Artificial intelligence is being explored as a screening platform for myriad ocular conditions, and now a new study suggests an automated algorithm based on intraretinal and subretinal fluid images may help clinicians assess the anti-VEGF treatment success of their diabetic macular edema (DME) patients.

This post hoc analysis of the Diabetic Retinopathy Clinical Research Network trial assessed 6,945 OCT volume scans of 570 eyes of 570 patients with DME. Individuals were treated with aflibercept, ranibizumab or bevacizumab with or without deferred laser. Slightly more than half of the patients were male (53%), 65% were white and the average age was about 43.

The study reported aflibercept and ranibizumab were associated with a greater reduction of intraretinal fluid than bevacizumab. The researchers found no difference among anti-VEGF agents regarding reduction of subretinal fluid.

At baseline, the mean fluid volumes in the central 3mm were 448.6nL for intraretinal fluid and 36.9nL for subretinal fluid. At 12 months, the intraretinal fluid volume was 161.2nL and the subretinal fluid volume was 4.4nL.

The researchers noted the presence of subretinal fluid at baseline was associated with a worse baseline BCVA Early Treatment Diabetic Retinopathy Study (ETDRS) score of 63.2 (approximate Snellen equivalent of 20/63) in eyes with subretinal fluid vs. 66.9 (approximate Snellen equivalent of 20/50) without subretinal fluid. Additionally, the study noted a greater gain in the ETDRS score (0.5) every four weeks during follow-up in eyes with subretinal fluid compared with baseline vs. 0.4 in eyes without subretinal fluid compared with baseline when adjusted for initial BCVA.

Aflibercept was associated with a greater reduction of intraretinal fluid volume compared with bevacizumab after the first injection (difference: 79.8nL) and every following four weeks (a difference of 10.4nL). Also of note: ranibizumab was associated with a greater reduction of intraretinal fluid after the first injection compared with bevacizumab (a difference of 75.2nL).

Roberts PH, Vogl WD, Gerendas BS, et al. Quantification of fluid resolution and visual acuity gain in patients with diabetic macular edema using deep learning a post hoc analysis of a randomized clinical trial. JAMA Ophthalmology. July 23, 2020. [Epub ahead of print].