Accelerated EPR imaging using deep learning denoising

IF 3 3区医学 Q2 RADIOLOGY, NUCLEAR MEDICINE & MEDICAL IMAGING

Magnetic Resonance in Medicine Pub Date : 2025-03-17 DOI:10.1002/mrm.30473

Irene Canavesi, Navin Viswakarma, Boris Epel, Alan McMillan, Mrignayani Kotecha

{"title":"Accelerated EPR imaging using deep learning denoising","authors":"Irene Canavesi, Navin Viswakarma, Boris Epel, Alan McMillan, Mrignayani Kotecha","doi":"10.1002/mrm.30473","DOIUrl":null,"url":null,"abstract":"<div>\n \n \n <section>\n \n <h3> Purpose</h3>\n \n <p>Trityl OXO71-based pulse electron paramagnetic resonance imaging (EPRI) is an excellent technique to obtain partial pressure of oxygen (pO<sub>2</sub>) maps in tissues. In this study, we used deep learning techniques to denoise 3D EPR amplitude and pO<sub>2</sub> maps.</p>\n </section>\n \n <section>\n \n <h3> Methods</h3>\n \n <p>All experiments were performed using a 25 mT EPR imager, JIVA-25®. The MONAI implementation of four neural networks (autoencoder, Attention UNet, UNETR, and UNet) was tested, and the best model (UNet) was then enhanced with joint bilateral filters (JBF). The training dataset was comprised of 227 3D images (56 in vivo and 171 in vitro), 159 images for training, 45 for validation, and 23 for testing. UNet with 1, 2, and 3 JBF layers was tested to improve image SNR, focusing on multiscale structural similarity index measure and edge sensitivity preservation. The trained algorithm was tested using acquisitions with 15, 30, and 150 averages in vitro with a sealed deoxygenated OXO71 phantom and in vivo with fibrosarcoma tumors grown in a hind leg of C3H mice.</p>\n </section>\n \n <section>\n \n <h3> Results</h3>\n \n <p>We demonstrate that UNet with 2 JBF layers (UNet+JBF2) provides the best outcome. We demonstrate that using the UNet+JBF2 model, the SNR of 15-shot amplitude maps provides higher SNR compared to 150-shot pre-filter maps, both in phantoms and in tumors, therefore, allowing 10-fold accelerated imaging. We demonstrate that the trained algorithm improves SNR in pO<sub>2</sub> maps.</p>\n </section>\n \n <section>\n \n <h3> Conclusions</h3>\n \n <p>We demonstrate the application of deep learning techniques to EPRI denoising. Higher SNR will bring the EPRI technique one step closer to clinics.</p>\n </section>\n </div>","PeriodicalId":18065,"journal":{"name":"Magnetic Resonance in Medicine","volume":"94 1","pages":"436-446"},"PeriodicalIF":3.0000,"publicationDate":"2025-03-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"Magnetic Resonance in Medicine","FirstCategoryId":"3","ListUrlMain":"https://onlinelibrary.wiley.com/doi/10.1002/mrm.30473","RegionNum":3,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"Q2","JCRName":"RADIOLOGY, NUCLEAR MEDICINE & MEDICAL IMAGING","Score":null,"Total":0}

引用次数: 0

Abstract

Purpose

Trityl OXO71-based pulse electron paramagnetic resonance imaging (EPRI) is an excellent technique to obtain partial pressure of oxygen (pO₂) maps in tissues. In this study, we used deep learning techniques to denoise 3D EPR amplitude and pO₂ maps.

Methods

All experiments were performed using a 25 mT EPR imager, JIVA-25®. The MONAI implementation of four neural networks (autoencoder, Attention UNet, UNETR, and UNet) was tested, and the best model (UNet) was then enhanced with joint bilateral filters (JBF). The training dataset was comprised of 227 3D images (56 in vivo and 171 in vitro), 159 images for training, 45 for validation, and 23 for testing. UNet with 1, 2, and 3 JBF layers was tested to improve image SNR, focusing on multiscale structural similarity index measure and edge sensitivity preservation. The trained algorithm was tested using acquisitions with 15, 30, and 150 averages in vitro with a sealed deoxygenated OXO71 phantom and in vivo with fibrosarcoma tumors grown in a hind leg of C3H mice.

Results

We demonstrate that UNet with 2 JBF layers (UNet+JBF2) provides the best outcome. We demonstrate that using the UNet+JBF2 model, the SNR of 15-shot amplitude maps provides higher SNR compared to 150-shot pre-filter maps, both in phantoms and in tumors, therefore, allowing 10-fold accelerated imaging. We demonstrate that the trained algorithm improves SNR in pO₂ maps.

Conclusions

We demonstrate the application of deep learning techniques to EPRI denoising. Higher SNR will bring the EPRI technique one step closer to clinics.

查看原文本刊更多论文

使用深度学习去噪加速EPR成像。

目的：基于三氧基氧71的脉冲电子顺磁共振成像（EPRI）是一种获得组织中氧分压（pO2）图的优秀技术。在这项研究中，我们使用深度学习技术对3D EPR振幅和pO2图进行降噪。方法：所有实验均使用25mt EPR成像仪JIVA-25®进行。对四种神经网络（自动编码器、注意力UNet、UNETR和UNet）的MONAI实现进行了测试，然后用联合双边滤波器（JBF）增强了最佳模型（UNet）。训练数据集包括227张3D图像（56张在体内，171张在体外），159张用于训练，45张用于验证，23张用于测试。为了提高图像的信噪比，对具有1、2和3层JBF的UNet进行了测试，重点关注多尺度结构相似度指标的度量和边缘灵敏度的保持。训练后的算法在体外用密封的脱氧OXO71模型和体内用C3H小鼠后腿生长的纤维肉瘤肿瘤分别获得15、30和150个平均值进行测试。结果：我们证明具有2个JBF层的UNet （UNet+JBF2）提供了最好的结果。我们证明，使用UNet+JBF2模型，无论是在幻影还是肿瘤中，15次振幅图的信噪比都比150次预滤波图提供更高的信噪比，因此可以实现10倍的加速成像。我们证明了训练后的算法提高了pO2映射的信噪比。结论：我们展示了深度学习技术在EPRI去噪中的应用。更高的信噪比将使EPRI技术更接近临床。

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分钟内获得全文求助全文

来源期刊

Magnetic Resonance in Medicine 医学-核医学

CiteScore

6.70

自引率

24.20%

发文量

376

审稿时长

2-4 weeks

期刊介绍： Magnetic Resonance in Medicine (Magn Reson Med) is an international journal devoted to the publication of original investigations concerned with all aspects of the development and use of nuclear magnetic resonance and electron paramagnetic resonance techniques for medical applications. Reports of original investigations in the areas of mathematics, computing, engineering, physics, biophysics, chemistry, biochemistry, and physiology directly relevant to magnetic resonance will be accepted, as well as methodology-oriented clinical studies.