Abstract
Introduction: The aim of the study was to assess the diagnostic yield of transoesophageal endoscopic ultrasound-guided needle aspiration (EUS-NA) in lung cancer (LC). Material and methods: Real time EUS-NA was performed under local anaesthesia and sedation in consecutive LC patients. All negative EUS-NA results in NSCLC patients were verified by transcervical extended bilateral mediastinal lymphadenectomy (TEMLA). Results: In 146 patients there were 206 biopsies performed in lymph node stations: subcarinal (7):124, left lower paratracheal (4L):70, paraoesophageal (8):9 and pulmonary ligament (9):3. A mean short axis of punctured node was 10 ± 6.3 (95% CI) mm. Lymph node biopsy was technically successful in 95.6% and was diagnostic in 40.1% of LC patients. In NSCLC staging, the sensitivity of EUS-NA calculated on the per-patient basis was 85.5%, specificity 100%, accuracy 93.6% and negative predictive value (NPV) 89.7% in stations accessible for EUS-NA, but in all mediastinal stations it was 70.7%, 100%, 84.3% and 74.7, respectively (p = 0.009). The sensitivity of EUS-NA in NSCLC staging patients, calculated on the per-biopsy basis was 88.6%, specificity 100%, accuracy 95.4% and NPV 91.4%. A diagnostic yield of EUS-NA on the per-biopsy basis was higher for station 4L than 7, but the difference was not significant (χ2 p = 0.4). Conclusions: The diagnostic value of EUS-NA in LC is high. In NSCLC staging EUS-NA is insufficient and should be complemented by other invasive techniques, especially those that give access to the right paratracheal region.