Diagnostic Performance of Extrahepatic Protein Induced by Vitamin K Absence in the Hepatocellular Carcinoma: A Systematic Review and Meta-Analysis

Background and Objectives: the early diagnosis of hepatocellular carcinoma (HCC) benefits from the use of alpha-fetoprotein (AFP) together with imaging diagnosis using abdominal ultrasonography, CT, and MRI, leading to improved early detection of HCC. A lot of progress has been made in the field, but some cases are missed or late diagnosed in advanced stages of the disease. Therefore, new tools (serum markers, imagistic technics) are continually being reconsidered. Serum alpha-fetoprotein (AFP), protein induced by vitamin K absence or antagonist II (PIVKA II) diagnostic accuracy for HCC (global and early disease) has been investigated (in a separate or cumulative way). The purpose of the present study was to determine the performance of PIVKA II compared to AFP. Materials and Methods: systematic research was conducted in PubMed, Web of Science, Embase, Medline and the Cochrane Central Register of Controlled Trials, taking into consideration articles published between 2018 and 2022. Results: a total number of 37 studies (5037 patients with HCC vs. 8199 patients—control group) have been included in the meta-analysis. PIVKA II presented a better diagnostic accuracy in HCC diagnostic vs. alpha-fetoprotein (global PIVKA II AUROC 0.851 vs. AFP AUROC 0.808, respectively, 0.790 vs. 0.740 in early HCC cases). The conclusion from a clinical point of view, concomitant use of PIVKA II and AFP can bring useful information, added to that brought by ultrasound examination.


Introduction
Hepatocellular carcinoma (HCC) is the most widespread histological subtype of primary liver cancer (accounting for approximately 90% of all cases [1]), with an increasing incidence [2]; at the same time, it is currently recognized as the third most common cause of death worldwide [3][4][5]. Unfortunately, at the time of diagnosis, a small percentage of patients are eligible for curative treatment, the most common cause being an advanced tumor stage [6].
Studies in the literature have shown that chronic viral hepatitis B and C, autoimmune hepatitis, nonalcoholic steatohepatitis, and genetic and epigenetic changes are the main risk factors for the development of hepatocellular carcinoma HCC [7][8][9][10][11]. The prognosis of patients with advanced liver disease or cirrhosis (regardless of etiology), even in those responding to antiviral treatment, is influenced by the appearance of HCC [4,5]. and possible recurrence of HCC [12,22,[43][44][45]. What differentiates PIVKA II from AFP is that the value of the former is not affected by liver disease activity [12].
In view of the above, the difficulty of performing adequate screening for HCC (to detect early cases), new screening methods are being examined. Current studies aim at comparative and summative evaluation of different methods, with Japan and other countries [46,47] implementing simultaneous determination of PIVKA II and AFP as a screening method to monitor patients at high risk of developing hepatocellular carcinoma [48].
To date, results on the diagnostic performance of PIVKA II in comparison to or in combination with AFP are conflicting. The available data come mainly from studies involving Asian patients [46,49], with results from Western studies limited by a relatively small sample size. Published studies (with the exception of the most recently published ones) have been systematized into past published meta-analyses, evaluating the accuracy of HCC detection by serum determination of PIVKA II and AFP biomarkers, alone or in combination in patients at risk of tumor development [44,45,48,49].
In the present one, the most recently published studies for establishing the role of PIVKA-II versus AFP (globally, but also in a relationship with the HCC stage) were taken into consideration.
Knowledge of this topic is needed for better screening and diagnosis of at-risk HCC patients. The aim of this work was to extend the knowledge of comparative evaluation of PIVKA II and AFP HCC diagnostic values, especially in early HCC patients.

