Negative EBUS-TBNA Predicts Very Low Prevalence of Mediastinal Disease in Staging of Non-Small Cell Lung Cancer.

Background: Confirmation of mediastinal disease (N2/3) in non-small cell lung cancer (NSCLC) generally precludes curative surgical management. Endobronchial ultrasound-guided transbronchial needle aspiration (EBUS-TBNA) has become a routine first test in mediastinal staging of NSCLC; however, it remains unclear whether a negative EBUS-TBNA should be followed by mediastinoscopy before proceeding to surgery. Understanding the prevalence of metastases in lymph nodes with benign findings on EBUS-TBNA will inform decision making following negative EBUS-TBNA.

Methods: We examined a retrospective cohort of patients who underwent EBUS-TBNA before resection with mediastinal lymph node sampling for NSCLC between December 2009 and June 2014 in 3 hospitals in Melbourne, Australia. All patients had integrated positron emission tomography/computed tomography (PET/CT) before EBUS-TBNA.

Results: Eighty-two matched mediastinal lymph node stations were sampled in 57 patients by both EBUS-TBNA and surgical resection, 47 nodes in patients staged cN0/1 by PET/CT and 35 nodes in patients staged cN2/3. All patients had a negative EBUS-TBNA. Four malignant nodes were identified surgically (4.9% of lymph nodes). The mean size of malignant deposits was 5.5 mm. Per-node negative predictive value was 78/82=0.95. All malignant nodes were located in patients with moderate-high risk disease (cN2/3), giving a disease prevalence in cN2/3 patients of 11%, and 0% in cN0/1. In patients staged cN2, per-node NVP was 0.89.

Conclusion: The prevalence of mediastinal nodal disease following negative EBUS-TBNA is very low, at 4.9%. The per-node NVP of EBUS-TBNA is 0.95, decreasing to 0.89 in moderate-high risk patients. We suggest that a negative EBUS-TBNA of mediastinal nodes does not need to be confirmed by mediastinoscopy of those nodal stations, regardless of PET/CT findings.