The diagnostic value of PIVKA-II, AFP, AFP-L3, CEA, and their combinations in primary and metastatic hepatocellular carcinoma.

Abstract Background Early diagnosis decreases the mortality of hepatocellular carcinoma (HCC). We aimed to investigate the usefulness of PIVKA‐II, AFP, AFP‐L3, CEA, and their combinations in the diagnosis of primary and metastatic HCC. Methods One hundred and twenty patients with primary HCC (PHC), 115 with metastatic HCC (MHC), 89 with chronic liver disease (CLD), and 116 healthy volunteers were included. The diagnostic values of each marker and their combinations for HCC diagnosis were represented by ROC curve analyses. Results PIVKA‐II, AFP, and AFP‐L3 levels in PHC group were higher than that in normal control, CLD, and MHC groups. CEA levels in MHC group were higher than that in the other three groups. When the four markers were analyzed individually, PIVKA‐II showed the highest positive rate in PHC group (76.7%) and CEA showed the highest positive rate in MHC group (69.6%). PIVKA‐ II showed the largest area under ROC curve (AUC = 0.835) to discriminate PHC group from CLD group. Combined PIVKA‐II with AFP‐L3 increased the AUC to 0.910. CEA showed the highest AUC (0.849) to discriminate MHC group from CLD group. Combined CEA with PIVKA‐II increased the AUC to 0.866. AFP‐L3 alone showed the highest AUC (0.890) to discriminate MHC group from PHC group. Combined PIVKA‐II with AFP‐L3, and CEA increased the AUC to 0.957. Conclusion PIVKA‐II, AFP‐L3, AFP, and CEA are effective biomarkers for the diagnosis of PHC and MHC. Their combinations could improve the diagnostic performance compared with each marker used alone in detecting PHC and MHC.


| INTRODUC TI ON
Hepatocellular carcinoma (HCC) is one of the most common malignancies worldwide. 1 Every year there are more than 500 000 newly diagnosed HCC cases. 2 HCC has a variable geographical distribution, and the incidence in developing countries is two to three times higher than that in Western countries. In Eastern Asia and Middle Africa, the age-adjusted incidence rate ranges from 20 to 28 cases per 105 people in men and around 55% of cases are in China. 3 However, even in the western countries such as the United States (USA), the incidence of HCC is increasing, 4 it is predicted to continue to increase in the following ten years. 5 Hepatocellular carcinoma generally has no clinical symptoms in the early stage, and 2/3 of cases are found in the middle and advanced stage, when only about 20% of patients are suitable for surgical resection. 6 Vascular invasion or extrahepatic tumor spread lead to no curative treatment options available. Metastatic HCC (MHC) is mainly determined by progression of the underlying liver disease rather than by the extrahepatic metastases. 7 In China and Africa patients who present with symptoms usually die within 4 months 8 but longer survival is possible in western countries. 9 Therefore, screening and early diagnosis are vital to reduce the high mortality of HCC.
Several methods are available for HCC diagnosis. The recommended noninvasive methods include imaging techniques such as magnetic resonance imaging and the use of tumor markers such as AFP. However, in up to 40% of HCC patients, AFP levels are normal, especially in the early stage of the disease, which reflects low sensitivity. In order to improve the clinical outcome of patients, it is necessary to determine more reliable serum biomarkers.
PIVKA-II is also known as des gamma carboxy prothrombin (DCP) and is an abnormal prothrombin molecule that is generated due to acquired defect in the posttranslational carboxylation of the prothrombin precursor in malignant cells. In 1984, Liebman et al 10 used a radioimmunoassay to detect serum PIVKA-II levels in patients with HCC and found that 91% of patients showed a significant increase. This finding has since been successfully applied in clinical practice. 11 PIVKA-II combined with Golgi protein 73 (GP73) showed higher accuracy than AFP in early HCC diagnosis. 12 PIVKA-II has been included in markers for auxiliary diagnosis of HCC Serum carcino-embryonic antigen (CEA) is a relatively non-specific antigen used in the clinical diagnosis of gastrointestinal cancer. 14 The elevated level of CEA is found in some patients with poor HCC prognosis. 15 Therefore, this marker may be useful for classification of prognosis. Many studies have reported the value of different tumor markers in the diagnosis of HCC, but there is currently no study using PIVKA-II, AFP, AFP-L3, CEA, and their combinations for the differential diagnosis of primary and metastatic HCC. Because of the high prevalence of HCC cases in China, seeking a set of diagnostic markers for HCC with better sensitivity and specificity than those in current use is particularly important. The aim of this study was to find an effective method to improve the early detection and accurate diagnosis of HCC, thus benefiting the prognosis of patients.

