion ( $P=$ 0.034), ALB ( $P=$ 0.019), ALT ( $P=$ 0.001), AST ( $P<$ 0.001), AKP ( $P=$ 0.050), and $\gamma$ -GT ( $P=$ 0.003) were statistically different between two groups. By performing PSM, all clinical characteristics were well balanced between groups divided by the SII and APRI (Tables 3 and 4, all $P>$ 0.05).

Table 3
Baseline characteristics of the patients classified by the SII before and after propensity score-matching

Characteristics	Before matching			After matching
	SII 0 ( $N=$ 131)	SII 1 ( $N=$ 89)	$P$ value	SII 0 ( $N=$ 50)	SII 1 ( $N=$ 50)	$P$ value
Age, years ( $\leqslant$ 65/ $>$ 65)	112/19	78/11	0.308	41/9	42/8	0.790
Sex (female/male)	37/94	21/68	0.442	13/37	15/35	0.381
HBV infection (absence/presence)	26/105	25/64	0.155	16/34	11/39	0.371
HCV infection (absence/presence)	129/2	88/1	0.800	49/1	50/0	/
HBeAg (negative/positive)	101/30	76/13	0.128	44/6	39/11	0.274
HBA-DNA load, IU/mL ( $\leqslant$ 100/ $>$ 100)	98/33	62/27	0.400	34/16	34/16	1.000
Ascites (absence/presence)	119/12	81/8	0.965	47/3	45/5	0.461
Liver cirrhosis (absence/presence)	54/77	40/49	0.584	17/33	20/30	0.536
Diabetes mellitus (absence/presence)	128/8	84/5	0.880	47/3	47/3	0.399
Child-Pugh grade (A vs. B)	107/24	72/17	0.884	44/6	41/9	0.400
AJCC TNM-8 stage (I/II vs. III/IV)	105/26	46/43	$<$ 0.001 ${}^{*}$	31/19	29/21	1.000
Tumor size, cm ( $\leqslant$ 5/ $>$ 5)	95/36	32/57	$<$ 0.001 ${}^{*}$	26/24	26/24	0.840
Tumor number (single/multiple)	95/36	62/27	0.646	36/14	34/16	0.817
Tumor capsule (complete/incomplete)	50/81	29/60	0.397	19/31	19/31	1.000
Microvascular invasion (absence/presence)	110/21	70/19	0.316	43/7	39/11	0.790
Macrovascular invasion (absence/presence)	125/6	79/10	0.062	46/4	46/4	0.294
Lymph node involvement (absence/presence)	125/6	74/15	0.002 ${}^{*}$	44/6	44/6	0.461
Tumor differentiation (1/2 vs. 3/4)	92/39	68/21	0.313	36/14	36/14	1.000
Blood loss, mL ( $<$ 1000/ $\geqslant$ 1000)	126/5	82/7	0.194	48/2	46/4	0.307
Blood transfusion (no/yes)	123/8	73/16	0.006 ${}^{*}$	46/4	46/4	0.399
Portal vascular occlusion (no/yes)	74/57	35/54	0.012 ${}^{*}$	23/27	21/29	0.539
Adjuvant TACE (no/yes)	112/19	64/25	0.013 ${}^{*}$	36/14	37/13	0.817
ALB, g/L ( $>$ 35/ $\leqslant$ 35)	123/8	83/6	0.850	49/1	48/2	0.646
TB, umol/L ( $<$ 34/ $\geqslant$ 34)	116/15	81/8	0.558	47/3	44/6	0.461
ALT, U/L ( $<$ 40/ $\geqslant$ 40)	81/50	62/27	0.232	32/18	34/16	0.667
AST, U/L ( $<$ 35/ $\geqslant$ 35)	70/61	45/44	0.675	25/25	25/25	0.686
AKP, U/L ( $<$ 135/ $\geqslant$ 135)	117/14	70/19	0.030 ${}^{*}$	42/8	41/9	0.779
$\gamma$ -GT, U/L ( $<$ 50/ $\geqslant$ 50)	60/71	24/65	0.005 ${}^{*}$	16/34	19/31	0.671
AFP, ng/mL ( $<$ 400/ $\geqslant$ 400)	112/19	65/24	0.022 ${}^{*}$	41/9	43/7	0.424
CEA, ug/L ( $<$ 5/ $\geqslant$ 5)	117/14	69/20	0.018 ${}^{*}$	40/10	43/7	0.585
CA19-9, ku/L ( $<$ 37/ $\geqslant$ 37)	101/30	59/30	0.077	36/14	35/15	0.817
APRI (0 vs. 1)	45/86	35/54	0.452	20/30	18/32	0.407

Abbreviations: SII: systemic immune-inflammation index; HBV: hepatitis B virus; HCV: hepatitis C virus; HBeAg: hepatitis B virus e antigen; AJCC TNM-8: the 8 ${}^{\text{th}}$ edition of American Joint Committee on Cancer TNM staging; TACE: transarterial chemoembolization; ALB: albumin; TB, total bilirubin; ALT: alanine aminotransferase; AST: aspartate aminotransferase; AKP: alkaline phosphatase; $\gamma$ -GT: $\gamma$ -glutamyl transpeptidase; AFP: Alfa-Fetoprotein; CEA: carcinoembryonic antigen; CA19-9: carbohydrate antigen 19-9; APRI: aspartate aminotransferase to platelet ratio index. ${}^{*}$ Statistically significant.

