To determine the frequency of c-Kit staining in desmoids and optimize an assay for clinical use, we stained 19 desmoids from various sites at various dilutions with 2 commonly used rabbit polyclonal, anti-c-Kit antibodies (A4502, DAKO, Carpinteria, CA; C-19, Santa Cruz Biotechnology, Santa Cruz, CA), with and without heat-induced epitope retrieval (HIER) in citrate buffer. Approdpriate external and internal control samples were evaluated for each test condition. At dilutions of 1:50 both antibodies stained substantial numbers of desmoids: with/without HIER, A4502, 89%/63%; C-19, 37%/74%. The staining was cytoplasmic without cell membrane accentuation. However, background stromal staining and nonspecific staining of endothelium and smooth and striated muscle were problematic with both antibodies at 1:50. At higher dilutions, C-19 stained no desmoid; however, diminished staining of external and internal control samples made it unreliable. A4502 similarly stained many fewer desmoids at higher dilutions. However, it retained strong staining of both external and internal control samples and showed much less nonspecific staining. Best results were achieved at 1:250 without HIER; only weak focal staining was present in 1 desmoid. With a simple immunohistochemical method optimized for clinical use, desmoid can be regarded as a c-Kit-negative tumor.