Purpose: To quantitate long-term risk of local treatment failure after proton irradiation of choroidal/ciliary body melanomas and to evaluate risk of metastasis-related deaths after local failure.
Methods: We followed prospectively 1,922 patients treated at the Harvard Cyclotron between January 1975 and December 1996 for local recurrences of their tumors. Mortality surveillance was completed through June 1999. For analysis, patient follow-up continued until tumor regrowth was detected or, in patients without recurrence, until the date of the last dilated examination prior to April 1998. Actuarial methods were used to calculate rates of recurrence and metastatic deaths. Cox regression models were constructed to evaluate risk factors for these outcomes.
Results: Median ocular follow-up after irradiation was 5.2 years. Local recurrence was documented in 45 patients by ultrasound and/or sequential fundus photographs; in 17 more patients, the eye was enucleated due to suspected but unconfirmed tumor growth. Recurrences were documented between 2 months and 11.3 years after irradiation. The 5- and 10-year rates of regrowth, including suspected cases, were 3.2% (95% confidence interval [CI], 2.5%-4.2%), and 4.3% (95% CI, 3.3%-5.6%). Among the 45 documented recurrences, about one half (21) occurred at the margin, presumably due to treatment planning errors. The remaining cases represented extrascleral extensions (nine cases), ring melanomas (six cases), or uncontrolled tumor (nine cases). Recurrence of the tumor was independently related to risk of tumor-related death.
Conclusion: These data, based on relatively long-term follow-up, demonstrate that excellent local control is maintained after proton therapy and that patients with recurrences experience poorer survival.