2021
DOI: 10.1161/circimaging.120.011951
|View full text |Cite
|
Sign up to set email alerts
|

Automated Left Ventricular Dimension Assessment Using Artificial Intelligence Developed and Validated by a UK-Wide Collaborative

Abstract: Background: Artificial intelligence (AI) for echocardiography requires training and validation to standards expected of humans. We developed an online platform and established the Unity Collaborative to build a dataset of expertise from 17 hospitals for training, validation, and standardization of such techniques. Methods: The training dataset consisted of 2056 individual frames drawn at random from 1265 parasternal long-axis video-loops of patients und… Show more

Help me understand this report

Search citation statements

Order By: Relevance

Paper Sections

Select...
2
2

Citation Types

0
26
0

Year Published

2021
2021
2023
2023

Publication Types

Select...
5
1
1

Relationship

1
6

Authors

Journals

citations
Cited by 20 publications
(26 citation statements)
references
References 18 publications
0
26
0
Order By: Relevance
“… 32 , 33 With fine-tuning on site-specific data, our model compares favorably with prior state-of-the-art approaches to assessing ventricular wall thickness and hypertrophy on open benchmarks. 16 Expanding on previous work, 21 we collaborated with stakeholders across Stanford Medicine to release our data set of 12 000 deidentified PLAX echocardiogram videos as a resource for the medical machine learning community for future comparison and validation of deep learning models. This expands the prior data set of 10 030 apical 4-chamber videos 21 to a total of 22 030 echocardiogram videos made publicly available, which to our knowledge, is the largest data set release of labeled medical videos with matched clinician annotations.…”
Section: Discussionmentioning
confidence: 99%
See 2 more Smart Citations
“… 32 , 33 With fine-tuning on site-specific data, our model compares favorably with prior state-of-the-art approaches to assessing ventricular wall thickness and hypertrophy on open benchmarks. 16 Expanding on previous work, 21 we collaborated with stakeholders across Stanford Medicine to release our data set of 12 000 deidentified PLAX echocardiogram videos as a resource for the medical machine learning community for future comparison and validation of deep learning models. This expands the prior data set of 10 030 apical 4-chamber videos 21 to a total of 22 030 echocardiogram videos made publicly available, which to our knowledge, is the largest data set release of labeled medical videos with matched clinician annotations.…”
Section: Discussionmentioning
confidence: 99%
“…Nevertheless, quantification of ventricular thickness remains subject to significant intra-and inter-provider variability across imaging modalities 8,9 . Even with the high image resolution and signal-to-noise ratio of cardiac magnetic resonance imaging, there is significant test-retest variability due to the laborious, manual nature of wall thickness measurement 10,11 .…”
Section: Introductionmentioning
confidence: 99%
See 1 more Smart Citation
“…Moreover, with an expanding data set we will be able to train quality control CNNs to recognize specific types of quality and planning errors. In order to decrease subjectivity of this task, a collaborative initiative to build a consensus across a vast number of operators, similar to a recent one developed in the field of echo and AI ( 25 ), would be of great value. Lastly, our method now requires post-processing and is separated from the CMR scanner.…”
Section: Discussionmentioning
confidence: 99%
“…Deep learning (DL) allows automated contour detection offering the promise of faster and potentially more accurate and reproducible evaluation of LV volumes and EF by echocardiography (3,7,13,14). In this work, we developed a new DL algorithm for manual contouring based on a U-Net convolutional network architecture, using an anonymised database of echocardiographic images.…”
Section: Introductionmentioning
confidence: 99%