Serum CA19-9 Level is Correlated with Liver Inammation, Cirrhosis, and Poor Prognosis after Curative Resection in Hepatitis B-related Hepatocellular Carcinoma

Background: The relationship between serum carbohydrate antigen 19-9 (CA19-9) and hepatocellular carcinoma (HCC) is unclear. To explore the prognostic value of preoperative CA19-9 in α-fetoprotein (AFP)-positive and -negative HCC with hepatitis B virus (HBV) background (HBV-HCC). Methods: We evaluated HBV-HCC patients who underwent curative resection (Cohort 1). Recurrence-free survival (RFS) and overall survival (OS) were assessed, and the clinicopathological characteristics were compared according to a CA19-9 cutoff of ≥ 39 U/ml. Immunohistochemical staining of CA19-9 in HCC tumor tissue and background liver were quantied in another cohort of 216 patients with resected HCC (Cohort 2). Immunohistochemical staining or serum CA19-9 level was also compared between patients with intrahepatic cholangiocarcinoma (ICC) (Cohort 3). Results: In Cohort 1, CA19-9 ≥ 39 U/ml was an independent risk factor for RFS (HR=1.507, 95% CI=1.087-2.091, p=0.014) and OS (HR=1.646, 95% CI=1.146-2.366, p=0.007) in both AFP-positive and AFP-negative patients. CA19-9 ≥ 39 U/ml was also associated with signicantly higher incidence of macrovascular invasion compared with CA19-9 < 39 U/ml (23.0% vs. 7.2%, p=0.002), elevated aminotransferase and aspartate aminotransferase to platelet ratio index, and lower albumin. In contrast, no association was found between CA19-9 and systemic inammation. CA19-9 expression was found exclusively in the background liver but not in HCC tumor cells. In contrast, tumor tissue was the main source of CA19-9 in ICC patients. Conclusions: associated with lower and Comparison of clinicopathological factors between CA19-9 (+) and CA19-9(-) patients revealed that CA19-9 (+) patients tend to be older (mean age: 58.4 ± 10.4 years vs. 55.4 ± 10.6 years, p = 0.048), have higher incidence of liver cirrhosis (70.5% vs. 56.1%, p = 0.037), higher APRI (1.53 ± 1.61 vs. 0.72 ± 0.96, p < 0.001), longer PT (12.1 ± 1.0 vs. 11.3 ± 1.2, p < 0.001), elevated ALT (75.4 ± 77.3 U/L vs. 43.9 ± 59.5 U/L, p = 0.004), elevated AST (75.1 ± 69.0 U/L vs. 41.5 ± 46.0 U/L, p < 0.001), increased rGT (147.1 ± 162.4 U/L vs. 81.4 ± 99.4 U/L, P = 0.003), higher level of ALP (117.5 ± 70.2 ng/ml vs. 95.7 ± 52.9 ng/ml, p = 0.006), and lower level of albumin (39.8 ± 5.5 g/L vs. 42.0 ± 5.3 g/L, p = 0.002) (Table 3). All the factors except MaVI are related to liver cirrhosis. To The results showed that still liver ALT, AST,

Background Primary liver cancer is the sixth most commonly diagnosed cancer and the fourth leading cause of cancer-related death worldwide [1]. In general, primary liver cancer is classi ed into two types as hepatocellular carcinoma (HCC) and intrahepatic cholangiocarcinoma, with HCC being more common, accounting for 75%-85% of all cases. However, mixed HCC-ICC and other rare types have also been reported. Alpha-fetoprotein (AFP) and carbohydrate antigen 19 − 9 (CA19-9) are the most commonly used biomarkers for HCC and ICC. A higher AFP level is associated with poor outcomes after curative resection or liver transplantation [2][3][4]. CA19-9, also known as sialyl Lewis-a, is mainly used as a biomarker for malignancies of the hepatobiliary tract and pancreas [5]. However, serum CA19-9 levels may also be elevated in gastric, esophageal, and colonic cancers and in a number of non-malignant conditions that are associated with jaundice [6]. The serum CA19-9 level is elevated in approximately 60% of cholangiocarcinoma patients and in 30% of HCC patients [7]. It is also frequently elevated in patients with combined HCC-cholangiocarcinoma. Elevated preoperative serum CA19-9 levels have been reported to be associated with worse survival in HCC patients who had undergone resection (> 27 U/mL) or liver transplantation (> 100 U/ml) [8][9][10]. However, most of these studies mainly included patients with HCVrelated HCC who underwent resection or transplantation, and the underlying mechanism by which CA19-9 in uences prognosis remains unclear.
Thus, this study aimed to investigate the prognostic value of preoperative serum CA19-9 according to AFP status in HCC patients and in ICC patients with HBV background who underwent curative resection.

