Endobronchial ultrasound-guided transbronchial needle aspiration (EBUS-TBNA) is an important minimally invasive technique in lung cancer staging and diagnosis.

Aim A retrospective analysis of EBUS-TBNA performance in patients with suspected malignancy referred between November 2009 and December 2013 to a UK tertiary EBUS centre.

Methods We reviewed consecutive EBUS-TBNA cases with CT/PET positive mediastinal/hilar nodes detected in suspected malignancy. EBUS-TBNA was performed as previously documented [Medford A. et al. QJM 2009; 102:859–864]. No rapid on-site cytology was available.

Results 186 patients with suspected malignancy (intra/extrathoracic and lymphoproliferative) were referred for EBUS-TBNA. Mean age was 66(31–87) with a 3:2 male:female ratio. In this group the sensitivity of EBUS-TBNA was 95.5%, accuracy 96.2% and negative predictive value (NPV) 80.6%. The prevalence of malignancy was 84.4%.

159 patients (85%) were referred with suspected lung cancer. In this group the sensitivity of EBUS-TBNA was 94.7%, accuracy 95.6% and NPV 78.8%. The prevalence of lung cancer was 83.6%.

The performance of EBUS-TBNA by lung cancer stage was also analysed, (See Table).

Conclusions This study shows high EBUS-TBNA diagnostic accuracy for all lung cancer stages. The NPV may have been reduced by the high prevalence of lung cancer in our cohort.

The majority of EBUS-TBNA procedures were performed for radiological stage III-IV disease where confirmation of disease would select those suitable for palliative rather than radical treatment.