Cardiometabolic biomarker prediction based on retinal fundus image

<p dir="ltr">Diagnosing common noncommunicable diseases, such as cardiovascular disease and diabetes, typically relies on blood sample analysis for biomarker measurement. This process is invasive, time-consuming, and relatively expensive. To address these limitations, deep learning m...

Full description

Saved in:
Bibliographic Details
Main Author: Syed Abdullah Basit (18021787) (author)
Other Authors: Hamada R.H. Al-Absi (22928605) (author), Saleh Musleh (15279190) (author), Tanvir Alam (638619) (author)
Published: 2025
Subjects:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1864513525556183040
author Syed Abdullah Basit (18021787)
author2 Hamada R.H. Al-Absi (22928605)
Saleh Musleh (15279190)
Tanvir Alam (638619)
author2_role author
author
author
author_facet Syed Abdullah Basit (18021787)
Hamada R.H. Al-Absi (22928605)
Saleh Musleh (15279190)
Tanvir Alam (638619)
author_role author
dc.creator.none.fl_str_mv Syed Abdullah Basit (18021787)
Hamada R.H. Al-Absi (22928605)
Saleh Musleh (15279190)
Tanvir Alam (638619)
dc.date.none.fl_str_mv 2025-07-25T15:00:00Z
dc.identifier.none.fl_str_mv 10.1016/j.engappai.2025.111734
dc.relation.none.fl_str_mv https://figshare.com/articles/journal_contribution/Cardiometabolic_biomarker_prediction_based_on_retinal_fundus_image/30971449
dc.rights.none.fl_str_mv CC BY 4.0
info:eu-repo/semantics/openAccess
dc.subject.none.fl_str_mv Biomedical and clinical sciences
Cardiovascular medicine and haematology
Engineering
Biomedical engineering
Health sciences
Health services and systems
Information and computing sciences
Machine learning
Retinal fundus imaging
Deep learning
Cardiometabolic biomarkers
Non-invasive diagnostics
Qatar Biobank
Artificial intelligence
Cardiovascular disease
Diabetes
dc.title.none.fl_str_mv Cardiometabolic biomarker prediction based on retinal fundus image
dc.type.none.fl_str_mv Text
Journal contribution
info:eu-repo/semantics/publishedVersion
text
contribution to journal
description <p dir="ltr">Diagnosing common noncommunicable diseases, such as cardiovascular disease and diabetes, typically relies on blood sample analysis for biomarker measurement. This process is invasive, time-consuming, and relatively expensive. To address these limitations, deep learning methods were leveraged to estimate common cardiometabolic biomarkers using retinal fundus (RF) images. The study utilized 15,802 RF images from 5,653 participants in the Qatar Biobank (QBB), leading to the development of 19 deep-learning models to estimate biomarkers across seven categories: demographics and body composition, blood pressure, lipid profile, blood profile, hormones, kidney function, and metabolites. The proposed model outperformed existing models for the QBB-specific cohort across all biomarkers, achieving higher R-squared (R<sup>2</sup>) values, lower mean absolute error (MAE), and higher area under the curve (AUC). The proposed model achieved excellent performance in demographic predictions with age (MAE: 2.56, R<sup>2</sup> : 0.93) and gender (Accuracy: 96%, AUC: 0.94). For cardiovascular markers, it showed moderate predictability with systolic blood pressure (MAE: 8.02, R<sup>2</sup> : 0.49) and diastolic blood pressure (MAE: 6.06, R<sup>2</sup> : 0.45). For metabolic markers, the model demonstrated varying performance, with hemoglobin showing strong prediction (MAE: 0.79, R<sup>2</sup> : 0.60) while lipid markers showed moderate performance (total cholesterol MAE: 0.63, R<sup>2</sup> : 0.29). For creatinine, a kidney function marker, we achieved the best results with MAE: 9.00, R<sup>2</sup> : 0.33. Stratified analyses revealed systematic performance variations across gender, age, and disease-specific subgroups, with better predictions in males, young-agers, and non-diabetic participants. External validation of the CAD group confirms the effect of age, gender, and disease on prediction results, suggesting the need for personalized background in consideration for developing AI models. This study presents a promising approach for non-invasive biomarker estimation using retinal images, potentially revolutionizing early intervention and treatment planning in healthcare.</p><h2 dir="ltr">Other Information</h2><p dir="ltr">Published in: Engineering Applications of Artificial Intelligence<br>License: <a href="http://creativecommons.org/licenses/by/4.0/" target="_blank">http://creativecommons.org/licenses/by/4.0/</a><br>See article on publisher's website: <a href="https://dx.doi.org/10.1016/j.engappai.2025.111734" target="_blank">https://dx.doi.org/10.1016/j.engappai.2025.111734</a></p>
eu_rights_str_mv openAccess
id Manara2_de149d6245ee2f4a294f8c172d4f7206
identifier_str_mv 10.1016/j.engappai.2025.111734
network_acronym_str Manara2
network_name_str Manara2
oai_identifier_str oai:figshare.com:article/30971449
publishDate 2025
repository.mail.fl_str_mv
