MITEA: A dataset for machine learning segmentation of the left ventricle in 3D echocardiography using subject-specific labels from cardiac magnetic resonance imaging
Segmentation of the left ventricle (LV) in echocardiography is an important task for the quantification of volume and mass in heart disease. Continuing advances in echocardiography have extended imaging capabilities into the 3D domain, subsequently overcoming the geometric assumptions associated wit...
| Main Authors: | , , , , , , , , , , , , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Frontiers Media S.A.
2023-01-01
|
| Series: | Frontiers in Cardiovascular Medicine |
| Subjects: | |
| Online Access: | https://www.frontiersin.org/articles/10.3389/fcvm.2022.1016703/full |
| _version_ | 1797956457889529856 |
|---|---|
| author | Debbie Zhao Edward Ferdian Gonzalo D. Maso Talou Gina M. Quill Kathleen Gilbert Vicky Y. Wang Thiranja P. Babarenda Gamage João Pedrosa Jan D’hooge Timothy M. Sutton Boris S. Lowe Malcolm E. Legget Peter N. Ruygrok Peter N. Ruygrok Robert N. Doughty Robert N. Doughty Oscar Camara Alistair A. Young Alistair A. Young Martyn P. Nash Martyn P. Nash |
| author_facet | Debbie Zhao Edward Ferdian Gonzalo D. Maso Talou Gina M. Quill Kathleen Gilbert Vicky Y. Wang Thiranja P. Babarenda Gamage João Pedrosa Jan D’hooge Timothy M. Sutton Boris S. Lowe Malcolm E. Legget Peter N. Ruygrok Peter N. Ruygrok Robert N. Doughty Robert N. Doughty Oscar Camara Alistair A. Young Alistair A. Young Martyn P. Nash Martyn P. Nash |
| author_sort | Debbie Zhao |
| collection | DOAJ |
| description | Segmentation of the left ventricle (LV) in echocardiography is an important task for the quantification of volume and mass in heart disease. Continuing advances in echocardiography have extended imaging capabilities into the 3D domain, subsequently overcoming the geometric assumptions associated with conventional 2D acquisitions. Nevertheless, the analysis of 3D echocardiography (3DE) poses several challenges associated with limited spatial resolution, poor contrast-to-noise ratio, complex noise characteristics, and image anisotropy. To develop automated methods for 3DE analysis, a sufficiently large, labeled dataset is typically required. However, ground truth segmentations have historically been difficult to obtain due to the high inter-observer variability associated with manual analysis. We address this lack of expert consensus by registering labels derived from higher-resolution subject-specific cardiac magnetic resonance (CMR) images, producing 536 annotated 3DE images from 143 human subjects (10 of which were excluded). This heterogeneous population consists of healthy controls and patients with cardiac disease, across a range of demographics. To demonstrate the utility of such a dataset, a state-of-the-art, self-configuring deep learning network for semantic segmentation was employed for automated 3DE analysis. Using the proposed dataset for training, the network produced measurement biases of −9 ± 16 ml, −1 ± 10 ml, −2 ± 5 %, and 5 ± 23 g, for end-diastolic volume, end-systolic volume, ejection fraction, and mass, respectively, outperforming an expert human observer in terms of accuracy as well as scan-rescan reproducibility. As part of the Cardiac Atlas Project, we present here a large, publicly available 3DE dataset with ground truth labels that leverage the higher resolution and contrast of CMR, to provide a new benchmark for automated 3DE analysis. Such an approach not only reduces the effect of observer-specific bias present in manual 3DE annotations, but also enables the development of analysis techniques which exhibit better agreement with CMR compared to conventional methods. This represents an important step for enabling more efficient and accurate diagnostic and prognostic information to be obtained from echocardiography. |
| first_indexed | 2024-04-10T23:49:33Z |
| format | Article |
| id | doaj.art-c4cc2de0fc184bd4970484049d9348e2 |
| institution | Directory Open Access Journal |
| issn | 2297-055X |
| language | English |
| last_indexed | 2024-04-10T23:49:33Z |
| publishDate | 2023-01-01 |
| publisher | Frontiers Media S.A. |
| record_format | Article |
| series | Frontiers in Cardiovascular Medicine |
