Automated assessment of transthoracic echocardiogram image quality using deep neural networks

Labs, Robert B., Vrettos, Apostolos, Loo, Jonathan and Zolgharni, Massoud ORCID: https://orcid.org/0000-0003-0904-2904 (2022) Automated assessment of transthoracic echocardiogram image quality using deep neural networks. Intelligent Medicine. ISSN 2667-1026

[thumbnail of 1-s2.0-S2667102622000705-main.pdf]
Preview
PDF
1-s2.0-S2667102622000705-main.pdf - Accepted Version
Available under License Creative Commons Attribution Non-commercial No Derivatives.

Download (1MB) | Preview

Abstract

Background
Standard views in two-dimensional echocardiography are well established but the quality of acquired images are highly dependent on operator skills and are assessed subjectively. This study is aimed at providing an objective assessment pipeline for echocardiogram image quality by defining a new set of domain-specific quality indicators. Consequently, image quality assessment can thus be automated to enhance clinical measurements, interpretation, and real-time optimization.

Methods
We have developed deep neural networks for the automated assessment of echocardiographic frame which were randomly sampled from 11,262 adult patients. The private echocardiography dataset consists of 33,784 frames, previously acquired between 2010 and 2020. Unlike non-medical images where full-reference metrics can be applied for image quality, echocardiogram's data is highly heterogeneous and requires blind-reference (IQA) metrics. Therefore, deep learning approaches were used to extract the spatiotemporal features and the image's quality indicators were evaluated against the mean absolute error. Our quality indicators encapsulate both anatomical and pathological elements to provide multivariate assessment scores for anatomical visibility, clarity, depth-gain and foreshortedness, respectively.

Results
The model performance accuracy yielded 94.4%, 96.8%, 96.2%, 97.4% for anatomical visibility, clarity, depth-gain and foreshortedness, respectively. The mean model error of 0.375±0.0052 with computational speed of 2.52 ms per frame (real-time performance) was achieved.

Conclusion
The novel approach offers new insight to objective assessment of transthoracic echocardiogram image quality and clinical quantification in A4C and PLAX views. Also lays stronger foundations for operator's guidance system which can leverage the learning curve for the acquisition of optimum quality images during transthoracic exam.

Item Type: Article
Identifier: 10.1016/j.imed.2022.08.001
Keywords: Image quality, Echocardiography, Objective assessment, Deep learning, Ultrasound
Subjects: Computing > Intelligent systems
Related URLs:
Depositing User: Massoud Zolgharni
Date Deposited: 17 Nov 2022 15:42
Last Modified: 06 Feb 2024 16:12
URI: https://repository.uwl.ac.uk/id/eprint/9396

Downloads

Downloads per month over past year

Actions (login required)

View Item View Item

Menu