Gene Expression Signature Predicts Survival In Uveal Melanoma

Abstract

 

Purpose:

 

To correlate the gene expression profiles with histopathology data and the tumor related mortality in effort to identify patients at high risk for metastasis.

 

Methods:

 

Twenty–seven microarray samples of medium to large sized uveal melanoma from patients who had consented to undergo enucleation were used in the study. RNA was isolated using the RNAqueous kit (Ambion). Cy3– or Cy5–labeled cDNA was prepared by indirect incorporation. Posthybridization data were acquired with a GenePix 4000B laser scanner and GenePix Pro, version 5.0, software. The tumor cDNA array comprised a set of sequence–verified cDNA clones representing approximately 4000 genes involved in tumorigenesis.

 

Results:

 

The average age of the patients were 61, of which 13 were female and 16 were dead by the follow–up date. Two multivariate data mining techniques, the Hierarchical cluster analysis and the Multidimensional scaling were used to investigate the grouping structure in the gene expression data of the 27 tumor samples. Two distinct classes were found (Class 1 and Class2). When correlated with the metastatic data, the two molecular classes correspond very well to the survival data of the 27 samples. Cluster Analysis were repeatedly performed over a randomly selected subset of genes and cumulative contribution of all the genes in making the correct grouping were recorded. Gene Weights was the average gene selection rate, and 20% was used as the cutoff to select 44 genes that correctly defined the molecular classes. One unique gene alone (ectonucleotide pyrophosphatase/phosphodiesterase 2 (autotaxin)) could classify the molecular classes (Figure 1). The expression pattern was confirmed using real time quantitative PCR.

 

Conclusions:

 

Gene expression pattern of one gene alone (autotaxin) predicts survival in uveal melanoma.  

 

Keywords: melanoma • gene microarray • oncology