Title of Article: Accuracy of Artificial Intelligence in Estimating Best-Corrected Visual Acuity From Fundus Photographs in Eyes With Diabetic Macular Edema
What are the key takeaway points from this article?
Diabetic macular edema is an accumulation of fluid in the macula secondary to diabetes. A measurement used to guide management for this disease is the best-corrected visual acuity (BCVA). This requires a qualified worker to assess the patient’s BCVA using a chart and recording the patient’s results, a time consuming process. Researchers set out to develop an AI algorithm that could be utilized to detect BCVA from fundus photographs in patients undergoing treatment for diabetic macular edema, which may be advantageous by minimizing resources and time.
The study evaluated the mean absolute error (MEA) of the BCVA and the secondary outcome of the study consisted of the percentage of predictions that were within 10 letters. The study examined 459 participants and 7,185 fundus images. The baseline BCVA score ranged from 73 to 24 letters (approximate Snellen equivalent: 20/40 to 20/320). The MEA was 9.66 letters, with 33% of the values being within 0 to 5 letters and 28% being within 6 to 10 letters.
In summary, the study results showed that 33% of the estimated values were within 0 to 5 letters and 28% within 6 to 10 letters. This suggests that artificial intelligence may be used to estimate BCVA from fundus photographs in patients with DME, without refraction or subjective visual acuity measurements. It will typically accomplish this within 1 to 2 lines on an ETDRS chart. This technology could be used to minimize the amount of time and costs associated with travel for patients.
Publication Date: June 8, 2023
Reference:
Paul W, Burlina P, Mocharla R, et al. Accuracy of Artificial Intelligence in Estimating Best-Corrected Visual Acuity From Fundus Photographs in Eyes With Diabetic Macular Edema. JAMA Ophthalmol. Published online June 08, 2023. doi:10.1001/jamaophthalmol.2023.2271
Summary by: Daniel Lamoureux