Table 4

Baseline characteristics of the patients classified by the APRI before and after propensity score-matching

Characteristics	Before matching			After matching
	APRI 0 ( $N=$ 80)	APRI 1 ( $N=$ 140)	$P$ value	APRI 0 ( $N=$ 52)	APRI 1 ( $N=$ 52)	$P$ value
Age, years ( $\leqslant$ 65/ $>$ 65)	67/13	123/17	0.393	45/7	44/8	0.780
Sex (female/male)	24/56	34/106	0.355	12/40	13/39	0.819
HBV infection (absence/presence)	28/52	23/117	0.002 ${}^{*}$	10/42	13/39	0.478
HCV infection (absence/presence)	80/0	137/3	0.187	52/0	52/0	/
HBeAg (negative/positive)	69/11	108/32	0.101	43/9	44/8	0.791
HBA-DNA load, IU/mL ( $\leqslant$ 100/ $>$ 100)	60/20	100/40	0.567	38/14	40/12	0.651
Ascites (absence/presence)	75/5	125/15	0.268	48/4	49/3	0.696
Liver cirrhosis (absence/presence)	25/55	69/71	0.009 ${}^{*}$	20/32	15/37	0.299
Diabetes mellitus (absence/presence)	76/4	131/9	0.666	50/2	49/3	0.647
Child-Pugh grade (A vs. B)	74/6	105/35	0.001 ${}^{*}$	46/6	48/4	0.506
AJCC TNM-8 stage (I/II vs. III/IV)	60/20	91/49	0.124	42/10	40/12	0.631
Tumor size, cm ( $\leqslant$ 5/ $>$ 5)	49/31	78/62	0.424	34/18	28/24	0.230
Tumor number (single/multiple)	66/14	91/49	0.006 ${}^{*}$	41/11	39/13	0.642
Tumor capsule (complete/incomplete)	26/54	53/87	0.426	18/34	19/33	0.838
Microvascular invasion (absence/presence)	63/17	117/23	0.372	41/11	43/9	0.619
Macrovascular invasion (absence/presence)	79/1	125/15	0.009 ${}^{*}$	51/1	52/0	0.315
Lymph node involvement (absence/presence)	71/9	128/12	0.515	49/3	48/4	0.696
Tumor differentiation (1/2 vs. 3/4)	61/19	99/41	0.375	36/16	41/11	0.263
Blood loss, mL ( $<$ 1000/ $\geqslant$ 1000)	79/1	129/11	0.038 ${}^{*}$	51/1	51/1	1.000
Blood transfusion (no/yes)	76/4	120/20	0.034 ${}^{*}$	50/2	49/3	0.647
Portal vascular occlusion (no/yes)	37/43	72/68	0.460	26/26	23/29	0.674
Adjuvant TACE (no/yes)	64/16	112/28	1.000	42/10	38/14	0.352
Albumin, g/L ( $>$ 35/ $\leqslant$ 35)	79/1	127/13	0.019 ${}^{*}$	51/1	51/1	1.000
Total bilirubin, umol/L ( $<$ 34/ $\geqslant$ 34)	75/5	122/18	0.123	47/5	49/3	0.462
ALT, U/L ( $<$ 40/ $\geqslant$ 40)	63/17	80/60	0.001 ${}^{*}$	37/15	37/15	1.000
AKP, U/L ( $<$ 135/ $\geqslant$ 135)	73/7	114/26	0.050 ${}^{*}$	46/6	47/5	0.750
$\gamma$ -GT, U/L ( $<$ 50/ $\geqslant$ 50)	41/39	43/97	0.003 ${}^{*}$	23/29	23/29	1.000
AFP, ng/mL ( $<$ 400/ $\geqslant$ 400)	64/16	113/27	0.898	43/9	41/11	0.619
CEA, ug/L ( $<$ 5/ $\geqslant$ 5)	71/9	115/25	0.192	46/6	46/6	1.000
CA19-9, ku/L ( $<$ 37/ $\geqslant$ 37)	61/19	99/41	0.375	41/11	39/13	0.642
SII (0 vs. 1)	45/35	86/54	0.239	31/21	27/25	0.430

