Methods: Records of 131 patients (96 males, 35 females; mean age 58.3±8.4 years; range 42 to 75 years) with known primary or suspected malignancy who had enlarged and hypermetabolic hilar/mediastinal lymph nodes detected by thoracic CT and at PET-CT and in whom EBUS-TBNA performed for cytologic confirmation of the malignancy between October 2008 and April 2011 were retrospectively analyzed. More invasive procedures including mediastinoscopy/video-assisted thoracoscopic surgery (VATS) were performed in patients who did not receive definite diagnosis using EBUSTBNA. The maximum standardised uptake value (SUVmax) cut-off level of PET-CT was considered ≥3.0. The sensitivity, specificity, diagnostic accuracy, and negative and positive predictive values of PET-CT and EBUS-TBNA in diagnosis of malignant hilar/mediastinal lymph nodes were calculated. The results were compared with each other.
Results: A total of 191 lymph node stations of 131 patients were aspirated from the LN stations. Of the 142 lymph nodes, 134 were diagnosed with malignancy using EBUS-TBNA, while SUVmax value was ≥3.0 in 127 by PET-CT. The sensitivity, spesificity, diagnostic accuracy, and negative and positive predictive values of EBUS-TBNA and PET-CT were 94.3%, 100%, 95.8%, 85.9%, 100% and 89.4%, 18.3%, 71.2%, 37.5%, 76.0%, respectively. With combined use of EBUS-TBNA and PET-CT, the sensitivity increased to 100%.
Conclusion: The sensitivity, specificity, diagnostic accuracy, and negative predictive value and positive predictive value of EBUS-TBNA is higher than PET-CT. Based on conventional data, histological confirmation of PET-CT is necessary in mediastinal staging, due to high level of false positivity of PET-CT. EBUS-TBNA is an effective, reliable and minimally invasive method for histologic confirmation of PET-CT-positive malignant mediastinal/hilar lymph nodes.