医学
局灶性结节性增生
超声造影
超声波
对比度(视觉)
腺瘤
放射科
增生
前瞻性队列研究
病理
内科学
肝细胞癌
人工智能
计算机科学
作者
Jordan Swensson,Mary McCrate,Vivek Gowdra Halappa,Taylor Stethen,Fatih Akisik
出处
期刊:Ultrasound Quarterly
[Ovid Technologies (Wolters Kluwer)]
日期:2024-11-22
卷期号:40 (4)
标识
DOI:10.1097/ruq.0000000000000696
摘要
Magnetic resonance imaging (MRI) may be time-consuming, expensive, or poorly tolerated by patients with liver lesions. This is a prospective clinical trial designed to evaluate if contrast-enhanced ultrasound (CEUS) can be used to differentiate focal nodular hyperplasia (FNH) from hepatocellular adenoma (HCA) with similar accuracy compared with hepatobiliary agent MRI.Institutional review board approval was obtained (1805450097), and the trial was registered with ClinicalTrials.gov (NCT03652636). From 2018 through 2023, 40 patients who had lesions consistent with FNH or HCA on HBA-MRI underwent 1-time prospective CEUS of up to 2 hepatic lesions. Sonographic images obtained before and after intravenous administration of 2 mL sulfur hexafluoride lipid-type A microspheres (Lumason) per lesion totaling 59 lesions (27 FNHs/32 HCAs). Two blinded radiologists provided a diagnosis of FNH or HCA.Thirty-eight female and 2 male patients (age 36.7 ± 9.9) were scanned. Radiologists provided diagnosis of FNH or HCA with respective sensitivity (66.7/64.0%), specificity (71.9/90.6%), and accuracy (69.5/78.0%). For 38 lesions greater than or equal to 2 cm in size (17 FNHs/21 HCAs), readers had sensitivity (70.6/84.2%), specificity (70.6/84.2%), and accuracy (81.5/86.8%). Interobserver agreement for all lesions was fair (κ = 0.34), whereas agreement for lesions 2 cm or larger was substantial (κ = 0.67).Contrast-enhanced ultrasound can differentiate FNH from HCA with accuracy approaching that of hepatobiliary agent MRI for lesions 2 cm or greater. Interobserver agreement is improved with larger lesions. CEUS may have utility as an alternate diagnostic tool for FNH/HCA, especially in patients who cannot or do not desire to undergo MRI.
科研通智能强力驱动
Strongly Powered by AbleSci AI