Objective: To evaluate the diagnostic yield, the learning curve and the safety of endobronchial ultrasound-guided transbronchial needle biopsy (EBUS-TBNA) in mediastinal staging of patients with lung cancer.
Methods: Mediastinal staging was performed with EBUS-TBNA according to the Danish national guidelines in patients fulfilling one or more of the following criteria: (1) central tumour; (2) enlarged (>10 mm) mediastinal lymph nodes on computed tomography; or (3) positron emission tomography (PET)-positive mediastinal lymph nodes. The study period began in January 2006 when EBUS-TBNA was introduced in the department and ended in December 2007. All records were reviewed retrospectively. None of the four examiners had any previous experience with EBUS-TBNA or ultrasound when the study began. All examinations were performed under general anaesthesia. Patients without useful cytological material from the EBUS-TBNA were subjected to a supplementary standard cervical mediastinoscopy if the mediastinal lymph nodes were found to be enlarged (>10 mm), PET positive or if the examiner was insecure of the result of the EBUS-TBNA. Patients with mediastinal lymph node involvement, detected by EBUS-TBNA or standard cervical mediastinoscopy, were referred to oncological treatment, while those without mediastinal lymph node involvement underwent--if they were otherwise eligible for surgery--resection and systematic lymph node sampling either by thoracotomy or by video-assisted thoracoscopy. Final mediastinal staging was defined as positive if mediastinal lymph node involvement was detected by EBUS-TBNA, standard cervical mediastinoscopy or surgery, or defined as negative otherwise.
Results: A total of 157 patients were included in the study. N2/N3 disease was found in 67 patients (42.6%). EBUS-TBNA missed the mediastinal spread in 10 patients. Five of the ten patients had lymph node metastases in station 5, 6 or 8--out of reach of EBUS-TBNA or standard cervical mediastinoscopy. EBUS-TBNA had a sensitivity of 0.85 (0.74-0.93) and a negative predictive value of 0.90 (0.82-0.95). No complications occurred from EBUS-TBNA. The number of supplementary standard cervical mediastinoscopies decreased significantly in the study period.
Conclusion: The results of this study suggest that staging of the mediastinum with EBUS-TBNA is safe and easy to learn--even without previous experience with ultrasound. The diagnostic yield of EBUS-TBNA is in accordance with the yield of standard cervical mediastinoscopy reported in the literature. We do not find any indications in the present study of the recommended necessity for mediastinoscopy in all EBUS-TBNA-negative patients.