As the use of endoscopic ultrasonography for mediastinal diagnosis and/or staging is widely spread in Belgium, the investigators aimed to determine the number of mediastinoscopies needed to detect one additional mediastinal lymph node invasion during routine clinical practice in the staging of potentially resectable clinical stage III non-small cell lung cancer.

Background : The observation made by the ASTER investigators might be criticized as all procedures were performed in highly experienced centers. To date, the number of mediastinoscopies needed to detect one additional N2/3 disease in the routine clinical practice of chest physician performing endosonography for mediastinal staging is unknown. The investigators therefore seek to answer whether a negative endosonography should routinely be followed by mediastinoscopy in day to day clinical practice.

Aim : As the use of endoscopic ultrasonography for mediastinal diagnosis and/or staging is widely spread in Belgium, the investigators aimed to determine the number of mediastinoscopies needed to detect one additional mediastinal lymph node invasion during routine clinical practice.

Setting : centers in Belgium with EBUS-TBNA and/or EUS-FNA experience in at least 20 patients agreed to participate and will include their patients.

Design : Prospective national observational multicenter study. All patients with clinical N2/3 disease based on CT and/or PET requiring invasive mediastinal staging will primarily undergo invasive mediastinal staging with endosonography (EBUS +/- EUS). A subsequent cervical mediastinoscopy will be performed in case no mediastinal lymph node involvement was found with endosonography. Local surgeons perform these procedures according to their institutional practice. Thoracotomy with mediastinal lymph node dissection will be the gold standard for invasive mediastinal staging, in case no mediastinal lymph node metastases were found during clinical staging including endosonography and mediastinoscopy.

Patients : The study will include 255 patients, based on the calculation of 15 consecutive patients in each participating center, in order to validate the ASTER data.

Primary endpoint : The number of mediastinoscopies needed to detect one additional N2/3.

Secondary endpoints : The number of mediastinal lymph nodes stations sampled with endosonography ; the median size of largest mediastinal lymph node sampled; characteristics of mediastinal nodal disease missed by endosonography.

Inclusion Criteria:

• Consecutive patients with (suspected) NSCLC in whom invasive mediastinal staging is required based on presence of ACCP group B mediastinal lymph nodes and/or FDG-PET positive (visual interpretation of FDG uptake in mediastinal nodes as present) mediastinal lymph nodes (either ACCP group B or ACCP group D) in lymph node stations 2, 4, 7, 8 or 9 (see Appendix).
• Potentially operable and resectable disease.
• Radically treated previous extrathoracic malignancies are allowed whenever the extrathoracic malignancy is considered in remission, and a primary parenchymal lung tumour is present.
• Provision of a written informed consent.

Exclusion Criteria:

• Previous cervical mediastinoscopy.
• Uncorrected coagulopathy.
• Former treatment for a lung cancer.
• Patient unable to give a written informed consent.
• Absence of a primary parenchymal lung tumour.
• Distant metastases (cM1 disease) after routine clinical work-up.
• Clinical N2/3 disease only based on suspected mediastinal lymph nodes in stations 5 or 6.
• Patients belonging to ACCP groups A and C based on CT scan.