The Impact of Anatomic Racial Variations on Artificial Intelligence Analysis of Filipino Retinal Fundus Photographs Using an Image-Based Deep Learning Model
Carlo A. Kasala, MD1, Kaye Lani Rea B. Locaylocay, MD-MBA1, Paolo S. Silva, MD1,2,3
1Eye and Vision Institute, The Medical City, Pasig City, Philippines
2Beetham Eye Institue, Joslin Diabetes Center, Boston, Massachusetts, USA
3Department of Ophthalmology, Harvard Medical School, Boston, Massachusetts, USA
Correspondence: Carlo A. Kasala, MD
Office Address: Eye and Vision Institute Department Office, 4th Floor, The Medical City, Ortigas Avenue, Pasig City, Philippines 1602
Office Phone Number: +639171692275
Email Address: carlokasalamd@gmail.com
Disclosures: C.A.K. and K.L.R.B.L. report no financial conflict of interest. P.S.S. receive research support and/or honorarium from Optos plc, Eye Health, Kubota Vision that is not related to this paper.
ABSTRACT
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.
Keywords: artificial intelligence; deep learning; retinal imaging; dataset diversity; racial variations