Prognostic accuracy of the mean of the largest nucleoli, vascular patterns, and PC-10 in posterior uveal melanoma.

Abstract

This study aimed to compare the prognostic value and the predictive accuracy of the PC-10 cell cycling marker with the largest tumor diameter, the mean of the largest nucleoli, and vascular patterns in posterior uveal melanoma. The study design was a case-control study. Eyes enucleated for posterior uveal melanoma from patients who either died of metastatic melanoma or survived without signs of metastatic disease 10 years or more after surgery were studied. Three observers assessed the above prognostic indicators and standard histopathologic characteristics from microslides without access to survival data. Univariate and multivariate Cox models for survival were constructed, and a multiparameter prognostic index was calculated for each patient, based on covariates obtained from the final Cox model. The prognostic accuracy was determined by receiver operating characteristic curve analysis. The log PC-10 count, vascular networks, mean of the largest nucleoli, largest tumor diameter, age of patient, and prognostic index were independently associated with outcome. However, each of these indicators had no more than a poor-to-moderate predictive accuracy, and only the prognostic index was significantly better than the largest tumor diameter. The PC-10 count retains a prognostic value in uveal melanoma when adjusting for the effect of the mean of the largest nucleoli and diverse vascular patterns. A prognostic index combining two or more indicators may improve the predictive precision.