Introduction
Gallbladder cancer (GBC) originating from biliary duct epithelia is a highly malignant tumor ranked first in bile duct cancers and sixth in digestive duct cancers [
1]. Although the molecular mechanism is still unclear, risk factors of this lethal disease include chronic inflammation, aging, and the female gender [
2,
3]. The number of new GBC cases increases every year worldwide, but the treatment outcome is far from satisfactory, and such lack of efficiency mainly results from advanced diagnosis. Some GBC patients are unexpectedly diagnosed in the early stages of GBC during laparotomy or laparoscopy procedures for cholelithiasis, and could obtain good treatment effects. Therefore, at present, the treatment outcome of GBC patients mainly depends on early diagnosis.
However, the early diagnosis for GBC is very difficult in clinical practice. Around 20% of GBC patients’ cancer was confined to their gallbladder at the time of diagnosis, whereas the rest usually have infringed adjacent organs or the cancer has spread to distant sites [
4]. The possible reasons for this may be as follows. First, early symptoms are vague and non-specific; thus, we are usually not vigilant against GBC. Second, effective markers for early GBC diagnosis are lacking. Researchers conduct numerous studies to search for diagnosis markers to improve GBC prognosis. Considering the familial aggregation of GBC, some scholars conducted related investigations on the gene level [
1]. Jiao
et al. revealed that the exonic variants of xeroderma pigmentosum complementation group C (XPC) were associated with GBC risk in the Chinese population [
5]. In the mice model, interaction between liver X receptor-β, estrogen, and TGF-β signaling may play a crucial role in the malignant transformation of gallbladder epithelium [
6]. Additionally, single-nucleotide polymorphism of related genes is implicated with GBC risk [
5,
7-
12]. Nevertheless, these findings are far from routine utilization in clinical practice.
In view of the facts reviewed above, we attempted to determine effective markers for GBC diagnosis based on clinical data to prolong the survival time of GBC patients.
Materials and methods
Study population and samples
The clinical data of 144 GBC patients (34 with choledocholithiasis, 23.61%) and 116 cholelithiasis patients (19 with choledocholithiasis, 16.38%) treated at the First Affiliated Hospital, School of Medicine, Xi’an Jiaotong University from January 2009 to December 2013 were retrospectively reviewed. Approval of the ethics review committee of the First Affiliated Hospital, School of Medicine was obtained in advance. Eligibility criteria are stated in the following sentences. GBC patients must be confirmed by pathological results after surgery without other malignancies or other infections. The following tumor markers need to be detected: CA125 (carcinoembryonic antigen 125), CA199 (carbohydrate antigen 199), AFP (alpha-fetoprotein), and CEA (carcinoembryonic antigen). For cholelithiasis patients, tumor markers (CA125, CA199, AFP, and CEA) must be detected, and must be without malignancies or other infection diseases. Data of routine blood test and liver function test were collected, and these data were the first detections obtained when patients enrolled in the hospital. Blood samples were obtained by peripheral venous puncture before breakfast.
Statistical analysis
All statistical analyses were based on Statistical Package for Social Science (SPSS Inc. Chicago, IL, USA) Version 18.0. For categorical variables, Fisher’s exact test or Chi-square test was used to detect statistical differences. For continuous variables normally distributed, independent t test or one-way ANOVA was used to compare the mean difference, and if the variables were not normally distributed, a Wilcoxon rank test was conducted. A receiver operating characteristic (ROC) curve was performed to evaluate the sensitivity and specificity of different markers in GBC diagnosis. Logistic regression analysis was performed to evaluate the association between different factors and GBC. A probability (P) value less than 0.05 was considered statistically significant.
Results
Characteristics of subjects
Characteristics of the subjects are shown in Table 1. The entire cohort was composed of 116 cholelithiasis patients and 144 GBC patients. The two groups were not matched in terms of age (58.90±114.87 vs. 63.26±11.15, P = 0.008) and gender (P = 0.003). Additionally, the following factors exhibited significant statistical differences between the two groups: total bilirubin (TB), alkaline phosphatase (ALP), aspartate aminotransferase (AST), alanine aminotransferase (ALT), platelet count (PLT), CA125, and CA199 (P<0.05). No statistical significance was observed in terms of albumin, red blood count, hemoglobin, white blood count, neutrophil count, lymphocyte count, CEA, and AFP.
Evaluation of GBC risk factors
In univariate analysis, age (P = 0.009), gender (P = 0.003), gallstones (P<0.001), TB (P<0.001), PLT (P = 0.003), AST (P = 0.020), ALT (P = 0.040), ALP (P<0.0001), CA125 (P = 0.001), and CA199 (P = 0.010) were risk factors for GBC. In multiple analysis, age (P = 0.024), gender (P = 0.004), gallstones (P<0.001), TB (P = 0.042), ALP (P = 0.032), and CA125 (P = 0.011) were the independent influencing factors associated with GBC (Table 2).