Patients and study design
We retrospectively evaluated three patient cohorts as follows. between 2012 and 2013. In this cohort, CA19-9 (+) was de ned as serum CA19-9 ≥ 39 U/ml, whereas CA19-9 (-) was de ned as serum CA19-9 < 39 U/ml, according to the upper limit of serum CA19-9 in our hospital. AFP (+) was de ned as serum AFP > 20 ng/ml, whereas AFP (-) was AFP ≤ 20 ng/ml. Cohort 2 involved 216 patients with resected HCC in whom tissue microarray (TMA) samples were obtained. Patients with lymph node metastasis or distant metastasis were excluded to reduce the confounding factors. Cohort 3 included 136 ICC patients whose para n blocks were collected based on the availability of suitable formalin-xed, para n-embedded tissue and complete clinicopathologic and follow-up data of the patients.
All the patients underwent curative resection for HCC, de ned as complete macroscopic removal of the tumor. All tumors were staged according to the TNM classi cation system of International Union Against Cancer (8th edition) and the Barcelona Clinic Liver Cancer guidelines. Systemic factors included the neutrophil-to-lymphocyte ratio (NLR) and platelet-to-lymphocyte ratio (PLR). Liver factors included intraoperative detection of liver cirrhosis; alanine aminotransferase (ALT), aspartate aminotransferase (AST), gamma-glutamyl transpeptidase (rGT), alkaline phosphatase (ALP), total bilirubin, and albumin levels; prothrombin time (PT); and the aspartate aminotransferase-to-platelet ratio index (APRI) as the parameter most closely related to liver cirrhosis and brosis in both chronic hepatitis B [11][12][13] and hepatitis C [14,15]. NLR and PLR are both indicators of systemic in ammation and its relationship with the prognosis of several cancers has been identi ed [16][17][18][19].

Follow-up And Postoperative Treatment
After surgery, patients were routinely followed. Follow-up was censored on July 10, 2019, with the patients followed up for a median of 56.6 months. All patients were monitored prospectively according to serum AFP and carbohydrate antigen 19 − 9 levels and using abdomen ultrasonography every 2 months in the rst year and every 3 months after the rst year. Recurrence was con rmed using computed tomography and/or magnetic resonance imaging based on typical imaging appearance in the imaging scan and an elevated AFP level. The treatment modality after relapse varied among individuals.

Tma And Immunohistochemistry
TMAs were constructed as described previously [20]. TMAs sections were dewaxed and rehydrated in xylene and gradient ethanol, respectively. After antigen retrieval and endogenous peroxidase activity blocking, the slides were incubated with primary antibodies (4 °C for 14 h and 37 °C for 1 h) and HRPconjugated secondary antibody (37 °C for 1 h) in turn. Then the sections were visualized with 3,3'diaminobenzidine (ZLI-9017; Zhongshan Goldbridge) for 5 min and counterstained with Hematoxylin. Appropriate internal or external positive controls and negative controls were designed and used for each round [20]. The primary antibody used in the present study was anti-human CA19-9 (Zhongshan Company). Negative control slides omitting the primary antibodies were included in all assays. CA19-9 immunoreactivity was evaluated in a semiquantitative manner by evaluating both staining intensity and the percentage of immunopositive tumor cells.

