Endobronchial ultrasonography for mediastinal and hilar lymph node metastases of lung cancer

Study objectives: Conventional radiologic procedures are frequently unreliable in the diagnosis of mediastinal and hilar lymph node metastases of lung cancer. In order to improve diagnostic accuracy, we performed endobronchial ultrasonography (EBUS) during bronchofiberscopic examinations of patients with lung cancer. Methods and patients: To evaluate mediastinal and hilar lymph node metastases, EBUS was performed prospectively using a radial scanning probe of 20 MHz through a bronchofiberscope. Results: We observed hilar lymph nodes (10R, 11R superior, 11R inferior, 12R, 10L, 11L, 12L) in 20 of 37 patients who underwent EBUS, and we could clearly, identify whether direct invasion of the pulmonary artery by a lymph node bad occurred. Of the 27 patients who shoved no hilar lymph nodes on chest CT scan, lymph node swellings < 10 mm or ≥ 10 mm in diameter were identified by, EBUS in 9 patients and 2 patients, respectively. Interestingly, EBUS also revealed that the pulmonary artery was directly, invaded by an interlobar lymph node < 10 mm in diameter in one patient. In most patients, lymph node 7 was easily identified and was clearly differentiated from the surrounding esophagus, vessels, and mediastinal fat tissue by EBUS. However, fused lymph nodes or lymph nodes with low central density when visualized by chest CT scan were occasionally observed as independent lymph nodes by EBUS. When compared with the pathologic diagnosis of lymph node metastasis in 16 patients who underwent surgery, the most specific and sensitive method for identifying lymph node metastases were EBUS alone (92%) and EBUS in combination with CT scan (100%), respectively. The overall accuracy of EBUS was 94% for the diagnosis of direct invasion of the pulmonary, arteries by a hilar lymph node. Conclusions: EBUS in combination with conventional radiologic tools may contribute to improved staging, especially in surgical cases with hilar lymph node metastases.