Construction of the breast DCE-MRI images deduction model based on deep learning

WANG Pingping; NIE Pin; WANG Lifang; DANG Yanli; ZHU Kaiguo; CHEN Baoying

doi:10.19732/j.cnki.2096-6210.2021.05.004

您当前的位置：

首页 >

文章列表页 >

Construction of the breast DCE-MRI images deduction model based on deep learning　

更新时间：2025-12-15

- Construction of the breast DCE-MRI images deduction model based on deep learning　
- Vol. 30, Issue 5, Pages: 339-344(2021)
- 作者机构：
  
  1. 西安国际医学中心医院临床实验中心,陕西,西安,710700
  2. 西安国际医学中心医院影像诊疗中心,陕西,西安,710700
- 作者简介：
- 基金信息：
- DOI：10.19732/j.cnki.2096-6210.2021.05.004
  CLC：
- Published Online：28 October 2021，
  
  Published：28 October 2021
- 稿件说明：
移动端阅览
王苹苹,聂品,王丽芳,等. 基于深度学习的乳腺DCE-MRI图像推演模型的构建[J]. 肿瘤影像学, 2021, 30(5): 339-344 https://doi.

org/10.19732/j.cnki.2096-6210.2021.05.004
王苹苹,聂品,王丽芳,等. 基于深度学习的乳腺DCE-MRI图像推演模型的构建[J]. 肿瘤影像学, 2021, 30(5): 339-344 https://doi. DOI： 10.19732/j.cnki.2096-6210.2021.05.004.

org/10.19732/j.cnki.2096-6210.2021.05.004 DOI：

摘要

目的：

探究利用深度学习算法构建乳腺动态对比增强磁共振成像（dynamic contrast-enhanced magnetic resonance imaging，DCE-MRI）图像推演模型的可行性，寻找无须实际扫描即可获取DCE-MRI图像的新技术。

方法：

收集2020年1月2021年8月在西安国际医学中心医院行乳腺MRI检查的54例女性患者的临床信息及影像学资料，随机分为训练集（38例）、测试集1（8例）及测试集2（8例）。基于训练集中患者的T1加权成像（T1-weighted imaging，T1WI）及DCE序列第一期图像，利用深度学习Pix2Pix算法构建乳腺DCE-MRI图像推演模型。在测试集中合成DCE图像，采用峰值信噪比（peak signal to noise ratio，PSNR）、结构相似性（structural similarity，SSIM）、均方误差（mean square error，MSE）、平均绝对误差（mean absolute error，MAE）指标测试模型性能。另外，评估合成的DCE序列与原始序列间的相关性。

结果：

共合成DCE第一期图像760张。与原始图像相比，合成的图像具有良好的SSIM，较低的信息损失，其中SSIM为0.710.01，PSNR为23.701.41。另外，合成的DCE序列重建误差较小，MAE为0.0320.004，MSE为0.0060.002，且与原始DCE序列呈显著正相关（

=0.8720.038，95% CI：0.870~0.874，

=0.000）。

结论：

乳腺DCE-MRI图像推演模型能够自动生成DCE序列第一期图像，为补充非增强MRI序列中信息缺失提供了新思路，同时避免了造影剂的使用、缩短扫描时间，为乳腺MRI筛查的推广应用奠定了基础。

Abstract

Objective:

To investigate the feasibility of building the breast dynamic contrast-enhanced magnetic resonance imaging (DCE-MRI) images deduction model by using deep learning algorithms

with the aim of obtaining the DCE-MRI sequences w

ithout actual scanning.

Methods:

Clinical data and MRI images of 54 subjects with breast MRI examinations were collected retrospectively in Xian International Medical Center Hospital from January 2020 to August 2021 and divided into a training set (38 subjects)

a test set 1 (8 subjects) and a test set 2 (8 subjects). Here

we built the breast DCE-MRI images deduction model using the Pix2Pix algorithm based on the T1-weighted imaging (T1WI) and the first post-contrast phase of DCE sequence images in the training set. Furthermore

the model performance for synthesizing images and sequences was evaluated in the test set using several metrics such as peak signal to noise ratio (PSNR)

structural similarity index measure (SSIM)

mean squared error (MSE)

and mean absolute error (MAE). In addition

the correlation between the synthesized DCE sequences and the original DCE sequences was evaluated.

Results:

In total

760 images of the first post-contrast phase of DCE were synthesized using the model. The synthesized images had good structural similarity and low information loss compared with the original images

where the mean SSIM was 0.710.01 and PSNR was 23.701.41. In addition

we found that the synthesized DCE sequences had small reconstruction errors

MAE was 0.0320.004

MSE was 0.0060.002

and the synthesized first post-contrast phase of DCE sequences had a significantly positive correlation with the original sequences

(

=0.8720.038

95% CI: 0.870-0.874

=0.000).

Conclusion:

The model which we built can automatically synthesize the first post-contrast phase of DCE sequence images

which provides a new idea for supplementing the missing information in non-enhanced sequences. At the same time

it avoids the application of contrast agents

shortens the scanning time

and lays a solid foundation for the popularization and application of breast MRI screening.

关键词

Keywords

references

Views

1581

下载量

CNKI被引量

Alert me when the article has been cited

提交

Tools

Publicity Resources

Automated classification of breast MRI background parenchymal enhancement using deep learning and thresholding segmentation method

Diagnostic value of simplified MRI sequence in breast diseases

The value of diffusion kurtosis imaging combined with peripheral blood inflammatory markers in predicting the pathological grading of clear cell renal cell carcinoma

A flexible segmentation network for MRI: toward accurate tumor delineation and 3D reconstruction in brain glioma

Intelligent imaging for precision diagnosis and treatment of pancreatic tumors: current applications and future perspectives

Related Author

MA Mingming

ZHANG Yaofeng

WANG Xiangpeng

ZHANG Xiaodong

QIN Naishan

WANG Xiaoying

HUANG Xing

LIANG Yan

Related Institution

Department of Radiology, Peking University First Hospital

Beijing Smart-Imaging Technology Co. Ltd.

Department of Radiology, Jilin Provincial People's Hospital

Department of Medical Imaging, North China University of Science and Technology Affiliated Hospital

Department of Cardiothoracic Surgery, KaiLuan General Hospital

AI问答

⁰