One-Year Tumor Volume Regression Predicts Survival After I-125 Plaque Brachytherapy for Posterior Uveal Melanoma.

Abstract

To evaluate the prognostic significance of one-year tumor volume regression (TVR) following low-dose-rate iodine-125 episcleral plaque brachytherapy (EPBT) in patients with uveal melanoma. Retrospective single-center clinical cohort study. A total of 1,180 patients with uveal melanoma treated with low-dose-rate iodine-125 EPBT delivered at 52.8 cGy/h over 168 hours at a large academic medical center with an ocular oncology service between 1984 and 2022. Patients had tumor ultrasound measurements at diagnosis and one-year follow-up, with complete survival data available. Patients without complete survival data or ultrasound measurements at either timepoint were excluded. TVR was calculated using ultrasound-derived ellipsoid modeling at baseline and 1 year. Patients were stratified into delayed (≤33.3%), intermediate (33.3%-75%), and accelerated (≥75%) regression groups. Thresholds were selected using a log-rank test sweep across -100% to 100% regression in conjunction with locally weighted scatterplot smoothing (LOWESS) analysis. Inflection points were identified at 33.3% (p = .029) and 75% (p = .001). Kaplan-Meier (KM) survival analysis was used to compare overall survival (OS) among regression groups, with significance assessed by log-rank tests. Mean tumor volume decreased from 336.3 mm³ to 129.5 mm³ by one year, with 94.4% (N = 1,114) showing volume reduction. KM survival analysis demonstrated that patients with delayed and accelerated tumor regression had worse OS than those with intermediate-range regression across the full cohort (p < .0001, p < .0001). This association was conserved when tumors were stratified by Collaborative Ocular Melanoma Study (COMS) size and American Joint Committee on Cancer (AJCC) clinical stage guidelines. No difference was seen between the intermediate and accelerated regression groups for COMS small and cT1 tumors. TVR at 1 year following EPBT is a significant prognostic biomarker for OS in UM, with intermediate regression (33.33% to 75%) conferring the greatest survival benefit. These findings support 1 year tumor regression metrics as a practical tool for risk stratification and patient counseling.