repository.name.fl_str_mv
repository_id_str
rights_invalid_str_mv CC BY 4.0
spelling Cardiometabolic biomarker prediction based on retinal fundus imageSyed Abdullah Basit (18021787)Hamada R.H. Al-Absi (22928605)Saleh Musleh (15279190)Tanvir Alam (638619)Biomedical and clinical sciencesCardiovascular medicine and haematologyEngineeringBiomedical engineeringHealth sciencesHealth services and systemsInformation and computing sciencesMachine learningRetinal fundus imagingDeep learningCardiometabolic biomarkersNon-invasive diagnosticsQatar BiobankArtificial intelligenceCardiovascular diseaseDiabetes<p dir="ltr">Diagnosing common noncommunicable diseases, such as cardiovascular disease and diabetes, typically relies on blood sample analysis for biomarker measurement. This process is invasive, time-consuming, and relatively expensive. To address these limitations, deep learning methods were leveraged to estimate common cardiometabolic biomarkers using retinal fundus (RF) images. The study utilized 15,802 RF images from 5,653 participants in the Qatar Biobank (QBB), leading to the development of 19 deep-learning models to estimate biomarkers across seven categories: demographics and body composition, blood pressure, lipid profile, blood profile, hormones, kidney function, and metabolites. The proposed model outperformed existing models for the QBB-specific cohort across all biomarkers, achieving higher R-squared (R<sup>2</sup>) values, lower mean absolute error (MAE), and higher area under the curve (AUC). The proposed model achieved excellent performance in demographic predictions with age (MAE: 2.56, R<sup>2</sup> : 0.93) and gender (Accuracy: 96%, AUC: 0.94). For cardiovascular markers, it showed moderate predictability with systolic blood pressure (MAE: 8.02, R<sup>2</sup> : 0.49) and diastolic blood pressure (MAE: 6.06, R<sup>2</sup> : 0.45). For metabolic markers, the model demonstrated varying performance, with hemoglobin showing strong prediction (MAE: 0.79, R<sup>2</sup> : 0.60) while lipid markers showed moderate performance (total cholesterol MAE: 0.63, R<sup>2</sup> : 0.29). For creatinine, a kidney function marker, we achieved the best results with MAE: 9.00, R<sup>2</sup> : 0.33. Stratified analyses revealed systematic performance variations across gender, age, and disease-specific subgroups, with better predictions in males, young-agers, and non-diabetic participants. External validation of the CAD group confirms the effect of age, gender, and disease on prediction results, suggesting the need for personalized background in consideration for developing AI models. This study presents a promising approach for non-invasive biomarker estimation using retinal images, potentially revolutionizing early intervention and treatment planning in healthcare.</p><h2 dir="ltr">Other Information</h2><p dir="ltr">Published in: Engineering Applications of Artificial Intelligence<br>License: <a href="http://creativecommons.org/licenses/by/4.0/" target="_blank">http://creativecommons.org/licenses/by/4.0/</a><br>See article on publisher's website: <a href="https://dx.doi.org/10.1016/j.engappai.2025.111734" target="_blank">https://dx.doi.org/10.1016/j.engappai.2025.111734</a></p>2025-07-25T15:00:00ZTextJournal contributioninfo:eu-repo/semantics/publishedVersiontextcontribution to journal10.1016/j.engappai.2025.111734https://figshare.com/articles/journal_contribution/Cardiometabolic_biomarker_prediction_based_on_retinal_fundus_image/30971449CC BY 4.0info:eu-repo/semantics/openAccessoai:figshare.com:article/309714492025-07-25T15:00:00Z
spellingShingle Cardiometabolic biomarker prediction based on retinal fundus image
Syed Abdullah Basit (18021787)
Biomedical and clinical sciences
Cardiovascular medicine and haematology
Engineering
Biomedical engineering
Health sciences
Health services and systems
Information and computing sciences
Machine learning
Retinal fundus imaging
Deep learning
Cardiometabolic biomarkers
Non-invasive diagnostics
Qatar Biobank
Artificial intelligence
Cardiovascular disease
Diabetes
status_str publishedVersion
title Cardiometabolic biomarker prediction based on retinal fundus image
title_full Cardiometabolic biomarker prediction based on retinal fundus image
title_fullStr Cardiometabolic biomarker prediction based on retinal fundus image
title_full_unstemmed Cardiometabolic biomarker prediction based on retinal fundus image
title_short Cardiometabolic biomarker prediction based on retinal fundus image
title_sort Cardiometabolic biomarker prediction based on retinal fundus image
topic Biomedical and clinical sciences
Cardiovascular medicine and haematology
Engineering
Biomedical engineering
Health sciences
Health services and systems
Information and computing sciences
Machine learning
Retinal fundus imaging
Deep learning
Cardiometabolic biomarkers
Non-invasive diagnostics
Qatar Biobank
Artificial intelligence
Cardiovascular disease
Diabetes