Background: EUS-guided FNA has been well documented to aid in the diagnosis of subepithelial lesions by providing cytologic material. Studies to date evaluating the sensitivity of EUS-FNA for the diagnosis of GI stromal cell tumors (GIST) have been small, and few have relied on surgical histologic diagnosis as the reference standard.
Objective: Our purpose was to determine the diagnostic yield and sensitivity of EUS-FNA for the diagnosis of GIST and to identify EUS features of GIST that are predictive of the ability to obtain adequate tissue by EUS-FNA.
Design: All patients with histologically confirmed, c-kit-positive GIST who underwent EUS-FNA from 1998 to 2006 were reviewed. EUS images were examined for mass size, shape, location, wall layer, heterogeneity, echogenicity, cystic spaces, lobulation, ulceration, and central umbilication. Needle gauge, number of needle passes, and presence of a cytologist during the EUS-FNA were recorded.
Results: A total of 37 patients (29 with diagnostic FNA cytology; 8 with nondiagnostic cytology) met the inclusion criteria. The diagnostic yield and sensitivity of EUS-FNA cytology for the diagnosis of GIST was 78.4% (29/37). The sensitivity was 84.4% (27/32) for GISTs located in the stomach, but poor for lesions located in the duodenum because none of these tumors yielded diagnostic cytology (n = 3). An increase in size up to 10 cm, round/oval shape, and identification of the origin of GIST within a specific sonographic wall layer were statistically significant in their ability to predict adequate tissue yield.
Conclusions: The sensitivity of EUS-FNA cytology for the diagnosis of GIST is 78.4% and is influenced by size, location, shape, and layer of origin.