Abstract
Purpose :
To analyze optical coherence tomography (OCT) segmentation in naïve patients with diabetic macular edema (DME) using the artificial intelligence tool RetinAI Discovery Clinics, and to investigate the impact of hypertension (HTA) and intraretinal fluid (IRF) on visual acuity (AV), identifying significant tomographic biomarkers.
Methods :
A retrospective, descriptive analysis was performed on DME patients diagnosed over one year. OCT images were analyzed using the Discovery Clinics platform. Data included demographic information and thickness measurements of various retinal layers. Statistical analyses encompassed descriptive statistics, correlational analysis, subgroup comparisons (HTA and non-HTA groups), and linear regression modeling.
Results :
Out of 109 eyes from 72 patients (50 men, 22 women; mean age 66.1 ± 12.1 years), the mean VA was 0.42 ± 0.2. HTA patients showed significantly lower AV and higher IRF compared to non-HTA (p < 0.05). Regression analysis indicated a near-significant impact of PED on AV. Layer-wise, the mean thicknesses were: RNFL 50.7 ± 11.7𝛍m, GCL + IPL 66.2 ± 9.1 𝛍m, INL + OPL 57.2 ± 8.3 𝛍m, ONL 85.5 ± 12.3 𝛍m, PR+RPE 57.6 ± 3.8 𝛍m, and CC+CS 197.8 ± 31.2 𝛍m. A significant positive correlation was found between VA and PR+RPE thickness (p=0.039).
Conclusions :
AI tools like RetinAI Discovery Clinics are instrumental in characterizing DME patients. OCT thinning of the PR+RPE layer analyzed by AI is a marker of worse visual acuity. Additionally, hypertension and increased IRF are associated with reduced visual acuity, highlighting the importance of managing systemic conditions in DME. The data suggest that retinal changes, particularly in the PR+RPE layer, play a critical role in visual function.
This abstract was presented at the 2024 ARVO Annual Meeting, held in Seattle, WA, May 5-9, 2024.