Retinal microaneurysms (MAs) are among the earliest signs of diabetic retinopathy (DR) and can be classified in several subtypes by non-invasive multimodal retinal imaging. The main aim of the present study is to characterize retinal MAs perfusion properties and their blood flow network connectivity by means of Dense Automatic-RealTime (DART) OCTA technology, checking the relationship with the multimodal retinal imaging classification and testing the clinical impact of DART. A cross-sectional, observational study setting was chosen. Multimodal retinal imaging included confocal multicolour, OCT, OCTA and DART OCTA. We classified retinal MAs accordingly with the recently proposed multimodal retinal imaging classification and we tested the role of DART OCTA for detecting retinal MAs blood flow network connectivity. We also tested the relationship with clinical parameters. We included 206 retinal MAs of 36 DR eyes. We categorized retinal MAs as red (70; 34%), mixed (106; 51%) and green (30; 15%), corresponding to precise characteristics on structural OCT and both (regular) enface and DART OCTA images. The agreement between en-face and DART OCTA techniques for detecting MAs perfusion was very high (overall ICC 0.98; p < 0.01). However, DART OCTA provided clearer visualization than enface OCTA for detecting the blood flow network connectivity of retinal MAs, especially looking at the afferent and efferent MAs capillaries. Multimodal retinal imaging classification of retinal MAs provided significant correlations with DR duration, DR stage, and macular capillary non-perfusion. DART OCTA provided several new insights on retinal MAs characteristics and their blood flow network connectivity.
Read full abstract