Endoscopic ultrasonography (EUS) incorporates a high-frequency ultrasound transducer into the tip of an endoscope to provide high-resolution images of the gastrointestinal (GI) wall and structures in close proximity to the GI tract above and below the diaphragm. Transesophageal EUS imaging provides high-resolution images of the posterior mediastinum and multiple studies have considered the utility of EUS for detection/staging of mediastinal lymphadenopathy in lung cancer. Before the development of interventional EUS techniques allowing EUS-guided fine-needle aspiration (FNA), the focus was on studying echo features of lymph nodes to predict malignant invasion. Lymph nodes that were larger than 1 cm, round, hypoechoic, and with distinct margins were considered to be malignant.1 These EUS defined criteria that were initially described for GI cancers (eg, esophageal) were later applied to mediastinal lymph nodes in lung cancer. EUS echo features of mediastinal lymph nodes in lung cancer were shown in one study to have accuracy of 84% compared with computerized tomographic (CT) scan accuracy of 49%.2 However, the reliance on EUS echo features to diagnose malignant lymph node invasion in lung cancer has problems of interobserver variability, lack of standardization of frequencies used to study the lymph nodes, and lack of uniform criteria to label a lymph node as hypoechoic or with sharp, distinct margins.3 In addition, it has been shown in a study by Bhutani et al3 that although the presence of echo features described above could predict malignant invasion about 80% of the time, only 25% of nodes that had malignant invasion had all 4 echo features. This study also compared the accuracy of echo features of lymph nodes in patients with esophageal, pancreatic, and lung cancer to EUS-guided FNA. The authors found that EUS-guided FNA was a more reliable method for predicting lymph node invasion than echo features. Transesophageal EUS-guided real-time FNA of mediastinal lymph nodes has become a clinically useful minimally invasive method for detecting malignant lymph node invasion.3–7 EUS-FNA is well suited for evaluation of the mediastinum with lymph nodes in the subcarina, aorto-pulmonary window, para-esophageal area, and para-aortic area being the most amenable locations for EUS-guided FNA. One study is reporting 96% accuracy of EUS-FNA in patients with known nonsmall cell lung cancer who have enlarged mediastinal lymph nodes on CT scan.2 When a primary lung mass is seen on CT with enlarged mediastinal lymph nodes in the mediastinum, EUS-guided mediastinal FNA can provide a primary diagnosis and simultaneous staging information.6,7 Recent data suggest that EUS-guided FNA may detect advanced mediastinal disease and avoid un-necessary surgical exploration in a significant number of patients who have no evidence of enlarged mediastinal lymph nodes on CT scan.8–10 In a series of 104 patients the sensitivity, specificity, positive predictive value, negative predictive value, and accuracy of EUS-FNA were 92.5%, 100%, 100%, 94%, and 97%, respectively. EUS-FNA was more accurate and had a higher positive predictive value than the positron emission tomography (PET) or CT (P<0.001) scan in confirming cancer in the posterior mediastinal lymph nodes.11 EUS with FNA has also been found in some studies to be useful in sampling the left adrenal gland in nonsmall cell lung cancer.12,13 A few cost effectiveness and decision modeling studies have also been published showing EUS-FNA to be cost effective in staging nonsmall cell lung cancer.14–17 EUS is also useful in evaluating mediastinal lymphadenopathy of unknown origin with no primary lung mass on CT.7 If EUS with FNA can safely sample lymph nodes in the mediastinum, can it also potentially sample centrally located primary lung masses with acceptable accuracy and safety? In this issue of the J Clin Gastroenterol, Hernandez et al18 present data on a retrospective study on EUS-FNA as the primary diagnostic modality in centrally located lung cancers in 17 patients. The authors report an accuracy of 100%. A few other recent series have also investigated this potential utility of EUS. Varadarajulu et al19 in a retrospective series of 18 patients with primary lung masses abutting the esophagus showed a diagnostic yield of 100% with no complications. In the series by Annema et al,20 EUS-FNA yielded a diagnosis of lung cancer in 31 of 32 patients (97%) with centrally located lung tumors after a nondiagnostic bronchoscopy. These diagnoses were made with no procedure-related complications. If EUS with FNA can sample lymph nodes in the mediastinum, centrally located primary lung masses and the left adrenal, can it be applied as one of