Background Endoscopic ultrasound (EUS) can be used to guide the therapeutic plan for patients with gastric adenocarcinoma (GAC), but data on its use and accuracy remain poorly defined. We sought to define the use of EUS, as well as characterize the diagnostic accuracy of EUS among patients with GAC.
Study Design We identified 960 patients who underwent resection of GAC between 2000 and 2012 from 7 major academic institutions participating in the US Gastric Cancer Collaborative. Clinicopathologic and EUS data were collected and analyzed using chi and kappa statistics.
Results Of 960 patients, 223 (23.2%) underwent evaluation with preoperative EUS. Among patients who underwent EUS, 74 (33.2%) received neoadjuvant chemotherapy; 149 (66.8%) proceeded directly to resection. Among patients who did not receive neoadjuvant therapy and received curative intent gastric resection, the EUS T classifications were T1 (33.3%), T2 (35.6%), T3 (18.9%), T4 (12.1%) and the N classifications were N0 (68.1%) and N ≥ 1 (31.9%). In contrast, when tumor stage was based on the final surgical specimen, there was a higher proportion of cases with more advanced T stage (T1, 36.4%; T2, 14.4%; T3, 23.5%; T4, 25.7%) and N stage (N0, 51.3%; N ≥ 1, 48.7%). The agreement of preoperative EUS compared with surgical staging for T (kappa = 0.28, p < 0.001) and N (kappa = 0.33, p < 0.001) classification was only fair.
Conclusions Less than one-quarter of patients with GAC underwent preoperative EUS staging. In patients who did not receive preoperative chemotherapy, tumor stage on EUS often did not correlate with T stage and N stage on final pathologic analysis. Endoscopic ultrasound should be combined with other staging modalities to optimize staging of patients with GAC.