Distinguishing desmoplastic melanoma (DM) from scar tissue on routine microscopy can be difficult, especially in re-excision specimens, and S100 immunohistochemistry has been recommended as a useful adjunct. The purpose of this study is to evaluate the extent and nature of S100 positivity in scars. In this study, formalin-fixed paraffin archival tissues were evaluated with immunohistochemistry. Ten re-excision specimens of previously biopsied nonnevomelanocytic lesions were immunostained with the S100 and CD57 (Leu 7) antibodies. In 9 of the 10 cases, the scars contained S100-positive spindle cells, but there were no cases with CD57+ cells. Ten re-excised atypical nevi and 10 re-excised melanomas were also immunostained for the S100 protein, and all 20 cases contained S100-positive spindle cells within the scars. There was a trend toward quantitatively more S100-positive spindle cells in these nevomelanocytic re-excisions. To evaluate the nature of the spindle cells, scars from two of the nonnevomelanocytic re-excisions were further analyzed utilizing immunostains for glial fibrillary acidic protein, HMB-45, Melan-A, CD1a, factor XIIIa, and neuron specific enolase. In both scars, neuron specific enolase diffusely stained the fibroblast population, but the remaining immunostains were negative in the scar. The presence of S100-positive spindle cells in scars represents a potential diagnostic pitfall, particularly in the evaluation of re-excision specimens of DM.