Combined detection of CA19-9 and B7-H4 in the diagnosis and prognosis of pancreatic cancer

Abstract

BACKGROUND:

There were no specific indicators for the early detection of pancreatic cancer.

OBJECTIVE:

To analyze the diagnostic and prognostic value of CA19-9 plus B7-H4 detection in preoperative serum or surgical tissues of patients with pancreatic cancer.

METHODS:

One hundred and eighty-eight patients with pancreatic cancer and 25 controls were recruited. Their preoperative serum CA19-9 level was detected chemiluminescently, and B7-H4 expression in pancreatic cancer tissues was assessed immunohistochemically. The diagnostic and prognostic utility of detecting CA19-9, B7-H4 and their combination was evaluated.

RESULTS:

CA19-9 and B7-H4 levels were significantly upregulated in patients with pancreatic cancer compared with those in controls. The diagnostic value of combined CA19-9 plus B7-H4 detection was markedly better than that achieved from their separate detection analyzed with receiver operating characteristic curve. Sensitivity and specificity of the combined detection in the pancreatic cancer group was significantly increased compared with single detection. B7-H4 detection showed better prognostic value than detection of CA19-9. However, CA19-9 had high sensitivity and low specificity, while B7-H4 showed the opposite. The sensitivity of the combined prognosis was not significantly different to B7-H4 alone.

CONCLUSION:

The combined detection of CA19-9 and B7-H4 could become a new method for the clinical diagnosis and prognosis of pancreatic cancer.

1. Introduction

Pancreatic ductal adenocarcinoma refers to a malignant lesion of the pancreatic ductal epithelium under various carcinogenic factors, such as environment or heredity, and is one of the most malignant tumors of the digestive system. The incidence and mortality of pancreatic cancer in US are ranked 10th and 3th among malignant tumors, respectively [1]. Pancreatic cancers are typically diagnosed at an advanced stage, for which the 5-year survival rate is 3% [2]. Most of the patients are in the late stage of pancreatic cancer and mortality is very high. The incidence and mortality of pancreatic cancer are increasing in China [3].

There is still a lack of specific indicators for the early detection of pancreatic cancer. Invasion and metastasis are still the main causes of death in pancreatic cancer. More than 80% of patients had local invasion or distant metastasis at the time of diagnosis. In addition, most patients with pancreatic cancer are faced with a grim median survival of 5 to 6 months [2]. However, the 5-year survival of stage I pancreatic cancer is close to 70% [4]. Therefore, early diagnosis of pancreatic cancer can significantly improve patient prognosis, which has great clinical value [5]. Therefore, it is particularly important to identify biomarkers that are closely related to the growth and progression of pancreatic cancer.

Until now, serum carbohydrate antigen (CA) 19-9 was the most clinically effective diagnostic biomarker for pancreatic cancer. Serum CA19-9 levels were associated with prognosis, overall survival, sensitivity to chemotherapy, and postoperative recurrence in patients with pancreatic cancer. However, serum CA19-9 levels in patients with pancreatic cancer are less sensitive for diagnosis. Serum CA19-9 levels had no effect in screening asymptomatic individuals and were 79–81% sensitive and 82–90% specificity for the diagnosis of pancreatic cancer patients with symptomatic [6]. Therefore, the detection of serum CA19-9 level alone cannot achieve the purpose of diagnosis of pancreatic cancer. Combined detection of pancreatic cancer-related molecules may improve the early diagnosis rate and prognosis.

Recently, the relationship between abnormal expression of the B7 family immune costimulatory molecules and tumor formation has attracted attention. Studies found that abnormal expression of B7 family member, H4 (B7-H4) was involved in the development and progression of tumors in the digestive system (including esophagus [7], stomach [8], liver [9], pancreas [10], and bile duct [11]). However, there have been relatively few studies on B7-H4 in the diagnosis and prognosis of pancreatic cancer. In the present study, we evaluated the expression of CA19-9 and B7-H4 in different populations, and evaluated the clinical diagnosis and prognostic value of single and combined detection in patients with pancreatic cancer, aiming to improve the sensitivity and specificity of tumor markers in the diagnosis and prognosis of pancreatic cancer, and to assist clinical monitoring and evaluation of curative effects.

