RESUMO
OBJECTIVES: The aims of the study were to assess in patients with large gastric folds at endoscopy and endoscopic biopsies that tested negative for malignancy: (i) the predictive variables of malignancy in endoscopic ultrasonography (EUS), (ii) the impact of EUS. METHODS: Patients with this condition evaluated with radial EUS (following the standard study protocol of our Unit in these patients) during a 5-yr period were included. Nine variables were evaluated as possible predictors of malignancy (chi2, t-test, and logistic regression procedure) using pathology results, or a minimum of 2 yr of follow-up as gold standard. Likelihood ratios (LR), pre-EUS, and post-EUS probabilities for the diagnosis of malignancy were used to assess the clinical impact of the technique. RESULTS: Sixty-one patients were included (40 benign and 21 malignant). Predictive factors of malignancy were as follows: thickened gastric wall, thickened deep layers, impaired gastric distension, loss of the wall structure, and presence of ascites or lymph nodes, whereas the enlargement of superficial layers was a predictor of a benign condition. The enlargement of deep layers, as assessed by EUS was the only independent predictive factor for malignancy. When using this parameter, the pre-EUS probability of malignancy in our series (34%) increased up to 95% when EUS suggested malignancy (positive LR = 45), whereas it decreased to 4.7% when EUS precluded this diagnosis (negative LR = 0.102). CONCLUSIONS: (i) The enlargement of deep layers is the only independent predictive factor for malignancy in patients with large gastric folds at endoscopy and biopsies testing negative for malignancy and (ii) EUS has a high clinical impact in these patients.