Dose-volume predictors of radio-induced complications after radiosurgery for uveal melanoma

Abstract

Introduction Uveal melanoma (UM) is a life threatening intraocular malignant tumor. Gamma Knife Stereotactic Radiosurgery (GKSRS) is a well-assessed strategy for conservative treatment of UM providing good results for survival, local control and eye preservation. Purpose Our aim is to develop predictive models for radio-induced effects in UM patients treated with exclusive GKSRS. Materials and methods Medical records and 3D dosimetry data of critical structures of 66 patients were retrospectively reviewed. Cox’s proportional hazard model was used to identify clinical and dosimetric variables as independent risk factor for most clinically relevant GKSRS-related complications: radiation vasculopathy(RV), radiation papillopathy (RP) and neovascular glaucoma(NVG), visual acuity(VA) reduction >20% (VA20%) and complete loss of basal VA (VA100%). ROC curve analysis allowed predicting cutoff values of significant variables. Results The 2 years incidences were: RV 10%, RP 12%, NVG 14%, VA20% 59% and VA100% 27%. A clear relationship with D1% to optic nerve (ON) was found for RP and RV. Also, for RV, the prescription isodose-ON distance (PIOND) and the anterior to equator position of the tumor resulted as protective variables. The V20 of posterior segment and tumor thickness were predictive for NVG. Multivariate analyses resulted in two variables predictive model both for VA20% and VA100%, including tumor largest diameter and D1% to ON. A second predictive model, including PIOND, was found for VA100%. Conclusion We found clinical and dosimetric variables to predict the risk of main side effects after GKSRS for UM. These results may provide new dose constraints to critical structures potentially able to reduce toxicities. Disclosure Antonella del Vecchio has a contract as starter-up with Elekta Instrument AB. No other conflict of interest to declare for the other authors.