TCP and Dose Response after Brachytherapy for Choroidal Melanoma

Abstract

Brachytherapy is a commonly used eye-preserving treatment for choroidal melanomas. We hypothesized that the minimum tumor dose is related to risk of tumor recurrence. We studied consecutive patients with primary choroidal melanomas, treated with Ru-106 plaque brachytherapy from 2005 to 2014 at a single institution to an apical dose of 100 Gy. Plans were retrospectively created using 3D image-guided planning software. Pretreatment retinographies were used to contour the tumor; posttreatment retinographies to determine the accurate plaque position. Dose volume histograms were extracted. Patient and tumor characteristics, treatment details, and clinical outcomes were extracted from patient records. All patients were followed regularly until death or study cut-off (Jan 2018). Biologically effective doses (BED) were calculated using source half times, tissue specific factors (α/β=11.5 Gy, T1/2=1.5h), and a well-established model (Gagne et al, Med Phys, 2012) The relationship between tumor dose and risk of recurrence was examined using multivariate Cox regression modelling. Minimum physical dose to the tumor (D99%), transpupillary thermotherapy (TTT), tumor height, largest base dimension, incident eye, and gender were included in the model. Model robustness was assessed by considering alternative dose metrics, BED, and by taking the competing risk of death and TTT during follow-up into account (Aalen-Johansen estimator for cumulative incidence and Fine & Gray’s model for regression analysis). We included 227 patients (108 females, 110 right eyes), with median treatment time of 120h (IQR: 74-191). Median tumor height and largest base dimension were 4 mm (IQR: 3-6) and 11 mm (IQR: 9-13). Median follow-up time was 4.9 years. The estimated 3-year local control was 80% (95% CI: 75-86). The median D99% was 104 Gy (IQR: 63-138); this was the only significant factor associated with recurrence in multivariate modelling (p=0.0007). Hazard ratio (HR) for 10 Gy increase in D99% was 0.90 (95% CI:0.84-0.96). Estimated 3-year local control for D99%=50 Gy and D99%=100 Gy were 70% (95% CI: 58-85) and 81% (95% CI: 73-91), respectively. The median BED99% was 145 Gy (IQR: 73-196); using BED99% in the multivariate model gave no substantial differences in estimates (HR 0.93 (95% CI: 0.90-0.97, p=0.001). Robustness checks with D1-99% and BED1-99% showed D99% and BED99% to be the most significant dose metrics for recurrence. Competing risk analysis showed cumulative incidences at 3 years for recurrence, death, and TTT of 17% (95% CI: 12-22), 14% (95 % CI: 9-18), and TTT 6% (95% CI: 3-9). Accounting for death and TTT, D99% remained the only significant factor for recurrence (p=0.0004), with HR for 10 Gy increase in D99% of 0.90 (95% CI: 0.85-0.95). The minimum dose to the tumor correlated strongly with risk of tumor recurrence, with 100 Gy needed to ensure at least 80% local control at 3-years.