Isolated high-grade prostatic intraepithelial neoplasia (HGPIN) on needle biopsy confers an increased risk of prostate carcinoma (CaP) on follow-up biopsy. The aim of this study is to determine whether paraffin-section fluorescence in situ hybridization (FISH) of specific chromosome/oncogene copy number abnormalities (CNAs) in biopsy specimens with isolated HGPIN increases the predictive value for CaP on repeat biopsy. Cases were divided into 3 groups: controls (n=8) and sextant biopsy specimens with isolated HGPIN without CaP (group A; n=11) and with CaP (group B; n=14) on follow-up biopsy. Dual-color FISH assessing c-myc, HER-2/neu, chromosome region 7q31 (D7S486), and corresponding chromosome centromeres was performed. An amplification ratio (AR) for each marker centromere was derived for each biopsy specimen, with AR ranges designated as no/low, low-intermediate, and high. Also calculated for each marker were the percentage of cells with marker amplification, hyperdiploidy, and monosomy. A composite score for each biopsy specimen was calculated based on these parameters, with a possible range of 0 to 15. The specific chromosomal oncogene CNAs were as follows: for chromosome 7/7q31, 2 of 11 (18%) in group A and 6 of 14 (43%) in group B; for chromosome 8/c-myc, 4 of 11 (36%) in group A and 9 of 13 (69%) in group B; and for chromosome 17/HER-2/neu, 10 of 10 (100%) in group A and 13 of 14 (93%) in group B. The mean composite score was 0 for controls, 2.5 for group A, and 4.7 for group B. Composite scores > or =4 for the 3 groups were 0 of 9 (0%) for controls, 1 of 11 (12%) for group A, and 8 of 14 (57%) for group B. These differences were statistically significant (P=0.015). One group A patient with a high composite score (6) had atypical small glands on follow-up biopsy at <1 year. Chromosome/oncogene CNAs are uncommon in control patients, occurring with increasing frequency and magnitude in patients with isolated HGPIN without and with follow-up CaP. Chromosome/oncogene CNAs in HGPIN are mostly of the low to intermediate level and display intercellular heterogeneity. HER-2/neu amplification is common in HGPIN with and without follow-up CaP. Chromosome 7 and 8 aneusomy and 7q31 and c-myc amplification are greater in HGPIN with follow-up CaP. Patients with isolated HGPIN and high composite score without follow-up CaP are uncommon; these patients may have a small, unsampled CaP. Although patients with HGPIN without CaP are more likely to have a low composite score, a subset of patients with follow-up CaP have low composite score, suggesting (1) mutational pathways independent of chromosomes 7, 8, and 17 and HER-2/neu, c-myc, and chromosome region 7q31 CNAs; (2) CaP derived from an independent, unsampled focus of HGPIN; or (3) CaP not derived from HGPIN.