EUS staging of upper GI malignancies: results of a prospective randomized trial

Kai Matthes, Brenna C. Bounds, Kerry Collier, Alexandra Gutierrez, William R. Brugge

Research output: Contribution to journalArticlepeer-review

26 Scopus citations


Background: Electronic 270° transverse-array EUS (TA-EUS) provides high-quality cross-sectional images but cannot guide FNA. Linear EUS (L-EUS) provides longitudinal images of malignancies and the ability to guide FNA. Objective: We conducted a prospective randomized comparison of TA-EUS and L-EUS for the staging of upper-GI (UGI) malignancies. Design: Forty-three patients underwent L-EUS immediately followed by TA-EUS (N = 27, 63%) or TA-EUS immediately followed by L-EUS (N = 16, 37%). Patients: Forty-three subjects (mean age, 64 years; 37 men) with an UGI malignancy (4 stomach and 38 esophageal) were evaluated with both TA-EUS and L-EUS. Interventions: Abnormal lymph nodes were sampled by FNA for cytology. Results: There was agreement on the T stage by linear and radial techniques in 38 of 43 subjects (88%). Twenty-seven of 43 patients (63%) had abnormal lymph nodes by linear or transverse-array imaging. L-EUS demonstrated 66 abnormal lymph nodes in 27 subjects (average of 2.4 nodes/subject). TA-EUS demonstrated 90 abnormal lymph nodes in 27 subjects (average of 3.3 nodes/subject, P = .009, compared with L-EUS). In 16 of the 27 subjects, an FNA was performed, which was positive in 13 cases (81%) and negative in 3 cases (10%) for malignancy. Conclusions: TA-EUS and L-EUS provide similar results of T staging of UGI malignancies. However, the number of abnormal lymph nodes detected by TA-EUS was more than by L-EUS. These findings suggest that radial or transverse-array EUS imaging should be the primary method for staging of UGI malignancies.

Original languageEnglish
Pages (from-to)496-502
Number of pages7
JournalGastrointestinal endoscopy
Issue number4
StatePublished - Oct 2006


Dive into the research topics of 'EUS staging of upper GI malignancies: results of a prospective randomized trial'. Together they form a unique fingerprint.

Cite this