Education and experience improve the performance of transbronchial needle aspiration: a learning curve at a cancer center.

STUDY
OBJECTIVES: Transbronchial needle aspiration (TBNA) is an indispensable part of the pulmonologist's armamentarium, although it continues to be woefully underutilized despite its demonstrated safety and usefulness. We herein review our experience with the procedure.
MATERIALS AND METHODS: All TBNAs were conducted according to standard techniques using 21-gauge cytology needles or 19-gauge histology needles connected to a flexible bronchovideoscope. All procedures were conducted at a 180-bed cancer center, and results were analyzed retrospectively. The mediastinum and hilar lymph node mapping system proposed by Wang was followed exclusively.
RESULTS: From September 1999 to March 2003, inclusively, 90 of 549 patients undergoing diagnostic bronchoscopy were selected for TBNA: 66 patients for hilar-mediastinal lymphoadenopathies, and 24 patients for submucosal and/or peribronchial lesions. A total of 87 hilar-mediastinal lymph node stations were sampled, with a mean of 2.2 needle passes for each. Seventy-eight patients revealed a malignant diagnosis. TBNA provided positive results for malignancy for 59 patients. Sarcoidosis, mediastinal bronchogenic cyst, and mediastinal tuberculous adenitis were identified for another three patients, respectively. The diagnostic yield was 68.2% (45 of 66 patients) for hilar-mediastinal lesions, and 70.8% (17 of 24 patients) for submucosal and peribronchial lesions. The sensitivity was 75% (45 of 60 patients) for hilar-mediastinal lesions, and 80.9% (17 of 21 patients) for submucosal and peribronchial lesions. The overall accuracy of the procedure for returning a correct diagnosis was 75.9% (66 of 87 patients). Higher yields for patients featuring small-cell lung cancer were noted. Fifteen patients presenting mediastinal lesions attained to a specific pathologic diagnosis using TBNA despite normal-appearing airways. TBNA was the exclusive means of diagnostic sampling for 27 patients. Twenty-two patients had previously undergone a nondiagnostic bronchoscopy at other hospitals. Diagnosis and mediastinal staging was accomplished in one procedure for 19 patients exhibiting non-small cell lung cancer. The number of TBNA procedures performed per unit time rose steadily during the test period. The TBNA yield and sensitivity for the detection of hilar-mediastinal lymphoadenopathies increased significantly (p = 0.03) during the study period. The presence of the cytotechnologist during the TBNA procedure provided direct, immediate feedback pertaining to the quality of specimens acquired. With such rapid on-site examination of TBNA-derived specimens, there was a trend with borderline significance (p = 0.06) toward a decreasing frequency of inadequately acquired tissue specimens when using this technique.
CONCLUSIONS: TBNA performance was able to be improved over time. Increased specimen yield and sampling sensitivity over a 43-month period suggested the impact of enhanced training interventions and experience. Rapid on-site examination was also indispensable for the promotion of diagnostic accuracy. The progressive acquisition of skills as regards the use of cytology needles for TBNA purposes should precede the use of a histology needle for such biopsy purposes. For selected cases, the use of the 19-gauge histology needle increased the diagnostic yield of TBNA. It is to be hoped that increased experience with the TBNA technique and focused education regarding its performance will enhance its utilization by bronchoscopists and the spread of its acceptance.