Introduction and objectives The ASTER trial demonstrated that the sensitivity of combined endobronchial/endoscopic ultrasound [EBUS/EUS] is similar to that of mediastinoscopy.1 NICE guidelines now recommend combined EBUS/EUS for initial mediastinal staging as an alternative to surgical staging.2 Although surgical staging following negative endosonography is recommended, ASTER showed that 11 mediastinoscopies are required to identify one case of mediastinal disease. We aimed to determine the negative predictive value of EBUS/EUS for mediastinal staging in our practice.
Method We identified all patients who had undergone primary surgical resection with lymph node sampling for NSCLC between May 2012 and April 2014 previously staged with CT, PET-CT, EBUS+/-EUS and/or mediastinoscopy. Demographic, histological and surgical details were extracted from clinical records. Analysis was conducted on MedCalc software v13.3.1 and reviewed by an independent statistician.
Results 42 patients who underwent EBUS+/-EUS for mediastinal staging were found to have no evidence of N2/3 disease. In 3 cases subsequent mediastinoscopy was performed as a high degree of suspicion for mediastinal disease persisted. However, in all cases surgical staging correlated with endosonographic staging. At thoracotomy, 3 (other) patients were upstaged to N2 disease. In two cases, micrometastatic disease was present in a station 7 node and one case had positive station 5/6 not accessible at EBUS/EUS. Overall the NPV of EBUS+/-EUS was 93% (95% CI, 80%–98%). In 22 of 42 patients, the same nodal stations sampled on EBUS/EUS were removed at surgery. In this subset, EBUS/EUS had a NPV of 91% (95% CI, 71% to 99%).
Conclusion We have shown that in an experienced centre, mediastinal staging by EBUS+/-EUS can have a high NPV. In these circumstances, surgical staging following negative endosonography is probably not warranted unless a high degree of clinical suspicion remains following MDT discussion. Regular audit of NPV is recommended to ensure performance standards are maintained.
Annema et al. JAMA 2010;304:2245
NICE guidelines, 2011, Lung Cancer, CG121