Qiu Lixia, Xu Shanshan, Qiu Yundong, Liu Yali, Zhang Jing
Beijing Youan Hospital, Capital Medical University, Beijing, China.
Gu'an Hospital of Traditional Chinese Medicine, Langfang, China.
J Viral Hepat. 2023 Jun;30(6):559-566. doi: 10.1111/jvh.13829. Epub 2023 Mar 22.
Hepatitis C patients with advanced fibrosis or cirrhosis are at high risk of developing hepatocellular carcinoma (HCC) even after sustained virological response (SVR). Several HCC risk scores have been developed but which one is most suitable for this population is unclear. In this study, we compared the prediction ability of the aMAP model, THRI model, PAGE-B model and Models of HCV in a prospective hepatitis C cohort in order to propose better model(s) to clinical practice. Adult hepatitis C patients with baseline advanced fibrosis (141 cases), compensated cirrhosis (330 cases) and decompensated cirrhosis (80 cases) were included and followed up every 6 months for about 7 years or until HCC development. Demographic data, medical history and laboratory results were recorded. HCCs were diagnosed by radiography, AFP or liver histology. The median follow-up period was 69.93(60.99-74.93) months, during which 53 (9.62%) patients developed HCC. The areas under the receiver operating characteristic curve of aMAP, THRI, PAGE-B and Models of HCV scores were 0.74, 0.72, 0.70 and 0.63 respectively. The predictive power of the aMAP model score was comparable to that of THRI, PAGE-Band higher than that of Models of HCV (p < 0.05). Dividing patients into non-high-risk and high-risk groups, the cumulative incidence rates of HCC based on aMAP, THRI, PAGE-B and Models of HCV was 5.57% vs. 24.17%, 1.10% vs. 13.90%, 5.80% vs. 15.90% and 6.41% vs. 13.81% (all p < 0.05). The AUC of the four models were all below 0.7 in male while all were higher than 0.7 in female. The performance of all the models was not influenced by fibrosis stage. aMAP, THRI model and PAGE-B model were all performed well while THRI model and PAGE-B model were easier to calculate. There was no need to select score according to fibrosis stage but should be caution when explain the results in male patients.
即使获得持续病毒学应答（SVR），患有晚期纤维化或肝硬化的丙型肝炎患者发生肝细胞癌（HCC）的风险仍然很高。已经开发了几种HCC风险评分，但哪种评分最适合该人群尚不清楚。在本研究中，我们比较了aMAP模型、THRI模型、PAGE-B模型和丙型肝炎模型（Models of HCV）在一个前瞻性丙型肝炎队列中的预测能力，以便为临床实践提出更好的模型。纳入了具有基线晚期纤维化（141例）、代偿期肝硬化（330例）和失代偿期肝硬化（80例）的成年丙型肝炎患者，每6个月随访一次，持续约7年或直至发生HCC。记录人口统计学数据、病史和实验室检查结果。通过影像学、甲胎蛋白（AFP）或肝脏组织学诊断HCC。中位随访期为69.93（60.99 - 74.93）个月，在此期间53例（9.62%）患者发生了HCC。aMAP、THRI、PAGE-B和丙型肝炎模型评分的受试者工作特征曲线下面积分别为0.74、0.72、0.70和0.63。aMAP模型评分与THRI、PAGE-B的预测能力相当，且高于丙型肝炎模型（p < 0.05）。将患者分为非高危组和高危组，基于aMAP、THRI、PAGE-B和丙型肝炎模型的HCC累积发病率分别为5.57%对24.17%、1.10%对13.9%、5.80%对15.90%和6.41%对13.81%（均p < 0.05）。四个模型的曲线下面积（AUC）在男性中均低于0.7，而在女性中均高于0.7。所有模型的性能均不受纤维化分期的影响。aMAP、THRI模型和PAGE-B模型均表现良好，而THRI模型和PAGE-B模型计算更简便。无需根据纤维化分期选择评分，但在解释男性患者结果时应谨慎。
