HKU Data Repository

File(s) not publicly available

Multi-shot DWI with MUSE reconstruction improves image quality for the diagnostic evaluation of active Crohn’s disease

posted on 2020-08-20, 03:43 authored by Guangtao Chen
Diffusion-weighted imaging(DWI) can provide functional and quantitative information for the imaging assessment, and plays more and more important roles in the diagnostic evaluation of Crohn’s disease. Current DWI acquiring sequence is mainly based on single-shot method, such as single-shot diffusion-weighted echo planar imaging (ssDW-EPI). But single-shot method limits the spatial resolution and overall image quality of DWI. The conventional ssDW-EPI cannot effectively identify some penetrating complications or other active bowel inflammations, especially if the bowel is not well distended. Multi-shot diffusion-weighted echo planar image (msDW-EPI) sequence can produce high-resolution DWI images, but it is sensitive to motion-induced data inconsistency. A recently developed multiplexed sensitivity-encoding (MUSE) framework can deal with data inconsistency issue and produce msDW-EPI data with improved spatial resolution and reduced distortions. In this study, I proposed to improve the resolution and image quality of bowel DWI using two methods, one is using the MUSE framework with msDW-EPI, another is to use highly-accelerated SENSE (R=4) ssDW-EPI method, and evaluate their performance for diagnosing active Crohn’s disease.

This study include eighty-six patients diagnosed with Crohn’s disease. Each patient underwent standard contrast-enhanced MRE protocol and three type of DWI sequences, conventional ssDW-EPI, SENSE(R=4) ssDW-EPI and msDW-EPI. Sixty-three patients were diagnosed with active bowel inflammation from contrast-enhanced MRE, which was regarded as the standard reference. The other 23 patients did not have radiological evidence of active inflammation on contrast-enhanced MRE and served as the control group. The raw data of SENSE(R=4) ssDW-EPI and msDW-EPI acquisition were transferred to a workstation for off-line image reconstruction. The reconstruction pipeline is identical to original MUSE framework, including Nyquist ghost correction and measurement of inter-shot phase variation. The SENSE and MUSE reconstructed images were exported to DICOM format for further analysis. For both patient groups, diagnostic evaluation of active bowel inflammation and evaluation of image quality were performed for conventional ssDW-EPI, SENSE(R=4) ssDW-EPI and msDW-EPI data. For assessment of the image quality, 1) image resolution, 2) degree of geometric distortion, 3) impact of artifacts on image quality and diagnostic evaluation, and 4) overall image quality were assessed with a 5-point Likert scale. Images were independently reviewed by two board-certified radiologists who were blinded to the type of DWI.
The results showed that image resolution and overall image quality were rated higher for MUSE msDW-EPI compared with conventional ssDW-EPI and SENSE(R=4) ssDW-EPI by both raters. For geometric distortion, artifacts influence on image quality and artifacts influence on diagnostic evaluation, both raters rated MUSE msDW-EPI better than ssDW-EPI and SENSE(R=4) ssDW-EPI.

The results of this study demonstrated that MUSE msDW-EPI produces superior image quality and resolution compared to conventional ssDW-EPI and SENSE(R=4) ssDW-EPI, and shows promise in improving the accuracy in diagnosing active bowel inflammation in Crohn’s disease.