Accuracy of CT-guided transthoracic needle biopsy of lung lesions: Results of 612 consecutive procedures

Abstract

18022 Background: CT-guided transthoracic needle biopsy (TNB) is commonly used in diagnostic work-up of lung lesions. The availability of a on-site pathologist at the time of the procedure ameliorate its sensitivity, reduce the number of biopsies and false negatives. Methods: 612 procedures (608 patients with a CT-documented central or peripheral pulmonary lesion) performed at S. Luigi Hospital between November 2002 and August 2005 were prospectively analyzed; 66% males, median age was 66 years (range 29–87). Ineligibility criteria for the procedure included severe coagulopathy, previous contralateral pneumonectomy, lesions with a maximum diameter less than 5 mm or the impossibility to understand the procedure or to maintain the clinostatism for the time of the procedure. The on-site pathologist assigned to each specimen a semiquantitative score: 0 for bloody sample without other cells, 1 for aspecific benign or inflammatory cells, 2 for malignant cells without histotype characterization and 3 for well established benign or malignant histotype. Results: Most of the procedures was performed by fine needle aspiration biopsy, while in a minority of cases a tru-cut biopsy was requested. In 57.2% of the cases a single transthoracic access (range 1–4) was used and in 31% the procedure was repeated on the basis of the radiologist/pathologist judgment. In 154 patients a surgical resection was subsequently performed, while 454 were patients non-surgical. A score of 3 was obtained in 71% of cases (88% malignancies), 2 in 12.5%, 1 in 7.5% and 0 in 9%. A definitive diagnosis was made in 83.5% of procedures, while a score of 0–1 was assigned in 101 cases. Among 458 malignancies there were 411 lung cancer, 7 non-epithelial cancers and 40 metastases with only 1 false positive. The diagnostic accuracy for benign and malignant lesions was 67% and 92%, respectively (Pearson’s test p<0.005) with overall diagnostic accuracy of 83.3%. The variables affecting diagnostic accuracy were final diagnosis (benign 67%, malignant 92%, p<0.001) and lesion size (lesion 5 cm 78%, p<0.05). The presence of cavitation or necrotic areas and location of the lesion didn’t affect the diagnostic accuracy. Conclusion: In consecutive cases of CT-guided TNB final diagnosis and lesion size affect diagnostic accuracy. No significant financial relationships to disclose.