Abbreviations: APRI: aspartate aminotransferase to platelet ratio index; HBV: hepatitis B virus; HCV: hepatitis C virus; HBeAg: hepatitis B virus e antigen; AJCC TNM-8: the 8 ${}^{\text{th}}$ edition of American Joint Committee on Cancer TNM staging; TACE: transarterial chemoembolization; ALT: alanine aminotransferase; AST: aspartate aminotransferase; AKP: alkaline phosphatase; $\gamma$ -GT: $\gamma$ -glutamyl transpeptidase; AFP: Alfa-Fetoprotein; CEA: carcinoembryonic antigen; CA19-9: carbohydrate antigen 19-9; SII: systemic immune-inflammation index. ${}^{*}$ Statistically significant.

Figure 3.

Kaplan-Meier analysis of patients with combined hepatocellular-cholangiocarcinoma after propensity-score matching. (A) recurrence-free survival by the SII; (B) overall survival by the SII; (C) recurrence-free survival by the APRI; (D) overall survival by the APRI. ${}^{*}$ Statistically significant.

The PSM analysis generated 50 pairs of SII-matched patients and 52 pairs of APRI-matched patients, respectively. As shown in Fig. 3, after PSM, APRI 0 and APRI 1 groups showed no significantly survival differences ( $P=$ 0.722 for RFS and $P=$ 0.668 for OS, Fig. 3C and D). In contrast, the SII was significantly correlated with RFS and OS in SII-match cohort ( $P=$ 0.016 for RFS and $P=$ 0.001 for OS, Fig. 3A and B). Therefore, our findings demonstrated that the SII could serve as an accurate predictor of both RFS and OS in patients with cHCC-CCA.

3.5 Comparison of the discriminative ability between the inflammation-based scores

The time-dependent ROC curves at 1, 3 and 5 years were employed to compare the predictive values of the 6 inflammation-based scores (Table 5 and Fig. 4). The SII had 1-, 3- and 5-year area under the curve (AUC) values of 0.66 (95% CI: 0.59–0.72), 0.63 (95% CI: 0.56–0.71) and 0.65 (95% CI: 0.55–0.74), respectively, which significantly outperformed other scores. Moreover, the positive predictive value (PPV) and negative predictive value (NPV) of various inflammation-based scores were presented in Table 6. These results supported our finding that the SII was superior to other inflammation-based scores in predicting prognosis in cHCC-CCA.

Table 5
Comparison of the predictive performances between the inflammation-based scores

Scores	1-year AUC (95%CI)	3-year AUC (95%CI)	5-year AUC (95%CI)
SII	0.66 (0.59–0.72)	0.63 (0.56–0.71)	0.65 (0.55–0.74)
PNI	0.52 (0.50–0.55)	0.52 (0.47–0.57)	0.54 (0.45–0.63)
NLR	0.62 (0.55–0.68)	0.53 (0.45–0.62)	0.57 (0.46–0.68)
PLR	0.58 (0.51–0.65)	0.57 (0.48–0.67)	0.54 (0.41–0.68)
LMR	0.54 (0.50–0.59)	0.52 (0.44–0.59)	0.52 (0.42–0.63)
APRI	0.55 (0.48–0.61)	0.48 (0.39–0.57)	0.52 (0.39–0.65)

Abbreviations: AUC: area under the curve; CI: confidence interval; SII: systemic immune-inflammation index; PNI: prognostic nutritional index; NLR: neutrophil to lymphocyte ratio; PLR: platelet to lymphocyte ratio; LMR: lymphocyte to monocyte ratio; APRI: aspartate aminotransferase to platelet ratio index.

Table 6

The positive predictive value and negative predictive value of various scores

Scores	1-year		3-year		5-year
	PPV	NPV	PPV	NPV	PPV	NPV
SII	0.57	0.76	0.93	0.21	0.98	0.10
PNI	0.39	0.75	0.86	0.25	0.94	0.17
NLR	0.56	0.71	0.88	0.16	0.96	0.08
PLR	0.43	0.67	0.90	0.20	0.95	0.09
LMR	0.40	0.74	0.85	0.18	0.94	0.09
APRI	0.41	0.68	0.84	0.14	0.94	0.08

Abbreviations: PPV: positive predictive value; NPV: negative predictive value; SII: systemic immune-inflammation index; PNI: prognostic nutritional index; NLR: neutrophil to lymphocyte ratio; PLR: platelet to lymphocyte ratio; LMR: lymphocyte to monocyte ratio; APRI: aspartate aminotransferase to platelet ratio index.

Figure 4.

