Objective: To qualitatively and quantitatively assess the presentation of hepatic focal nodular hyperplasia (FNH) at diffusion-weighted magnetic resonance imaging (DWMRI) using multiple high b values.
Materials and Methods: Twenty-five patients with 27 FNHs had liver DWMRI at 1.5 T using free-breathing acquisition and 3 b values (0, 600, 1000 s/mm2). Focal nodular hyperplasias were evaluated qualitatively using visual analysis of diffusion-weighted magnetic resonance (DWMR) images and quantitatively using conventional apparent diffusion coefficient (ADC) and normalized ADC measurements.
Results: All FNHs (100%) were visible on b0 DWMR images; 26 of the 27 FNHs (96%), on b600 DWMR images; and 21 of the 27 FNHs (78%), on b1000 DWMR images. A total of 18 of the 27 FNHs (67%) exhibited a hyperintense central scar on the b0 DWMR images that remained visible on the b600 and b1000 DWMR images in 6 of the 27 FNHs (22%). Conventional ADC value of FNHs (1.318 × 10−3 mm2/±0.208) was significantly lower than that of adjacent hepatic parenchyma (1.414 × 10−3 mm2/s ± 1.95) (P = 0.0003), although a substantial overlap was found. The use of normalized ADC using the liver as reference organ resulted in a more restricted distribution of ADC values (variation coefficient, 5.3%).
Conclusions: Focal nodular hyperplasias show a wide range of morphological features at DWMRI using high b values. Further studies are needed to fully investigate as to what extent normalized ADC may result in better lesion characterization.
From the *Department of Abdominal Imaging, Hôpital Lariboisière, Assistance Publique-Hôpitaux de Paris; †Université Paris-Diderot, Sorbonne Paris Cité; ‡UMR INSERM 965-Paris 7 “Angiogenèse et recherche translationnelle”; §Department of Radiology, Hôpital Robert Debré, Reims Cedex; and ∥Department of Pathology, Hôpital Lariboisière, Assistance Publique-Hôpitaux de Paris, Paris, France.
Received for publication July 1, 2013; accepted August 12, 2013.
Reprints: Philippe Soyer, MD, PhD, Paris, France (e-mail: firstname.lastname@example.org).
The authors declare no conflict of interest.