Korean J Radiol.  2025 Jan;26(1):29-42. 10.3348/kjr.2023.1303.

Faster Acquisition and Improved Image Quality of T2-Weighted Dixon Breast MRI at 3T Using Deep Learning: A Prospective Study

• Wilpert C
• Schneider H
• Rau A
• Russe MF
• Oerther B
• Strecker R
• Nickel MD
• Weiland E
• Haeger A
• Benndorf M
• Mayrhofer T
• Weiss J
• Bamberg F
• Windfuhr-Blum M
• Neubauer J

• KMID: 2563376
• DOI: http://doi.org/10.3348/kjr.2023.1303

Abstract

Objective
The aim of this study was to compare image quality features and lesion characteristics between a faster deep learning (DL) reconstructed T2-weighted (T2-w) fast spin-echo (FSE) Dixon sequence with super-resolution (T2_DL) and a conventional T2-w FSE Dixon sequence (T2_STD) for breast magnetic resonance imaging (MRI).
Materials and Methods
This prospective study was conducted between November 2022 and April 2023 using a 3T scanner. Both T2_DL and T2_STD sequences were acquired for each patient. Quantitative analysis was based on region-of-interest (ROI) measurements of signal-to-noise ratio (SNR) and contrast-to-noise ratio (CNR). Qualitative analysis was performed independently by two radiologists using Likert scales to evaluate various image quality features, morphology, and diagnostic confidence for cysts and breast cancers. Reader preference between T2_DL and T2_STD was assessed via side-by-side comparison, and inter-reader reliability was also analyzed.
Results
Total of 151 women were enrolled, with 140 women (mean age: 52 ± 14 years; 85 cysts and 31 breast cancers) included in the final analysis. The acquisition time was 110 s ± 0 for T2_DL compared to 266 s ± 0 for T2_STD. SNR and CNR were significantly higher in T2_DL (P < 0.001). T2_DL was associated with higher image quality scores, reduced noise, and fewer artifacts (P < 0.001). All evaluated anatomical regions (breast and axilla), breast implants, and bone margins were rated higher in T2_DL (P ≤ 0.008), except for bone marrow, which scored higher in T2_STD (P < 0.001). Scores for conspicuity, sharpness/ margins, and microstructure of cysts and breast cancers were higher in T2_DL (P ≤ 0.002). Diagnostic confidence for cysts was improved with T2_DL (P < 0.001). Readers significantly preferred T2_DL over T2_STD in side-by-side comparisons (P < 0.001).
Conclusion
T2_DL effectively corrected for SNR loss caused by accelerated image acquisition and provided a 58% reduction in acquisition time compared to T2_STD. This led to fewer artifacts and improved overall image quality. Thus, T2_DL is feasible and has the potential to replace conventional T2-w sequences for breast MRI examinations.

Keyword

MRI methods; Artificial neural network/machine learning; Breast; Quality assurance/control and improvement; Cancer