Utility and Accuracy of Endobronchial Ultrasound As a Diagnostic and Staging Tool for the Evaluation of Mediastinal Adenopathy

Rebecca J Johnson, MD, J E Stephenson, MD FACS, C Schammel, PhD, B L Johnson, MS, K Banks, S Hutcheson, W Wall, W D Bolton, MD. Greenville Hospital System University Medical Center

 

Objective of Study: Endobronchial ultrasound (EBUS) is a technique for mediastinal evaluation which is less invasive than cervical mediastinoscopy, the traditional gold standard. The purpose of our study is to evaluate the utility and accuracy of EBUS as a diagnostic and staging tool at a regional teaching institution.
Methods: We retrospectively reviewed our institutional thoracic surgery database for patients undergoing EBUS from August 2008 to March 2011. All procedures were performed by a single surgeon, using a linear probe endobronchial ultrasound and onsite immediate pathology interpretation as well as cell block tissue evaluation.
Results: 190 patients underwent EBUS: 90 (47%) patients underwent EBUS for diagnosis only; 73 (38%) for staging only; and 27 (14%) for both diagnosis and staging. Diagnoses obtained by diagnostic EBUS included non-small cell lung cancer- n=27 (23%); other cancer- n=24 (21%); sarcoid/granulomatous- n=8 (7%); benign lymphoid tissue- n=57 (49%); and was nondiagnostic in only 1 case (1%). For staging EBUS, 27 (37%) patients had malignancy found in the lymph node specimen, while 46 (63%) patients had benign lymph node tissue. 103 patients therefore had a benign result at the time of EBUS. Fifty-six (54%) of these patients underwent subsequent mediastinal lymph node dissection or mediastinoscopy for tissue confirmation with the remainder undergoing follow up surveillance chest CT scans. Two patients had a false negative EBUS, one patient with a false negative diagnostic EBUS and one with a false-negative staging EBUS. Both false negative studies sampled levels 4L, 4R, and 7, with each having only one station that was found to be positive for malignancy. The overall false negative rate was 2% for all benign results, and 4% for those benign results confirmed with lymph node dissection or mediastinoscopy. The sensitivity and specificity of diagnostic EBUS was 97% and 100%. The sensitivity and specificity for staging EBUS was 98% and 100%. In those patients undergoing a staging EBUS, a mean of 2.6 nodal stations were sampled, with 71% (n=73) of these patients having three lymph node stations sampled and 29% (n=30) had two lymph node stations sampled. The mass itself was sampled in eight patients.
Conclusion: We found endobronchial ultrasound to be a highly accurate and minimally invasive manner in which to both diagnose mediastinal masses and stage the mediastinum for patients with non small cell lung cancer.
 


Session Number: ResFel – Residents/Fellows Scientific Session
Program Number: S137

« Return to SAGES 2012 abstract archive