Often the diagnosis of pancreas cancer needs to be established from limited cytology specimens or small biopsies. Most ductal adenocarcinomas are histologically well to moderately differentiated and mimicked closely by pancreatitis, and therefore the microscopic diagnosis can be difficult. In addition, there appears to be significant heterogeneity in the outcome of the patients with pancreatic cancer, which cannot be predicted accurately by current prognosticators such as the grade and stage of the tumor. Therefore, there is need for methods that can be used as adjuncts to routine diagnostic and prognostic parameters. This study was designed to test the utility of the fluorescent in situ hybridization (FISH) method in identifying the molecular alterations, particularly the ones that have been detected with relatively high frequency in pancreas cancer. Formalin-fixed and paraffin-embedded tissues of 10 cases were enumerated for chromosome 7, 8, 17, 18, and 20 copy numbers by using alpha-satellite probes, and for c-myc by using a gene-specific probe. The number of signals per nucleus (reflecting chromosomal copy number and status of c-myc amplification) were counted in more than two areas containing 50-500 cells. Because of tumor heterogeneity, monosomy (loss of one chromosome copy) was defined arbitrarily as one signal in >25% of nuclei. C-myc amplification was defined as more than two gene copies in >20% of the cells. The most frequent signal losses were found in chromosomes 8 (four of 10 cases) and chromosome 17 (four of 10), followed by 20 (three of 10) and 18 (two of 10). No loss of chromosome 7 was detected. In contrast, gains in chromosome copy number were identified in only one of 10 tumors, which showed gain of both chromosome 7 and 18. Amplification of c-myc gene was detected in two of 10 cases, but neither of the two had aneuploidy for chromosome 8, where the c-myc gene is located. In addition, loss in c-myc signal was observed in one case that also showed loss of chromosome 8 copy number. FISH can be used to detect chromosomal changes in pancreatic cancer; abundance of lytic enzymes in this organ is not an impediment for the applicability of this technique. Therefore it can potentially be used in the future as an adjunct to the conventional diagnostic and prognostic markers. This study confirms that loss of chromosomes, particularly chromosomes 17 and 18, which carry the p53 and DCC genes, are common in pancreas cancer. Chromosome 20 is also frequently lost. In addition, in this study, alterations of chromosome 8, which is seen commonly in prostatic adenocarcinoma but has not been previously documented in pancreatic cancer, also was detected in five of 10 tumors. Furthermore, amplification of the c-myc gene, which is located in chromosome 8, was found in the two of the remaining five cases. Further studies are needed to confirm this high incidence of chromosome 8 and c-myc alterations and their possible role in the pathogenesis of pancreatic adenocarcinoma.