The Impact of Anatomic Racial Variations on Artificial Intelligence Analysis of Filipino Retinal Fundus Photographs Using an Image-Based Deep Learning Model
Objectives: This study evaluated the accuracy of an artificial intelligence (AI) model in identifying retinal lesions, validated its performance on a Filipino population dataset, and evaluated the impact of dataset diversity on AI analysis accuracy. Methods: This cross-sectional, analytical, inst...
Saved in:
| Main Authors: | , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Philippine Academy of Ophthalmology
2024-12-01
|
| Series: | Philippine Journal of Ophthalmology |
| Subjects: | |
| Online Access: | https://paojournal.com/index.php/pjo/article/view/522 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1850226227632144384 |
|---|---|
| author | Carlo A. Kasala, MD Kaye Lani Rea B. Locaylocay, MD-MBA Paolo S. Silva, MD |
| author_facet | Carlo A. Kasala, MD Kaye Lani Rea B. Locaylocay, MD-MBA Paolo S. Silva, MD |
| author_sort | Carlo A. Kasala, MD |
| collection | DOAJ |
| description | Objectives: This study evaluated the accuracy of an artificial intelligence (AI) model in identifying retinal lesions, validated its performance on a Filipino population dataset, and evaluated the impact of dataset diversity on AI analysis accuracy.
Methods: This cross-sectional, analytical, institutional study analyzed standardized macula-centered fundus photos taken with the Zeiss Visucam®. The AI model's output was compared with manual readings by trained retina specialists.
Results: A total of 215 eyes from 109 patients were included in the study. Human graders identified 109 eyes (50.7%) with retinal abnormalities. The AI model demonstrated an overall accuracy of 73.0% (95% CI 66.6% – 78.8%) in detecting abnormal retinas, with a sensitivity of 54.1% (95% CI 44.3% – 63.7%) and specificity of 92.5% (95% CI 85.7% – 96.7%).
Conclusions: The availability and sources of AI training datasets can introduce biases into AI algorithms. In our dataset, racial differences in retinal morphology, such as differences in retinal pigmentation, affected the accuracy of AI image-based analysis. More diverse datasets and external validation on different populations are needed to mitigate these biases. |
| format | Article |
| id | doaj-art-bfbeb93861bf430fbb5d0c5318f16aaa |
| institution | OA Journals |
| issn | 0031-7659 |
| language | English |
| publishDate | 2024-12-01 |
| publisher | Philippine Academy of Ophthalmology |
| record_format | Article |
| series | Philippine Journal of Ophthalmology |
| spelling | doaj-art-bfbeb93861bf430fbb5d0c5318f16aaa2025-08-20T02:05:07ZengPhilippine Academy of OphthalmologyPhilippine Journal of Ophthalmology0031-76592024-12-01492130137522The Impact of Anatomic Racial Variations on Artificial Intelligence Analysis of Filipino Retinal Fundus Photographs Using an Image-Based Deep Learning ModelCarlo A. Kasala, MD0Kaye Lani Rea B. Locaylocay, MD-MBA1Paolo S. Silva, MD2Eye and Vision Institute, The Medical City, Pasig City, PhilippinesEye and Vision Institute, The Medical City, Pasig City, PhilippinesEye and Vision Institute, The Medical City, Pasig City, Philippines; Beetham Eye Institue, Joslin Diabetes Center, Boston, Massachusetts, USA; Department of Ophthalmology, Harvard Medical School, Boston, Massachusetts, USAObjectives: This study evaluated the accuracy of an artificial intelligence (AI) model in identifying retinal lesions, validated its performance on a Filipino population dataset, and evaluated the impact of dataset diversity on AI analysis accuracy. Methods: This cross-sectional, analytical, institutional study analyzed standardized macula-centered fundus photos taken with the Zeiss Visucam®. The AI model's output was compared with manual readings by trained retina specialists. Results: A total of 215 eyes from 109 patients were included in the study. Human graders identified 109 eyes (50.7%) with retinal abnormalities. The AI model demonstrated an overall accuracy of 73.0% (95% CI 66.6% – 78.8%) in detecting abnormal retinas, with a sensitivity of 54.1% (95% CI 44.3% – 63.7%) and specificity of 92.5% (95% CI 85.7% – 96.7%). Conclusions: The availability and sources of AI training datasets can introduce biases into AI algorithms. In our dataset, racial differences in retinal morphology, such as differences in retinal pigmentation, affected the accuracy of AI image-based analysis. More diverse datasets and external validation on different populations are needed to mitigate these biases.https://paojournal.com/index.php/pjo/article/view/522artificial intelligencedeep learningretinal imagingdataset diversityracial variations |
| spellingShingle | Carlo A. Kasala, MD Kaye Lani Rea B. Locaylocay, MD-MBA Paolo S. Silva, MD The Impact of Anatomic Racial Variations on Artificial Intelligence Analysis of Filipino Retinal Fundus Photographs Using an Image-Based Deep Learning Model Philippine Journal of Ophthalmology artificial intelligence deep learning retinal imaging dataset diversity racial variations |
| title | The Impact of Anatomic Racial Variations on Artificial Intelligence Analysis of Filipino Retinal Fundus Photographs Using an Image-Based Deep Learning Model |
| title_full | The Impact of Anatomic Racial Variations on Artificial Intelligence Analysis of Filipino Retinal Fundus Photographs Using an Image-Based Deep Learning Model |
| title_fullStr | The Impact of Anatomic Racial Variations on Artificial Intelligence Analysis of Filipino Retinal Fundus Photographs Using an Image-Based Deep Learning Model |
| title_full_unstemmed | The Impact of Anatomic Racial Variations on Artificial Intelligence Analysis of Filipino Retinal Fundus Photographs Using an Image-Based Deep Learning Model |
| title_short | The Impact of Anatomic Racial Variations on Artificial Intelligence Analysis of Filipino Retinal Fundus Photographs Using an Image-Based Deep Learning Model |
| title_sort | impact of anatomic racial variations on artificial intelligence analysis of filipino retinal fundus photographs using an image based deep learning model |
| topic | artificial intelligence deep learning retinal imaging dataset diversity racial variations |
| url | https://paojournal.com/index.php/pjo/article/view/522 |
| work_keys_str_mv | AT carloakasalamd theimpactofanatomicracialvariationsonartificialintelligenceanalysisoffilipinoretinalfundusphotographsusinganimagebaseddeeplearningmodel AT kayelanireablocaylocaymdmba theimpactofanatomicracialvariationsonartificialintelligenceanalysisoffilipinoretinalfundusphotographsusinganimagebaseddeeplearningmodel AT paolossilvamd theimpactofanatomicracialvariationsonartificialintelligenceanalysisoffilipinoretinalfundusphotographsusinganimagebaseddeeplearningmodel AT carloakasalamd impactofanatomicracialvariationsonartificialintelligenceanalysisoffilipinoretinalfundusphotographsusinganimagebaseddeeplearningmodel AT kayelanireablocaylocaymdmba impactofanatomicracialvariationsonartificialintelligenceanalysisoffilipinoretinalfundusphotographsusinganimagebaseddeeplearningmodel AT paolossilvamd impactofanatomicracialvariationsonartificialintelligenceanalysisoffilipinoretinalfundusphotographsusinganimagebaseddeeplearningmodel |