I congratulate Dooms et al1 on their important study recently published in CHEST (January 2015) that investigated the role of endosonography for mediastinal nodal staging in patients with operable and resectable lung cancer and suspicion of N1 involvement based on CT and PET imaging. To my surprise, the sensitivity of endosonography to detect N2 disease was only 38%, with a negative predictive value of 81%. First, the data show that the sensitivity of endosonography in subcentimeter fluorodeoxyglucose-PET nonavid nodes is inferior in comparison with the sensitivity of enlarged and/or fluorodeoxyglucose-PET avid nodes (94%).2 Second, of key interest are the 14 patients whose conditions were staged as false-negative. In only one of these patients, endobronchial ultrasound (EBUS) was followed by endoscopic ultrasound (EUS) or EUS performed with the EBUS scope (EUS-B). Importantly, in eight of the 14 patients, the missed metastases (located in stations 4L, 7, and 8) were located well within reach of EUS-B, and in two more patients, nodes in station 5 could have been well visualized and possibly sampled. In this scenario, as stated by the authors, a sensitivity of > 70% theoretically could have been reached.