A Limited Immunocytochemical Panel for the Distinction of Subepithelial Gastrointestinal Mesenchymal Neoplasms Sampled by Endoscopic Ultrasound–Guided Fine-Needle Aspiration
American Journal of Clinical Pathology
A b s t r a c t We studied the use of immunocytochemical analysis with material procured by endoscopic ultrasoundguided fine-needle aspiration (EUS-guided FNA) for the diagnosis of subepithelial intramural gastrointestinal (GI) mesenchymal neoplasms (SIGIMNs). We identified all EUS-guided FNA specimens of SIGIMNs that had undergone immunocytochemical analysis. Results were compared with follow-up histologic diagnoses. There were 95 aspirates that were diagnosed as GI mesenchymal tumors (GI
... mal tumors (GI stromal tumors [GISTs], n = 46), leiomyomas (n = 38), peripheral nerve sheath tumors (n = 5), and other neoplasms by cytologic examination. Immunoreactivity with antibodies to CD117 always predicted GIST at follow-up; 15 of 16 cases immunoreactive with antibodies to CD34 were found to be GISTs at follow-up. Strong immunoreactivity with antibodies to smooth muscle actin or desmin usually predicted a leiomyoma at follow-up aside from a single glomus tumor and a case with apparent nonneoplastic smooth muscle contaminant. When sufficient material is present, immunocytochemical analysis used with material obtained by EUS-guided FNA is highly predictive of final pathologic diagnosis. Endoscopic ultrasound-guided fine-needle aspiration (EUS-guided FNA) has been shown to be a safe and effective method for the sampling of gastrointestinal (GI) and adjacent lesions, including subepithelial intramural GI mesenchymal neoplasms (SIGIMNs). 1-8 The distinction of the various types of mesenchymal neoplasia can be important because these lesions may have different prognoses and may be amenable to different treatments (eg, gastrointestinal stromal tumors [GISTs] can be treated with tyrosine kinase inhibition). Although retrospective studies have described the cytologic features and immunocytochemical profiles of the various lesions, we are not aware of a large study that investigated the positive predictive value of cytologic diagnosis coupled with immunocytochemical analysis. 1,3-9 This study reviews our experience with the use of immunocytochemical analysis applied to samples of SIGIMNs obtained by EUSguided FNA. Materials and Methods The cytologic files of Hennepin County Medical Center, Minneapolis, MN, were searched for all SIGIMNs sampled by EUS-guided FNA that had sufficient cell block material to allow for ancillary immunocytochemical studies. The original interpretations of the immunocytochemical results were used for all cases. With almost all cases, results for each stain were recorded as positive or negative. Positive staining was usually described as strong and diffuse immunoreactivity, whereas negative staining was described as "nonreactive." The rare cases that were described as having weak and focal staining were considered negative for the purposes of this study.