患有肝细胞癌（HCC）的患者进行激进切除术是一种治愈性治疗方法，但是复发率仍然很高。我们的目标是通过术后随访期间对维生素K缺失引起的蛋白（PIVKA-II），甲胎蛋白（AFP）和凝集素反应性AFP（AFP-L3）的纵向监测来探索预测HCC复发的性能。纳入了2015年1月至2020年12月在宁波医学中心丽晖莉医院接受激进切除的HCC患者。所有纳入的患者在术后随访期间定期监测PIVKA-II，AFP，AFP-L3，每3个月一次。比较了PIVKA-II，AFP，AFP-L3在随访期间预测HCC复发的监测性能。使用广义估计方程（GEE）分析肿瘤生物标志物的趋势和与时间的交互作用。计算接收者操作特征曲线（AUROC）下的面积，最佳分界值，敏感性和特异性，以评估三个生物标志物的性能。通过Kaplan-Meier曲线和对数秩检验分析具有任何升高的生物标志物的患者的无复发生存期（RFS）和总生存期（OS）。使用多元逻辑回归模型分析复发的潜在危险因素。GEE分析表明，复发患者的PIVKA-II，AFP，AFP-L3在随访期间较未复发患者高，PIVKA-II和AFP在复发前6个月呈上升趋势。在预测复发方面，PIVKA-II，AFP，AFP-L3及其组合的AUROC分别为0.885，0.754，0.781和0.885，PIVKA-II，AFP，AFP-L3的最佳分界值分别为29.5 mAU/ml，10.7 ng/L，1.5%。预测复发的敏感性分别为75.0%，54.7%，57.8%和79.7%。在随访期间其任意升高的生物标志物的患者的RFS和OS显著短于无升高生物标志物的患者（P < 0.001）。多变量分析显示，任意升高的生物标志物是复发的独立危险因素。PIVKA-II，AFP和AFP-L3的纵向监测可以有效预测手术后的HCC复发。版权所有©2023 Wolters Kluwer Health, Inc. 保留所有权利。

Radical resection is a curative treatment for patients with hepatocellular carcinoma (HCC), but the incidence of recurrence remains high. We aimed to explore the performance of predicting HCC recurrence by longitudinal surveillance of the protein induced by vitamin K absence (PIVKA-II), alpha- fetoprotein (AFP), and lectin-reactive AFP (AFP-L3) during postoperative follow-up.Patients who underwent radical resection for HCC at the Ningbo Medical Centre Lihuili Hospital between January 2015 and December 2020 were included. All enrolled patients regularly monitor PIVKA-II, AFP, AFP-L3 every 3 months during postoperative follow-up. The surveillance performance of PIVKA-II, AFP, AFP-L3 during follow-up for the prediction of HCC recurrence was compared in patients. The generalized estimation equation (GEE) was used to analyze the trends of the tumor biomarkers and interactions with time. Area under the receiver operator characteristic (AUROC) curves, the optimal cut-off value, the sensitivity and specificity were calculated to evaluate the performance of the three biomarkers. The recurrence-free survival (RFS) and overall survival (OS) of patients with any of the elevated biomarkers was analyzed by Kaplan-Meier curves and the log-rank test. Multivariate logistic regression models were used to analyze potential risk factors for recurrence.The GEE analysis indicated that PIVKA-II, AFP, AFP-L3 in the recurrence patients were higher than the no recurrence patients during follow-up, PIVKA-II and AFP showed increasing trends from 6 months before recurrence. In predicting recurrence, the AUROCs for PIVKA-II, AFP, AFP-L3 and their combination were 0.885, 0.754, 0.781 and 0.885 respectively, the optimal cut-off value for PIVKA-II, AFP, AFP-L3 was 29.5 mAU/ml, 10.7 ng/L, 1.5% respectively. The sensitivity in predicting recurrence for PIVKA-II, AFP, AFP-L3 and combination were 75.0, 54.7, 57.8 and 79.7% respectively. The RFS and the OS of patients with any of the biomarkers elevated during the follow-up was significantly shorter than that without elevated biomarkers (P < 0.001). Multivariate analysis showed that any of the biomarkers elevated was the independent risk factor of recurrence.Longitudinal surveillance of PIVKA-II, AFP and AFP-L3 can effectively predict recurrence of HCC after operation.Copyright © 2023 Wolters Kluwer Health, Inc. All rights reserved.