The time-dependent receiver operating characteristic (ROC) curves of inflammation-based prognostic scores for predicting overall survival. (A) ROC curves for predicting 1-year overall survival; (B) ROC curves for predicting 3-year overall survival; (C) ROC curves for predicting 5-year overall survival.

Figure 5.

The discrimination efficacy of the SII in different subgroups based on age, sex, serum AFP level and TNM stage. (A) recurrence-free survival by the SII in subgroups based on age, sex, serum AFP level and TNM stage. (B) overall survival by the SII in subgroups based on age, sex, serum AFP level and TNM stage. ${}^{*}$ Statistically significant.

3.6 Subgroup analysis

To further verify the predictive value of the SII in patients with different clinical characteristics, we divided the entire cohort into various subgroups based on age, sex, serum AFP level and TNM stage. Our results showed that the SII could significantly discriminate both RFS (Fig. 5A) and OS (Fig. 5B) in all these subgroups ( $P<$ 0.05), which indicating a consistent discrimination efficacy of the SII in cHCC-CCA patients.

4. Discussion

cHCC-CCA is a rare subtype of primary liver cancer, which tends to appear more aggressive behaviors and poorer outcome than HCC [5, 24]. However, the prognostic characteristics of cHCC-CCA still remains undetermined. Due to the rarity of cHCC-CCA, most of the previous clinical researches were derived from small patient cohort, which were statistically inadequate for identifying definitive prognostic predictors for cHCC-CCA [24, 25, 26]. To bridge this gap, in present study, we for the first time compared various inflammation-based prognostic scores in a large cohort of 220 patients with Allen type C cHCC-CCA. Our results indicated that the SII could independently predict RFS and OS in cHCC-CCA, with superior predictive performance to other inflammation-based indices.

In this study, ROC analysis was applied to determine the optimal cut-offs of different inflammation-based prognostic scores (Supplementary Table 1 and Supplementary Fig. 1). The univariate and multivariate analyses demonstrated that the SII and APRI could independently predict both RFS and OS in patients with cHCC-CCA (Table 2). However, after adjusting for cofounding variables, the APRI failed to discriminate survival between low- and high APRI groups ( $P=$ 0.722 for RFS and $P=$ 0.668 for OS, Fig. 3C and D). In contrast, patients with high-SII levels still presented significantly shorter RFS and OS compared with those with low-SII levels ( $P=$ 0.016 for RFS and $P=$ 0.001 for OS and, Fig. 3A and B). Furthermore, ROC analysis showed the SII had the largest 1-, 3- and 5-year AUCs in comparison with other inflammation-based prognostic scores (Table 5 and Fig. 4). All these data demonstrated that the SII could serve as a preferable index for predicting prognosis in patients with cHCC-CCA.

The SII is an integrated index based on peripheral neutrophil, platelet and lymphocyte counts, which was initially developed for prognosis prediction in HCC patients undergoing curative resection [13]. Recently, its prognostic value has been unveiled in several other cancers including iCCA [15], gastric cancer [27], renal cell cancer [28] and breast cancer [29]. The physiological roles of the three important blood cells may be correlated with tumor malignant behaviors. Experimental studies have showed that neutrophils play a significant promoting-role in tumor development, progression and metastasis [30, 31]. One important mechanism is that neutrophils can promote tumor angiogenesis through several mitogenic and pro-angiogenic molecules including elastase, prokineticin-2 (PROK2) and matrix metalloproteinase-9 (MMP9) [30]. In addition, immunosuppression is another predominant mechanism for neutrophils to facilitate cancer cells proliferation and invasion, achieved by the production of specific cytokines and amino acid depletion [32]. Likewise, activated platelets can bind tumor cells through the adhesion receptor P selectin and promote tumor progression by producing vascular endothelial growth factor, tissue factor, the fibroblast growth factor and platelet-derived growth factor [33, 34]. Moreover, platelet-derived transforming growth factor- $\beta$ (TGF- $\beta$ ) and direct platelet-tumor cell contacts may contribute to a mesenchymal-like transition and enhance metastasis through TGF- $\beta$ /Smad and NF-kB pathways in cancer cells [35]. In contrast, lymphocytes mainly play a role in anti-tumor immunity and immune surveillance. It’s recognized that lymphocytes can induce cytotoxic cell death by secreting cytokines including interferon- $\gamma$ (IFN- $\gamma$ ) and tumor necrosis factor- $\alpha$ (TNF- $\alpha$ ), and inhibit tumor proliferation and migration [36]. Based on these biological mechanisms and our findings, the SII should be regarded as an integrated immune-inflammation index for prognosis prediction in cHCC-CCA.

