Clinical Efficacy Of PIVKA-? As A Novel Biomarker For Hepatocellular Carcinoma: A Multicenter Study

Background and Aim:Hepatocellular carcinoma(HCC)is one of the most prevalent malignancies worldwide,and HCC-related death ranks the second,with extremely poor prognosis.The initial symptom of HCC is unapparent,so that most of HCC sufferers are diagnosed when the tumor stage is intermediate or advanced,which means the optimal time for curative treatment is missing.As a result,early diagnosis and therapy are the critical method to improve the prognosis of HCC.Alpha fetoprotein(AFP)is the most frequently used serum biomarker to detect HCC.Given its undesirable diagnostic sensitivity for HCC,especially for early-stage HCC,and it is likely to be elevated among patients with hepatitis and cirrhosis,thus it is in urgent need to find a novel biomarker with excellent efficiency for detecting HCC.Protein induced by vitamin K absence or antagonist-II(PIVKA-II)has been widely used in Japan and western countries,but its diagnostic efficiency is not proved with a large-scale sample in China,limiting its further clinical application.The objective of the part one study is to identify the efficiency of PIVKA-II in diagnosing HCC,with comprehensive comparison to other commonly used biomarkers.Besides,hepatectomy is one of the most effective therapy against HCC,but postoperative recurrence remains to be a serious issue.Microvascular invasion(MVI)is demonstrated to be an independent risk factor of postoperative recurrence,therefore,the risk of MVI evaluated by preoperative examinations is helpful to give immediate intraand postoperative intervention to decrease the possibility of recurrence among high-risk patients and improve their prognosis.Given there are close correlation between PIVKA-II and the tumor progress and invasion of HCC,such as local metastasis and vascular invasion.Moreover,researchers from western countries and Japan has listed PIVKA-II as an excellent preoperative predictor of MVI,but such correlation has not been fully illustrated in China.Therefore,the objective of the part two study is to illuminate the correlation between PIVKA-II and MVI among patients undergoing curative hepatectomy for HCC,and evaluate whether it could an independent risk factor of MVI,so as to provide the surgeons with available clinical decisions.MethodsConsecutive patients with liver diseases from 11 hospitals in China between 2016-2017 were enrolled.Patients from 2016.05 to 2016.06 were defined as pilot study group,which was set to evaluate the diagnostic efficiency of several HCC biomarkers such as PIVKA-II?AFP?AFP-L3?alpha-L-fucosidase(AFU)and gamma glutamyl transpeptidase(?-GT).Patients from 2016.07 to 2017.09 were defined as study group to select biomarkers with superior efficiency from pilot study into subsequent comprehensive analysis.The range of PIVKA-II among healthy candidates as well as the optimal cut-off value of PIVKA-II in differentiating HCC and non-HCC were determined.The diagnostic efficiency of the cut-off value was then compared with clinically used positive value.Subsequently,comparisons were performed between PIVKA-II and AFP in differentiating HCC and non-HCC,HCC and cirrhosis,HCC and cirrhosis + chronic hepatitis.The main parameters were sensitivity,specificity,area under the receiver operating characteristic(AUROC)curves.Additionally,the diagnostic performance of PIVKA-II in subgroup was determined,including HCC patients with various etiologies and tumor stages.The performance of combined diagnosis(PIVKA-II and AFP)was based on the establishment of statistical model.Patients undergoing curative hepatectomy during 2016-2017 for HCC(with the examination of PIVKA-II)from our hospital were enrolled into the part two study.The clinical characteristics and their relevance with the value of PIVKA-II were analyzed.The optimal cut-off value of PIVKA-II in predicting MVI was determined by receiver operating characteristic(ROC)curves.Uni-and multivariable Logistic regressions were conducted to identify the independent factors of MVI.ResultsBased on the Chinese multicenter database,we firstly the superiority of PIVKA-II and AFP in diagnosing HCC compared with AFP-L3?AFU and ?-GT from the pilot study group,and they were enrolled into study group for further analysis.We found that among Chinese population,both PIVKA-II and AFP showed excellent discrimination in HCC and other liver diseases,and the diagnostic efficiency of PIVKA-II was significantly better than AFP: when the cut-off value of PIVKA-II and AFP were defined as 40mAU/mL and 20ng/mL,the sensitivity and specificity of PIVKA-II were 81.7% and 90.3%,superior than that of AFP(61.6% and 91.1%,respectively).Moreover,the AUC of PIVKA-II was significantly higher than AFP(0.907 vs 0.841,P<0.001).In subgroup analysis,the diagnostic efficiency of PIVKA-II in diagnosing early stage HCC(both BCLC A and TNM I),asymptomatic HCC,well differentiated HCC and HBV-/HCV infected HCC,which could be expected as an excellent serum biomarker for the surveillance and detection among at-risk candidates.In addition,the combination of PIVKA-II and AFP based on the Logistic-regression analysis could further improve the diagnostic efficiency compared to the single used PIVKA-II or AFP.The sensitivity could be elevated to 87.0% from 81.7% when the specificity remained stable,which was much more favorable than currently used diagnostic strategy.Regarding the part two study,there were 957 patients who underwent curative hepatectomy for HCC enrolled.Patients with positive value of PIVKA-II(?40mAU/mL)were associated with worse tumor stage and more severe liver inflammation.As for tumor features,patients with positive value of PIVKA-II showed larger maximum tumor size,more multiple tumors and higher incidences of macrovascular invasion and MVI than patients with negative PIVKA-II.ROC curves determined PIVKA-II?160mAU/mL as the optimal cut-off value predicting MVI and it was included into further Logistic-regression analysis.Univariable Logistic-regression analysis revealed age?60 years and diabetes were protective factors of MVI,while preoperative aspartate aminotransferase(AST)?40U/L,alanine aminotransferase(ALT)?40U/L,AFP?400ng/mL,PIVKA-II?160mAU/m L and maximum tumor size ?5cm were risk factors.Further multivariable analysis demonstrated only AFP?400ng/mL(OR=2.250,95%CI=1.666-3.039,P<0.001)and PIVKA-II?160mAU/mL(OR=1.597,95%CI=1.183-2.155,P=0.002)were independent risk factors of MVI.ConclusionIn summary,the superior performance of PIVKA-II in diagnosing HCC among Chinese patients was proved in the present study.In addition,the combined diagnostic model based on PIVKA-II and AFP was established and demonstrated,which showed improved diagnostic efficiency than single used biomarkers.The novel biomarker PIVKA-II and combined model with great diagnostic performance could gain possibility for at-risk patients to be diagnosed at an early stage,who could be subject to curative therapy,thereafter improving the long-term prognosis of HCC patients.Moreover,significant association between the value of PIVKA-II and tumor stage,as well as liver inflammation was observed in this study.In addition,PIVKA-II?160mAU/mL and AFP?400ng/mL were demonstrated as the independent risk factors of MVI in patients undergoing curative hepatectomy for HCC.The study highlighted the risk assessment of MVI according to the preoperative objective and stable examinations,and thus the surgeons could perform more immediate intra-and postoperative interventions among high-risk patients.Such strategy could decrease the recurrence rate by removing the tumor lesions as much as possible,and improve the long-term prognosis of HCC patients.