Microarray analysis of formalin-fixed and paraffin-embedded (FFPE) tissue seems to be of importance for the detection of molecular marker sets in prostate cancer (PC). The compromised RNA integrity of FFPE tissue results in a high degree of variability at the probe level of microarray data as shown by degradation plot. We tested methods that reduce the variability by including all probes within 300 nucleotides, within 600 nucleotides, or up to a calculated breakpoint with reference to the 3'-end. Accepted PC pathways such as the Wnt signaling pathway could be observed to be significantly regulated within FFPE microarray datasets. The best representation of PC gene expression, as well as better comparability to meta-analysis and fresh-frozen microarray data, could be obtained with a 600-nucleotide cutoff. Beyond the specific impact for PC microarray data analysis we propose a cutoff of 600 nucleotides for samples for which the integrity of the RNA cannot be guaranteed.