Materials and Methods
Search strategy: literature screening for meta-analysis. A systematic search was conducted for the interval from 1 January 2018 to 4 September 2022. Searches for relevant studies were mainly conducted in PubMed, Web of Science, Embase, Medline and the Cochrane Central Register of Controlled Trials.
Rigorous research of the papers was performed. Two main investigators performed independent literature research in order to identify the previously published papers. All useful papers were read by both investigators, even those with negative results.
Duplicates were removed. Only articles written in English that had abstracts were taken into consideration. Articles presented just as abstracts or conference presentations, reviews, systematic reviews, meta-analyses, editorials and in vitro studies were excluded. The quality assessment of diagnostic accuracy studies (QUADAS) was applied to evaluate the selected studies from a quality point of view.
The following data were extracted from the articles studied: title, authors, year of publication, study identification item, country, number of locations where the study was conducted, number of patients included (with HCC vs. without HCC, respectively, early HCC cases), study design, etiology of liver disease; for both PIVKA II and AFP, the AUROC (overall and for early HCC cases), sensitivity and specificity were followed.
A flow diagram of the literature search strategy and study selection process is summarized in Figure 1. According to the literature, at this moment, two tumor staging systems are used to define the extent of HCC-BCLC (Barcelona clinic liver cancer staging) staging [50,51], respectively, the 8th edition American Joint Committee on Cancer tumor-node-metastasis (TNM) staging [52]. BCLC stage 0 is defined as the tumor being less than 2 cm, performance status = 0 and the liver working normally (Child-Pugh A). BCLC stage A is defined in patients presenting single tumors of any size or 3 nodules < 3 cm in diameter, performance status = 0 and Child-Pugh class A or B.
In this meta-analysis, early-stage HCC was defined as BCLC stage 0/A and/or 8th edition TNM stage I (depending on the data reported by the included studies).

Statistical Analysis
MedCalc software version 20.115 (Ostend, Belgium) was used for performing the meta-analysis. Using for every study each AUC value and the corresponding standard error (SE), the weighted summary AUC (sAUC) was calculated. Most of the studies did not report the standard error for AUROC. The formula used for SE (AUC) calculation was the one proposed by Hanley and McNeil (1982)-presented in Formula (1).
The publication bias was assessed using funnel plots. Forrest plots showing the overall effect were constructed. Taking into consideration the presence or absence of heterogeneity, a fixed or random effects model was preferred. An I2 value >25% was considered indicative of heterogeneity.
A p value < 0.05 was considered statistically significant.

Results
A total number of 37 studies were included in the meta-analysis. Overall, 13,236 patients were included: 5037 patients with HCC (case group) vs. 8199 patients (the control group). The control group was represented by healthy patients (without previous liver diseases), chronic hepatitis B or C, liver cirrhosis or at-risk condition patients. Patients with HCC were divided depending on their HCC stage-1513 early HCC. Complete data about the included studies are presented in Table 1.
For each study included, the performances of PIVKA II and AFP were reported in Tables 2 and 3 (global and in early HCC cases). Sensibility and specificity for PIVKA II and AFP were also reported.
The sAUC of AFP, respectively PIVKA II for the discrimination between patients with HCC and those without, were 0.808 (95% CI 0.782 to 0.834) vs. 0.851 (95% CI 0.823 to 0.878)-data were reported in Figure 2. Considering that the studies showed heterogeneity (in both cases), random effects models were applied.
Some of the studies reported at the same time for AFP and PIVKA II; also, there were some studies reporting global for HCC, but also for early HCC; AFP = alpha-fetoprotein; PIVKA II = protein induced by vitamin K absence or antagonist-II.

