SESSION TITLE: Bronchoscopic Training and Approaches
SESSION TYPE: Original Investigation Slide
PRESENTED ON: Sunday, October 27, 2013 at 10:45 AM - 11:45 AM
PURPOSE: Many studies have shown that EBUS-TBNA affords excellent diagnostic results with a sensitivity of 69%-99.1%. The NPV of 11%-98.9% suggests that EBUS-TBNA still has a relatively high false negative rate compared with mediastinoscopy. Our study reviewed certain factors thought to have an influence on the accuracy of EBUS TBNA cytology in patients suspected to have lung cancer with radiographically positive mediastinal/hilar lymph nodes.
METHODS: A retrospective analysis of patients who underwent EBUS TBNA during a 2-year period (January 2010 - December 2011) was performed. Inclusion criteria for this study required that patients had radiographically enlarged mediastinal/hilar lymph nodes and underwent a CT and/or PET/CT within 40 days of having EBUS TBNA performed.
RESULTS: EBUS was performed on 246 patients, and we examined the 145 patients with a suspicion of lung cancer. Of these 145 patients, 83 were male (57.2%) with an average age of 67.3±11.5 years. Eighty-seven patient needle aspirates (60%) were diagnostic of lung cancer, and 45(31%) were negative. Of the 45 negative results, further testing revealed malignancy in 15 (33.3%) cases. Of these 15 false negative cases, the maximal SUV value on PET CT scans was much lower than the true lesions of EBUS TBNA (p=0.01, 4.8±3.1 vs 7.7±5.8). Type of sedation, mean HU of each lymph node LN, nodal station, size, ROSE, and number of needle passes did not appear to impact the likelihood of a false negative sample.
CONCLUSIONS: When EBUS TBNA is performed in patients with mediastinal lymphadenopathy with a high risk for lung cancer, false negative results are associated with lower values of maximal SUV.
CLINICAL IMPLICATIONS: Improved patient selection for mediastinal assessment with EBUS TBNA vs mediastinoscopy utilizing variations found in PET SUV avidity may further reduce EBUS TBNA’s negative predictive value.
DISCLOSURE: The following authors have nothing to disclose: KiHyun Seo, Carla Lamb, Caleb Mackey, Saatian Bahman, Ghazwan Acash, Sara Shadchehr
No Product/Research Disclosure Information