Impact of super-resolution deep learning-based reconstruction for hippocampal MRI: A volunteer and phantom study

Sentaro Takada, Takeshi Nakaura, Naofumi Yoshida, Hiroyuki Uetani, Kaori Shiraishi, Naoki Kobayashi, Kensei Matsuo, Kosuke Morita, Yasunori Nagayama, Masafumi Kidoh, Yuichi Yamashita, Ryohei Takayanagi, Toshinori Hirai,
Impact of super-resolution deep learning-based reconstruction for hippocampal MRI: A volunteer and phantom study,
European Journal of Radiology,
Volume 191,
2025,
112289,
ISSN 0720-048X,
https://doi.org/10.1016/j.ejrad.2025.112289.
(https://www.sciencedirect.com/science/article/pii/S0720048X25003754)
Abstract: Background and Purpose
To evaluate the effects of super-resolution deep learning-based reconstruction (SR-DLR) on thin-slice T2-weighted hippocampal MR image quality using 3 T MRI, in both human volunteers and phantoms.
Materials and Methods
Thirteen healthy volunteers underwent hippocampal MRI at standard and high resolutions. Original (standard-resolution; StR) images were reconstructed with and without deep learning-based reconstruction (DLR) (Matrix = 320 × 320), and with SR-DLR (Matrix = 960 × 960). High-resolution (HR) images were also reconstructed with/without DLR (Matrix = 960 × 960). Contrast, contrast-to-noise ratio (CNR), and septum slope were analyzed. Two radiologists evaluated the images for noise, contrast, artifacts, sharpness, and overall quality. Quantitative and qualitative results are reported as medians and interquartile ranges (IQR). Comparisons used the Wilcoxon signed-rank test with Holm correction. We also scanned an American College of Radiology (ACR) phantom to evaluate the ability of our SR-DLR approach to reduce artifacts induced by zero-padding interpolation (ZIP).
Results
SR-DLR exhibited contrast comparable to original images and significantly higher than HR-images. Its slope was comparable to that of HR images but was significantly steeper than that of StR images (p < 0.01). Furthermore, the CNR of SR-DLR (10.53; IQR: 10.08, 11.69) was significantly superior to the StR-images without DLR (7.5; IQR: 6.4, 8.37), StR-images with DLR (8.73; IQR: 7.68, 9.0), HR-images without DLR (2.24; IQR: 1.43, 2.38), and HR-images with DLR (4.84; IQR: 2.99, 5.43) (p < 0.05). In the phantom study, artifacts induced by ZIP were scarcely observed when using SR-DLR.
Conclusion
SR-DLR for hippocampal MRI potentially improves image quality beyond that of actual HR-images while reducing acquisition time.
Keywords: Image reconstruction; Magnetic resonance imaging; Super-resolution; Deep learning; Hippocampus