| Clinical data collection and examination methods
Three microliters of venous blood was collected from the patients and healthy controls. The serum was separated and stored in −80°C freezer until use. PIVKA-II, AFP, and CEA levels were measured in microparticle chemiluminescence instrument (Abbott I2000). AFP-L3 was measured by enzyme-linked immunosorbent assay (ELISA) (Shanghai Jonln Biotechnology). The cutoff values for the four markers were as follows: PIVKA-II >39.54 mAu/mL, AFP >8.78 ng/mL, AFP-L3 >7.26 ng/mL, and CEA >5.0 ng/mL.

| Statistical analysis
Data were analyzed using SPSS 25.0 statistical software (IBM Corp.).
Kruskal-Wallis H Test was used for the comparison of the marker levels in four groups; the Bonferroni method was used for comparisons between each two groups, and P ≤ .05 was considered statistically significant. The diagnostic value of each index for PHC or MHC was represented by receiver operating characteristic (ROC) curve, and these were used to calculate the sensitivity, specificity, Youden index, and other indexes.

| Demographic data of subjects
The baseline measurements for the four groups in this study are shown in Table 1. All of the measurements showed a significant difference between the four groups. Patients in PHC and MHC groups were older than those in CLD group and normal control group. It was reported that primary HCC can occur at any age, with the most patients between 40 and 59 years old and a male-to-female ratio of 2-5:1. About 60% patients were male in chronic liver disease group while about 76% patients were male in PHC group. Thus, there was a slight deviation in patient age. The serum levels of various liver function indicators including ALT, AST, GGT, and ALB in patients with HCC were higher than those with chronic liver disease. The decreases in blood cell indexes including Plt, RBC, and Hb were also observed in patients with HCC.

| Comparison of the diagnostic values of four biomarkers in different groups
The levels of PIVKA-II, AFP, and AFP-L3 in PHC group were significantly higher than those in normal control, CLD, and MHC groups (P < .05; Table 2). The levels of CEA in MHC group were significantly higher than those in other three groups. The levels of PIVKA-II in MHC group were significantly higher than those in normal control and CLD groups. AFP-L3 and AFP levels in MHC group were significantly higher than those in normal control group (P < .05). When the four indexes were analyzed individually (Table 3), PIVKA-II showed the highest positive rate in PHC group (76.7%) and CEA showed the highest positive rate in MHC group (69.6%). In the combined tests, PIVKA-II or AFP-L3, PIVKA-II or AFP, and AFP or AFP-L3 increased the positive rate of to 92.5%, 91.7%, and 91.7%, respectively, in PHC group.

| Evaluation of the diagnostic values of four biomarkers and their combinations in PHC group
The ROC curve analyses of the four markers in PHC group (compared with chronic liver disease group) were shown in Figure 1. When the four biomarkers were analyzed individually, PIVKA-II, AFP, AFP-L3, and CEA showed the area under ROC curve (AUC) of 0.835, 0.810, 0.807, and 0.625, respectively (Table 4) (Table 4). These findings indicate that the combination of PIVKA-II, AFP-L3, and AFP could improve their abilities to discriminate patients with PHC and chronic liver diseases.