Discussion
In our days, the neoplastic diseases show an increasing prevalence, with HCC being more and more frequently diagnosed, even in young patients. Lifestyle changes with an increased incidence of nonalcoholic steatohepatitis, chronic viral hepatitis and autoimmune hepatitis increase the risk of HCC, being responsible for HCC appearance.
A lot of progress has been made in HCC diagnosis, but some cases are missed or late diagnosed in advanced stages of the disease. Therefore, new tools (serum markers, performant imagistic technics) are continually being reconsidered.
For decades, AFP has been widely used as a tumor marker in the surveillance of populations at high risk of developing HCC, but some limitations are well known. The reported sensitivity and specificity of this biomarker (40-65%, 76-96%, respectively) differ significantly depending on the characteristics of the studied group [22,86]. AFP serum values were often elevated in patients with chronic liver disease or cirrhosis without HCC [22].
All current guidelines recommend the additional use of imaging diagnosis in order to improve the diagnostic accuracy. Ultrasonography, CT or MRI present limitations, sometimes encountering difficulties in small lesion diagnosis. In view of these data, AFP and ultrasonography have been used together to improve diagnostic sensitivity in medical practice [3,[86][87][88][89][90], but accuracy for the moment remains uncertain [91]. Under these conditions, HCC screening can be improved to detect neoplastic lesions at early stages. To date, several promising serum tumor markers with the potential for early diagnosis and surveillance of HCC have been proposed [3,21,44,45,[86][87][88][89][90]92,93] of which PIVKA II appears to be the most promising, with recently published data on its performance (alone or in combination with AFP or ultrasonography).
No clear PIVKA II cut-offs for HCC, respectively, for early HCC diagnosis were already established. Supplementary, different methods are used for biomarker determination-so, for clarifying these aspects, more data need to be published. To this moment, to our best knowledge, clinical and laboratory factors influencing the PIVKA II values have not been exhaustively investigated. The current meta-analysis brings to attention new data about the usefulness and ability of PIVKA II to detect HCC. Literature is scarce in revealing the role of PIVKA II versus AFP. The paper provides an overview of recently published data about the role of PIVKA II vs. AFP in HCC diagnosis. In this meta-analysis, PIVKA II presented greater accuracy for HCC diagnosis, taking into consideration all cases (0.851 vs. 0.808), but also in early HCC cases (0.790 vs. 0.740). The reported results (the better discriminatory value of PIVKA II) are in line with those reported by Caviglia [3] December 20, 2017). Also, the reported calculated AUROCs were approximately similar to those reported in previously mentioned studies. There also has been published some meta-analysis evaluating just one of the two biomarkers (PIVKA II or AFP), the results regarding the AUROC values being consistent with the results of this study [89,96,97].
A novel perspective brought to attention a parallel with the standard marker used (AFP)-higher accuracy for PIVKA II being found in the early diagnosis of HCC. Similar data have been published by Xing [87]. The results of this study highlight the possible role of PIVKA II in providing new data, useful for daily medical practice. A recently published study (just a few days ago, not included in the meta-analysis) also revealed that PIVKA II had a better predictive performance vs. AFP global and in early-HCC (the reported registered values being approximately similar to these ones [22]). The results of the study are supporting others' recommendations that PIVKA II can have usefulness in early HCC diagnosis in the incipient moments [66].
It was impossible to determine the PIVKA II and AFP performances, depending on the etiology of the liver diseases, with mixed etiology being taken into consideration. Frequently, the studies do not make a difference according to the liver disease etiology; in most of the cases, the reported results are globally calculated.
The 0 and A HCC BCLC classes were taken into consideration in a unitary way, which must be mentioned. Of course, that is a discrepancy between the two classes regarding the following treatment, but for the moment it was not possible to perform a detailed, stratified analysis.
Due to the heterogeneity of the reported studies (determined by the diversity of study populations in different countries, methodology used and sample size), these findings might not be representative of all populations-further research is needed.
Stratified analysis depending on the gender, ethnicity, age or liver disease type and stage represents an area to be explored further. More data should be published regarding the cut-off values, for a unitary approach regarding HCC diagnosis.
This study provides the backbone for a future meta-analysis in order to evaluate the accuracy of PIVKA II and AFP association in HCC diagnosis. Of the listed studies, just a few of them reported combined accuracy. In addition, future studies on the topic are recommended to determine the serum values of PIVKA II after HCC treatment (surgical or chemotherapy), theoretically bringing useful information for monitoring treatment results, for predicting diagnosis, relapse and survival.

Conclusions
These results provide a significant step toward the diagnosis of HCC by determining the serum value of vitamin-K-dependent proteins used as tumor biomarkers, along with other paraclinical examinations.
From a clinical and practical point of view, the use of PIVKA II concomitantly or instead of AFP is bringing useful information, added to those reported by ultrasound examination. Probably the emerging role of PIVKA II is in patients with previous hepatic diseases (hepatitis, cirrhosis) where AFP limitations are well-known.
This study provides the backbone for future studies on the relationship with an earlier diagnosis of hepatocarcinoma. Institutional Review Board Statement: "Not applicable"-study has been performed using published article.