the first tests in patients with lung cancer? A recent prospective study by Singh et al21 applied EUS as a first test for diagnosis and staging of lung cancer. Consecutive patients with a lung mass on CT scan were enrolled for EUS and results were compared with CT and PET scans. Among 113 subjects with lung cancer, EUS was performed as a first test (after CT) in 93 (82%). EUS-FNA provided a tissue diagnosis in 70% of cases. EUS-FNA, CT, and PET detected metastases to the mediastinal lymph nodes with accuracies of 93, 81, and 83%, respectively. Metastases to lymph nodes at the celiac axis, a predictor of poor survival in nonsmall cell lung cancer, were found in 11% of cases. The yield of EUS-FNA for detection of celiac axis metastases was 100% and it was 50% for CT (P<0.05). EUS-FNA has some advantages for imaging and sampling of the mediastinum above and beyond the ability of cervical mediastinoscopy. Lower para-esophageal lymph nodes (stations 8 and 9) are difficult if not impossible to reach without thoracoscopy or thoracotomy. Endoscopy with EUS-FNA allows for routine evaluation of these lymph node stations in a minimally invasive fashion. However, EUS has the disadvantage of not being able to image anterior to the trachea and left/right bronchi due to air in the tracheobronchial tree interfering with transesophageal EUS imaging. Therefore, up until recently one has had to rely on more invasive traditional techniques such as mediastinoscopy and video-assisted thoracoscopic surgery when tissue sampling in this region is required. With the recent development of endobronchial ultrasound (EBUS)-guided real time transbronchial needle aspiration (TBNA) one is now able to image and perform ultrasound-guided FNA anterior to the trachea/bronchi in a minimally invasive fashion as well. EBUS also affords access to the interlobar and intralobar lymph nodes that are otherwise not easily accessible by standard mediastinoscopy techniques. Therefore, EBUS and EUS are complementary techniques that can allow nearly complete imaging and staging of the mediastinum in patients with lung cancer. In a recent study by Rintoul et al22 the authors reported their initial experience of using a prototype EBUS probe with TBNA of paratracheal and hilar lymph nodes in 20 patients selected by CT scan. In 7 cases, sequential EUS was used to assess posterior mediastinal lymph nodes. Cytology results confirmed N2/N3 nodal disease in 11 out of 18 EBUS-TBNA cases. EUS provided additional information in all cases. Sensitivity, specificity, and accuracy for EBUS-TBNA were 85%, 100%, and 89%, respectively. The authors suggested that the combination of EBUS and EUS would allow investigation of the majority of the mediastinum. Vilmann et al23 also recently reported a combined approach of EUS-FNA and EBUS-TBNA in 33 patients with either known nonsmall cell lung cancer (N=20) or suspected lung cancer (n=13). EBUS-TBNA and EUS-FNA were not successful in 1 patient each. The final diagnoses were confirmed in the 28 of 31 patients at thoracotomy or clinical follow-up. Fifty-nine lesions were sampled by EUS-FNA and 60 by EBUS-TBNA. EUS-FNA and EBUS-TBNA demonstrated cancer in 26 and 28 lesions, respectively. Eleven additional cancer diagnoses and 3 samples with suspicious cells were procured by EBUS-TBNA that had not been obtained by EUS-FNA. On the other hand, EUS-FNA diagnosed 12 additional cancers that had not been obtained by EBUS-TBNA. The accuracy of combined EUS-FNA and EBUS-TBNA for the diagnosis of mediastinal involvement was 100%. The authors suggested that a combined approach with both EUS-FNA and EBUS-TBNA may be able to replace more invasive methods for evaluating the mediastinum in patients with known or suspected lung cancer. So where do we go from here? The combination of EBUS-TBNA and EUS-FNA has the potential to diagnose and completely stage the mediastinum in patients with suspected lung cancer without the need for more invasive techniques. However, the actual algorithm for combining these techniques (if both are available at an institution) is unclear at this time. Should EBUS-TBNA be done first and then EUS-FNA done in selected patients when additional information is needed or vice versa? Alternatively, is EBUS-TBNA with EUS-FNA done as a combined procedure under the same sedation the most cost effective, minimally invasive and expeditious method for diagnosis and staging of suspected lung cancer? Furthermore, should surgeons discard mediastinoscopy in favor of a completely nonsurgical staging? Large-scale, prospective, randomized clinical trials that incorporate both EBUS-TBNA and EUS-FNA with randomization to various combinations are needed to answer these compelling questions.