Diagnostic yield of N3 hilar staging by endobronchial ultrasonography (EBUS) in lung cancer.

Introduction: Systematic lung cancer staging with EBUS has proven to be equivalent to cervical mediastinoscopy. Nevertheless, in the daily practice it is common to explore and sample negative PET-CT hilar N3 lymph nodes (LN). This study aims to explore if there is enough evidence to support this clinical practice. Material and Methods: Retrospective study from our database including 1,013 explorations over the last 5 years. Including criteria were patients with lung cancer staged by PET-CT and EBUS-TBNA. Mediastinal and hilar N3 LN with a short axis ≥ 5 mm were sampled assessed by rapid on site evaluation. A single nuclear medicine expert reviewed blindly all PET-CT scans and determined the SUVmax of every LN. Those that were ≥ 5 SUVmax by PET-CT and/or ≥ 10mm in short axis by EBUS were considered abnormal. Results: 87 patients were included. The final histopathology diagnoses were adenocarcinoma (46%), squamous cell carcinoma (39%) and other histology (14%). None of the 61 normal hilar and normal mediastinum N3 LN, and none of the 7 normal N3 hilar LN with abnormal mediastinal LN (3 by PET-CT, 3 by EBUS and 1 for both) resulted positive for lung cancer. Of the 19 patients with abnormal N3 hilar LN (6 by PET-CT, 8 by EBUS and 6 for both) malignancy was found in 16.7%, 25% and 60% for both techniques, respectively. Conclusions: In absence of abnormal N3 hilar LN (PET: SUVmax<5; EBUS<10mm in short axis) it seems there is not enough evidence to sample them, regardless of N3 mediastinal status.