Prediction model may help manage patient expectations of treatment of ME-CRVO

A machine learning (ML) algorithm was able to predict visual and anatomic outcomes in patients with macular edema secondary to central retinal vein occlusion (ME-CRVO) after undergoing treatment with aflibercept, according to a presentation at ASRS 2022.

The model also predicted treatment frequency with high accuracy.

Patients with ME-CRVO from the COPERNICUS (n = 107) and GALILEO (n = 91) trials treated with monthly intravitreal aflibercept injection (IAI) for 24 weeks before switching to pro-re-nata (PRN) dosing through 52 weeks, were included in this study.

The algorithm predicted the actual observed values through the end of the study period with strong correlation for absolute BCVA, change in BCVA from baseline, and gain of ≥15 letters.

Change in CST from baseline was predicted but there was no correlation between predicted CST and observed absolute CST.

At weeks 16, 20, and 24, BVCA was a predictor of absolute BCVA, while change in BCVA was predicted by BCVA at baseline (≥15 letter gain), week 20 (≥15 letter gain), and week 24. CST and BCVA at baseline were predictors of change in CST.

Thicker CST at baseline and at week 4 were important factors in predicting >2 injections between week 24 and week 52.

Reference
Modi Y, et al. Predicting Outcomes and Treatment Frequency After Monthly Aflibercept for Macular Edema Secondary to Central Retinal Vein Occlusion: A Machine-Learning Model Approach. Presented at: ASRS 2022.