Surgical resection is the most widely used curative approach for cHCC-CCA [4]. However, it’s reported that the recurrence rate and long-term survival of cHCC-CCA patients after resection are poor [4, 6, 37]. Therefore, more effective therapies should be performed for those patients with high risk of recurrence and poor estimated survival. In present study, patients with high-SII levels presented significantly lower 1-, 3- and 5-year RFS rates (41.2%, 20.5% and 7.7% vs. 71.7%, 42.3% and 29.6%, $P<$ 0.001) and 1-, 3- and 5-year OS rates (59.0%, 21.0% and 5.3% vs. 90.1%, 56.1% and 38.8%, $P<$ 0.001) than patients with low-SII levels (Figs 1 and 2). Moreover, the discrimination efficacy of the SII also remains in patients at TNM stage I & II or with AFP levels $<$ 400 ng/mL (Fig. 5). These results indicate that the SII can serve as an effective additional tool to help clinicians identify cHCC-CCA patients at high risks of recurrence and enable targeted rational adjuvant therapy after surgical resection. Moreover, there is also a potential for the SII to be applied as a biomarker for treatment response surveillance [13].

There are several limitations in this study. First, as a single-center based retrospective study, potential selection bias is inevitable. To reduce selective bias, the PSM was performed to balance baseline variables. Second, the cut-off values for inflammation-based indices in this study was determined by ROC analysis based on survival data, further validation studies determining optimal threshold values for these inflammation-based indices are needed. Third, peripheral neutrophil, platelet and lymphocyte counts may be influenced by other pathological conditions including ischemia, infection, metabolic syndrome and renal dysfunction. Hence, the clinical significance of the SII should be cautiously explained with full consideration of concurrent diseases. Fourth, it’s inadequate to use a single biomarker to guide clinical decision making, therefore, the SII should be regarded as an additional tool in clinical practice.

5. Conclusion

To sum up, our study identifies the SII as an important prognostic predictor in cHCC-CCA. The low cost, easy access and reproducibility make the SII a promising tool for predicting cHCC-CCA prognosis in future clinical practice. However, large-scale external validation studies are still warranted.

Footnotes

Acknowledgments

This work was supported by the National Natural Science Foundation of China [grant number: 81972889] and Excellent Backbone Foundation of Zhongshan hospital, Fudan University [grant number: 2019ZSGG20].

Conflict of interest

The authors declare that they have no competing interests.

Authors’ contributions

Conception: Feng Zhang, Ying-Hong Shi and Xin Yin.

Interpretation or analysis of data: Feng Zhang, Ke-Shu Hu, Shen-Xin Lu and Miao Li.

Preparation of the manuscript: Feng Zhang, Ke-Shu Hu, Shen-Xin Lu and Miao Li.

Revision for important intellectual content: Rong-Xin Chen, Zheng-Gang Ren and Xin Yin.

Supervision: Rong-Xin Chen, Zheng-Gang ren, Ying-Hong Shi and Xin Yin.

Supplementary data

Table S1

Inflammation-based prognostic scores

Criteria	Score
Systemic immune-inflammation index (SII)
Platelet count ( $\times$ 10 ${}^{9}$ /L) $\times$ neutrophil count ( $\times$ 10 ${}^{9}$ /L)/lymphocyte count ( $\times$ 10 ${}^{9}$ /L) $<$ 331	0
Platelet count ( $\times$ 10 ${}^{9}$ /L) $\times$ neutrophil count ( $\times$ 10 ${}^{9}$ /L)/lymphocyte count ( $\times$ 10 ${}^{9}$ /L) $\geqslant$ 331	1
Prognostic nutritional index (PNI)
Albumin (g/L) $+$ 5 $\times$ total lymphocyte count ( $\times$ 10 ${}^{9}$ /L) $\geqslant$ 57.5	0
Albumin (g/L) $+$ 5 $\times$ total lymphocyte count ( $\times$ 10 ${}^{9}$ /L) $<$ 57.5	1
Neutrophil to lymphocyte ratio (NLR)
Neutrophil count ( $\times$ 10 ${}^{9}$ /L)/lymphocyte count ( $\times$ 10 ${}^{9}$ /L) $<$ 3.2	0
Neutrophil count ( $\times$ 10 ${}^{9}$ /L)/lymphocyte count ( $\times$ 10 ${}^{9}$ /L) $\geqslant$ 3.2	1
Platelet to lymphocyte ratio (PLR)
Platelet count ( $\times$ 10 ${}^{9}$ /L)/lymphocyte count ( $\times$ 10 ${}^{9}$ /L) $<$ 84.28	0
Platelet count ( $\times$ 10 ${}^{9}$ /L)/lymphocyte count ( $\times$ 10 ${}^{9}$ /L) $\geqslant$ 84.28	1
Lymphocyte to monocyte ratio (LMR)
Lymphocyte count ( $\times$ 10 ${}^{9}$ /L)/monocyte count ( $\times$ 10 ${}^{9}$ /L) $>$ 6.25	0
Lymphocyte count ( $\times$ 10 ${}^{9}$ /L)/monocyte count ( $\times$ 10 ${}^{9}$ /L) $\leqslant$ 6.25	1
Aspartate aminotransferase to platelet ratio index (APRI)
Aspartate aminotransferase (U/L)/platelet count ( $\times$ 10 ${}^{9}$ /L) $<$ 0.24	0
Aspartate aminotransferase (U/L)/platelet count ( $\times$ 10 ${}^{9}$ /L) $\geqslant$ 0.24	1

Figure S1.

