Accurate staging of the mediastinum in non-small cell lung cancer is essential to appropriately determine if patients should be candidates for surgical resection, neo-adjuvant therapy, or for definitive chemoradiation. It has been shown that patients with persistent, bulky N2 disease following induction therapy have poor survival if they go on to surgery. Thus, re-staging patients after induction therapy is becoming even more important; however, the ideal method of staging the mediastinum is still under debate. With the continued development of new minimally invasive techniques, such as endobronchial ultrasound (EBUS) guided transbronchial needle biopsy and VATS staging, the question is sure to evolve. Cervical mediastinoscopy is the current gold standard for mediastinal staging. Pooling data from several studies, Toloza et al, found that cervical mediastinoscopy had a sensitivity of 81% and specificity approaching 100%; for a positive predictive value of 1.0 and negative predictive value of 0.91.1 In contrast, PET had an overall positive predictive value of 0.79 and negative predictive value of 0.93.2 The main risk with PET is that some patients with pathologically negative mediastinal disease may continue to have FDG-avid lymph nodes. Thus, some patients with potentially curable disease will be denied surgical intervention. Standard cervical mediastinoscopy is not able to sample aortopulmonary window, paraesophageal, and pulmonary ligament nodes. Other drawbacks include the need for general anesthesia, the risk of bleeding, and risk of recurrent laryngeal nerve injury. A left anterior parasternal mediastinotomy, or Chamberlain procedure, provides access to the subaortic space, however, generally requires division of costal cartilage. The sensitivity and specificity are 74% and 100% respectively. VATS also allows sampling of the AP window nodes. If performed at the time of definitive resection, an initial VATS sampling of ipsilateral mediastinal lymph nodes with frozen section may prevent the surgeon from going on to perform an unnecessary resection. There have been relatively little data comparing the accuracy of VATS staging to either PET or conventional mediastinoscopy. Like mediastinoscopy, these procedures require general anesthesia and carry a risk of bleeding. Endoscopic ultrasound guided fine-needle aspiration (EUS-FNA) allows access to left sided paratracheal nodes, as well as para-aortic, subcarinal, para-esophageal, and sometimes inferior pulmonary ligament nodes. For these lymph nodes, EUS-FNA has been shown to have comparable sensitivity and specificity to CT and PET. The diagnostic accuracy is further increased if there is lymph node enlargement. Endobronchial ultrasound (EBUS) is a relatively newer method of staging that uses an ultrasound transducer mounted on a videobronchoscope to guide FNA. It can be used to biopsy paratracheal, subcarinal, and peribronchial nodes; in experienced hands, even some AP window nodes can be biopsied safely. In the few published case series, the sensitivity and specificity are roughly 85% and 100% respectively. Both EUS and EBUS can be done relatively quickly, without general anesthesia, and with a minimal complication rate. Combined, EUS and EBUS have the potential to accurately stage the entire mediastinum with low morbidity. The role for invasive staging of the mediastinum continues to evolve as new technologies emerge. Randomized trials comparing the newer methods with established methods are necessary. Although it is yet unclear which method or methods will become the new standard, all these invasive approaches have one distinct advantage over non-invasive methods like CT or PET - access to tissue. For patients with mediastinal disease, these procedures can be used to make a definitive tissue diagnose at the same time as staging, which is required for initiation of induction therapy. In addition, as molecular profiling of tumors becomes more important, being able to analyze the biology of the tumor (both pre- and post-induction) may add another dimension to individually tailored treatment regimens.