Comparison of FNA and fine-needle biopsy for EUS-guided sampling of suspected GI stromal tumors

Background and Aims Subepithelial lesions are found in about 1% of all EGD procedures, and GI stromal tumors are a type of subepithelial lesion commonly encountered. Although the majority of subepithelial lesions are benign, GI stromal tumors have malignant potential, making a definitive diagnosis important. Currently, the criterion standard for the diagnosis of GI stromal tumors is EUS-directed FNA (EUS-FNA). The definitive diagnosis of GI stromal tumors relies on immunohistochemical staining, which depends on enough tissue being submitted to the pathologist. Achieving adequate tissue acquisition from suspected GI stromal tumors by EUS-FNA remains a limitation. Advancements in needle design, however, have improved tissue acquisition and therefore may improve the definitive diagnosis of GI stromal tumors by EUS-FNA. The goal of this study is to compare a fine-needle biopsy (FNB) needle (SharkCore, Medtronics) with FNA needles in definitively diagnosing suspected GI stromal tumors. Methods This is a retrospective, single-center study of consecutive patients with suspected GI stromal tumors by EUS characterization who underwent EUS-FNA or EUS-FNB. Results A total of 106 patients (53 men, mean [± standard deviation {SD}] age 62.19 ± 16.33 years) were included in the study undergoing EUS-FNA or EUS-FNB of suspected GI stromal tumors. The needle size that was used most often was 22 gauge in both groups. The average size of the lesions was 27.68 ± 15.70 mm; 71.7% of lesions were located in stomach, 19.8% in the esophagus, 5.7% in the duodenum, and 2.8% in the rectosigmoid colon. Ninety-one patients underwent EUS-FNA and 15 patients underwent EUS-FNB. Adequate tissue was procured, allowing immunohistochemical staining in 59 (64.8%) patients in the FNA group and 15 (100%) patients in the FNB group; P =.006. A diagnosis was reached by immunohistochemical staining in 48 (52.7%) patients in the FNA group and 13 (86.7%) patients in the FNB group; P =.01. Tissue was insufficient to make a cytologic diagnosis in 22 (24.2%) patients in the FNA group compared with none in the FNB group; P =.03. Adequate tissue was procured on the first pass of the FNB needle in the majority of patients (83.3%), whereas only 23.5% of patients had adequate tissue on the first pass by the FNA needle, with a median of 3 passes; P =.00. Tissue was insufficient to perform immunohistochemical staining, and thus a diagnosis could not be confirmed before surgery in 8 of the 34 surgical patients in the FNA group. Ten of 15 patients in the EUS-FNB group underwent surgery, all of whom were correctly diagnosed by FNB. There were no reported immediate adverse events or technical difficulties in either group. Conclusions EUS-FNB by using a SharkCore needle for suspected GI stromal tumors is technically similar and equally safe as FNA, with better tissue acquisition, which was achieved with fewer needle passes and an improved diagnostic yield by immunohistochemical staining.