Purpose: To build and validate a prognostic model that predicts long-term overall survival (OS) in metastatic choroidal and ciliary body melanoma (CCBM) to facilitate patient counseling and planning, reporting, and interpreting clinical trials.
Design: Retrospective cohort study with validation.
Methods: We analyzed predictors of intermediate (IMT; 25-<42 months) and long-term (LT; ≥42 months) OS in a Finnish nationwide cohort of 330 patients with metastatic CCBM. Short-term (<25 months), IMT, and LT survival were compared with pairwise and ordinal logistic regression. A single-center cohort of 259 patients from Italy was used for validation. Models were compared with a deviance test.
Results: Median OS was 12 and 17 months in the building and validation datasets, respectively; 40 (12%) and 31 (9%) compared with 44 (17%) and 32 (12%) patients were IMT and LT survivors, respectively. Alkaline phosphatase or lactate dehydrogenase level never exceeded 2 times the upper normal limit (UNL) in either LT cohort. Conditional to both being ≤2 times the UNL, distant metastasis-free interval (DMFI) >42 months (odds ratio [OR] 4.09-4.64; P < .001) paired with age <60 years (OR 3.23; P = .002), having no symptoms (OR 4.19; P = .005), and the largest diameter of the largest metastasis <30 mm (Tumor, Node, Metastasis stage M1a; OR 3.05; P = .001) independently predicted higher odds of surviving longer (IMT or LT) without model preference. These results were confirmed in the validation dataset.
Conclusions: Alkaline phosphatase or lactate dehydrogenase >2 times the UNL essentially precluded LT survival. The most robust predictor otherwise was DMFI >42 months, followed by age <60 years, absence of symptoms, and Tumor, Node, Metastasis stage M1a.
Copyright © 2022 The Author(s). Published by Elsevier Inc. All rights reserved.