A tale of two clones: Caldesmon staining in the differentiation of cutaneous spindle cell neoplasms

Background We sought to compare the sensitivity and specificity of 2 different caldesmon antibodies in differentiating leiomyosarcoma from other cutaneous spindle cell neoplasms. Methods Representative cutaneous spindle cell neoplasms were identified, including leiomyosarcoma, atypical fibroxanthoma, dermatomyofibroma and spindle cell squamous cell carcinoma. Immunohistochemistry was performed with antibodies directed toward caldesmon, smooth‐muscle actin (SMA) and desmin. Sensitivity and specificity were calculated using grades from 3 independent observers. Results The sensitivity of caldesmon (Ventana) was 100% (95% CI 78.2%‐100%) and the specificity was 8.3% (2.8%‐18.4%). Because this stain appeared to be non‐specific, additional testing was performed on the same set of specimens using a second caldesmon clone (H‐caldesmon, Dako), which had a sensitivity of 53.9% (25.1%‐80.8%) and specificity of 96.6% (88.1%‐99.6%). The sensitivity and specificity of SMA were 85.7% (57.2%‐98.2%) and 84.5% (72.6%‐92.7%), respectively. The sensitivity of desmin was 53.3% (26.6%‐78.7%) with a specificity of 100% (94.0%‐100%). Conclusions The Ventana caldesmon clone is not specific to smooth muscle, a potential pitfall to laboratories using this clone. The staining pattern, sensitivity and specificity of the Dako H‐caldesmon antibody clone are similar to results from prior studies. The sensitivity and specificity of the Dako clone support its use in smooth muscle identification as an additional marker in challenging cases.