Automation of a Rule-based Workflow to Estimate Age from Brain MR Imaging of Infants and Children Up to 2 Years Old Using Stacked Deep Learning
-
- Wada Akihiko
- Department of Radiology, Juntendo University School of Medicine
-
- Saito Yuya
- Department of Radiology, Juntendo University School of Medicine
-
- Fujita Shohei
- Department of Radiology, Juntendo University School of Medicine
-
- Irie Ryusuke
- Department of Radiology, Juntendo University School of Medicine
-
- Akashi Toshiaki
- Department of Radiology, Juntendo University School of Medicine
-
- Sano Katsuhiro
- Department of Radiology, Juntendo University School of Medicine
-
- Kato Shinpei
- Department of Radiology, Juntendo University School of Medicine
-
- Ikenouchi Yutaka
- Department of Radiology, Juntendo University School of Medicine
-
- Hagiwara Akifumi
- Department of Radiology, Juntendo University School of Medicine
-
- Sato Kanako
- Department of Radiology, Juntendo University School of Medicine
-
- Tomizawa Nobuo
- Department of Radiology, Juntendo University School of Medicine
-
- Hayakawa Yayoi
- Department of Radiology, Juntendo University School of Medicine
-
- Kikuta Junko
- Department of Radiology, Juntendo University School of Medicine
-
- Kamagata Koji
- Department of Radiology, Juntendo University School of Medicine
-
- Suzuki Michimasa
- Department of Radiology, Juntendo University School of Medicine
-
- Hori Masaaki
- Department of Radiology, Juntendo University School of Medicine
-
- Nakanishi Atsushi
- Department of Radiology, Juntendo University School of Medicine
-
- Aoki Shigeki
- Department of Radiology, Juntendo University School of Medicine
この論文をさがす
抄録
<p>Purpose: Myelination-related MR signal changes in white matter are helpful for assessing normal development in infants and children. A rule-based myelination evaluation workflow regarding signal changes on T1-weighted images (T1WIs) and T2-weighted images (T2WIs) has been widely used in radiology. This study aimed to simulate a rule-based workflow using a stacked deep learning model and evaluate age estimation accuracy.</p><p>Methods: The age estimation system involved two stacked neural networks: a target network-to extract five myelination-related images from the whole brain, and an age estimation network from extracted T1- and T2WIs separately. A dataset was constructed from 119 children aged below 2 years with two MRI systems. A four-fold cross-validation method was adopted. The correlation coefficient (CC), mean absolute error (MAE), and root mean squared error (RMSE) of the corrected chronological age of full-term birth, as well as the mean difference and the upper and lower limits of 95% agreement, were measured. Generalization performance was assessed using datasets acquired from different MR images. Age estimation was performed in Sturge–Weber syndrome (SWS) cases.</p><p>Results: There was a strong correlation between estimated age and corrected chronological age (MAE: 0.98 months; RMSE: 1.27 months; and CC: 0.99). The mean difference and standard deviation (SD) were −0.15 and 1.26, respectively, and the upper and lower limits of 95% agreement were 2.33 and −2.63 months. Regarding generalization performance, the performance values on the external dataset were MAE of 1.85 months, RMSE of 2.59 months, and CC of 0.93. Among 13 SWS cases, 7 exceeded the limits of 95% agreement, and a proportional bias of age estimation based on myelination acceleration was exhibited below 12 months of age (P = 0.03).</p><p>Conclusion: Stacked deep learning models automated the rule-based workflow in radiology and achieved highly accurate age estimation in infants and children up to 2 years of age.</p>
収録刊行物
-
- Magnetic Resonance in Medical Sciences
-
Magnetic Resonance in Medical Sciences 22 (1), 57-66, 2023
日本磁気共鳴医学会