Synthesis of MR fingerprinting information from magnitude-only MR imaging data using a parallelized, multi network U-Net convolutional neural network.

Frontiers in radiology Pub Date : 2024-12-16 eCollection Date: 2024-01-01 DOI:10.3389/fradi.2024.1498411
Kiaran P McGee, Yi Sui, Robert J Witte, Ananya Panda, Norbert G Campeau, Thomaz R Mostardeiro, Nahil Sobh, Umberto Ravaioli, Shuyue Lucia Zhang, Kianoush Falahkheirkhah, Nicholas B Larson, Christopher G Schwarz, Jeffrey L Gunter
{"title":"Synthesis of MR fingerprinting information from magnitude-only MR imaging data using a parallelized, multi network U-Net convolutional neural network.","authors":"Kiaran P McGee, Yi Sui, Robert J Witte, Ananya Panda, Norbert G Campeau, Thomaz R Mostardeiro, Nahil Sobh, Umberto Ravaioli, Shuyue Lucia Zhang, Kianoush Falahkheirkhah, Nicholas B Larson, Christopher G Schwarz, Jeffrey L Gunter","doi":"10.3389/fradi.2024.1498411","DOIUrl":null,"url":null,"abstract":"<p><strong>Background: </strong>MR fingerprinting (MRF) is a novel method for quantitative assessment of <i>in vivo</i> MR relaxometry that has shown high precision and accuracy. However, the method requires data acquisition using customized, complex acquisition strategies and dedicated post processing methods thereby limiting its widespread application.</p><p><strong>Objective: </strong>To develop a deep learning (DL) network for synthesizing MRF signals from conventional magnitude-only MR imaging data and to compare the results to the actual MRF signal acquired.</p><p><strong>Methods: </strong>A U-Net DL network was developed to synthesize MRF signals from magnitude-only 3D <i>T</i> <sub>1</sub>-weighted brain MRI data acquired from 37 volunteers aged between 21 and 62 years of age. Network performance was evaluated by comparison of the relaxometry data (<i>T</i> <sub>1</sub>, <i>T</i> <sub>2</sub>) generated from dictionary matching of the deep learning synthesized and actual MRF data from 47 segmented anatomic regions. Clustered bootstrapping involving 10,000 bootstraps followed by calculation of the concordance correlation coefficient were performed for both <i>T</i> <sub>1</sub> and <i>T</i> <sub>2</sub> MRF data pairs. 95% confidence limits and the mean difference between true and DL relaxometry values were also calculated.</p><p><strong>Results: </strong>The concordance correlation coefficient (and 95% confidence limits) for <i>T</i> <sub>1</sub> and <i>T</i> <sub>2</sub> MRF data pairs over the 47 anatomic segments were 0.8793 (0.8136-0.9383) and 0.9078 (0.8981-0.9145) respectively. The mean difference (and 95% confidence limits) were 48.23 (23.0-77.3) s and 2.02 (-1.4 to 4.8) s.</p><p><strong>Conclusion: </strong>It is possible to synthesize MRF signals from MRI data using a DL network, thereby creating the potential for performing quantitative relaxometry assessment without the need for a dedicated MRF pulse sequence.</p>","PeriodicalId":73101,"journal":{"name":"Frontiers in radiology","volume":"4 ","pages":"1498411"},"PeriodicalIF":0.0000,"publicationDate":"2024-12-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC11686891/pdf/","citationCount":"0","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"Frontiers in radiology","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.3389/fradi.2024.1498411","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"2024/1/1 0:00:00","PubModel":"eCollection","JCR":"","JCRName":"","Score":null,"Total":0}
引用次数: 0

Abstract

Background: MR fingerprinting (MRF) is a novel method for quantitative assessment of in vivo MR relaxometry that has shown high precision and accuracy. However, the method requires data acquisition using customized, complex acquisition strategies and dedicated post processing methods thereby limiting its widespread application.

Objective: To develop a deep learning (DL) network for synthesizing MRF signals from conventional magnitude-only MR imaging data and to compare the results to the actual MRF signal acquired.

Methods: A U-Net DL network was developed to synthesize MRF signals from magnitude-only 3D T 1-weighted brain MRI data acquired from 37 volunteers aged between 21 and 62 years of age. Network performance was evaluated by comparison of the relaxometry data (T 1, T 2) generated from dictionary matching of the deep learning synthesized and actual MRF data from 47 segmented anatomic regions. Clustered bootstrapping involving 10,000 bootstraps followed by calculation of the concordance correlation coefficient were performed for both T 1 and T 2 MRF data pairs. 95% confidence limits and the mean difference between true and DL relaxometry values were also calculated.

