Correlation of limited-early-response status with 12-month CST, BVA, and machine learning-quantified retinal fluid in diabetic macular oedema in routine clinical practice

Anti-VEGF treatment response in DMO has been measured by changes in the central subfield thickness (CST) and best visual acuity (BVA) outcomes at 3 months after initial treatment, termed early or limited early response (ER/LER). This study correlates LER with 12-month BVA, CST, and retinal fluid volumes quantified by a machine learning algorithm on optical coherence tomography (OCT).

The study included treatment naïve DMO patients ≥ 18 years with OCT scans at baseline (M0), M3, M6, and M12. The 220 patients were categorized as limited early responders (LER) if they had ≤ 10% CST reduction and/or < 5 ETDRS letter gain at M3. BVA, CST, and subretinal (SRF), intraretinal (IRF), and total retinal (TRF) fluid volumes quantified by a machine learning algorithm were compared between groups and across time.

At M12, the anatomic LER (aLER), defined solely by CST, had significantly worse BVA and CST versus the anatomic ER (aER) group (p < 0.001). Retinal fluid M12 outcomes did not significantly vary between all LER and ER groups. No significant BVA, CST, TRF, and IRF variance across time for LER was found (p > 0.1).

BVA and CST M12 outcomes vary by aLER/aER status indicating that CST may be a strong predictor of treatment outcomes, while retinal fluid volumes were not predicted by LER status.