1 The Extreme Cardiac MRI Analysis Challenge under Respiratory Motion CMRxMotion

2025-04-27 0 0 374.29KB 4 页 10玖币

侵权投诉

The Extreme Cardiac MRI Analysis Challenge

under Respiratory Motion (CMRxMotion)

Shuo Wang1,2*, Chen Qin3†, Chengyan Wang4‡, Kang Wang1,2, Haoran Wang1,2, Chen Chen5, Cheng Ouyang5,

Xutong Kuang5, Chengliang Dai5, Yuanhan Mo6, Zhang Shi7, Chenchen Dai7, Xinrong Chen1,2, He Wang8, and

Wenjia Bai4

1Digitial Medical Research Center, Fudan Univeristy, China

2Shanghai Key Laboratory of MICCAI, China

3Department of Electrical and Electronic Engineering, Imperial College London, UK

4Human Phenome Institute, Fudan University, China

5Department of Computing, Imperial College London, UK

6School of Engineering, University of Oxford, UK

7Department of Radiology, Zhongshan Hospital Afﬁliated to Fudan University, China

8Institute of Science and Technology for Brain-inspired Intelligence, Fudan University, China

(Summary of CMRxMotion Challenge Design)

Abstract—The quality of cardiac magnetic resonance (CMR)

imaging is susceptible to respiratory motion artifacts. The model

robustness of automated segmentation techniques in face of real-

world respiratory motion artifacts is unclear. This manuscript

describes the design of extreme cardiac MRI analysis challenge

under respiratory motion (CMRxMotion Challenge). The chal-

lenge aims to establish a public benchmark dataset to assess the

effects of respiratory motion on image quality and examine the

robustness of segmentation models. The challenge recruited 40

healthy volunteers to perform different breath-hold behaviors

during one imaging visit, obtaining paired cine imaging with

artifacts. Radiologists assessed the image quality and annotated

the level of respiratory motion artifacts. For those images with

diagnostic quality, radiologists further segmented the left ventri-

cle, left ventricle myocardium and right ventricle. The images

of training set (20 volunteers) along with the annotations are

released to the challenge participants, to develop an automated

image quality assessment model (Task 1) and an automated

segmentation model (Task 2). The images of validation set (5

volunteers) are released to the challenge participants but the

annotations are withheld for online evaluation of submitted

predictions. Both the images and annotations of the test set (15

volunteers) were withheld and only used for ofﬂine evaluation of

submitted containerized dockers. The image quality assessment

task is quantitatively evaluated by the Cohen’s kappa statistics

and the segmentation task is evaluated by the Dice scores and

Hausdorff distances.

Index Terms—Cardiac magnetic resonance, image quality

assessment, image segmentation, respiratory motion artifacts,

model robustness.

I. INTRODUCTION

Cardiac magnetic resonance (CMR) imaging is the cur-

rent gold-standard modality for evaluating cardiac structure

and function [1]. Machine learning-based approaches have

* Task 1 contact: shuowang@fudan.edu.cn

†Task 2 contact: c.qin15@imperial.ac.uk

‡CMR acquisition contact: wangcy@fudan.edu.cn

achieved remarkable performance in the CMR image seg-

mentation task [2], [3]. However, the model performance is

still challenged by inconsistent imaging environments (e.g.,

vendors and protocols) [4], population shifts (normal v.s.

pathological cases) [5] and unexpected human behaviors (e.g.,

body movements) in clinical practice. It is useful to investigate

potential failure modes [6] by exposing a trained segmentation

model to extreme cases in a ‘stress test’. To date, existing

challenges focus on the vendor variability [4] and anatomical

structure variations [5] while the implications of human be-

haviors are less explored. For CMR image analysis, respiration

motion is one of the major problems [7]. Patients may not

be able to follow the breath-hold instructions well, which is

common in heart failure cases or children. The poor breath-

hold behaviors result in degraded image quality and inaccurate

analysis [8]. Therefore, we launch the extreme cardiac MRI

analysis challenge under respiratory motion (CMRxMotion

Challenge) 1, to establish a public benchmark dataset to assess

the effects of respiratory motion on CMR imaging quality and

examine the robustness [9] of automated segmentation models.

To curate such an extreme CMR dataset with respiratory

motion artifacts, one retrospective way is to screen the images

stored in hospital database and identify those problematic ones.

But this requires considerable human efforts and may also

bring about confounding factors such as vendors, scan protocol

and pathology. Instead, we design a prospective study that

health volunteers are recruited to perform different breath-

hold behaviors during one imaging visit. As the confounding

factors of MRI equipment and scan protocols are controlled,

this extreme CMR dataset is established in speciﬁc to respi-

ratory motion artifacts. The rest of the manuscript provides

a summary of CMRxMotion challenge design, including the

data acquisition, annotation, tasks, evaluation metric, ranking

1http://cmr.miccai.cloud

arXiv:2210.06385v1 [eess.IV] 12 Oct 2022

文档加载中……请稍候！
如果长时间未打开，您也可以点击刷新试试。

下载文档到电脑，查找使用更方便

10 玖币 0人已下载

立即下载

摘要：

1TheExtremeCardiacMRIAnalysisChallengeunderRespiratoryMotion(CMRxMotion)ShuoWang1,2*,ChenQin3y,ChengyanWang4z,KangWang1,2,HaoranWang1,2,ChenChen5,ChengOuyang5,XutongKuang5,ChengliangDai5,YuanhanMo6,ZhangShi7,ChenchenDai7,XinrongChen1,2,HeWang8,andWenjiaBai41DigitialMedicalResearchCenter,FudanUniveri...

展开>> 收起<<

1 The Extreme Cardiac MRI Analysis Challenge under Respiratory Motion CMRxMotion.pdf

共4页,预览1页

还剩页未读，继续阅读

声明：本站为文档C2C交易模式，即用户上传的文档直接被用户下载，本站只是中间服务平台，本站所有文档下载所得的收益归上传人(含作者)所有。玖贝云文库仅提供信息存储空间，仅对用户上传内容的表现方式做保护处理，对上载内容本身不做任何修改或编辑。若文档所含内容侵犯了您的版权或隐私，请立即通知玖贝云文库，我们立即给予删除！

1 The Extreme Cardiac MRI Analysis Challenge under Respiratory Motion CMRxMotion

相关推荐

开通VIP享超值会员特权

作者详情

相关内容

热门标签

举报选择: