Quantifying subclinical and longitudinal microvascular changes following episcleral plaque brachytherapy (EPB) using spectral-domain OCT angiography

Background I-125 episcleral plaque brachytherapy (EPB) is standard-of-care for globe-conserving treatment of medium-sized choroidal melanomas. Radiation retinopathy is a potential consequence of treatment, characterized by deleterious effects on retinal microvasculature. We investigated the application of Optical Coherence Tomography Angiography (OCTA) for detecting and longitudinally monitoring I-125 episcleral plaque brachytherapy induced radiation retinopathy.

Methods High resolution OCTA of the central 3×3mm macula were obtained from I-25 episcleral plaque brachytherapy treated and untreated fellow eyes of 62 patients. Capillary density (vessel skeleton density, VSD) and caliber (vessel diameter index, VDI) were quantified using previously validated semi-automated algorithms. Nonperfusion was also quantified as flow impairment regions (FIR). Exams from treated and fellow eyes obtained pre-treatment and at 6-month, 1-year, and 2-year intervals were compared using generalized estimating equation linear models. Dosimetry maps were used to evaluate spatial correlation between radiation dose and microvascular metrics.

Results Mean time from treatment to last follow-up was 10.8 months. Mean±SD and median radiation dose at the fovea were 64.5 ± 76 Gy and 32.0 Gy, respectively. Preoperative logMAR (Snellen) mean visual acuity was 0.26 ± 0.05 (∼20/35) and 0.08 ± 0.02 (∼20/25) in treated and fellow eyes, respectively. At 6 months, treated eyes had significantly lower VSD (0.147 ± 0.003 vs 0.155 ± 0.002; p = 0.023) and higher FIR (1.95 ± 0.176 vs 1.45 ± 0.099; p = 0.018) compared to fellow eyes. There was a significant decrease in VSD and a corresponding increase in FIR even for treated eyes without clinically identifiable retinopathy at 6 months. VDI was significantly higher in treated eyes than in fellow eyes at 2 years (2.93 ± 0.022 vs 2.84 ± 0.016; p = 0.002). Microvascular changes were spatially correlated with a radiation gradient of 85-250 Gy across the fovea.

Conclusions OCTA can be used to quantify and monitor EPB induced radiation, and can detect vascular abnormalities even in the absence of clinically observable retinopathy. OCTA may therefore be useful in investigating treatment interventions that aim to delay EPB-induced radiation retinopathy.