Related Experiment Video
Updated: Jul 12, 2026
15:48
Tracking the Mammary Architectural Features and Detecting Breast Cancer with Magnetic Resonance Diffusion Tensor Imaging
Published on: December 15, 2014
Impact of deep-learning image reconstruction on multiplexed sensitivity encoding diffusion-weighted imaging in the
Elaine Yuen Phin Lee1, Chia-Wei Li2, Grace Ho3
1Department of Diagnostic Radiology, Clinical School of Medicine, Li Ka Shing Faculty of Medicine, The University of Hong Kong, Hong Kong SAR, China.
Quantitative Imaging in Medicine and Surgery
|July 11, 2026
Summary
Deep-learning reconstruction (DLRecon) significantly enhances multiplexed sensitivity encoding (MUSE) diffusion-weighted imaging (DWI) in the female pelvis, improving image quality and lesion detection. This method rivals higher-shot sequences while maintaining accurate ADC quantification.
Area of Science:
- Medical Imaging
- Radiology
- Artificial Intelligence in Medicine
Background:
- Pelvic diffusion-weighted imaging (DWI) often suffers from artifacts, degrading image quality.
- Multiplexed sensitivity encoding (MUSE) DWI is a technique used for pelvic MRI.
- Deep learning reconstruction (DLRecon) is an emerging technology for improving MRI quality.
Purpose of the Study:
- To evaluate the impact of DLRecon on MUSE DWI image quality in the female pelvis.
- To compare the effectiveness of DLRecon with standard MUSE DWI acquisition.
- To assess the influence of DLRecon on lesion conspicuity and quantitative parameters.
Main Methods:
- Prospective recruitment of female patients undergoing pelvic MRI with 2-shot MUSE DWI.
- Qualitative assessment of image quality, artifacts, lesion conspicuity, and sharpness by radiologists.
- Quantitative analysis of signal-to-noise ratio (SNR), contrast-to-noise ratio (CNR), and apparent diffusion coefficient (ADC).
- Comparison of 2-shot MUSE DWI with and without DLRecon, and with 4-shot MUSE DWI.
Main Results:
- DLRecon significantly improved overall image quality, reduced artifacts, and enhanced lesion conspicuity and sharpness (P<0.001).
- Image quality with 2-shot MUSE DWI plus DLRecon was comparable to 4-shot MUSE DWI without DLRecon (P=0.079-0.225).
- DLRecon increased SNR and CNR (P<0.050) but preserved ADC quantification stability for normal tissues and lesions (P>0.1).
Conclusions:
- DLRecon effectively improves MUSE DWI image quality in the female pelvis by mitigating artifacts and enhancing image features.
- DLRecon offers a potential solution to achieve high-quality pelvic DWI comparable to longer acquisition times.
- DLRecon enhances diagnostic confidence by improving lesion visualization without compromising quantitative ADC measurements.