Improving Negative Predictive Value of Endobronchial Ultrasound Guided Transbronchial Needle Aspiration (EBUS-TBNA) Staging of Mediastinal Lymph Nodes
CB Marshall, B Jacob, J Stewart. University of Texas MD Anderson Cancer Center, Houston, TX
Background: Endobronchial ultrasound-guided transbronchial needle aspiration (EBUS-TBNA) is a well established and cost effective method to stage patients with thoracic cancers. Previously reported negative predictive values range between 76% and 86%.
Design: A retrospective study was performed to identify patients who underwent EBUS-TBNA for pre-operative staging of lung carcinoma or mesothelioma between October 2006 and July 2009 at a large cancer center. All cytology samples obtained by a single thoracic surgeon who is highly experienced in EBUS-TBNA were selected for review. All samples had immediate on-site assessment for adequacy by a senior cytotechnologist. Samples lacking lymphoid tissue were reported as non-diagnostic, rather than negative. The cytology diagnosis for each lymph node station was correlated with subsequent histology.
Results: Of 315 EBUS-TBNAs performed, 95 samples from 53 patients had subsequent surgical resection within 24 days. The underlying diagnosis was mesothelioma (47%) or non-small cell lung carcinoma (53%). The average size of sampled lymph nodes was 0.8 cm (range 0.3-2.1 cm). The most common lymph node stations to have histologic follow-up were stations 7, 4R, and 4L (92% of samples). Diagnostic material was obtained in 92 of 95 (97%) samples. The cytologic diagnoses for this group contained 8 false negative diagnoses (8%) and 1 false positive diagnosis (1%). The sensitivity, specificity, positive predictive value, and negative predictive value were 56%, 99%, 83%, and 91%, respectively. Unanticipated findings included: each of the 8 false negative results occurred in mesothelioma patients, the lymph node station that was most prone to a false negative was station 7 (7 of 8 samples), and the majority of false negative samples were reported to contain adequate evidence of lymph node sampling (7 of 8 samples).
Conclusions: The negative predictive value of EBUS-TBNA can approach that of mediastinoscopy through cooperation across medical disciplines. An experienced bronchoscopist is necessary to provide diagnostic samples even on lymph nodes less than 1 cm. An experienced cytotechnologist or pathologist is needed to quickly and accurately evaluate specimens for evidence of lymph node sampling and/or presence of malignant cells to provide immediate feedback during the procedure. Even with multidisciplinary collaboration false negative diagnoses occur, and appear to be a result of sampling rather than misinterpretation.
Wednesday, March 24, 2010 1:00 PM
Poster Session VI # 51, Wednesday Afternoon