Multiphasic Dynamic MR Imaging of Focal Nodular Hyperplasia

The purpose of this study is to evaluate the plain and multiphasic dynamic Gd-enhanced MR imaging features of hepatic focal nodular hyperplasia (FNH) with breath-hold MR sequences. Retrospectively, sixteen lesions of FNH in fifteen patients (ten male and five female) were reviewed. Among the sixteen lesions, three were proved by surgery, four by percutaneous biopsy, and the other nine by clinical follow-up. Patient's age ranged from 21 to 62 years old (mean 33.1). MR examinations were performed with a 1.5-T superconducting whole body imager and phased-array body coil. All patients in pre-contrast scan were imaged with breath-hold T1-weighted fast low angle shot (FLASH), T2-weighted half-Fourier acquisition single-shot turbo-spin echo (HASTE) and T2-weighted fast spine echo (FSE) sequences. Multiphasic dynamic gadolinium (Gd)-enhanced MR images were obtained at 30 seconds, 70 seconds 3 minutes, 6 minutes and 10 minutes following a bolus of intravenous injection of 0.1 mmole/kg Gd-DTPA with fat-saturation breath-hold T1-weighted FLASH sequence. FNH could be detected on T1-weighted FLASH images as hypointense lesions (n = 10, 63%) and T2-weighted FSE or HASTE images as hyperintense lesions (n = 11, 69%). Gd-enhanced dynamic MR imaging demonstrated homogeneous and complete enhancement (except central scar if present) on arterial-phase imaging (n = 16, 100%). Portal-venous-phase and delayed-phase imaging showed isointense (n = 8, 50%) or slightly hyperintense (n = 8, 50%). Central scar could be identified in twelve lesions and showed to be hypointense on T1-weighted (n = 9, 75%) and hyperintense on T2-weighted images (n = 9, 75%). In dynamic study, central scar demonstrated to be hypointense (n = 12, 100%) in arterial-phase images. Delayed enhancement of central scar could be depicted in portal-venous-phase images (n = 2, 17%) or a series of delayed scans: 3 minutes (n = 4, 33%), 6 minutes (n = 3, 25%), and 10 minutes (n = 3, 25%). Size of the central scar in seven lesions was smaller than 5 mm (58%), four was 5mm to 1cm (33%), and one was 2 cm in diameter (8%). Only one particular lesion demonstrated capsule formation. Breath-holding plain and dynamic multiphasic MR imaging have been used broadly in the diagnosis of FNH. In the present study, both arterial-and delayed-phase images with Gd-enhanced T1-weighted sequences were thought to be superior to plain images in lesion detection. We also demonstrated that delayed scan longer than 6 minutes is essential to depict the tiny central scar. Typical MR imaging features of FNH included homogeneous solid mass with iso- or slightly hypointense on T1-weighted, iso- or slightly hyperintense on T2-weighted images as compared to surrounding normal liver parenchyma. Central scar presented as markedly hypointense on T1-weighted and iso- or hyperintense on T2-weighted images as compared to tumor. Dynamic study of main tumor demonstrated strong enhancement during arterial-phase imaging but central scar depicted strong enhancement during delayed-phase imaging.