CT screening for lung cancer: implication of lung biopsy recommendations

The purpose of this article is to address the implications of invasive diagnostic procedures recommended by a lung cancer screening protocol. In particular, we assess how many invasive procedures were recommended for benign nodules.

Between 2003 and 2009, 4782 high-risk current and former smokers were enrolled in a lung cancer screening study. A helical low-dose CT of the chest was performed. Morphologic features targeted were parenchymal nodules. The indication for biopsy was made according to the diagnostic algorithm provided by the International Early Lung Cancer Action Program. We recorded the time points of biopsy recommendation; shape, size, and growth of nodules; types of diagnostic procedures; complication rates; and final pathologic diagnosis.

A total of 128 diagnostic biopsies were recommended for suspicious nodules, and 127 biopsies were performed, including 110 percutaneous CT-guided fine-needle aspiration biopsies (FNABs), nine video-assisted thoracoscopic surgery (VATS) resections, seven bronchoscopies, and one ultrasound-guided biopsy of a lymph node. Of 110 FNABs, 24 had unsatisfactory results, 13 of which were referred for secondary diagnostic VATS resection. The indication for biopsy was made on the basis of shape in 48% of cases (62/128), growth on follow-up in 40% of cases (51/128), and the appearance of new nodules in 12% of cases (15/128). In total, 104 of 124 biopsies (84%) were correctly indicated (true-positive recommendation) for malignancy, 20 were benign (false-positive) (16%), and final results are pending for four cases. The overall false-positive recommendation rate was 0.42% (20/4782); 11.6% of FNABs (16/128) and 3.6% of VATS (5/128) revealed benign nodules, corresponding to an overall false-positive rate of 0.33% for FNAB (16/4782) and 0.10% for VATS (5/4782).

The recommended biopsy procedures for screen-detected suspicious pulmonary nodules resulted in a low intervention rate for benign nodules. This rate is minimal when we followed a research protocol that relies on shape and growth.