钆塞酸增强MRI诊断<10 mm肝细胞癌的不同标准比较
本研究回顾性纳入308例肝癌高危患者共352个<10 mm肝脏病灶(含147个肝细胞癌),在钆塞酸增强MRI上比较典型血管模式、LI-RADS 4类、APASL及KLCA-NCC四种诊断标准。LR-4类准确度最高但总体表现仍有限,提示需要优化<10 mm肝细胞癌的诊断标准。
要点
- LR-4类敏感度最高(0.816),适合作为筛查性诊断阈值以减少漏诊。
- 典型血管模式特异度最高但敏感度最低,报告时应说明其排除价值。
- 四种标准对<10 mm肝细胞癌的诊断效能均有限,阅片时不宜过度依赖单一标准。
核心结论
- 四种诊断标准在敏感度与特异度之间各有优劣,整体诊断效能对<10 mm肝细胞癌仍属中等水平。
- LR-4类准确度最高(0.756,95%CI 0.708-0.798),敏感度亦最高(0.816,95%CI 0.746-0.871),但特异度仅为中等(0.712,95%CI 0.647-0.770)。
- 典型血管模式特异度最高但敏感度最低,APASL与KLCA-NCC的敏感度与LR-4类无显著差异(均p>0.05)。
研究设计与样本
- 回顾性研究,纳入308例肝癌高危患者(平均年龄51.69±11.17岁,210例男性),共352个<10 mm肝脏病灶,其中147个为肝细胞癌。
- 采用钆塞酸增强MRI,按病灶为单位评估诊断效能,参考标准为病理或临床随访诊断。
- 使用广义估计方程进行指南间两两比较,并根据是否有肝硬化进行亚组分析。
关键影像表现/征象
- 钆塞酸增强MRI上评估<10 mm病灶的动脉期强化、门静脉期及延迟期廓清等血管征象以区分肝细胞癌与其他病变。
- 亚组分析显示无论有无肝硬化,各标准的诊断效能模式与总体一致。
诊断标准与关键数值
- LR-4类:准确度0.756(95%CI 0.708-0.798),敏感度0.816(95%CI 0.746-0.871),特异度0.712(95%CI 0.647-0.770)。
- KLCA-NCC特异度0.654,APASL特异度0.551,两者比较p=0.055,差异接近但未达显著。
- LR-4类与APASL特异度比较:0.712对0.551,p<0.001,差异有统计学意义。
临床应用与阅片要点
- 阅片时对<10 mm病灶可优先参考LR-4类标准以提高检出敏感度。
- 典型血管模式虽特异度高但敏感度低,不宜单独用于排除<10 mm肝细胞癌。
- 报告<10 mm病灶时应注明所用诊断标准的局限,并结合患者肝硬化状态综合判断。
局限性
- 回顾性单中心研究设计,存在选择偏倚可能。
- 四种标准对<10 mm肝细胞癌的总体诊断效能均有限,提示现有标准在该尺寸范围适用性不足。
以上解读依据论文摘要(全文未开放获取或许可不允许翻译转载)。
英文原文(摘要)· 点击展开
OBJECTIVES: To evaluate and compare the diagnostic performance of typical vascular pattern, the Liver Imaging Reporting and Data System (LI-RADS) category 4 (LR-4 category), the Asian Pacific Association for the Study of the Liver (APASL) and the Korean Liver Cancer Association-National Cancer Center (KLCA-NCC) for diagnosing hepatocellular carcinoma (HCC) < 10 mm using gadoxetic acid-enhanced MRI.
METHODS: This retrospective study included 308 patients (mean age, 51.69 ± 11.17 years; 210 men) at high risk for HCC with 352 hepatic lesions < 10 mm on gadoxetic acid-enhanced MRI, including 147 HCC. Diagnostic performance was assessed on a per-lesion basis. Generalized estimating equations with pairwise comparisons were used to compare guidelines. Subgroup analyses were performed according to cirrhosis status.
RESULTS: The LR-4 category achieved the highest accuracy (0.756, 95 % CI: 0.708-0.798) among the four diagnostic criteria and also yielded the highest sensitivity (0.816, 95 % CI: 0.746-0.871), although its specificity remained modest (0.712, 95 % CI: 0.647-0.770). Sensitivity did not differ significantly among the APASL, KLCA-NCC, and LR-4 category (all p > 0.05). Compared with APASL, KLCA-NCC showed slightly higher specificity (0.551 vs. 0.654, p = 0.055), whereas LR-4 category demonstrated significantly higher specificity (0.551 vs. 0.712, p < 0.001). The typical vascular pattern provided the highest specificity but the lowest sensitivity. Similar diagnostic performance patterns were also observed in both the cirrhotic and non-cirrhotic subgroups.
CONCLUSION: The evaluated diagnostic criteria showed different strengths and limitations in balancing sensitivity and specificity, but their overall diagnostic performance for HCC < 10 mm remained modest, highlighting the need for optimized diagnostic criteria for HCC < 10 mm.
引用
- Xiao, Gu, Zhou, et al. Comparison of diagnostic criteria for hepatocellular carcinoma < 10 mm on gadoxetic acid-enhanced MRI. Eur J Radiol. 2026. PMID: 42705138. doi:10.1016/j.ejrad.2026.113223
供执业医师工作参考,不替代临床判断;数值与分级以指南原文为准。