| spelling | doaj.art-c4cc2de0fc184bd4970484049d9348e22023-01-10T21:01:38ZengFrontiers Media S.A.Frontiers in Cardiovascular Medicine2297-055X2023-01-01910.3389/fcvm.2022.10167031016703MITEA: A dataset for machine learning segmentation of the left ventricle in 3D echocardiography using subject-specific labels from cardiac magnetic resonance imagingDebbie Zhao0Edward Ferdian1Gonzalo D. Maso Talou2Gina M. Quill3Kathleen Gilbert4Vicky Y. Wang5Thiranja P. Babarenda Gamage6João Pedrosa7Jan D’hooge8Timothy M. Sutton9Boris S. Lowe10Malcolm E. Legget11Peter N. Ruygrok12Peter N. Ruygrok13Robert N. Doughty14Robert N. Doughty15Oscar Camara16Alistair A. Young17Alistair A. Young18Martyn P. Nash19Martyn P. Nash20Auckland Bioengineering Institute, University of Auckland, Auckland, New ZealandDepartment of Anatomy and Medical Imaging, University of Auckland, Auckland, New ZealandAuckland Bioengineering Institute, University of Auckland, Auckland, New ZealandAuckland Bioengineering Institute, University of Auckland, Auckland, New ZealandAuckland Bioengineering Institute, University of Auckland, Auckland, New ZealandAuckland Bioengineering Institute, University of Auckland, Auckland, New ZealandAuckland Bioengineering Institute, University of Auckland, Auckland, New ZealandInstitute for Systems and Computer Engineering, Technology and Science (INESC TEC), Porto, PortugalDepartment of Cardiovascular Sciences, KU Leuven, Leuven, BelgiumCounties Manukau Health Cardiology, Middlemore Hospital, Auckland, New ZealandGreen Lane Cardiovascular Service, Auckland City Hospital, Auckland, New ZealandDepartment of Medicine, University of Auckland, Auckland, New ZealandGreen Lane Cardiovascular Service, Auckland City Hospital, Auckland, New ZealandDepartment of Medicine, University of Auckland, Auckland, New ZealandGreen Lane Cardiovascular Service, Auckland City Hospital, Auckland, New ZealandDepartment of Medicine, University of Auckland, Auckland, New ZealandDepartment of Information and Communication Technologies, Universitat Pompeu Fabra, Barcelona, SpainDepartment of Anatomy and Medical Imaging, University of Auckland, Auckland, New ZealandDepartment of Biomedical Engineering, King’s College London, London, United KingdomAuckland Bioengineering Institute, University of Auckland, Auckland, New Zealand0Department of Engineering Science, University of Auckland, Auckland, New ZealandSegmentation of the left ventricle (LV) in echocardiography is an important task for the quantification of volume and mass in heart disease. Continuing advances in echocardiography have extended imaging capabilities into the 3D domain, subsequently overcoming the geometric assumptions associated with conventional 2D acquisitions. Nevertheless, the analysis of 3D echocardiography (3DE) poses several challenges associated with limited spatial resolution, poor contrast-to-noise ratio, complex noise characteristics, and image anisotropy. To develop automated methods for 3DE analysis, a sufficiently large, labeled dataset is typically required. However, ground truth segmentations have historically been difficult to obtain due to the high inter-observer variability associated with manual analysis. We address this lack of expert consensus by registering labels derived from higher-resolution subject-specific cardiac magnetic resonance (CMR) images, producing 536 annotated 3DE images from 143 human subjects (10 of which were excluded). This heterogeneous population consists of healthy controls and patients with cardiac disease, across a range of demographics. To demonstrate the utility of such a dataset, a state-of-the-art, self-configuring deep learning network for semantic segmentation was employed for automated 3DE analysis. Using the proposed dataset for training, the network produced measurement biases of −9 ± 16 ml, −1 ± 10 ml, −2 ± 5 %, and 5 ± 23 g, for end-diastolic volume, end-systolic volume, ejection fraction, and mass, respectively, outperforming an expert human observer in terms of accuracy as well as scan-rescan reproducibility. As part of the Cardiac Atlas Project, we present here a large, publicly available 3DE dataset with ground truth labels that leverage the higher resolution and contrast of CMR, to provide a new benchmark for automated 3DE analysis. Such an approach not only reduces the effect of observer-specific bias present in manual 3DE annotations, but also enables the development of analysis techniques which exhibit better agreement with CMR compared to conventional methods. This represents an important step for enabling more efficient and accurate diagnostic and prognostic information to be obtained from echocardiography.https://www.frontiersin.org/articles/10.3389/fcvm.2022.1016703/full3D echocardiography (3DE)machine learning (ML)segmentation (image processing)left ventricle (LV)multimodal imagingcardiac magnetic resonance (CMR) imaging |
| spellingShingle | Debbie Zhao Edward Ferdian Gonzalo D. Maso Talou Gina M. Quill Kathleen Gilbert Vicky Y. Wang Thiranja P. Babarenda Gamage João Pedrosa Jan D’hooge Timothy M. Sutton Boris S. Lowe Malcolm E. Legget Peter N. Ruygrok Peter N. Ruygrok Robert N. Doughty Robert N. Doughty Oscar Camara Alistair A. Young Alistair A. Young Martyn P. Nash Martyn P. Nash MITEA: A dataset for machine learning segmentation of the left ventricle in 3D echocardiography using subject-specific labels from cardiac magnetic resonance imaging Frontiers in Cardiovascular Medicine 3D echocardiography (3DE) machine learning (ML) segmentation (image processing) left ventricle (LV) multimodal imaging cardiac magnetic resonance (CMR) imaging |