2. Materials and methods

2.1 Patients and tissue specimens

This study analyzed tissue samples from 188 patients pathologically diagnosed with pancreatic cancer between January 2011 and December 2014 at The Second Affiliated Hospital, Zhejiang University School of Medicine. All the cases included were adenocarcinoma. Their median age was 65, and ranged from 35 to 82; 112 were male, and 76 were female. Clinical parameters included patient demographics, tumor differentiation, tumor stage, and treatment modality. Furthermore, 25 patients with benign pancreatic disease, such as pancreatitis and cysts, were enrolled as a control group, and were accrued in the same period from January 2011 to December 2014 (their median age was 55, and ranged from 39 to 76; 16 were male, and 9 were female). All tissue samples were fixed in 10% buffered formalin and embedded in paraffin. All archival hematoxylin and eosin (H&E)-stained sections were reviewed by two pathologists. The serum samples were collected in separation gel vacuum collective tubes. The present study was approved by the Ethics Committee of The Second Affiliated Hospital, Zhejiang University School of Medicine (No. 2018-024).

2.2 The detection of serum CA19-9

The CA19-9 level in serum samples was detected by a chemiluminescent method, using a Beckman Coulter UniCel DxI 800 analyzer (Brea, CA, USA) and supporting reagents. The normal reference value was 0–37 kU/L.

2.3 Immunohistochemical staining

Immunohistochemical staining was performed using a two-step EnVisiontexttrademark method (Dako, Glostrup, Denmark). Briefly, paraffin-embedded tissues were cut into 5- $\mu$ m serial sections, transferred onto adhesive slides, and dried at 65 ${}^{\circ }$ C for 2 h. The sections were deparaffinized with xylene and rehydrated through a graded alcohol series. Endogenous peroxidase activity was blocked using 0.3% hydrogen peroxide solution for 30 min at room temperature and antigen retrieval was performed at 100 ${}^{\circ }$ C for 30 min in citrate buffer (10 mmol/L; pH 6.0). After washing three times with phosphate-buffered saline (PBS) for 5 minutes each, sections were incubated with 10% normal goat serum to block nonspecific binding. Sections were then incubated with a rabbit anti-human B7-H4 monoclonal antibody (1:400 dilution; clone number EP1165; Abcam, MA, USA) at 4 ${}^{\circ }$ C overnight, followed by immunodetection using the DAKO EnVision detection system (K5007). Slides were counterstained with Mayer’s hematoxylin, dehydrated in graded alcohol, and mounted with neutral resin. Negative control staining was performed by replacing the primary antibody with PBS. Human tonsil tissue was used as a positive control sample for B7-H4 expression.

2.4 Evaluation of B7-H4 staining

The sections were examined and scored by two pathologists without knowledge of the patient’s clinical record. Five tumor fields at 400 $\times$ magnification were randomly selected. Cytoplasm and/or tumor cell membrane staining were considered to indicate positive expression. A previous semiquantitative system [12] was used to determine B7-H4 expression. The proportion of positive cells was scored as follows: 0 ( $<$ 5%); 1 (6–25%); 2 (26–50%); 3 (51–75%); 4 ( $>$ 75%). Staining intensity was evaluated as follows: 0 (no staining); 1 (weak staining, light yellow); 2 (moderate staining, yellowish brown); 3 (strong staining, brown). The sum score was determined by multiplying the positive proportion score by the intensity score: 0 (negative); 1–4 (weakly positive); 5–8 (moderately positive); and 9–12 (strongly positive), as seen in Fig. 1. In the present study, B7-H4 expression was considered as low expression when the score was $<$ 9, and high expression when the score was $⩾$ 9.

Figure 1.

B7-H4 expression in pancreatic cancer tissue. Representative images of negative staining (staining score $=$ 0) in normal/benign pancreas tissue (A); weak staining (staining score: 1–4) (B); moderate staining (staining score: 5–8) (C); strong staining (staining score: 9–12) (D). B7-H4 expression was considered as low expression when the score was $<$ 9, and high expression when the score was $⩾$ 9. Original magnification $\times$ 200 (for all photomicrographs).

