Vitreous hemorrhage after plaque radiotherapy for uveal melanoma

Abstract

Purpose: The purpose of this article is to determine the incidence, etiology, management, and outcome of vitreous hemorrhage (VH) after plaque radiotherapy for uveal melanoma.

Methods: Retrospective review of medical records.

Results: Of 3,707 eyes treated with plaque radiotherapy for uveal melanoma, VH developed in 4.1% at 1 year, 15.1% at 5 years, and 18.6% at 10 years by Kaplan-Meier analysis. Presumed causes of VH included tumor necrosis (29%), proliferative radiation retinopathy (24%), posterior vitreous detachment (16%), vascular occlusion (5%), and unknown (19%). Tumor necrosis was the most common cause of VH early in the follow-up period (3% at 1 year), while proliferative radiation retinopathy was the most common source of VH later (6.2% at 15 years). The most common initial management was conservative observation for resolution in 48%, laser photocoagulation in 24%, and vitrectomy in 18%. After a mean follow-up period of 5 years, the VH was completely resolved in 41%, partially resolved in 19%, unresolved in 20%, worsened in 5%, and enucleation was necessary in 15%. By multivariable analysis, risk factors for development of VH were the presence of diabetic retinopathy at first visit (relative risk, 6.64), shorter tumor distance to the optic disc (relative risk, 1.07), greater initial tumor thickness (relative risk, 1.1), and break in the Bruch membrane (relative risk, 2.93). The rates of local tumor recurrence, extraocular extension, and distant metastasis in 74 patients who underwent vitrectomy for VH removal after tumor regression were similar to those in patients who did not have vitrectomy for VH.

Conclusion: Vitreous hemorrhage occurs after plaque radiotherapy for uveal melanoma in 15.1% of the patients by 5 years. The main factors predictive of VH included underlying diabetic retinopathy, closer tumor proximity to the disc, greater tumor thickness, and break in the Bruch membrane. After tumor regression, vitrectomy for blood removal appears to be safe.