Xia Feng, Yan Jing, Liu Xulin, Huang Zhao, Qiu Zhancheng, Zhang Qiao, Wu Zhenheng, Huang Zhiyuan, Wei Renjie, Lin Li, Liu Liping, Han Shuangqin, Yuan Yulin, Yin Huaxuan, Xia Guobing, Wan Yunyan, Xiao Shuo, Li Jian, Zhou Guoxiang, Xia Xiafei, Sun Huapeng, Wang Shuai, Zheng Jun, Gao Hengyi, Zheng Jiang, Ren Li, Mo Ali, Fu Junquan, Ye Lin, Ruan Shun, Wang Yuefan, Chen Xiaoping, Cheng Qi, Zhang Bixiang, Zhu Peng
Department of Hepatic Surgery, Tongji Hospital, Tongji Medical College of Huazhong University of Science and Technology, Wuhan, 430030, China.
Department of Ultrasound in Medicine, The Second Affiliated Hospital of Zhejiang University School of Medicine, Hangzhou, 310009, China.
Sci China Life Sci. 2025 Sep 10. doi: 10.1007/s11427-025-3045-1.
This study aimed to develop and validate a novel pathological scoring system, the MSE score (microvascular invasion, satellite foci, Edmondson-Steiner grade), for predicting postoperative recurrence in hepatocellular carcinoma (HCC) patients at BCLC stage 0-A following liver resection. The MSE score was compared to the established microvascular invasion/satellitosis (mVI/S) scoring system, which combines microvascular invasion and satellite nodules, both of which are significant predictors of aggressive recurrence and mortality. A total of 3,338 patients from 28 centers, who underwent curative liver resection between January 2017 and January 2020, were included. These patients were divided into a training cohort (n=2,337) and a validation cohort (n=1,001). Univariate and multivariate Cox regression models, along with Kaplan-Meier survival analysis, receiver operating characteristic (ROC) curves, and random forest modeling, were used to assess recurrence factors and compare predictive performance. The results showed that the MSE score had superior discriminatory power compared to mVI/S in predicting recurrence and overall survival. Kaplan-Meier analysis indicated a significantly better separation of low- and high-risk groups with the MSE score (P<0.001), while ROC analysis revealed higher AUC values for the MSE score at 1, 3, and 5 years. Multivariate Cox analysis demonstrated higher predictive power for the MSE score (HR=2.900 (1.808-4.652)) compared to mVI/S (HR=2.134 (1.567-3.043)). Additionally, logistic regression confirmed that the MSE score had stronger predictive power for aggressive recurrence (OR=2.992 (1.816-4.862) vs. mVI/S OR=1.847 (1.105-3.470)). SHAP analysis highlighted that the MSE score was more significant than mVI/S. The MSE score also outperformed traditional BCLC staging, TNM staging, and preoperative serum AFP levels in predicting recurrence. In conclusion, the MSE score provides reliable predictions for postoperative aggressive recurrence and survival in BCLC stage 0-A HCC, offering superior risk stratification compared to mVI/S and promising utility in guiding postoperative management.
本研究旨在开发并验证一种新的病理评分系统,即MSE评分(微血管侵犯、卫星灶、Edmondson-Steiner分级),用于预测BCLC 0-A期肝细胞癌(HCC)患者肝切除术后的复发情况。将MSE评分与既定的微血管侵犯/卫星灶(mVI/S)评分系统进行比较,后者结合了微血管侵犯和卫星结节,这两者都是侵袭性复发和死亡的重要预测指标。纳入了28个中心的3338例患者,这些患者在2017年1月至2020年1月期间接受了根治性肝切除术。这些患者被分为训练队列(n = 2337)和验证队列(n = 1001)。采用单因素和多因素Cox回归模型,以及Kaplan-Meier生存分析、受试者工作特征(ROC)曲线和随机森林建模,来评估复发因素并比较预测性能。结果表明,在预测复发和总生存方面,MSE评分比mVI/S具有更高的鉴别力。Kaplan-Meier分析表明,MSE评分能更显著地区分低风险和高风险组(P < 0.001),而ROC分析显示MSE评分在1年、3年和5年时具有更高的AUC值。多因素Cox分析表明,与mVI/S(HR = 2.134(1.567 - 3.043))相比,MSE评分具有更高的预测能力(HR = 2.900(1.808 - 4.652))。此外,逻辑回归证实,MSE评分对侵袭性复发具有更强的预测能力(OR = 2.992(1.816 - 4.862) vs. mVI/S OR = 1.847(1.105 - 3.470))。SHAP分析突出表明,MSE评分比mVI/S更具显著性。在预测复发方面,MSE评分也优于传统的BCLC分期、TNM分期和术前血清AFP水平。总之,MSE评分为BCLC 0-A期HCC患者术后侵袭性复发和生存提供了可靠的预测,与mVI/S相比具有更好的风险分层,在指导术后管理方面具有广阔的应用前景。