2.5 Statistical analysis

Statistical analysis was performed using SPSS v17.0 (IBM Corp., Armonk, NY, USA)and MedCalc 15.2.2 software. The data were expressed as X $\pm$ SD, and $P<$ 0.05 was considered statistically significant. Variance analysis and rank sum tests were used for data discrepancy analysis. Based on CA19-9 values and B7-H4 staining scores, a logistic regression model was established for the receiver operating characteristic curve (ROC curve) to evaluate the value of the separate detection of CA19-9 and B7-H4, and a combination of the two in the diagnosis of pancreatic cancer. The area under the curve (AUC), critical value, 95% Confidence Interval (95% CI), specificity and sensitivity were also calculated.

3. Results

3.1 General and experimental data comparison

This study showed that compared with the benign pancreatic disease group, the serum CA19-9 value (1974.70 $\pm$ 3651.36 kU/L vs. 154.65 $\pm$ 439.19 kU/L) and B7-H4 staining score (5.10 $\pm$ 2.97 vs. 0.86 $\pm$ 0.75) were significantly higher in the pancreatic cancer group ( $P<$ 0.05). The characteristics of the patient population are summarized in Table 1.

Table 1

Characteristics of patients with pancreatic cancer and experimental data comparison

Characteristic	Pancreatic cancer group	Benign pancreatic disease group
Age (years)
Median	65	55
Range	35–82	39–76
Gender (number)
Male	112	16
Female	76	9
CA19-9 (kU/L)	1974.70 $\pm$ 3651.36 ${}^{\ast }$	154.65 $\pm$ 439.19
B7-H4 (staining score)	5.10 $\pm$ 2.97 ${}^{\ast \ast }$	0.86 $\pm$ 0.75

*: vs. benign pancreatic disease group, $P<$ 0.05; **: vs. benign pancreatic disease group, $P<$ 0.01.

3.2 Correlation between CA19-9 and B7-H4 expression in pancreatic cancer

There was no significant correlation between the expression level of B7-H4 in the tumor tissues of patients with pancreatic cancer (5.10 $\pm$ 2.97) and the CA19-9 expression in preoperative serum (1974.70 $\pm$ 3651.36 kU/L) using Pearson’s test ( $P>$ 0.05). According to Spearman’s test, the positive rate of B7-H4 high expression in the tumor tissues of patients with pancreatic cancer (16.5%) was also not correlated with the positive rate of preoperative serum CA19-9 expression (81.9%, $P>$ 0.05).

3.3 Combined detection of CA19-9 and B7-H4 in the diagnosis of pancreatic cancer

According to the CA19-9 values and B7-H4 staining scores of the two patient groups, a logistic regression model was established and a ROC curve was used to fit the combined diagnostic curve to evaluate the diagnostic value of CA19-9 plus B7-H4 in pancreatic cancer (Fig. 2). The larger the area under the curve (AUC), the better the diagnostic performance. AUC $<$ 0.5 was considered to have no diagnostic significance and AUC $>$ 0.9 was considered to have a high diagnostic accuracy. The AUC of CA19-9 alone was 0.869 (95% CI: 0.815–0.912). The AUC of B7-H4 alone was 0.958 (95% CI: 0.920–0.981). The AUC of CA19-9 plus B7-H4 was 0.976 (95% CI: 0.944–0.992), which was larger than the AUCs of CA19-9 alone and B7-H4 alone ( $P<$ 0.05).

Figure 2.

Combined detection of CA19-9 and B7-H4 in the diagnosis of pancreatic cancer. A receiver operating characteristic (ROC) curve was used to fit the combined diagnostic curve to evaluate the diagnostic value of CA19-9 plus B7-H4 in pancreatic cancer, compared with CA19-9 alone and B7-H4 alone. The larger the area under the curve, the better the diagnostic performance.

The threshold was the value of the point on the curve when the Youden index (the sum of sensitivity and specificity minus 1) was the greatest. A value greater than the threshold may be considered as a possibility of pancreatic cancer.

