Abstract
Purpose:
Almost 50% of patients with choroidal melanoma develop metastatic disease despite successful eradication of the primary tumor. Accurate prognostication for patients with choroidal melanoma enhances care planning and improves patient quality of life, even when the prognosis is found to be poor. However, univariate prognostication based on anatomic, histologic or genetic predictors is not accurate enough to be relevant to individual patients. Further, the accuracy of various prognostic models is diminished by missing data.<br /> The Liverpool Uveal Melanoma Prognosticator Online (LUMPO) performs multivariate analysis of genetic, histologic and anatomical data, taking the patient’s age and sex into account and compensating for any missing data, providing prognostications that are accurate enough to be relevant to individual patients. This tool has been validated in the United Kingdom, but it remains to be seen whether it is generalizable to other patient populations. In this study, we seek to validate LUMPO in a group of patients who received care for choroidal melanoma at the University of California, San Francisco.
Methods:
A retrospective chart review was performed for all patients who were treated for choroidal melanoma at the University of California, San Francisco between 2002 and 2007. Tumor characteristics were recorded including largest basal tumor diameter and thickness, extraocular extension, and ciliary body involvement. In patients who underwent enucleation, histopathological data was gathered, including the presence of epithelioid cells, vascular loops, and mitotic count. Survival data nonspecific to cause of death were gathered from the UCSF Cancer Center tumor database. The data were analyzed using the method previously outlined by Eleuteri et al.
Results:
A total of 409 patients were compared to the established database of patients in the United Kingdom. Patient characteristics including age at diagnosis, race, gender, and laterality were found to be similar across both patient groups. Time from diagnosis to death was as well similar between the two groups. LUMPO accurately predicted survival in our cohort of patients, even when information was missing including histopathological and genetic data.
Conclusions:
These results confirm that LUMPO is a valuable method of prognostication for choroidal melanoma and is generalizable to patients in the United States.