目的 探究细胞周期蛋白依赖性激酶抑制因子3（CDKN3）和血清壳多糖酶3样蛋白1（CHI3L1）、微小RNA-106b（miR-106b）联合检测在肝癌转移中的预测价值。 方法 选取2019年9月至2023年12月期间医院收治的肝癌患者100例作为研究对象，纳入肝癌组，将同期87例肝硬化患者作为肝硬化组，96例慢性乙肝患者作为慢性乙肝组、常规体检健康人员98例纳入健康对照组，对比各组血清CHI3L1、细胞周期分子CDKN3、miR-106b水平。将肝癌组按照随访1年期间是否转移分为26例转移组、63例未转移组、11例失随访，对比两组患者基线资料、细胞周期分子CDKN3、miR-106b、血清CHI3L1水平。利用多因素logistic分析肝癌转移的风险因素，利用ROC曲线分析miR-106b、细胞周期分子CDKN3、血清CHI3L1联合检测对肝癌转移的预测价值。 结果 与健康对照组相比，肝癌组、慢性乙型组、肝硬化组的CHI3L1、细胞周期分子CDKN3、血清miR-106b水平升高，且慢性乙型组 < 肝硬化组 < 肝癌组，随肝脏疾病程度不断增加，其呈现升高趋势，差异有统计学意义（P < 0.05）。与未转移组相比，转移组谷丙转氨酶（ALT）、肿瘤> 3 cm占比、CHI3L1、细胞周期素A2（CCNA2）、肿瘤分期 > Ⅲ期占比、细胞周期分子CDKN3、异常凝血酶原、谷氨酰转肽酶（GGT）、miR-106b、总胆汁酸（TBA）、甲胎蛋白（AFP）、肿瘤数量多发占比、谷草转氨酶（AST）水平升高，差异有统计学意义（P < 0.05）。多因素logistic分析结果显示，肝癌转移的风险因素包括ALT、CCNA2、肿瘤 > 3 cm、AST、GGT、肿瘤分期 > Ⅲ期、AFP、CHI3L1、多发肿瘤、miR-106b、细胞周期分子CDKN3、TBA。ROC曲线分析结果显示：miR-106b、血清CHI3L1、细胞周期分子CDKN3联合检测敏感度为93.65%、准确度为84.27%、AUC值为0.776高于单项检测敏感度、准确度、AUC值，差异有统计学意义（P < 0.05）。 结论 细胞周期分子CDKN3和血清CHI3L1、miR-106b水平升高为肝癌转移的风险因素，可用于预测肝癌转移发生情况，且联合检测预测效能显著。

Objective To investigate the predictive value of the combination of cell cycle protein-dependent kinase inhibitor 3（CDKN3） and serum chitosanase 3-like protein 1（CHI3L1）， and microRNA-106b（miR-106b） assays in hepatocellular carcinoma metastasis. Methods One hundred patients with hepatocellular carcinoma admitted to the hospital between September 2019 and December 2023 were selected as study subjects and included in the hepatocellular carcinoma group， 87 patients with cirrhosis were included as the cirrhosis group during the same period， 96 patients with chronic hepatitis B as the chronic hepatitis B group， and 98 healthy persons undergoing routine physical examination were included in the healthy control group， and the groups were compared with respect to the serum CHI3L1， the cell-cycle molecule CDKN3， and the miR-106b levels. The hepatocellular carcinoma group was divided into 26 metastatic cases， 63 non-metastatic cases， and 11 cases lost to follow-up according to whether they metastasized during the 1-year follow-up period， and baseline data， cell cycle molecules CDKN3， miR-106b， and serum CHI3L1 levels were compared between the two groups. Risk factors for hepatocellular carcinoma metastasis were analyzed using multifactorial logistic， and the predictive value of the combined detection of miR-106b， the cell cycle molecule CDKN3， and serum CHI3L1 for hepatocellular carcinoma metastasis was analyzed using ROC curves. Results Compared with the healthy control group， the levels of CHI3L1， cell cycle molecule CDKN3， and serum miR-106b were elevated in the hepatocellular carcinoma group， chronic type B group， and cirrhosis group， and the chronic type B group < cirrhosis group < hepatocellular carcinoma group， which showed a tendency to be elevated with the increasing degree of liver disease， and the difference was statistically significant（P<0.05）. Compared with the non-metastatic group， the metastatic group had higher levels of alanine aminotransferase（ALT）， percentage of tumors >3 cm， CHI3L1， cytokinin A2（CCNA2）， percentage of tumors with stage > Ⅲ， the cell cycle molecule CDKN3， abnormal plasminogen， glutamyltransferase（GGT）， miR-106b， total bile acids（TBA）， alpha-fetoprotein（AFP）， number of tumors percentage of multiple occurrences， and glutamyltransferase（AST） levels were elevated， and the difference was statistically significant（P<0.05）. The results of multifactorial logistics analysis showed that the risk factors for hepatocellular carcinoma metastasis included ALT， CCNA2， tumor > 3 cm， AST， GGT， tumor stage > Ⅲ， AFP， CHI3L1， multiple tumors， abnormally high blood plasminogen， miR-106b， cell cycle molecule CDKN3， TBA. the results of the ROC curve analysis showed that miR -106b， serum CHI3L1， cell cycle molecule CDKN3 combined detection sensitivity was 93.65%， accuracy was 84.27%， and AUC value was 0.776 higher than the sensitivity， accuracy， and AUC value of single detection， and the difference was statistically significant（P < 0.05）. Conclusions Elevated levels of cell cycle molecule CDKN3， serum CHI3L1， and miR-106b are risk factors for liver cancer metastasis. It can be used to predict the occurrence of liver cancer metastasis， and the combined detection has significant predictive efficacy.