Amplification of MYCN in neuroblastoma is associated with a poor prognosis. However, methods for estimating the number of MYCN genes based on pooled cells do not address copy number heterogeneity at the cell level and can underestimate or even miss amplification. MYCN copy number can be directly assessed by fluorescence in situ hybridization, but evaluation of tissue histology is next to impossible. We have used a chromogenic method for in situ hybridization (CISH) that enables determination of MYCN copy number using routine light microscopy on routinely processed paraffin sections. Of 41 cases studied, CISH identified 100% of the 18 cases that were determined to be amplified by other techniques and was more sensitive than Southern blotting or quantitative DNA polymerase chain reaction. Because the technique evaluates individual tumor cells, heterogeneity of MYCN copy number was apparent from cell to cell. When defined as 50% or greater variation in copy number between cells in amplified tumors, almost 30% of cases were scored as heterogeneous. Heterogeneity reflects different tumor clones and its role has likely been under-recognized and underestimated in neuroblastoma biology. CISH will provide a valuable tool to assess this phenomenon in conjunction with other morphologic parameters in neuroblastoma specimens, to further our understanding of the biology of this childhood tumor.