Sequence-Type Classification of Brain MRI for Acute Stroke Using a Self-Supervised Machine Learning Algorithm

Na, Seongwon; Ko, Yousun; Ham, Su Jung; Sung, Yu Sub; Kim, Mi-Hyun; Shin, Youngbin; Jung, Seung Chai; Ju, Chung; Kim, Byung Su; Yoon, Kyoungro; Kim, Kyung Won

doi:10.3390/diagnostics14010070

Open AccessArticle

Sequence-Type Classification of Brain MRI for Acute Stroke Using a Self-Supervised Machine Learning Algorithm

by

Seongwon Na

^1,2,†,

Yousun Ko

^3,†

,

Su Jung Ham

³,

Yu Sub Sung

^4,5

,

Mi-Hyun Kim

^6,7,

Youngbin Shin

²,

Seung Chai Jung

³,

Chung Ju

^8,9,

Byung Su Kim

⁸,

Kyoungro Yoon

^1,10,* and

Kyung Won Kim

^3,*

¹

Department of Computer Science and Engineering, Konkuk University, Seoul 05029, Republic of Korea

²

Biomedical Research Center, Asan Institute for Life Sciences, Asan Medical Center, Seoul 05505, Republic of Korea

³

Department of Radiology and Research Institute of Radiology, Asan Medical Center, University of Ulsan College of Medicine, Seoul 05505, Republic of Korea

⁴

Clinical Research Center, Asan Medical Center, Seoul 05505, Republic of Korea

⁵

Department of Convergence Medicine, University of Ulsan College of Medicine, Seoul 05505, Republic of Korea

⁶

Trialinformatics Inc., Seoul 05505, Republic of Korea

⁷

Department of Radiation Science & Technology, Jeonbuk National University, Jeonju 56212, Republic of Korea

⁸

Shin Poong Pharm. Co., Ltd., Seoul 06246, Republic of Korea

⁹

Graduate School of Clinical Pharmacy, CHA University, Pocheon-si 11160, Republic of Korea

¹⁰

Department of Smart ICT Convergence Engineering, Konkuk University, Seoul 05029, Republic of Korea

^*

Authors to whom correspondence should be addressed.

^†

These authors contributed equally to this work.

Diagnostics 2024, 14(1), 70; https://doi.org/10.3390/diagnostics14010070

Submission received: 21 November 2023 / Revised: 18 December 2023 / Accepted: 22 December 2023 / Published: 27 December 2023

(This article belongs to the Section Medical Imaging and Theranostics)

Download

Browse Figures

Versions Notes

Abstract

We propose a self-supervised machine learning (ML) algorithm for sequence-type classification of brain MRI using a supervisory signal from DICOM metadata (i.e., a rule-based virtual label). A total of 1787 brain MRI datasets were constructed, including 1531 from hospitals and 256 from multi-center trial datasets. The ground truth (GT) was generated by two experienced image analysts and checked by a radiologist. An ML framework called ImageSort-net was developed using various features related to MRI acquisition parameters and used for training virtual labels and ML algorithms derived from rule-based labeling systems that act as labels for supervised learning. For the performance evaluation of ImageSort-net (ML_virtual), we compare and analyze the performances of models trained with human expert labels (ML_humans), using as a test set blank data that the rule-based labeling system failed to infer from each dataset. The performance of ImageSort-net (ML_virtual) was comparable to that of ML_human (98.5% and 99%, respectively) in terms of overall accuracy when trained with hospital datasets. When trained with a relatively small multi-center trial dataset, the overall accuracy was relatively lower than that of ML_human (95.6% and 99.4%, respectively). After integrating the two datasets and re-training them, ML_virtual showed higher accuracy than ML_virtual trained only on multi-center datasets (95.6% and 99.7%, respectively). Additionally, the multi-center dataset inference performances after the re-training of ML_virtual and ML_humans were identical (99.7%). Training of ML algorithms based on rule-based virtual labels achieved high accuracy for sequence-type classification of brain MRI and enabled us to build a sustainable self-learning system.

Keywords: magnetic resonance image; machine learning; metadata

Share and Cite

MDPI and ACS Style

Na, S.; Ko, Y.; Ham, S.J.; Sung, Y.S.; Kim, M.-H.; Shin, Y.; Jung, S.C.; Ju, C.; Kim, B.S.; Yoon, K.; et al. Sequence-Type Classification of Brain MRI for Acute Stroke Using a Self-Supervised Machine Learning Algorithm. Diagnostics 2024, 14, 70. https://doi.org/10.3390/diagnostics14010070

AMA Style

Na S, Ko Y, Ham SJ, Sung YS, Kim M-H, Shin Y, Jung SC, Ju C, Kim BS, Yoon K, et al. Sequence-Type Classification of Brain MRI for Acute Stroke Using a Self-Supervised Machine Learning Algorithm. Diagnostics. 2024; 14(1):70. https://doi.org/10.3390/diagnostics14010070

Chicago/Turabian Style

Na, Seongwon, Yousun Ko, Su Jung Ham, Yu Sub Sung, Mi-Hyun Kim, Youngbin Shin, Seung Chai Jung, Chung Ju, Byung Su Kim, Kyoungro Yoon, and et al. 2024. "Sequence-Type Classification of Brain MRI for Acute Stroke Using a Self-Supervised Machine Learning Algorithm" Diagnostics 14, no. 1: 70. https://doi.org/10.3390/diagnostics14010070

APA Style

Na, S., Ko, Y., Ham, S. J., Sung, Y. S., Kim, M.-H., Shin, Y., Jung, S. C., Ju, C., Kim, B. S., Yoon, K., & Kim, K. W. (2024). Sequence-Type Classification of Brain MRI for Acute Stroke Using a Self-Supervised Machine Learning Algorithm. Diagnostics, 14(1), 70. https://doi.org/10.3390/diagnostics14010070

Note that from the first issue of 2016, this journal uses article numbers instead of page numbers. See further details here.

Article Menu

Sequence-Type Classification of Brain MRI for Acute Stroke Using a Self-Supervised Machine Learning Algorithm

Abstract

Share and Cite

Article Metrics

Article Access Statistics

Further Information

Guidelines

MDPI Initiatives

Follow MDPI