Diagnostic Value Of A Volumetric Interpolated Breath-hold Examination MR Imaging Sequence In Evaluating Focal Hepatic Lesions

Objective:To investigate the diagnostic value of a fast gradient-echo (GRE)three-dimensional isotropic volumetric interpolated breath-hold examination (3D-VIBE) MR sequence capable of Gadolinium (Gd) enhanced multi-phasic acquisitions in evaluating focal liver lesions , by comparing withGd-enhanced two-dimensional GRE T1W sequence. Materials and Methods:Conventional spin-echo T2W, 2D GRE T1W plain scan, Gd-enhanced3D-VIBE multi-phasic acquisitions, and Gd-enhanced 2D GRE T1W scanning were prospectively performed for 128 consecutive patients suspected of having focal liver lesions on clinic or other imaging examination basis. The major acquisition parameters of 3D-VIBE sequence were TR/TE=4.2/1.8ms, flip angle 12°, bandwidth=488Hz/pixel, slice thickness 1.5~2 mm without gap. Native T2W and 2D GRE T1W were acquired first, then 3D-VIBE fast scanning at early arterial, late arterial and portal venous phases respectively, followed by 2D GRE T1W scanning at the equilibriumphase, hi 51 patients with a total of 90 focal liver lesions the final diagnosis was confirmed either surgical-pathologically or by integrating clinical data, imaging and laboratory findings, and follow-up observations, thus constituting the study group. Among them, there were 25 cases of hepatocellular carcinoma (42 lesions), 7 cases of neoplasm metastasis (21 lesions), 12 cases of cavernous hemangioma (15 lesions) and 7 cases of simple cyst (12 lesions). The diameters of liver lesions ranged from 0.2-14 cm, mean 3.11±2.71 cm.(D The MR signal characteristics (signal-to-noise ration, SNR and contrast-to-noise ratio, CNR) and morphological features of liver lesions at different phases were observed; (2) The reconstructed vascular structures of the liver and the lesions generated from 3D-VIBE data were delineated; (D The sensitivities, specificities, accuracies, positive and negative predictive values of 3D-VIBE and 2D GRE T1W for various focal liver lesions werecalculated. Results:1. The SNRs and CNRs of different focal liver lesions varied from3D-VIBE images of different phases. Also the Gd enhancement patterns and the evolution of enhancement varied from different kinds of focal liver lesions. Most of HCCs were slightly lowintense on Tl-weighted images and slightly highintense on T2-weighted images, which enhanced mainly on early arterial phases. Most of neoplasm metastasis were lowintense on Tl-weighted images and slightly highintense on T2-weighted images, which enhanced mainly on late arterial and portal venous phases. Cavernous hemangioma and simple cyst all presentedlowintense on Tl-weighted images and highintense on T2-weighted images, cavernous hemangioma enhanced mainly on portal venous phases and presented highintense or isointense on equilibrium. But simple cyst had non-enhancement on all phases.2. The diagnostic accuracies of 3D-VIBE and 2D GRE T1W for lesion detection and characterization were 98.9% (89/90), 91.1% (82/90) and 86.7% (78/90), 75.6% (68/90) respectively, the differences being of statistical significance (p<0.05). The sensitivities, specificities, accuracies, positive and negative predictive values of 3D-VIBE and 2D GRE T1W for characterizing the natures of focal liver lesions were 91.1%, 95.7%, 93.5%,95.3%, 91.8%fn 75.6%, 91.8%, 84%, 89.5%, 80.4% respectively.3. Using the data generated from 3D-VIBE sequence, the hepatic arterial anatomy was well delineated in 44 patients (86.3%), while the portal venous structures were clearly depicted in 46 cases (90.2 %). The feeding arteries of 17 lesions and the draining veins of 14 lesions of hepatocellular carcinoma were clearly displayed through MIP and MPRreconstruction techniques. Conclusion:The performance of 3D-VIBE is better than 2D GRE T1W not only inlesion detection and characterization for patients with focal liver lesions, but also in being able to provide much more information about the details of lesion hemodynamics and hepatic vascular anatomy.