Objective: To validate the performance of a multi-omics combined test for early screening of high-risk liver cancer populations. Methods: 173 high-risk patients with liver cancer were prospectively screened in a real-world setting, and 164 cases were finally enrolled. B-ultrasound, alpha-fetoprotein (AFP), and HCC screens were conducted in all patients. A multi-omics early screening test was performed for liver cancer in combination with multi-gene methylation, TP53/TERT/CTNNB1 mutations, AFP, and abnormal prothrombin (PIVKA-II). Differences in rates were compared using the chi-square test, adjusted chi-square test, or Fisher's exact probability method for count data. A non-parametric rank test (Mann-Whitney) was used to compare the differences between the two groups of data. Results: The HCCscreen detection had a sensitivity of 100% for liver cancer screening, 93.8% for liver cancer and precancerous diseases, 34.1% for positive predictive value, 99.2% for negative predictive value, and 0.89 for an area under the curve (AUC). Parallel detection of AFP, AFP+B-ultrasound, and methylation+mutation had a sensitivity/specificity and AUC of 31.3%/88.5% (AUC=0.78), 56.3%/88.2% (AUC=0.86), and 81.3%/82.4 % (AUC=0.84). At the same time, the disease severity range was significantly correlated with the methylation+mutation score, HCCscreen score, or positive detection rate (PDR). There was no significant correlation between AFP serum levels and methylation+mutation or HCCscreen scores, while there was a significant linear correlation between methylation+mutation scores and HCCscreen scores (r = 0.73, P < 0.001). Conclusion: In real-world settings, HCCscreen shows high sensitivity for screening opportunistic, high-risk liver cancer populations. Furthermore, it may efficaciously detect liver cancer and precancerous diseases, with superior performance to AFP and AFP+ultrasound. Hence, HCCscreen has the potential to become an effective screening tool that is superior to existing screening methods for high-risk liver cancer populations.
目的： 拟验证一项多组学联合检测在肝癌高危人群进行肝癌早筛的性能。 方法： 在真实世界环境下前瞻性筛选173例肝癌高危患者，最后入组164例患者。所有患者均行B超、甲胎蛋白（AFP）和HCCscreen检测。HCCscreen检测是一项合并多基因甲基化、TP53/TERT/CTNNB1突变、AFP和异常凝血酶原（PIVKA-II）的多组学肝癌早筛检测。计数资料使用χ(2)检验、校正χ(2)检验或Fisher精确概率法比较率的差异；等级资料使用非参数检验（Mann-Whitney）法比较两组数据间的差异。 结果： HCCscreen检测对肝癌的筛查灵敏度为100%，对肝癌及癌前疾病的筛查灵敏度为93.8%，阳性预测值34.1%，阴性预测值99.2%，曲线下面积（AUC）为0.89。平行检测的AFP、AFP+B超和甲基化+突变的灵敏度/特异度及AUC分别为31.3%/88.5%（AUC=0.78）、56.3%/88.2%（AUC=0.86）和81.3%/82.4%（AUC=0.84）。同时，疾病严重程度与甲基化+突变评分、HCCscreen评分或阳性检出率存在显著相关性。虽然AFP血清水平与甲基化+突变或HCCscreen评分之间没有显著相关性，但甲基化+突变评分与HCCscreen评分之间存在显著的线性相关性(r = 0.73, P < 0.001)。 结论： 在真实世界机会性筛查条件下，HCCscreen显示出对肝癌高危人群筛查高灵敏度，可有效发现肝癌及癌前疾病，其表现优于AFP及AFP+B超。HCCscreen有潜力成为优于现有筛查方法的肝癌高危人群的有效筛查工具。.
Keywords: DNA methylation; Early diagnosis; Gene mutation; Hepatocellular carcinoma; Screening; Tumor biomarker.