Identification of the optimal cut-off values of different inflammation-base prognostic scores by using receiver operating characteristic (ROC) analysis regarding overall survival. (A) systemic immune-inflammation index (SII); (B) prognostic nutritional index (PNI); (C) neutrophil to lymphocyte ratio (NLR); (D) platelet to lymphocyte ratio (PLR); (E) lymphocyte to monocyte ratio (LMR); (F) aspartate aminotransferase to platelet ratio index (APRI).

References

Nagtegaal

I.D.

Odze

R.D.

Klimstra

Paradis

Rugge

Schirmacher

Washington

K.M.

Carneiro

Cree

I.A.

and Edito

W.H.O.C.T.

, The 2019 WHO classification of tumours of the digestive system, Histopathology, 7.

Brunt

Aishima

Clavien

P.A.

Fowler

Goodman

Gores

Gouw

Kagen

Klimstra

Komuta

Kondo

Miksad

Nakano

Nakanuma

Paradis

Park

Y.N.

Quaglia

Roncalli

Roskams

Sakamoto

Saxena

Sempoux

Sirlin

Stueck

Thung

Tsui

W.M.S.

Wang

X.W.

Wee

Yano

Yeh

Zen

Zucman-Rossi

and Theise

, cHCC-CCA: consensus terminology for primary liver carcinomas with both hepatocytic and cholangiocytic differentation, Hepatology 68 (2018), 113–126.

Yin

Zhang

B.H.

Qiu

S.J.

Ren

Z.G.

Zhou

Chen

X.H.

Zhou

and Fan

, Combined hepatocellular carcinoma and cholangiocarcinoma: clinical features, treatment modalities, and prognosis, Annals of Surgical Oncology 19 (2012), 2869–2876.

Leoni

Sansone

De Lorenzo

Ielasi

Tovoli

Renzulli

Golfieri

Spinelli

and Piscaglia

, Treatment of combined hepatocellular and cholangiocarcinoma, Cancers 12 (2020).

Yap

A.Q.

Chen

C.L.

Yong

C.C.

Kuo

F.Y.

Wang

S.H.

Lin

C.C.

Liu

Y.W.

Lin

T.L.

W.F.

Millan

C.A.

and Wang

C.C.

, Clinicopathological factors impact the survival outcome following the resection of combined hepatocellular carcinoma and cholangiocarcinoma, Surgical Oncology-Oxford 22 (2013), 55–60.

Yamashita

Aishima

Nakao

Yoshizumi

Nagano

Kuroki

Takami

Ide

Ohta

Takatsuki

Nanashima

Ishii

Kitahara

Iino

Beppu

Baba

and Eguchi

, Clinicopathological characteristics of combined hepatocellular cholangiocarcinoma from the viewpoint of patient prognosis after hepatic resection: High rate of early recurrence and its predictors, Hepatology Research, 8.

Hanahan

and Weinberg

R.A.

, Hallmarks of cancer: the next generation, Cell 144 (2011), 646–674.

Chan

S.L.

Wong

L.L.

Chan

K.C.A.

Chow

Tong

J.H.M.

Yip

T.C.F.

Wong

G.L.H.

Chong

C.C.N.

Liu

P.H.

Chu

C.M.

Wong

V.W.S.

K.F.

Reeves

H.L.

and Chan

A.W.H.

, Development of a novel inflammation-based index for hepatocellular carcinoma, Liver Cancer 9 (2020), 167–181.

Mano

Shirabe

Yamashita

Harimoto

Tsujita

Takeishi

Aishima

Ikegami

Yoshizumi

Yamanaka

and Maehara

, Preoperative neutrophil-to-lymphocyte ratio is a predictor of survival after hepatectomy for hepatocellular carcinoma a retrospective analysis, Annals of Surgery 258 (2013), 301–305.

10.

Peng

Zhu

W.J.

Wen

T.F.

Yan

L.N.

Wang

W.T.

and Yang

J.Y.

, Prognostic value of the platelet to lymphocyte ratio change in liver cancer, Journal of Surgical Research 194 (2015), 464–470.

11.

Lin

Z.X.

Ruan

D.Y.

D.H.

X.K.

Chen

Z.H.

Wang

T.T.

