High resolution multi-arterial phase MRI improves lesion contrast in chronic liver disease

Abstract

Purpose: To determine the reliability of arterial phase capture and evaluate hypervascular lesion contrast kinetics with a combined view-sharing and parallel imaging dynamic contrast-enhanced acquisition, DIfferential Sub-sampling with Cartesian Ordering (DISCO), in patients with known chronic liver disease.Methods: A retrospective review of 3T MR images from 26 patients with known chronic liver disease referred for hepatocellular carcinoma surveillance or post-treatment follow up was performed. After administration of a gadolinium-based contrast agent, a multiphasic acquisition was obtained in a 28 s breath-hold, from which seven sequential post-contrast image volumes were reconstructed.Results: The late arterial phase was successfully captured in all cases (26/26, 95% CI 87-100%). Images obtained 26 s post-injection had the highest frequency of late arterial phase capture (20/26) and lesion detection (23/26) of any individual post-contrast time; however, the multiphasic data resulted in a significantly higher frequency of late arterial phase capture (26/26, p=0.03) and a higher relative contrast (5.37+/-0.97 versus 7.10+/-0.98, p < 0.01). Conclusion:Multiphasic acquisition with combined view-sharing and parallel imaging reliably captures the late arterial phase and provides sufficient temporal resolution to characterize hepatic lesion contrast kinetics in patients with chronic liver disease while maintaining high spatial resolution.

View‐Sharing Artifact Reduction With Retrospective Compressed Sensing Reconstruction in the Context of Contrast‐Enhanced Liver MRI for Hepatocellular Carcinoma (HCC) Screening

“…Although only a single hepatic arterial phase is required, recent advances in pulse sequence capabilities now permit rapid acquisition of multiple arterial phases within a single breath‐hold. Acquiring multiple arterial phases improves the likelihood that a diagnostic hepatic arterial phase will be captured and accommodates variability in conspicuity of lesions based on timing of arterial enhancement …”

CT during celiac artery angiography for localization of clinically suspected small insulinomas

“…There are pros and cons for each method, for example, the sensitivity of non-invasive preoperative imaging, for localization of lesions, is between 73.08 and 90% [ 5 – 8 ]. Conventional enhanced images sometimes are not able to accurately capture the most prominent phase of enhancement of a particular lesion, which makes it possible to get an equivocal results at conventional imaging [ 9 ]. Especially for small lesions, non-invasive imaging is not sufficient for preoperative localization of insulinomas.…”

Combined parenchymal and vascular imaging: High spatiotemporal resolution arterial evaluation of hepatocellular carcinoma

“…For arterially enhancing observations greater than 1 cm in maximal dimension, regions of interest (ROIs) were placed over the observation and adjacent hepatic parenchyma using OsiriX (T.A.H.). To minimize signal variations due to coil‐related signal intensity variations, the adjacent parenchymal ROI was placed in a similar AP plane as the lesion . These ROIs were duplicated across all imaged phases and on precontrast imaging.…”

“…To minimize signal variations due to coil-related signal intensity variations, the adjacent parenchymal ROI was placed in a similar AP plane as the lesion. 22 These ROIs were duplicated across all imaged phases and on precontrast imaging. The ratio of signal intensity of the lesion to the adjacent parenchyma (lesion-to-adjacent) as well as the signal intensity of the lesion at each phase relative to precontrast imaging (lesion-pre) was calculated for each imaged phase.…”

“…Recently, several novel methods for contrast-enhanced multiphase liver imaging have been introduced with high temporal resolution without compromising spatial resolution or coverage. [18][19][20][21][22] A high spatiotemporal resolution technique, DISCO (Differential Subsampling with Cartesian Ordering), was described that combines a dual-echo SPGR sequence with pseudo-random k-space segmentation and a view-sharing reconstruction. 18 It provides reliable fat-water separated images with multiple postcontrast phases acquired in a single breath-hold.…”