The clinical application of PIVKA‐II in hepatocellular carcinoma and chronic liver diseases: A multi‐center study in China

Abstract Background Due to the absence of specific symptoms and low survival rate, efficient biomarkers for hepatocellular carcinoma (HCC) diagnosis are urgently required. The purpose of this study was to evaluate the diagnostic performance of protein induced by vitamin K absence or antagonist‐II (PIVKA‐II) and to determine the optimal cutoff values for HBV infection‐related HCC. Methods We conducted a cross‐sectional, multi‐center study in China to ascertain the cutoff value for HCC patients in the context of CHB‐ and HBV‐related cirrhosis. The receiver operating characteristic curve (ROC) and the area under the curve (AUC) were used to evaluate the diagnostic performance of PIVKA‐II. Results This study enrolled 784 subjects and demonstrated that PIVKA‐II had a sensitivity of 84.08% and a specificity of 90.43% in diagnosis HCC from chronic liver diseases. PIVKA‐II at a cutoff of 37.5 mAU/mL yielded an AUC of 0.9737 (sensitivity 91.78% and specificity 96.30%) in discriminating HCC from chronic hepatitis B (CHB) patients. PIVKA‐II at a cutoff of 45 mAU/mL yielded an AUC of 0.9419 (sensitivity 77.46% and specificity 95.12%) in discriminating HCC‐ from HBV‐related cirrhosis patients. Furthermore, using a cutoff value of 40 mAU/mL for PIVKA‐II as an HCC marker, only 4.81% (15/312) was positive in chronic hepatitis and 12.80% (37/289) in cirrhosis patients, revealing the satisfactory specificity of PIVKA‐II in chronic liver disease of different etiologies. Conclusion Our data indicated that PIVKA‐II had satisfactory diagnostic efficiencies and could be used as a screening or surveillance biomarker in HCC high‐risk population.


| INTRODUC TI ON
In 2018, liver cancer ranked the sixth most common cancer and the fourth leading cause of cancer deaths in the worldwide. Every year, there are about 841,000 new cases and 782,000 deaths globally. 1 Primary liver cancer can be divided into hepatocellular carcinoma (HCC), cholangiocarcinoma (ICC), and mixed type according to the pathological types, while HCC accounts for about 90% of all cases. 2 As a high-risk area of HCC, China contributes to more than half of the new incidence and mortality related to HCC annually, and the morbidity and mortality rates ranked the fourth and third in China, respectively. 3 Due to the absence of specific symptoms and early diagnosis, 70%-80% of patients are often diagnosed at an advanced stage with a 5-year survival rate less than 40% and a recurrence rate exceeding 60%. 4 Thus, effective tools for HCC diagnosis and prognosis prediction are urgently required.
Although there are regional differences, the main risk factors for

HCC include virus infection (hepatitis B virus [HBV], hepatitis C virus
[HCV]), aflatoxin-contaminated foodstuffs, alcoholism, obesity, and type 2 diabetes. 1 HBV contributes the largest proportion to liver cancer mortality in East Asia, at 41%. 5 The current HBV infection rate in China is 6.1%, accounting for almost 30% of the global number. 6 Consequently, the key determinant of HCC in China is HBV infection. Simultaneously, liver fibrosis, developed from chronic hepatitis, is also a high-risk factor with progressive liver function damage. 7 According to guidelines for the prevention and treatment of chronic hepatitis B (version 2019) formulated by the Chinese Society of Infectious Diseases, the Chinese Society of Hepatology, and the Chinese Medical Association, it is recommended that populations with HBV infection or liver cirrhosis should be monitored HCC risk at least every 3-6 months. 8 Alpha-fetoprotein (AFP), synthesized by approximately half of HCC, 9 is currently the most widely used serological marker for HCC detection. However, AFP concentration may remain normal in up to 40% of the patients, especially in the early stages. 10 In contrast, an elevation of AFP concentration can be detected in embryonic carcinomas, gastric cancer, and lung cancer. 11 Increased AFP concentrations can also be frequently noted in benign liver diseases, such as 15%-45.6% in HBsAg-positive patients 12,13 and 11%-47% in liver cirrhosis. 14,15 Since high risk of HCC is almost always presented in patients with chronic liver diseases, these characteristics limit its utility. Consequently, the AFP assay was recalled from the diagnostic criteria recommended by the "American Association for the Study of Liver Disease" (AASLD) and the "European Association for Study of the Liver" (EASL). 16,17 Protein induced by vitamin K absence or antagonist-II (PIVKA-II), also known as des-gamma carboxyprothrombin (DCP), was first introduced as a novel HCC biomarker in 1984. 18 With the development of research over the years, PIVKA-II's clinical significance in primary liver cancer diagnosis and management has been emphasized. 19,20 Furthermore, PIVKA-II showed less tendency to be elevated in other chronic liver diseases, yielded a specificity around 90%-95% when compared HCC with cirrhosis and chronic hepatitis. 21,22 Thus, monitoring HCC risk in CHB or cirrhosis patients by PIVKA-II may be more specific than AFP. Therefore, we conducted a cross-sectional, multi-center study to analyze the diagnostic performance of PIVKA-II in China. In addition, we ascertained the cutoff value for HCC patients in the context of CHB-and HBV-related cirrhosis.

