Impact of Endoscopic Ultrasound (EUS) on Management of Patients with Large Gastric Folds and Endoscopic Biopsies Negative for Malignancy (LGF-NB)

Abstract

Background: The management of patients with large gastric folds is a challenge in clinical practice. Moreover, it is known that endoscopic biopsies negative for malignancy do not exclude this possibility. Aim: 1) To evaluate the clinical impact of EUS in the diagnosis of malignancy in patients with LGF-NB and 2) To assess the EUS predictive features of malignancy. Patients and methods: We included all patients with LGF-NB referred for EUS evaluation between March 97 and December 02. EUS was performed with a radial echoendoscope (Olympus GF UM20). 12 variables were registered from endoscopic and EUS findings. We used pathology of resected specimens or clinical follow-up (range 12-68 months) as gold standard. Likelihood ratios (LR) and a logitic regression model were used to assess the clinical impact and independent EUS predictive factors for malignancy respectively. Results: we included 64 patients (35M/29F; mean age 58+17). Definitive diagnosis disclosed benignity in 42 patients (66%) and malignancy in 22 (34%). The Table below shows the performance characteristics for EUS compared with upper endoscopy for diagnosis of malignancy. The pre-test probability of malignancy in our series (34%) increased up to 91% when EUS suggested malignancy (positive LR=19) whereas decreased to 2,6% when EUS precluded this diagnosis (negative LR=0.053). Predictive factors in the univariate analysis were: thickened wall, involvement of deep layers, impaired distensibility, and the presence of ascites or lymph nodes. The only independent predictive factor of malignancy in the multivariate analysis was the presence of lymph nodes. Conclusions: 1) EUS is a very accurate technique with a high clinical impact for diagnosis of malignancy in patients with LGF-NB 2) The only independent predictive factor for malignancy is the presence of lymph nodes.