Zhihao Xue, Sicheng Zhu, Fan Yang, Juan Gao, Hao Peng, Chao Zou, Hang Jin, Chenxi Hu
{"title":"用于高度加速三维冠状动脉磁共振血管造影的深度图像先验和压缩传感混合重建方法","authors":"Zhihao Xue, Sicheng Zhu, Fan Yang, Juan Gao, Hao Peng, Chao Zou, Hang Jin, Chenxi Hu","doi":"10.3389/fcvm.2024.1408351","DOIUrl":null,"url":null,"abstract":"IntroductionHigh-resolution whole-heart coronary magnetic resonance angiography (CMRA) often suffers from unreasonably long scan times, rendering imaging acceleration highly desirable. Traditional reconstruction methods used in CMRA rely on either hand-crafted priors or supervised learning models. Although the latter often yield superior reconstruction quality, they require a large amount of training data and memory resources, and may encounter generalization issues when dealing with out-of-distribution datasets.MethodsTo address these challenges, we introduce an unsupervised reconstruction method that combines deep image prior (DIP) with compressed sensing (CS) to accelerate 3D CMRA. This method incorporates a slice-by-slice DIP reconstruction and 3D total variation (TV) regularization, enabling high-quality reconstruction under a significant acceleration while enforcing continuity in the slice direction. We evaluated our method by comparing it to iterative SENSE, CS-TV, CS-wavelet, and other DIP-based variants, using both retrospectively and prospectively undersampled datasets.ResultsThe results demonstrate the superiority of our 3D DIP-CS approach, which improved the reconstruction accuracy relative to the other approaches across both datasets. Ablation studies further reveal the benefits of combining DIP with 3D TV regularization, which leads to significant improvements of image quality over pure DIP-based methods. Evaluation of vessel sharpness and image quality scores shows that DIP-CS improves the quality of reformatted coronary arteries.DiscussionThe proposed method enables scan-specific reconstruction of high-quality 3D CMRA from a five-minute acquisition, without relying on fully-sampled training data or placing a heavy burden on memory resources.","PeriodicalId":12414,"journal":{"name":"Frontiers in Cardiovascular Medicine","volume":null,"pages":null},"PeriodicalIF":2.8000,"publicationDate":"2024-09-12","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":"{\"title\":\"A hybrid deep image prior and compressed sensing reconstruction method for highly accelerated 3D coronary magnetic resonance angiography\",\"authors\":\"Zhihao Xue, Sicheng Zhu, Fan Yang, Juan Gao, Hao Peng, Chao Zou, Hang Jin, Chenxi Hu\",\"doi\":\"10.3389/fcvm.2024.1408351\",\"DOIUrl\":null,\"url\":null,\"abstract\":\"IntroductionHigh-resolution whole-heart coronary magnetic resonance angiography (CMRA) often suffers from unreasonably long scan times, rendering imaging acceleration highly desirable. Traditional reconstruction methods used in CMRA rely on either hand-crafted priors or supervised learning models. Although the latter often yield superior reconstruction quality, they require a large amount of training data and memory resources, and may encounter generalization issues when dealing with out-of-distribution datasets.MethodsTo address these challenges, we introduce an unsupervised reconstruction method that combines deep image prior (DIP) with compressed sensing (CS) to accelerate 3D CMRA. This method incorporates a slice-by-slice DIP reconstruction and 3D total variation (TV) regularization, enabling high-quality reconstruction under a significant acceleration while enforcing continuity in the slice direction. We evaluated our method by comparing it to iterative SENSE, CS-TV, CS-wavelet, and other DIP-based variants, using both retrospectively and prospectively undersampled datasets.ResultsThe results demonstrate the superiority of our 3D DIP-CS approach, which improved the reconstruction accuracy relative to the other approaches across both datasets. Ablation studies further reveal the benefits of combining DIP with 3D TV regularization, which leads to significant improvements of image quality over pure DIP-based methods. Evaluation of vessel sharpness and image quality scores shows that DIP-CS improves the quality of reformatted coronary arteries.DiscussionThe proposed method enables scan-specific reconstruction of high-quality 3D CMRA from a five-minute acquisition, without relying on fully-sampled training data or placing a heavy burden on memory resources.