Evaluation of diagnosis values among different risk factors for GBC
Risk factors with OR>1 in multivariate analysis were selected to evaluate the diagnosis significance for GBC, as shown in Fig. 1. The area under curve (AUC) of CA125, ALP, TB, and age were 0.72, 0.71, 0.64, and 0.57, respectively. The sensitivity and specificity of each factor are shown in Table 3. CA125 had the best sensitivity (88.2%), and TB had the best specificity (94.8%). Additionally, ALP had better sensitivity (68.8%) and specificity (64.7%). We further investigated whether a distribution difference of these risk factors was observed between early stages and advanced stages, such as stages 1 and 2 vs. stages 3 and 4, high and moderate differentiation vs. low differentiation. Unfortunately, there was no statistical difference between different groups. Fig. 2 shows that although the ALP level exhibited an increasing tendency from early stage (T1 and T2) to advanced stage (T3 and T4), there was no statistical difference among the different groups.
Establishment of GBC diagnosis model combined with independent risk factors
We established a model to aid GBC diagnosis. The components of this model are shown in Table 4. According to the cut-off value determined by the ROC curve, each risk factor was subdivided into two categories. Then, each patient was scored as follows: age>57 years old, female gender, TB>147.615 U/L, ALP>113.885 U/L, and CA125>13.665 U/L were scored as 1; age≤57 years old, male gender, TB≤147.615 U/L, ALP≤113.885 U/L, and CA125≤13.665 U/L were scored as 0. We then calculated the score sum of the five risk factors as the final score for each patient. The diagnosis model has a sensitivity of 76.4%, specificity of 66.4%, and an AUC of 0.791.
Discussion
The incidence of GBC worldwide is relatively low, but its high mortality rate has attracted increased public attention [
13]. Early diagnosis is essential for reducing GBC mortality. Based on the clinical data in our institution, we conducted related investigation in an attempt to find effective diagnosis markers for GBC. Our study would be significant for GBC diagnosis.
Our results showed that age and female gender were risk factors for GBC, which were consistent with previous studies [
6,
14]. The incidence of GBC exhibited marked gender bias, and females are several times more susceptible than males [
14]. Therefore, some researchers speculated that estrogen might play an important role in GBC development [
15]. If patients are aged>57 years old, female, and have a long period history of cholelithiasis, they are at high risk of developing GBC.
AFP, CA125, CA199, and CEA are important markers for digestive duct cancer. AFP has been detected regularly to predict hepatocellular cancer. Elevated CEA level indicates colon and gastric cancers. However, the diagnosis significance of these tumor markers for GBC remains to be determined. AFP was not associated with GBC risk. CA199, the independent prognostic marker for GBC in a previous study [
16], did not exhibit an obvious relationship to GBC risk in our study. CA125 was commonly used as a tumor marker in ovarian carcinoma. According to our investigation, CA125 was not only notably correlated with GBC risk but is also the most valuable diagnosis marker for GBC, and this finding was consistent with those obtained in previous studies. Therefore, regarding the GBC diagnosis value of these tumor biomarkers, we believe that CA125 is superior to AFP, CA199, and CEA.
ALP as a stable serum marker is routinely tested for liver function evaluation. Aberrant expression of ALP is associated with many cancers, such as hepatocellular carcinoma [
17], renal cell cancer [
18], nasopharyngeal carcinoma [
19], and esophageal cancer [
20]. However, the diagnosis implication of ALP in GBC has not been reported. ALP as a risk factor has significant diagnosis value only second to CA125 for GBC. The possible mechanism by which ALP promotes tumor development is unclear at present. According to literature, in cancer cells, ALP activity was significantly elevated in the nucleolus and changes in localization during the cell cycle; a high level of the ALP reaction product may be related to the high level of proliferation of cancer cells [
21]. In spite of this, the signal pathway between ALP and GBC needs further investigation.
No significant difference was observed between the early and advanced stages in terms of the distribution of these factors (data were not shown except for Fig. 2). Therefore, the value of the early diagnosis of these factors for GBC is limited.
Inspired by previous studies [
22,
23], we established a GBC diagnosis model based on independent risk factors. This model has higher sensitivity and specificity compared with any sole risk factor. This model provides a new tool for identifying GBC when comparatively considering all risk factors.
Some limitations are present in this study. First, this study was retrospective in nature. Second, the study population was only selected from Han of China; hence, the cut-off value (such as the age of participants,>57 years old) may not represent the entire population. Therefore, our findings need to be confirmed by conducting similar investigations in other parts of the world. Such investigations need to include larger samples and multiple ethnic populations.
In conclusion, age, gender, TB, ALP, and CA125 can be used as the auxiliary diagnosis factors, and the risk model would facilitate GBC diagnosis when comprehensively considering the various risk factors.
Higher Education Press and Springer-Verlag Berlin Heidelberg
PDF
(202KB)