Results: The concordance correlation coefficient (and 95% confidence limits) for T 1 and T 2 MRF data pairs over the 47 anatomic segments were 0.8793 (0.8136-0.9383) and 0.9078 (0.8981-0.9145) respectively. The mean difference (and 95% confidence limits) were 48.23 (23.0-77.3) s and 2.02 (-1.4 to 4.8) s.

Conclusion: It is possible to synthesize MRF signals from MRI data using a DL network, thereby creating the potential for performing quantitative relaxometry assessment without the need for a dedicated MRF pulse sequence.

查看原文
分享 分享
微信好友 朋友圈 QQ好友 复制链接
本刊更多论文
利用并行化的多网络U-Net卷积神经网络从仅震级的MR成像数据中合成MR指纹信息。
背景:磁共振指纹是一种新的定量评估体内磁共振弛豫测量的方法,具有较高的精密度和准确性。然而,该方法需要使用定制的、复杂的采集策略和专用的后处理方法进行数据采集,从而限制了其广泛应用。目的:建立一个深度学习(DL)网络,用于从常规磁共振成像数据中合成磁共振信号,并将结果与实际获得的磁共振信号进行比较。方法:开发U-Net DL网络,从37名年龄在21岁至62岁之间的志愿者获得的仅三维t1加权脑MRI数据中合成MRF信号。通过将深度学习合成的字典匹配生成的松弛测量数据(t1, t2)与47个分割解剖区域的实际MRF数据进行比较,评估网络性能。对t1和t2 MRF数据对进行了涉及10,000个bootstrap的聚类bootstrap,然后计算了一致性相关系数。还计算了95%置信限和真实松弛测量值与DL松弛测量值之间的平均差值。结果:47个解剖节段的t1和t2 MRF数据对的一致性相关系数(及95%置信限)分别为0.8793(0.8136 ~ 0.9383)和0.9078(0.8981 ~ 0.9145)。平均差异(95%置信限)为48.23 (23.0 ~ 77.3)s和2.02 (-1.4 ~ 4.8)s。结论:使用DL网络从MRI数据合成MRF信号是可能的,从而创造了在不需要专用MRF脉冲序列的情况下进行定量松弛测量评估的潜力。
本文章由计算机程序翻译,如有差异,请以英文原文为准。
求助全文
约1分钟内获得全文 去求助
来源期刊
CiteScore
1.20
自引率
0.00%
发文量
0
期刊最新文献
Current state and promise of user-centered design to harness explainable AI in clinical decision-support systems for patients with CNS tumors. DreamOn: a data augmentation strategy to narrow the robustness gap between expert radiologists and deep learning classifiers. Editorial: Advances in artificial intelligence and machine learning applications for the imaging of bone and soft tissue tumors. Synthesis of MR fingerprinting information from magnitude-only MR imaging data using a parallelized, multi network U-Net convolutional neural network. Pneumothorax detection and segmentation from chest X-ray radiographs using a patch-based fully convolutional encoder-decoder network.
×
引用
GB/T 7714-2015
复制
MLA
复制
APA
复制
导出至
BibTeX EndNote RefMan NoteFirst NoteExpress
×
×
提示
您的信息不完整,为了账户安全,请先补充。
现在去补充
×
提示
您因"违规操作"
具体请查看互助需知
我知道了
×
提示
现在去查看 取消
×
提示
确定
0
微信
客服QQ
Book学术公众号 扫码关注我们
反馈
×
意见反馈
请填写您的意见或建议
请填写您的手机或邮箱
已复制链接
已复制链接
快去分享给好友吧!
我知道了
×
扫码分享
扫码分享
Book学术官方微信
Book学术文献互助
Book学术文献互助群
群 号:481959085
Book学术
文献互助 智能选刊 最新文献 互助须知 联系我们:info@booksci.cn
Book学术提供免费学术资源搜索服务,方便国内外学者检索中英文文献。致力于提供最便捷和优质的服务体验。
Copyright © 2023 Book学术 All rights reserved.
ghs 京公网安备 11010802042870号 京ICP备2023020795号-1