To examine the performance of our large pulmonary transthoracic fine needle aspiration/core biopsy (FNA/CB) practice over time, we performed a retrospective analysis of data from 333 consecutive procedures performed in 1996-1998 and 568 consecutive procedures performed in 2003-2005. Fluoroscopic guidance was performed more frequently in the earlier cohort, while a larger majority of procedures in the later cohort were by computed tomography (CT-guidance). A follow-up histologic diagnosis of cancer or clinical evidence of disease was considered the gold-standard. FNA/CB procedures during the later time period were performed on smaller lesions overall (3.60 cm versus 2.97 cm; P = 0.003) and malignant lesions also tended to be smaller (3.87 cm versus 3.14 cm; P = 0.006). Minimal improvements in sensitivity (94% versus 91%), specificity (99% versus 95%), diagnostic accuracy (95% versus 92%), negative predictive value (NPV) (80% versus 74%), and positive predictive value (PPV) (100% versus 99%) were noted during 2003-2005 when compared with 1996-1998 in all lesions. Larger improvements in sensitivity (94% versus 73%), diagnostic accuracy (95% versus 79%), and NPV (79% versus 50%) were identified in very small lesions (<1 cm) in the later patient cohort in comparison to the earlier patient cohort, as well as a significant decrease in total procedure complications. CT-guided transthoracic FNA/CB continues to be a very effective tool in our practice assessing lung lesions and performance has improved considerably at our institution for very small lesions.
Copyright © 2011 Wiley Periodicals, Inc.