In treatment-naive diabetic macular oedema, no structural OCT biomarker was independently associated with proliferative disease after accounting for fellow-eye correlation, centre effects, mutual adjustment, and multiplicity, though posterior hyaloid detachment gave the most consistent nominal signal.
Eyes with PDR had significantly worse baseline visual acuity than eyes with NPDR in treatment-naive DMO.
In univariable clustered comparisons, no OCT biomarker differed between NPDR and PDR after false discovery rate correction.
In the joint multivariable model, no structural biomarker was independently associated with PDR after Holm correction.
Posterior hyaloid detachment showed a nominal association with PDR that did not survive multiplicity correction.
VRIA was not independently associated with PDR after adjustment for posterior hyaloid status, suggesting the vitreoretinal interface variables carry overlapping information.
Centre-adjusted central macular thickness did not differ between NPDR and PDR eyes.
In secondary analyses, PDR, disrupted EZ, higher CMT and VRIA were each associated with worse BCVA.
This study looked at eye scans (optical coherence tomography, or OCT) from 818 eyes of 534 people with diabetes who had swelling in the central retina (diabetic macular oedema) and had not yet been treated. Researchers compared eyes with earlier-stage disease (non-proliferative) to eyes with more advanced disease (proliferative) to see whether particular features on the scans differed between the two groups. Data came from nine hospitals in a retrospective review. Eyes with proliferative disease had worse vision on average than those with non-proliferative disease. However, once the analysis accounted for the fact that many patients contributed both eyes, differences between hospitals, the overlap between different scan features, and the number of comparisons made, none of the scan features (such as retinal thickness, integrity of the photoreceptor layer, or the state of the vitreous gel's attachment to the retina) was independently linked to proliferative disease. Detachment of the vitreous gel from the retina showed a possible link, but this was not strong enough to hold up after statistical correction. This research suggests that these common OCT features may not reliably distinguish proliferative from non-proliferative disease in untreated patients with macular swelling. Features of the retina–vitreous boundary appeared to overlap in the information they provided. Because the study captured a single point in time, the authors stress that the findings do not show how patients will fare over time.
Özsoy Saygın I, Kumral Türkseven E, Kaplan F, Sevik M, Aydoğan Gezginaslan T, Önder Tokuç E, et al.. (2026). Comparison of Optical Coherence Tomography Biomarkers in Nonproliferative and Proliferative Diabetic Retinopathy: Marmasia Study Group Report No. 2.. Turkish journal of ophthalmology. https://doi.org/10.4274/tjo.galenos.2026.24196