| Study design and participants
This was a cross-sectional, multi-center study. The study protocol was approved by the institutional review board of Eastern

| Specimen collection and testing methods
The serum samples were collected at the time of diagnosis without treatment and stored at −20°C or lower temperature immediately.

| Reproducibility
Four concentration sample pools including low concentration (≤40 mAU/mL), medium concentration (41-200 mAU/mL), high concentration (201-1000 mAU/mL), and super-high concentration (1001-30,000 mAU/mL) were prepared and measured ten times on the first day to calculate their coefficient of variation (CV%) at three research centers respectively. The 4-level samples were measured six times again on days 3 and 7 after storage at 4°C; and days 14 and 30 after storage at −20°C to satisfy the different storage needs of clinical samples.

| Dilution linearity
Fourteen high concentration (20,000-30,000 mAU/mL) sample pools were prepared to evaluate linearity of the assay. Double dilution methods were performed by diluting specimens (with the specimen diluent of Lumipulse or the Calibrator A (0 mAU/mL) of Architect) until 1:32 dilution. Linearity was evaluated as the coefficient of determination (R 2 ) between measured and expected values for the respective samples. 23

| Statistical analysis
All data were analyzed by GraphPad Prism 6 (GraphPad Software, La Jolla, CA) and Medcalc (MedCalc, Ostend, Belgium) software programs. The Mann-Whitney U test was used for the differential analysis of PIVKA-II concentration. Discrete variables were compared by contingency table analysis of the chi-square test or Fisher's Exact Test.
Correlation coefficients were calculated to observe the correlation between two assays. The receiver operating characteristic curve (ROC) and the area under the curve (AUC) were used to evaluate the diagnostic performance of PIVKA-II. The optimal cutoff value was determined by Youden's index. Two-sided settings were used for statistical analysis and p ≤ 0.05 was considered as statistically significant.

| Participates characteristics
As shown in Figure 1, 950 individuals were included in the study, and 166 were excluded due to the incomplete information or other malignancies.
Eventually, 784 participates were enrolled in the study. The basic characteristics of every participant were summarized in Table 1. Patients with liver cancer and cirrhosis were predominantly male (p < 0.0001), while the gender difference was not significant in patients with chronic hepatitis.
Simultaneously, patients with liver cancer and cirrhosis were older than those with chronic hepatitis (p < 0.0001). According to the results of previous large-scale studies, 24

| Satisfactory diagnostic performance of PIVKA-II
As shown in Figure 2 and Table 2 Table 3. Figure 4 and Table 4  We also analyzed the diagnostic performance of AFP. AFP had a sensitivity of 61.33% and a specificity of 91.15% when distinguishing HCC from disease controls at the cutoff value of 20 ng/mL (Table 5).

| The distribution of PIVKA-II among chronic liver diseases
To better explore the specificity of PIVKA-II, we observed the distribution of PIVKA-II among chronic liver diseases with different etiologies in Figure 5. The mean value of PIVKA-II in CH was 25

| The optimal cutoff values of PIVKA-II for diagnosing HCC in patients with CHB or cirrhosis
Since the key determinant of HCC in China is HBV infection, HCC surveillance in patients with these contexts was significantly important. Therefore, we inquired the optimal cutoff values of PIVKA-II in patients with CHB-or HBV-related cirrhosis ( Figure 6). PIVKA-II

| DISCUSS ION
As the most widely used HCC marker, AFP always has a falsenegative rate from 30% to 40%, 26

| CON CLUS ION
In conclusion, our data indicated that PIVKA-II had satisfactory diagnostic performances in patients with liver diseases of various etiologies and demonstrated that PIVKA-II can be used as a screening or surveillance biomarker in HCC high-risk populations. Furthermore, this research revealed that PIVKA-II was less affected by the patients with HBV-related liver disease, showing its superior diagnostic efficacy than AFP. We believe that the combination of AFP and PIVKA-II will improve the diagnosis and the recurrence detection rate of HCC patients, increase the proportion of early treatment, ultimately improve the survival rate and prolong the life of HCC patients.