Introduction

Magnetic resonance elastography of the liver (MREL) has been shown to aid in the non-invasive diagnosis and staging of liver fibrosis [1]. Two-dimensional (2D) gradient-recalled-echo (GRE) MRE employing: 1) a mechanical driver frequency of 60 Hz, 2) motion encoding gradients (MEG) applied along a single direction, and 3) inflow saturation, acquired during breath-hold (BH) is currently the standard of care (SC) sequence for the assessment of liver fibrosis [2]. Four transverse MREL images through mid-liver are typically acquired in 4 consecutive breath-holds of ~14 – 18sec (SC_2D_4BH) [3]. Given the motion sensitive nature of MREL, inadequate breath-holds lead to inaccurate stiffness estimation and/or failed studies. MREL images can be obtained in less than half the breath-hold time (2D_CS_HBH) e.g. <7s, with identical spatial resolution and field of view by setting the repetition time to 1.5 the period of applied mechanical frequency to encode motion using wave polarity inversion, plus data under-sampling using compressed Sensitivity Encoding (C-SENSE) [4]. The purpose of this study is to compare liver stiffness measurements obtained using SC_2D_4BH technique to those obtained by the 2D_CS_HBH.

Methods

In this prospective HIPAA compliant IRB approved study, 19 participants (age 36.9±12.6 years, range 20-62 years, 8 males) underwent MREL imaging with written informed consent. All imaging was performed on a Philips Ingenia 1.5T scanner (Best, The Netherlands), using 28-channel torso coil. MREL images were obtained using both the commercially available SC_2D_4BH MREL sequence (Fig. 1 A) and 2D_CS_HBH MREL sequence with (Fig. 1 B) and without inflow saturation (Fig. 1 C). Commercially available SENSE (based on regular undersampling and coil sensitivity information, and spatial solution space constraint based on prior knowledge of the image extent) and C-SENSE (combines spatial domain variable-density pseudorandom undersampling of k-space with the SENSE reconstruction algorithm using iterative reconstruction and sparsity constraints) were employed [5]. All image reconstructions and stiffness map (using multimodal direct inversion algorithm with 95% confidence interval mask overlays [6]) computations were performed inline in real-time available on the scanner. All the MREL data, including magnitude, phase, and wave images as well as shear stiffness maps, were exported to a post-processing workstation (with IntelliSpace Portal v10.1; ISP, Philips). Manual liver shear stiffness measurements were made by a single trained image analyst (>1-year experience), under the supervision of a board-certified Pediatric Radiologist (>10-years post-fellowship experience). On four slices from each acquisition, a single freehand region of interest (ROI) was drawn to encompass as much of the right hepatic lobe as possible while remaining within the boundaries of the 95% confidence mask and avoiding the liver capsule, large blood vessels, dilated bile ducts, and areas of artifact, including hot-spots immediately under the passive driver and respiratory motion in cases of failed breath holds (Fig. 2). Additionally, to compare the shear stiffness values in the same tissue across the techniques, identically sized and positioned ROIs were drawn on both sets of shear stiffness maps. Overall participant-specific liver shear stiffness was calculated as the ROI area weighted mean liver shear stiffness of the estimates extracted from each slice. Two-sided paired t test, intra-class correlation coefficient (ICC), , and Bland-Altman analysis were used to compare and assess agreement in the ROI sizes and shear stiffness estimates between the sequences.

Results

MREL image acquisition and reconstruction was performed successfully in all the participants. Breath-hold times were 13.3 ± 0.2 s per slice with SC_2D_4BH MRE, 6.3 ± 0.2 s with interleaved flow saturation, and 4.9 ± 0.2 s without flow saturation for 2D_CS_HBH techniques. Descriptive and comparative statistics of the liver stiffness values and area within 95% confidence mask are presented in Tables (1-2) and Fig. (3). Area weighted mean liver stiffness estimates for identical tissue ROI for the SC_2D_4BH (2.17 ± 0.68 kPa, range 1.60 – 4.53 kPa) and 2D_CS_HBH with flow saturation (2.04 ± 0.66 kPa, range 1.46 – 4.39 kPa), and without flow saturation (2.03 ± 0.61 kPa, range 1.40 – 4.15 kPa), MRE sequences correlated very strongly (ICC>0.96).

Discussion

Liver shear stiffness values obtained by 2D_CS_HBH, with or without flow saturation, correlate very strongly (ICC > 0.96) with stiffness values obtained by SC_2D_4BH, underestimating stiffness significantly (p<0.03) with bias <0.15 kPa (<6%). While all the techniques had identical acquired spatial resolution and coverage, the difference in total ROI area across 4 slices was less than 17 cm². Liver stiffness values estimated by 2D_CS_HBH with or without flow saturation were comparable (p>0.07). 2D_CS_HBH MREL has the potential to markedly reduce breath hold times, allowing liver stiffness mapping in participants who are only able to perform consistent 5 to 6 second breath-holds. Alternatively, in participants who are able to perform consistent 10 to 12 second breath-hold, the 2D_CS_HBH technique enables either reduced number of breath-holds, or increased coverage relative to the standard 2D_SC_4BH MREL approach.

Conclusion

Liver stiffness values measured in less than half (<5-7 s) the standard (>13 s) breath-hold time per slice achieved by the combined use of wave polarity-inversion motion encoding and C-SENSE (2D_CS_HBH) were comparable to liver stiffness values estimated by the standard of care MREL sequence. 2D-CS_HBH MREL has potential clinical benefits in participants with limited and/or compromised breath-holding capabilities.

Acknowledgements

No acknowledgement found.

References

[1] Singh S et al. Diagnostic performance of magnetic resonance elastography in staging liver fibrosis: a systematic review and meta-analysis of individual participant data. Clinical gastroenterology and hepatology, 2015: 13(3), 440–451. [2] Venkatesh SK. et al. Hepatic MR elastography: Clinical performance in a series of 1377 consecutive examinations. Radiology, Magnetic resonance elastography of liver: technique, analysis, and clinical applications. J Magn Reson Imaging. 2013: 37, 544–555. [3] Yin, M. et al. Hepatic MR elastography: Clinical performance in a series of 1377 consecutive examinations. Radiology, 2016: 278, 114–124. [4] Wang H. et al. Fusing acceleration and saturation techniques with wave amplitude labeling of time-shifted zeniths MR elastography. Magn Reson in Med. 2021: 85(3), 1552-1560, doi: 10.1002/mrm.28488. [5] Geerts-Ossevoort L. et al. A. Compressed SENSE. Speed done right. Every time. Philips Field Strength Magazine, 2018: 278, 6619. [6] Silva AM. et al. A. Magnetic resonance elastography: Evaluation of new inversion algorithm and quantitative analysis method. Abdom Imaging, 2015: 40, 810-817.