The specificity and sensitivity of the combined detection of CA19-9 and B7-H4 were 95.2% and 92.3%, respectively. The specificity and sensitivity of serum CA19-9 alone were 90.5% and 75.7%, respectively. The specificity and sensitivity of B7-H4 were 100.0% and 85.1%, respectively. Thus, the sensitivity of the combined detection was significantly higher than that of either single test. The specificity between CA19-9 and B7-H4 is shown in Table 2.

Table 2

Diagnostic effectiveness of CA19-9, B7-H4, and combined detection in pancreatic cancer

Project	AUC (95% CI)	Threshold	Sensitivity (%)	Specificity (%)
CA19-9	0.869 (0.815–0.912)	72.3 kU/L	75.7	90.5
B7-H4	0.958 (0.920–0.981)	2.125	85.1	100.0
CA19-9 plus B7-H4	0.976 (0.944–0.992)	2.0241*	92.3	95.2

*: Combined predictors obtained by logistic regression equation, which was B7-H4+CA19-9/1972. AUC, area under the receiver operating characteristic (ROC) curve; CI, confidence interval.

3.4 Combined detection of CA19-9 and B7-H4 in the prognosis of pancreatic cancer

The ROC curve was also used to fit the combined prognostic curve to evaluate the prognostic value for six-month survival of CA19-9 plus B7-H4 in pancreatic cancer (Fig. 3). The AUC of CA19-9 alone was 0.575 (95% CI: 0.499–0.648), and the AUC of B7-H4 alone was 0.708 (95% CI: 0.636–0.773), which was larger than that of CA19-9 alone ( $P<$ 0.05). The AUC of CA19-9 plus B7-H4 was 0.717 (95% CI: 0.646–0.782), which was larger than that of CA19-9 alone ( $P<$ 0.05), but not significantly different to that of B7-H4 alone.

Figure 3.

Combined detection of CA19-9 and B7-H4 in prognosis of pancreatic cancer. A receiver operating characteristic (ROC) curve was used to fit the combined prognostic curve to evaluate the prognostic value of CA19-9 plus B7-H4 in pancreatic cancer, compared with CA19-9 alone and B7-H4 alone. The larger the area under the curve, the better the diagnostic performance.

The specificity and sensitivity of serum CA19-9 alone in prognosis for six-months survival were 26.2% and 88.2%, respectively. The specificity and sensitivity of B7-H4 alone were 65.4% and 68.6%, respectively, as shown in Table 3. The sensitivity and sensitivity of the combined prognosis were not significantly different to those of B7-H4 alone.

Table 3

Prognostic effectiveness of CA19-9, B7-H4, and combined detection in pancreatic cancer (6-month survival)

Project	AUC (95% CI)	Threshold	Sensitivity (%)	Specificity (%)
CA19-9	0.575 (0.499–0.648)	61.05 kU/L	88.2	26.2
B7-H4	0.708 (0.636–0.773)	4.25	68.6	65.4

AUC, area under the receiver operating characteristic (ROC) curve; CI, confidence interval.

4. Discussion

In recent years, the incidence of pancreatic cancer has increased worldwide [1, 13]. How to evaluate the curative effect, evaluate prognosis, and predict postoperative recurrence and survival has attracted more and more attention [14, 15]. The expression of B7-H4 and CA19-9 has an important application in assessing pancreatic cancer metastasis, monitoring recurrence, and the clinical staging of tumors [10, 16]. In the present study, we analyzed the diagnostic and prognostic value of pancreatic cancer serum tumor marker CA19-9 and negative immunostimulatory molecule B7-H4 in pancreatic cancer. Our findings suggested that the diagnostic value of CA19-9 combined with B7-H4 for pancreatic cancer was higher than that of CA19-9 or B7-H4 alone. In addition, the prognostic value of CA19-9 combined with B7-H4 for pancreatic cancer was higher compared with that of CA19-9, but was not different to that of B7-H4. The optimal threshold values of CA19-9 and B7-H4 determined by ROC curve analysis could improve the diagnosis of pancreatic cancer when used together, resulting in their optimal application and the promotion of clinical screening and diagnosis of pancreatic cancer.

