Multimodal imaging characteristics in eyes with vitreoretinal lymphoma treated with intravitreal rituximab.

Abstract

To characterize the imaging features in eyes with vitreoretinal lymphoma (VRL) using ultra-widefield fundus photography (UWF-FP), swept-source optical coherence tomography (SSOCT) and fundus autofluorescence (FAF) that are correlated to ongoing treatment with intravitreal Rituximab(IVR). Retrospective observational imaging-based study of 15 treatment-naive eyes with VRL treated with IVR. All patients with primary VRL underwent vitreous biopsy using 23/25G microincision vitrectomy system for confirmation of diagnosis. All eyes received monthly IVR (1mg/0.1mL) injections till disease remission. Baseline clinical characteristics, treatment details, outcomes, and sequential imaging features on UWF-FP, FAF, and SSOCT were analyzed. Baseline features and changes in UWF-FP, FAF patterns, and SSOCT features in response to treatment RESULTS: Clinically, patients presented with sub-RPE deposits (n = 15), superficial retinal hemorrhages (n = 2), 'giant' RPE (retinal pigment epithelium) holes (n = 2), and anterior segment reaction (n = 1). Eyes were treated with mean 5.7 IVR injections (median: 5; range 1-13) over a mean 7.2 ± 4.9months. During the course of treatment, two eyes developed superficial retinal hemorrhages with spontaneous resolution, 2 eyes developed CME, and 4 eyes developed characteristic 'leopard skin' pigmentation. Hyper-autofluorescence corresponding to areas of active lesions decreased with each treatment cycle and was finally replaced by hypo-autofluorescence. Serial OCTs showed regression of sub-RPE/subretinal deposits (n = 15), ellipsoid zone disruption (n = 9), and its resolution with treatment (n = 3), epiretinal membrane (ERM; n = 6), choroidal hyperreflective foci (HRF; n = 4), disorganization of retinal inner layers (DRIL; n = 3), RPE-rip (n = 2), cystoid macular edema (CME; n = 2), and hyperreflective lesions in the choroid (n = 1). Complete resolution was observed in all eyes with extensive hypo-AF. The central foveal thickness decreased from 237 ± 113μ to 182 ± 114μ (p = 0.1) and subfoveal choroidal thickness decreased from 258 ± 66μ to 220 ± 64μ (p = 0.12) at final follow-up. The mean baseline BCVA was logMAR 0.9 ± 0.9 that deteriorated to mean logMAR 1 ± 1 final visit (p = 0.7). The mean recurrence-free follow-up was 5.9 ± 5.1months CONCLUSION: Multimodal imaging provides novel insights into features of VRL, a better understanding of regression patterns, and prognostication of outcomes when treated with intravitreal rituximab. Larger, multicentric studies with longer follow-up will help unravel imaging biomarkers to understand these aspects better.