Lin

Wen

J.Y.

and Wu

X.Y.

, Lymphocyte-to-monocyte ratio predicts survival of patients with hepatocellular carcinoma after curative resection, World Journal of Gastroenterology 21 (2015), 10898–10906.

12.

Pinato

D.J.

North

B.V.

and Sharma

, A novel, externally validated inflammation-based prognostic algorithm in hepatocellular carcinoma: the prognostic nutritional index (PNI), British Journal of Cancer 106 (2012), 1439–1445.

13.

Yang

X.R.

Sun

Y.F.

Sun

Guo

Zhang

Wang

W.M.

Qiu

S.J.

Zhou

and Fan

, Systemic immune-inflammation index predicts prognosis of patients after curative resection for hepatocellular carcinoma, Clinical Cancer Research 20 (2014), 6212–6222.

14.

Shen

S.L.

S.J.

Chen

Kuang

S.Q.

Hua

Y.P.

Liang

L.J.

Guo

Hao

Y.T.

and Peng

B.G.

, Preoperative aspartate aminotransferase to platelet ratio is an independent prognostic factor for hepatitis b-induced hepatocellular carcinoma after hepatic resection, Annals of Surgical Oncology 21 (2014), 3802–3809.

15.

Tsilimigras

D.I.

Moris

Mehta

Paredes

A.Z.

Sahara

Guglielmi

Aldrighetti

Weiss

Bauer

T.W.

Alexandrescu

Poultsides

G.A.

Maithel

S.K.

Marques

H.P.

Martel

Pulitano

Shen

Soubrane

Koerkamp

B.G.

Endo

and Pawlik

T.M.

, The systemic immune-inflammation index predicts prognosis in intrahepatic cholangiocarcinoma: an international multi-institutional analysis, HPB: the official journal of the International Hepato Pancreato Biliary Association (2020).

16.

Wang

C.W.

Yuan

Y.C.

Zhang

Y.P.

Zou

R.H.

Liao

Y.D.

Liu

W.W.

Yang

Z.W.

Zuo

D.L.

Qiu

J.L.

Zheng

B.K.

and Yuan

Y.F.

, Comparison of the prognostic value of inflammation-based scores in early recurrent hepatocellular carcinoma after hepatectomy, Liver International 40 (2020), 229–239.

17.

Imai

Maeda

Shimokawa

Wang

H.L.

Yoshiya

Takeishi

Itoh

Harada

Ikegami

Yoshizumi

and Mori

, Prognostic nutritional index is superior as a predictor of prognosis among various inflammation-based prognostic scores in patients with hepatocellular carcinoma after curative resection, Hepatology Research 50 (2020), 101–109.

18.

Mandaliya

Jones

Oldmeadow

and Nordman

I.I.C.

, Prognostic biomarkers in stage IV non-small cell lung cancer (NSCLC): neutrophil to lymphocyte ratio (NLR), lymphocyte to monocyte ratio (LMR), platelet to lymphocyte ratio (PLR) and advanced lung cancer inflammation index (ALI), Translational Lung Cancer Research 8 (2019), 886–894.

19.

Nagtegaal

I.D.

Odze

R.D.

Klimstra

Paradis

Rugge

Schirmacher

Washington

K.M.

Carneiro

Cree

I.A.

and Edito

W.H.O.C.T.

, The 2019 WHO classification of tumours of the digestive system, Histopathology (2019).

20.

Allen

R.A.

and Lisa

J.R.

, Combined liver cell and bile duct carcinoma, The American journal of pathology 25 (1949), 647–655.

21.

Liu

C.L.

Fan

S.T.

C.M.

I.O.L.

Lam

C.M.

Poon

R.T.P.

and Wong

, Hepatic resection for combined hepatocellular and cholangiocarcinoma, Archives of Surgery 138 (2003), 86–90.

22.

Yin

Zhang

Wang

Y.H.

Zhang

B.H.

Gan

Y.H.

N.L.

Chen

L.X.

and Ren

Z.G.

, Transcatheter arterial chemoembolization combined with radiofrequency ablation delays tumor progression and prolongs overall survival in patients with intermediate (BCLC B) hepatocellular carcinoma, BMC Cancer 14 (2014), 849.

23.

Ren

Z.G.

Lin

Z.Y.

Xia

J.L.

S.L.

Z.C.

Q.H.

Qin

L.X.

Z.Q.

Fan

and Tang

Z.Y.

, Postoperative adjuvant arterial chemoembolization improves survival of hepatocellular carcinoma patients with risk factors for residual tumor: a retrospective control study, World J Gastroenterol 10 (2004), 2791–2794.

24.