Statistical analysis
Kaplan-Meier method and Log-rank test with survival curves were used in univariate analyses of cumulative survival. Multivariate analyses were based on the Cox proportional hazards regression model.
For the comparison of individual variables, χ 2 tests and Student's t-tests were used as appropriate. All statistical analyses were performed using SPSS software (SPSS v22.0, Chicago, IL). A two-tailed P value of < 0.05 was considered statistically signi cant.
CA19-9 was associated with higher incidence of MaVI and a trend toward multiple tumors  CA19-9 was associated with more severe liver cirrhosis and liver in ammation but not with systemic in ammation Comparison of clinicopathological factors between CA19-9 (+) and CA19-9(-) patients revealed that CA19-9 (+) patients tend to be older  Table 3). All the factors except MaVI are related to liver cirrhosis.
To exclude the confounding effect of MaVI, we excluded patients with MaVI. The results showed that CA19-9 was still correlated with liver cirrhosis, APRI, ALT, AST, rGT, ALP, and albumin (data not shown). Furthermore, multivariate analysis showed that CA19-9 (+) and MaVI (+) were both independent risk factors for RFS.
In the current study, CA19-9 was not correlated to NLR or PLR, indicating that CA19-9 was not correlated to systemic in ammation.

Immunohistochemical staining
To determine the source of CA19-9, we examined its expression in TMA samples of HCC patients. Immunohistochemical staining of CA19-9 in both tumor tissue and non-tumor liver parenchyma specimens from HCC patients was also assessed. The results showed that none of the HCC tumor cells express CA19-9, and CA19-9 was expressed in non-tumor liver parenchyma in all patients.
Immunohistochemical staining of CA19-9 in both tumor and non-tumor liver parenchyma samples from Cohort 3 revealed that CA19-9 was expressed in 64% (87/136) of ICC tumors and 4.4% (6/136) of nontumor liver parenchyma. Serum CA19-9 was positive (≥ 39 U/ml) in 58.1% and negative (< 39 U/ml) in 41.9% of the patients with ICC. The results that immunohistochemical staining of CA19-9 was positive only in 4.4% of ICCs indicate that serum CA19-9 mainly derives from the tumor tissue of patients with ICC, which is distinct from the dominant expression of CA19-9 in the background liver in HCC patients.