Most researchers believed that serum CA19-9 is a better marker of digestive system tumors [15, 17], especially pancreatic cancer. CA19-9 is not only a diagnostic indicator, but also a predictor of the therapeutic effect and prognosis of pancreatic cancer [6]. However, CA19-9 is not very sensitive for pancreatic cancer [18]. Therefore, CA19-9 should be combined with other tests to diagnose pancreatic cancer to help in the early diagnosis of pancreatic cancer, together with pathologist’s qualitative analysis of tumors as an auxiliary diagnosis. Postoperative recurrence and metastasis are the main causes of death in patients with pancreatic cancer. To evaluate the probability of recurrence and metastasis of pancreatic cancer, and the progress of the disease, more comprehensively and accurately, it is important to detect a combination of specific markers of pancreatic cancer.

B7-H4 is a potential clinical predictor in the diagnosis or prediction of clinical outcomes for tumors [19]. A high correlation between the serum levels of B7-H4 and the development and progression of tumors was observed. In addition, B7-H4 can be a prognostic marker for tumors [20, 21, 22], which was consistent with the results of our study. However, B7-H4 is also highly expressed in other gastrointestinal tumors [7, 8, 9, 10, 11]. In terms of a single marker for the diagnosis of pancreatic cancer, B7-H4 has higher specificity with a relatively low sensitivity. CA19-9 has higher sensitivity with relatively low specificity. Thus, neither of them can be used alone as an effective tumor marker to identify pancreatic cancer. Our data showed that the combined use of CA19-9 and B7-H4 could improve the diagnosis of pancreatic cancer, which indicated that the discriminating capabilities of CA19-9 for pancreatic cancer could be significantly improved by the detection of B7-H4.

Previously, we identified that the expression of B7-H4 in pancreatic cancer tissues correlated with tumor stage, and patients with advanced stage pancreatic cancer tended have higher levels of B7-H4 than did those with early stage cancer [10]. As most advanced cases of pancreatic cancer were metastatic tumors, these results were consistent with the previous report that B7-H4 was involved in the process of pancreatic cancer metastasis [12, 23]. B7-H4 has important prognostic value for multiple tumors, such as renal cell carcinoma [20], intrahepatic cholangiocarcinoma [11], non-small cell lung cancer [24], breast cancer [25], and cervical cancer [26]. Moreover, high levels of B7-H4 expression were associated with poor prognosis of patients with pancreatic cancer. B7-H4 [10] and preoperative serum CA19-9 levels [14] were significantly higher in patients with pancreatic cancer with distant metastasis than in those without. Multivariate survival analysis using the Cox proportional hazards model showed that B7-H4 and CA19-9 were independent prognostic factors of pancreatic cancer [10, 27]. In the present study, our data indicated that B7-H4 was an exceedingly promising diagnostic and prognostic biomarker for pancreatic cancer. The performance of B7-H4 in prognosing pancreatic cancer as an individual biomarker was significant, particularly in survival prediction. Therefore, B7-H4 could improve the performance of CA19-9 in pancreatic cancer.

Our study predicted the diagnostic value of B7-H4 in pancreatic cancer, although the expression of serum B7-H4 is sure more diagnostic. It was reported that soluble B7-H4 expression could be detected in the serum of some tumor patients, and the expression level of serum B7-H4 was associated with tumor progression [20]. We had prepared a variety of monoclonal antibodies against B7-H4 [28], and we will prepare a reagent for detecting serum B7-H4 to analyze the expression and clinical significance of B7-H4 in serum of patients with pancreatic cancer.

In summary, the addition of B7-H4 detection could improve the performance of CA19-9 detection in the diagnosis of pancreatic cancer. The present study suggested that B7-H4 was a potential diagnostic and prognostic biomarker complementing CA19-9 to detect pancreatic cancer. Combined detection of B7-H4 and CA19-9 may prove to be useful for the diagnosis of pancreatic cancer and to predict prognosis.