Degenerate oligonucleotide-primed polymerase chain reaction (DOP-PCR) is often applied to small amounts of DNA from microdissected tissues in the analyses of chromosomal copy number with comparative genomic hybridization (CGH). The sensitivity and specificity in CGH analyses largely depend on the unbiased amplification and labeling of probe DNA, and the sensitivity and specificity should be high enough to detect one-copy changes in aneuploid cancer cells when accurate assessment of chromosomal instability is needed. The present study was designed to assess the effects of DOP-PCR and labeling method on the sensitivity of metaphase- and array-based CGHs in the detection of one-copy changes in near-tetraploid Kato-III cells. By focusing on several chromosomes whose absolute copy numbers were determined by FISH, we first compared the green-to-red ratio profiles of metaphase- and array-based CGH to the absolute copy numbers using the DNA diluted with varying proportions of lymphocyte DNA, with and without prior DOP-PCR amplification, and found that the amplification process scarcely affected the sensitivity but gave slightly lower specificity. Second, we compared random priming (RP) labeling with nick translation (NT) labeling and found that the RP labeling gave fewer false-positive gains and fewer false-negative losses in the detection of one-copy changes. In array CGH, locus-by-locus concordance between the DNAs with and without DOP-PCR amplification was high (nearly 100%) in the gain of three copies or more and the loss of two copies or more. This suggests that we could pinpoint the candidate genes within large-shift losses-gains that are detected with array CGH in microdissected tissues.