Yield of transbronchial needle aspiration in diagnosis of mediastinal lesions.

Abstract

STUDY OBJECTIVES: To determine the transbronchial needle aspiration (TBNA) yield for procedures performed by fellows in training, and the predictors of positive TBNA yield at our center. DESIGN: Retrospective chart review. SETTING: A tertiary teaching hospital. PATIENTS: One hundred seventy patients who underwent fiberoptic bronchoscopy with TBNA of mediastinal lesions from January 1991 to July 1999. RESULTS: Final diagnoses were available for 166 patients. TBNA was diagnostic in 104 patients (61%) and nondiagnostic in 66 patients (39%). Of 170 cases, 123 patients (72%) had malignancies, 30 patients (18%) had benign disease, and 13 patients (8%) were normal. Of 123 malignancies, 85 patients (69%) had a positive result by TBNA. Of 30 cases with benign disease, 11 patients (37%) had positive TBNA findings. Eight of 13 patients (62%) with a normal diagnosis had diagnostic TBNA (normal lymphoid tissue). There were statistically significant correlations between TBNA result and cell type of the lesion (p < 0.001), size of the lesion (p < 0.05), and type of malignancy (small cell carcinoma more than non-small cell carcinoma more than lymphoma, p < 0.05). We did not find any significant difference for aspiration yield between carinal and tracheal sites (p > 0.05). Logistic regression analysis indicated that the presence of malignancy is the major determinant of TBNA yield (p = 0.009). In addition, lesion size does affect yield after being adjusted for diagnosis (one-sided p = 0.04). CONCLUSIONS: TBNA is a minimally invasive diagnostic technique with a high yield, even in hands of less experienced operators. Malignancy, lesion size, and type of malignancy are major determinants of TBNA yield.