Yamashita

Aishima

Nakao

Yoshizumi

Nagano

Kuroki

Takami

Ide

Ohta

Takatsuki

Nanashima

Ishii

Kitahara

Iino

Beppu

Baba

and Eguchi

25.

Wakizaka

Yokoo

Kamiyama

Ohira

Kato

Fujii

Sugiyama

Okada

Ohata

Nagatsu

Shimada

Orimo

Kamachi

and Taketomi

, Clinical and pathological features of combined hepatocellular-cholangiocarcinoma compared with other liver cancers, Journal of Gastroenterology and Hepatology 34 (2019), 1074–1080.

26.

C.B.

Mao

Y.Z.

Lao

X.M.

S.P.

and Lin

X.J.

, Neutrophil-to-lymphocyte ratio predicts overall survival of patients with combined hepatocellular cholangiocarcinoma, Oncology Letters 15 (2018), 4262–4268.

27.

Wang

Diao

F.Y.

Z.J.

Zhang

X.H.

Zhai

E.T.

Ren

Y.L.

Cai

S.R.

and Chen

J.H.

, Prognostic value of systemic immune-inflammation index in patients with gastric cancer, Chinese Journal of Cancer 36 (2017), 7.

28.

Shao

Y.-X.

Yang

Z.-Q.

Dou

W.-C.

Xiong

S.-C.

and Li

, e systemic immune-inflammation index predicts prognosis of patients with non-metastatic renal cell carcinoma: a propensity score-matched analysis, Cancer Cell International 20 (2020), 222–222.

29.

Hua

Long

Z.-Q.

Zhang

Y.-L.

Wen

Guo

Xia

Zhang

W.-W.

and Lin

H.-X.

, Prognostic value of preoperative systemic immune-inflammation index in breast cancer: a propensity score-matching study, Frontiers in Oncology 10 (2020).

30.

Coffelt

S.B.

Wellenstein

M.D.

and de Visser

K.E.

, Neutrophils in cancer: neutral no more, Nature Reviews Cancer 16 (2016), 431–446.

31.

Mantovani

Cassatella

M.A.

Costantini

and Jaillon

, Neutrophils in the activation and regulation of innate and adaptive immunity, Nature Reviews Immunology 11 (2011), 519–531.

32.

Gabrilovich

D.I.

Ostrand-Rosenberg

and Bronte

, Coordinated regulation of myeloid cells by tumours, Nature Reviews Immunology 12 (2012), 253–268.

33.

Coupland

L.A.

Chong

B.H.

and Parish

C.R.

, Platelets and p-selectin control tumor cell metastasis in an organ-specific manner and independently of NK cells, Cancer Research 72 (2012), 4662–4671.

34.

Carr

B.I.

Cavallini

D’Alessandro

Refolo

M.G.

Lippolis

Mazzocca

and Messa

, Platelet extracts induce growth, migration and invasion in human hepatocellular carcinoma in vitro , Bmc Cancer 14 (2014), 9.

35.

Labelle

Begum

and Hynes

R.O.

, Direct signaling between platelets and cancer cells induces an epithelial-mesenchymal-like transition and promotes metastasis, Cancer Cell 20 (2011), 576–590.

36.

Ferrone

and Dranoff

, Dual roles for immunity in gastrointestinal cancers, Journal of Clinical Oncology 28 (2010), 4045–4051.

37.

Jiang

X.-X.

Huang

X.-T.

Huang

C.-S.

Chen

L.-H.

Liang

L.-J.

and Yin

X.-Y.

, Long-term outcome and prognostic factors of combined hepatocellular carcinoma and cholangiocarcinoma after curative resection, Gastroenterology Report 8 (2020), 134–142.

Prognostic significance of preoperative systemic immune-inflammation index in combined hepatocellular-cholangiocarcinoma

Abstract

BACKGROUND:

OBJECTIVE:

METHODS:

RESULTS:

CONCLUSIONS:

Keywords

1. Introduction

2. Materials and methods

2.1 Patients selection

2.2 Data collection

2.3 Inflammation-based prognostic scores

2.4 Treatment and follow-up

2.5 Statistical analysis

2.6 Ethics approval statement

3. Results

3.1 Patient characteristics

Table 1 Clinical characteristics of enrolled patients ( N = 220)

3.3 Univariable and multivariate analyses

3.4 Propensity score-matching analysis

Table 3 Baseline characteristics of the patients classified by the SII before and after propensity score-matching

Table 5 Comparison of the predictive performances between the inflammation-based scores

4. Discussion

5. Conclusion

Footnotes

Acknowledgments

Conflict of interest

Authors’ contributions

Supplementary data

References

Table 1
Clinical characteristics of enrolled patients ( $N=$ 220)

Table 3
Baseline characteristics of the patients classified by the SII before and after propensity score-matching

Table 5
Comparison of the predictive performances between the inflammation-based scores