Read our publications

MR-self Noise2Noise: self-supervised deep learning–based image quality improvement of submillimeter resolution 3D MR images

Authors

   


  • Woojin Jung(AIRS Medical, Seoul, Republic of Korea)  & Hyun-Soo Lee(Siemens Healthineers Ltd, Seoul, Republic of Korea)& Minkook Seo(Department of Radiology, Seoul St. Mary’s Hospital, College of Medicine, The Catholic University of Korea, Seoul 06591, Republic of Korea) & Yoonho Nam(Division of Biomedical Engineering, Hankuk University of Foreign Studies, Yongin, Gyeonggi-do, Republic of Korea) & Yangsean Choi(Department of Radiology, Seoul St. Mary’s Hospital, College of Medicine, The Catholic University of Korea, Seoul 06591, Republic of Korea) & Na-Young Shin(Department of Radiology, Seoul St. Mary’s Hospital, College of Medicine, The Catholic University of Korea, Seoul 06591, Republic of Korea) & Kook-Jin Ahn(Department of Radiology, Seoul St. Mary’s Hospital, College of Medicine, The Catholic University of Korea, Seoul 06591, Republic of Korea)&Bum-soo Kim(Department of Radiology, Seoul St. Mary’s Hospital, College of Medicine, The Catholic University of Korea, Seoul 06591, Republic of Korea) & Jinhee Jang(Department of Radiology, Seoul St. Mary’s Hospital, College of Medicine, The Catholic University of Korea, Seoul 06591, Republic of Korea)




European radiology (2022)


Abstract

Objectives

The study aimed to develop a deep neural network (DNN)–based noise reduction and image quality improvement by only using routine clinical scans and evaluate its performance in 3D high-resolution MRI.

Methods

This retrospective study included T1-weighted magnetization-prepared rapid gradient-echo (MP-RAGE) images from 185 clinical scans: 135 for DNN training, 11 for DNN validation, 20 for qualitative evaluation, and 19 for quantitative evaluation. Additionally, 18 vessel wall imaging (VWI) data were included to evaluate generalization. In each scan of the DNN training set, two noise-independent images were generated from the k-space data, resulting in an input-label pair. 2.5D U-net architecture was utilized for the DNN model. Qualitative evaluation between conventional MP-RAGE and DNN-based MP-RAGE was performed by two radiologists in image quality, fine structure delineation, and lesion conspicuity. Quantitative evaluation was performed with full sampled data as a reference by measuring quantitative error metrics and volumetry at 7 different simulated noise levels. DNN application on VWI was evaluated by two radiologists in image quality.

Results

Our DNN-based MP-RAGE outperformed conventional MP-RAGE in all image quality parameters (average scores = 3.7 vs. 4.9, p < 0.001). In the quantitative evaluation, DNN showed better error metrics (p < 0.001) and comparable (p > 0.09) or better (p < 0.02) volumetry results than conventional MP-RAGE. DNN application to VWI also revealed improved image quality (3.5 vs. 4.6, p < 0.001).

Conclusions

The proposed DNN model successfully denoises 3D MR image and improves its image quality by using routine clinical scans only.

Key Points

  • Our deep learning framework successfully improved conventional 3D high-resolution MRI in all image quality parameters, fine structure delineation, and lesion conspicuity.
  • Compared to conventional MRI, the proposed deep neural network-based MRI revealed better quantitative error metrics and comparable or better volumetry results.

  • Deep neural network application to 3D MRI whose pulse sequences and parameters were different from the training data showed improvement in image quality, revealing the potential to generalize on various clinical MRI.



Read More


AIRS Medical Inc.


13,14F, Keungil Tower, 223, Teheran-ro, Gangnam-gu, Seoul, 06142, Republic of Korea

E-mail: support@airsmed.com

Tel: (+82) 70 7777 3186


Copyright ⓒ 2022 AIRS Medical Inc. All rights reserved.