钆酸
肝细胞癌
对比度(视觉)
钆DTPA
细胞外
医学
放射科
化学
磁共振成像
内科学
计算机科学
生物化学
人工智能
作者
Yuyao Xiao,Peng Huang,Cheng Wang,Changwu Zhou,Fei Wu,Zeyang Wang,Haoran Dai,Xinyue Liang,Xi Jia,Chun Yang,Mengsu Zeng
出处
期刊:Liver cancer
[Karger Publishers]
日期:2025-08-01
卷期号:: 1-15
摘要
Introduction: The optimal imaging modality and diagnostic criteria for accurately detecting and characterizing subcentimeter hepatocellular carcinoma (HCC) remain uncertain, and this study aims to compare performance of gadoxetic acid-enhanced MRI (EOB-MRI) and extracellular contrast agent-enhanced MRI (ECA-MRI) in detecting and characterizing subcentimeter HCC. Methods: A total of 1,022 patients at risk of HCC (mean age, 53.80 ± 11.24, 732 men) with 1,210 subcentimeter hepatic lesions were retrospectively included. Lesion detection rate and HCC characterization performance were calculated and compared between EOB-MRI and ECA-MRI sets using generalized estimating equation method. Results: Consensually, EOB-MRI demonstrated significantly higher sensitivity for detecting subcentimeter hepatic lesions compared to ECA-MRI (0.995 vs. 0.953, p < 0.001). EOB-MRI and ECA-MRI showed comparable performance in characterizing subcentimeter HCC based on typical vascular pattern (sensitivity, 0.382 vs. 0.457, p = 0.064; specificity 0.941 vs. 0.933, p = 0.462). After applying modified criteria, the sensitivities (EOB-MRI: 0.382 vs. 0.812, p < 0.001; ECA-MRI: 0.457 vs. 0.574, p < 0.001) were significantly increased on both MRIs by consensus reading, while specificities did not differ a lot (EOB-MRI: 0.859 vs. 0.941, p = 0.012; ECA-MRI: 0.894 vs. 0.933, p = 0.084). And compared with ECA-MRI, EOB-MRI exhibited significantly higher sensitivity (0.812 vs. 0.574, p < 0.001) based on modified criteria, without a substantial loss of specificity (0.859 vs. 0.894, p = 0.162). Conclusion: EOB-MRI with modified criteria exhibited superior detection and characterization performance of subcentimeter HCC when compared with ECA-MRI in patients at risk of HCC, thus offering clinicians more opportunities to accurately identify high-risk subcentimeter lesions.
科研通智能强力驱动
Strongly Powered by AbleSci AI