AJR: Algorithm reduces benign lung biopsy rate
Specifically, an algorithm that relies on shape, size and growth of nodules will limit the impact of screen-detected nodules, according to Heidi C. Roberts, MD, of the University Health Network in Toronto, and colleagues.
“It is well known that low-dose CT screening detects a high rate of lung nodules and also that almost all of the screen-detected nodules are benign,” wrote the authors. “However, all screen-detected lung nodules require follow-up or management to highlight the few small early-stage lung cancers within the abundant number of benign nodules.”
Researchers looked at 4,782 high-risk, current and former smokers enrolled in a lung cancer screening study between 2003 and 2009. Indication for biopsy was made according to the diagnostic algorithm from the International Early Lung Cancer Action Program.
After screening, a total of 128 diagnostic biopsies were recommended for suspicious nodules and 127 were actually performed. This included 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.
Forty-eight percent of biopsy recommendations were made on the basis of shape, 40 percent were based on growth on follow-up and the rest were based on the appearance of new nodules.
Using shape, growth and appearance of new nodules as a guide, 84 percent of biopsies were correctly indicated for malignancy and 16 percent were benign. The overall false-positive recommendation rate for the entire study group was 0.4 percent.
“Our data show that a simple but well-defined algorithm to follow up screen-detected nodules results in a very low rate of interventional diagnostic procedures for benign lesions,” wrote the authors. “Adherence to a similarly stringent algorithm has been successfully implemented in other screening studies, such as the NELSON trial, in which the false-positive rate was only 7.9 percent.”
Roberts et al pointed out that the study protocol used provided suggestions only and left final decisions about follow-up to the patient and physicians. In an experienced environment, this process yields a high positive prediction of malignant nodules.
The authors wrote that the rate of intervention for benign nodules can be further reduced by taking into consideration cavitations and very slow or very fast growth rates. A purely diagnostic invasive procedure may not be required at all.
“We speculate that with careful assessment of shape and size, in combination with imaging surveillance and growth assessment as well as functional imaging such as PET or perfusion imaging, we would be able to predict the malignant nature sufficiently reliably to refer the individual for surgical treatment without a cytologic diagnosis,” wrote Roberts et al.