\",\"PeriodicalId\":12414,\"journal\":{\"name\":\"Frontiers in Cardiovascular Medicine\",\"volume\":null,\"pages\":null},\"PeriodicalIF\":2.8000,\"publicationDate\":\"2024-09-12\",\"publicationTypes\":\"Journal Article\",\"fieldsOfStudy\":null,\"isOpenAccess\":false,\"openAccessPdf\":\"\",\"citationCount\":\"0\",\"resultStr\":null,\"platform\":\"Semanticscholar\",\"paperid\":null,\"PeriodicalName\":\"Frontiers in Cardiovascular Medicine\",\"FirstCategoryId\":\"3\",\"ListUrlMain\":\"https://doi.org/10.3389/fcvm.2024.1408351\",\"RegionNum\":3,\"RegionCategory\":\"医学\",\"ArticlePicture\":[],\"TitleCN\":null,\"AbstractTextCN\":null,\"PMCID\":null,\"EPubDate\":\"\",\"PubModel\":\"\",\"JCR\":\"Q2\",\"JCRName\":\"CARDIAC & CARDIOVASCULAR SYSTEMS\",\"Score\":null,\"Total\":0}","platform":"Semanticscholar","paperid":null,"PeriodicalName":"Frontiers in Cardiovascular Medicine","FirstCategoryId":"3","ListUrlMain":"https://doi.org/10.3389/fcvm.2024.1408351","RegionNum":3,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"Q2","JCRName":"CARDIAC & CARDIOVASCULAR SYSTEMS","Score":null,"Total":0}
引用次数: 0
摘要
导言高分辨率全心冠状动脉磁共振血管造影术(CMRA)往往存在扫描时间过长的问题,因此非常需要成像加速。用于 CMRA 的传统重建方法依赖于手工创建的先验或监督学习模型。为了应对这些挑战,我们引入了一种无监督重建方法,将深度图像先验(DIP)与压缩传感(CS)相结合,以加速 3D CMRA。该方法结合了逐片 DIP 重建和三维总变异(TV)正则化,在显著加速的情况下实现了高质量重建,同时确保了切片方向的连续性。我们使用回顾性和前瞻性欠采样数据集,将我们的方法与迭代 SENSE、CS-TV、CS-小波和其他基于 DIP 的变体进行了比较,从而对我们的方法进行了评估。消融研究进一步揭示了将 DIP 与 3D TV 正则化相结合的优势,与纯 DIP 方法相比,这种方法显著提高了图像质量。对血管锐利度和图像质量评分的评估表明,DIP-CS 提高了重新格式化冠状动脉的质量。 讨论所提出的方法可以在不依赖全采样训练数据或对内存资源造成沉重负担的情况下,从五分钟的采集数据中重建特定扫描的高质量 3D CMRA。
A hybrid deep image prior and compressed sensing reconstruction method for highly accelerated 3D coronary magnetic resonance angiography
IntroductionHigh-resolution whole-heart coronary magnetic resonance angiography (CMRA) often suffers from unreasonably long scan times, rendering imaging acceleration highly desirable. Traditional reconstruction methods used in CMRA rely on either hand-crafted priors or supervised learning models. Although the latter often yield superior reconstruction quality, they require a large amount of training data and memory resources, and may encounter generalization issues when dealing with out-of-distribution datasets.MethodsTo address these challenges, we introduce an unsupervised reconstruction method that combines deep image prior (DIP) with compressed sensing (CS) to accelerate 3D CMRA. This method incorporates a slice-by-slice DIP reconstruction and 3D total variation (TV) regularization, enabling high-quality reconstruction under a significant acceleration while enforcing continuity in the slice direction. We evaluated our method by comparing it to iterative SENSE, CS-TV, CS-wavelet, and other DIP-based variants, using both retrospectively and prospectively undersampled datasets.ResultsThe results demonstrate the superiority of our 3D DIP-CS approach, which improved the reconstruction accuracy relative to the other approaches across both datasets. Ablation studies further reveal the benefits of combining DIP with 3D TV regularization, which leads to significant improvements of image quality over pure DIP-based methods. Evaluation of vessel sharpness and image quality scores shows that DIP-CS improves the quality of reformatted coronary arteries.DiscussionThe proposed method enables scan-specific reconstruction of high-quality 3D CMRA from a five-minute acquisition, without relying on fully-sampled training data or placing a heavy burden on memory resources.
期刊介绍:
Frontiers? Which frontiers? Where exactly are the frontiers of cardiovascular medicine? And who should be defining these frontiers?
At Frontiers in Cardiovascular Medicine we believe it is worth being curious to foresee and explore beyond the current frontiers. In other words, we would like, through the articles published by our community journal Frontiers in Cardiovascular Medicine, to anticipate the future of cardiovascular medicine, and thus better prevent cardiovascular disorders and improve therapeutic options and outcomes of our patients.