Discussion
In the current study, it is revealed that CA19-9 is associated with lower OS and RFS in both AFP (+) and AFP (-) patients. CA19-9 is associated with higher incidence of macrovascular invasion and more severe liver cirrhosis and in ammation. Importantly, immunostaining of CA19-9 showed that it is secreted by the background liver, but not by tumor cells in patients with HCC.
CA19-9 is generally recognized as a tumor marker for gastrointestinal and especially hepatobiliary cancers [21], however, CA19-9 is frequently elevated in biliary obstruction [22]. In hepatocellular carcinoma, CA19-9 elevation may stand for a subgroup of patients with partial cholangiocytic differentiation [23]. However, the signi cance of CA19-9 in pure HCC is still far from clear. In the current study, we excluded mixed HCC-ICC as well as cholangiocytic differentiation by two independent pathologists, and immunohistochemical staining of CA19-9 in HCC and ICC tumor tissues as well as nontumor liver parenchyma con rmed that CA19-9 is secreted exclusively by the background liver but not by tumor cells in patients with HCC. In contrast, CA19-9 was frequently expressed in ICC tumors (64%, n = 87/136), but rarely in the non-tumor liver parenchyma (4.4%, 6/136).
In Cohorts 1 and 2, CA19-9 ≥ 39 U/ml was associated with a high incidence of MaVI, elevated rGT, ALT, AST, PT, and APRI; and decreased albumin. Furthermore, there were no signi cant difference in NLR and PLR, as biomarkers of systemic in ammation, between CA19-9 (+) and CA19-9 (-) patients, indicating that CA19-9 is not associated with systemic in ammation. Previous reports have shown that elevated ALT, AST, and rGT levels are correlated to liver brosis, cirrhosis [24,25], and recurrence [26,27]. Thus, we con rmed that CA19-9 is a liver biomarker that higher serum CA19-9 level indicated more severe liver in ammation and liver cirrhosis rather than a tumor factor in HCC.
Macrovascular invasion and multiple tumor nodules were more common in CA19-9 (+) patients. These ndings can be attributed to two reasons. First, chronic in ammatory diseases are the key etiological risk factors for HCC [28], and an elevated ALT/AST/APRI in patients with elevated CA19-9 indicated more severe liver cirrhosis and an in amed liver background [29,30], which is closely related to de novo tumor pathogenesis and multicentric recurrence [31,32]. Second, liver in ammation may promote intravenous metastasis of HCC tumor cells and has been reported as an independent risk factor for early tumor recurrence in patients with HCC [33][34][35].
Further, a preoperative CA19-9 value of > 27 U/mL was associated with poor prognosis after resection for HCC. However, the authors did not compare the clinicopathological factors related to liver in ammation and liver cirrhosis (e.g., ALT, AST, ALP, rGT, and albumin) according to a CA19-9 cutoff of ≤ 27 U/mL. Wan et al. [10] also showed that preoperative serum AFP levels of > 400 ng/ml and CA19-9 > 100 U/ml predicted survival after liver transplantation in patients with HCC. However, the study only included patients with liver transplantation and did not exclude patients with ICC or cHCC-CC. Hsu et al. [8] investigated 145 cases of HCV-related HCC and found that an elevated serum CA19-9 level of ≥ 100 U/ml is an independent predictor of poor OS. However, 59% (85/145) of the HCC patients had underlying hepatitis C virus, and only 16% had hepatitis B virus. Lu et al. [36] evaluated 750 cases of AFP-negative HCC and found that a preoperative CA19-9 level of > 32.6 U/ml predicted poor prognosis and can thus be used as a prognostic marker in AFP-negative HCC.
The current study is the only study including purely patients with HBV-related HCC. At a cut-off value of 39 U/mL, 16.1% of patients (61/380) were found to be CA19-9 positive, and CA19-9 ≥ 39 U/ml predicted worse OS and RFS in both AFP(-) and AFP (+) patients. These ndings indicate that patients with CA19-9 ≥ 39 U/ml should be closely followed up considering the higher incidence of tumor recurrence and mortality rate caused by more severe liver cirrhosis and in ammation in this population. Antiin ammatory medications may be helpful in reducing the risk of tumor recurrence.
Our study has several limitations. First, it is a single-center study of retrospective cohorts, and only the serum level and immunohistochemical expression of CA19-9 were evaluated. Second, the precise mechanism by which CA19-9 promotes macroscopic vascular invasion is still unclear and thus further studies are needed to elucidate the underlying mechanism.

Conclusions
In conclusion, CA19-9 is associated with lower OS and RFS in both AFP (+) and AFP (-) patients.
Importantly, CA19-9 is secreted by the background liver, but not by tumor cells in patients with HCC. Thus, CA19-9 is not a tumor biomarker, but a biomarker for liver cirrhosis and in ammation and a risk factor for worse OS and RFS in HCC.
APRI, Aspartate aminotransferase to Platelet Ratio Index.
AST, Aspartate Aminotransferase.  AFP, CA19-9 and combination to predict OS and RFS for HCC patients after curative resection.

Figure 2
Immunohistochemical staining of CA19-9 in HCC and ICC in Cohort 2 and 3. A, HCC tumor tissue was negative for CA19-9. B, HCC non-tumor liver parenchyma was positive for CA19-9 in the portal area. C, ICC tumor tissue was positive for CA19-9. D, Positive staining of CA19-9 in the portal area in non-tumor liver parenchyma of ICC. CA19-9 in HCC and ICC in Cohort 2 and 3. a, comparison of serum CA19-9 level in ICC and HCC patients.