| title | MITEA: A dataset for machine learning segmentation of the left ventricle in 3D echocardiography using subject-specific labels from cardiac magnetic resonance imaging |
| title_full | MITEA: A dataset for machine learning segmentation of the left ventricle in 3D echocardiography using subject-specific labels from cardiac magnetic resonance imaging |
| title_fullStr | MITEA: A dataset for machine learning segmentation of the left ventricle in 3D echocardiography using subject-specific labels from cardiac magnetic resonance imaging |
| title_full_unstemmed | MITEA: A dataset for machine learning segmentation of the left ventricle in 3D echocardiography using subject-specific labels from cardiac magnetic resonance imaging |
| title_short | MITEA: A dataset for machine learning segmentation of the left ventricle in 3D echocardiography using subject-specific labels from cardiac magnetic resonance imaging |
| title_sort | mitea a dataset for machine learning segmentation of the left ventricle in 3d echocardiography using subject specific labels from cardiac magnetic resonance imaging |
| topic | 3D echocardiography (3DE) machine learning (ML) segmentation (image processing) left ventricle (LV) multimodal imaging cardiac magnetic resonance (CMR) imaging |
| url | https://www.frontiersin.org/articles/10.3389/fcvm.2022.1016703/full |
| work_keys_str_mv | AT debbiezhao miteaadatasetformachinelearningsegmentationoftheleftventriclein3dechocardiographyusingsubjectspecificlabelsfromcardiacmagneticresonanceimaging AT edwardferdian miteaadatasetformachinelearningsegmentationoftheleftventriclein3dechocardiographyusingsubjectspecificlabelsfromcardiacmagneticresonanceimaging AT gonzalodmasotalou miteaadatasetformachinelearningsegmentationoftheleftventriclein3dechocardiographyusingsubjectspecificlabelsfromcardiacmagneticresonanceimaging AT ginamquill miteaadatasetformachinelearningsegmentationoftheleftventriclein3dechocardiographyusingsubjectspecificlabelsfromcardiacmagneticresonanceimaging AT kathleengilbert miteaadatasetformachinelearningsegmentationoftheleftventriclein3dechocardiographyusingsubjectspecificlabelsfromcardiacmagneticresonanceimaging AT vickyywang miteaadatasetformachinelearningsegmentationoftheleftventriclein3dechocardiographyusingsubjectspecificlabelsfromcardiacmagneticresonanceimaging AT thiranjapbabarendagamage miteaadatasetformachinelearningsegmentationoftheleftventriclein3dechocardiographyusingsubjectspecificlabelsfromcardiacmagneticresonanceimaging AT joaopedrosa miteaadatasetformachinelearningsegmentationoftheleftventriclein3dechocardiographyusingsubjectspecificlabelsfromcardiacmagneticresonanceimaging AT jandhooge miteaadatasetformachinelearningsegmentationoftheleftventriclein3dechocardiographyusingsubjectspecificlabelsfromcardiacmagneticresonanceimaging AT timothymsutton miteaadatasetformachinelearningsegmentationoftheleftventriclein3dechocardiographyusingsubjectspecificlabelsfromcardiacmagneticresonanceimaging AT borisslowe miteaadatasetformachinelearningsegmentationoftheleftventriclein3dechocardiographyusingsubjectspecificlabelsfromcardiacmagneticresonanceimaging AT malcolmelegget miteaadatasetformachinelearningsegmentationoftheleftventriclein3dechocardiographyusingsubjectspecificlabelsfromcardiacmagneticresonanceimaging AT peternruygrok miteaadatasetformachinelearningsegmentationoftheleftventriclein3dechocardiographyusingsubjectspecificlabelsfromcardiacmagneticresonanceimaging AT peternruygrok miteaadatasetformachinelearningsegmentationoftheleftventriclein3dechocardiographyusingsubjectspecificlabelsfromcardiacmagneticresonanceimaging AT robertndoughty miteaadatasetformachinelearningsegmentationoftheleftventriclein3dechocardiographyusingsubjectspecificlabelsfromcardiacmagneticresonanceimaging AT robertndoughty miteaadatasetformachinelearningsegmentationoftheleftventriclein3dechocardiographyusingsubjectspecificlabelsfromcardiacmagneticresonanceimaging AT oscarcamara miteaadatasetformachinelearningsegmentationoftheleftventriclein3dechocardiographyusingsubjectspecificlabelsfromcardiacmagneticresonanceimaging AT alistairayoung miteaadatasetformachinelearningsegmentationoftheleftventriclein3dechocardiographyusingsubjectspecificlabelsfromcardiacmagneticresonanceimaging AT alistairayoung miteaadatasetformachinelearningsegmentationoftheleftventriclein3dechocardiographyusingsubjectspecificlabelsfromcardiacmagneticresonanceimaging AT martynpnash miteaadatasetformachinelearningsegmentationoftheleftventriclein3dechocardiographyusingsubjectspecificlabelsfromcardiacmagneticresonanceimaging AT martynpnash miteaadatasetformachinelearningsegmentationoftheleftventriclein3dechocardiographyusingsubjectspecificlabelsfromcardiacmagneticresonanceimaging |