To the Editor.
—In his article in the July issue comparing enucleation with cobalt 60 irradiation of melanomas, Dr Gass graciously provided detailed information on the patients included in his study.1 I have analyzed these data using the Cox regression model and wish to report the following findings:
Of the various parameters determined for all patients, six showed independent correlation with death from uveal melanoma during ten years of follow-up (P <.05): patient age, tumor size at time of detection, largest dimension, volume and location of tumor at time of enucleation, and modality of treatment (cobalt 60 plaque vs enucleation).
The best independent predictors were tumor location at time of enucleation and tumor size at the time of detection (P <<.001).
For the numerous Cox regressions studied, the mode of treatment was not statistically significant (P >.05) when added to any regression that contained tumor location
Gamel JW. Ocular Melanoma. Arch Ophthalmol. 1985;103(9):1284. doi:10.1001/archopht.1985.01050090036023