Side effects of proton beam therapy of choroidal melanoma in dependence of the dose to the optic disc and the irradiated length of the optic nerve.

Abstract

To analyze the impact of the dose to the optic disc and the irradiated length of the optic nerve on radiation-induced optic neuropathy, radiation-induced retinopathy, iris neovascularization, secondary glaucoma, enucleation, and local tumor control after proton beam therapy (PBT) of choroidal melanoma. Retrospective analysis of 1129 patients, who received primary PBT for the treatment of choroidal melanoma with a dose of 60 cobalt gray equivalents (CGE) between 1998 and 2013 at the Helmholtz-Zentrum Berlin, Germany. Kaplan-Meier curves and logrank test have been used for time-to-event analyses. Adjustment for potential confounders was done using multiple Cox regression models with forward and backward selection. We found a significant correlation between the irradiated length of the optic nerve and the dose to the optic disc (correlation coefficient, 0.93). Multivariate Cox regression revealed the dose to the optic disc as an independent predictive risk factor for the development of radiation-induced optic neuropathy (p < 0.001, HR 1.023, 95 CI 1.016-1.029), iris neovascularization (p < 0.001, HR 1.013, 95% CI 1.008-1.019), secondary glaucoma (p < 0.001, HR 1.017, 95% CI: 1.011-1.023) and enucleation (p < 0.001, HR 1.037, 95% CI 1.020-1.053). The irradiated length of the optic nerve was not a statistically independent predictive risk factor in multivariate analysis. Our data implicate the predominance of the dose to the optic disc over the irradiated length of the optic nerve regarding radiation-induced optic neuropathy, iris neovascularization, secondary glaucoma, and enucleation.