| Evaluation of the diagnostic values of four biomarkers and their combinations in MHC group
The ROC curve analyses of the four markers in PHC group (compared with chronic liver disease group) were shown in Figure 2.  (Table 5). These findings indicate that CEA, PIVKA-II, and their combinations could improve their abilities to discriminate patients with MHC and chronic liver diseases.

| Evaluation of the diagnostic values of four biomarkers and their combinations in discriminating PHC and MHC groups
The ROC curve analyses of the four markers in MHC group (compared with PHC group) were shown in  (Table 6).
Adding AFP had no alteration in AUC although it increased the sensitivity but decreased the specificity. These findings indicate that the combination of PIVKA-II, AFP-L3, and CEA could improve their abilities to discriminate patients with MHC and PHC.

| D ISCUSS I ON
The early diagnosis of HCC is essential for curative interventions, which helps improve the prognosis and long-term survival of pa- PIVKA-II has been found to be superior to AFP or AFP-L3 in detecting PHC; this finding is consistent with those from several earlier studies. 20,21 We also compared the usefulness of AFP, AFP-L3, and PIVKA-II both individually and in combination in diagnosing PHC. We found that the combination of PIVKA-II and  with that reported in the literature 18,22 showing that the combination of PIVKA-II, AFP, or AFP-L3 has a superior detection of PHC with no significant decrease in specificity in Asian population. Although the combination of three or four markers works better, the cost of examination also increases, which may reduce its potential of use. 21 Other combinations of two or three markers did not provide superior diagnostic ability. Intriguingly, it has been reported that the changes of PIVKA-II in PHC are not associated with AFP. 23 Therefore, PIVKA-II can improve the positive rate of diagnosis for AFP-negative patients. The advantages of combined test have also been confirmed in clinical practice. For example, AFP, AFP-L3, and PIVKA-II are included as serum biomarkers in the clinical settings of Japan National Health Insurance. 16 In the future, the sample size will be further expanded, and the association of PIVKA-II with staging, curative effect and prognosis in PHC will be further studied.

TA B L E 2 The expression levels of four markers in different groups
For the diagnosis of metastatic HCC, CEA alone and PIVKA-II alone showed a better AUC and the combination of CEA and PIVKA-II showed the highest diagnostic accuracy. However, CEA had a better specificity than PIVA-II while PIVA-II had a better sensitivity than CEA in MHC. These findings indicate that PIVKA-II and CEA has an ability to discriminate patients with MHC and chronic liver disease. CEA is a broad-spectrum tumor marker that is combined with other biomarkers to diagnose primary HCC. 24 However, we found that CEA was significantly elevated in MHC, which is helpful for differential diagnosis between primary and metastatic HCC. This is consistent with the findings from Huang et al 25 showing that the patients with distant metastasis have high CEA levels than those without distant metastases. In addition, AFP-L3 showed a sensitivity of 0.890 for distinguishing MHC and PHC groups, which was slightly higher than that previously reported in the Chinese population. 26 The combination of AFP-L3 and CEA was better to discriminate patients with MHC and PHC. (AUC 0.945, sensitivity 0.887, specificity 0.908). Overall, the sensitivity, specificity, and AUC for combined use were higher than those used alone for distinguishing MHC and PHC groups.
In conclusion, a combination of four biomarkers including AFP, PIVKA-II, AFP-L3 and CEA showed better accuracy than either marker alone in distinguishing patients with MHC, PHC, and chronic liver disease. The detection methods are simple, stable and reliable, and thus these markers are suitable for application in hospitals at all levels.

ACK N OWLED G M ENTS
We thank the members of Department of Clinical Laboratory in Gansu Provincial Hospital for collecting clinical samples, providing technical assistance, and performing statistical analyses.

CO N FLI C T S O F I NTE R E S T
The authors have no conflicts of interest to be declared.