Radiology: CT volume growth model may identify lung malignancies
A normative model based on the variation in volume growth rates of stable lung nodules may be used in the surveillance and monitoring of lung nodules and aid in the differentiation of benign and malignant lesions, according to a study published online Dec. 8 in Radiology.

Although the ability to detect an increase in the size of a pulmonary nodule offers a useful clinical tool, current processes rely on manual 1D and 2D measurements. These measures may be less sensitive than 3D techniques for identifying change in nodule size.

Consequently, Jane P. Ko, MD, of New York University Langone Medical Center in New York City, and colleagues sought to determine the precision of a 3D volume method for measuring lung nodule growth. “While absolute accuracy or the ability to measure true volume of a nodule is of interest, in clinical practice, the precision or reproducibility of a measurement method is more relevant,” wrote the researchers.

Ko and colleagues analyzed 239 CT studies, 59 baseline and 180 follow-up exams, of 123 stable lung nodules in 59 individuals. CT follow-up lasted a mean of 6.4 years and included an average of 4.1 exams per patient.

Eight patients were diagnosed with a lung malignancy based on a change in size during the study period, with results confirmed by surgery in seven cases and biopsy in one.

Researchers calculated the volume of each nodule and applied a formula to measure the growth rate at any two points in time. When they plotted the growth rate values with all available time points, the analysis showed most values clustered around a zero growth rate, indicating that the nodules were clinically stable, according to Ko and colleagues.

The eight nodules determined to be malignant “had an abnormally high growth rate detected by using the method.” In addition, the model identified disease significantly earlier than conventional manual measurements in four cases and did not delay diagnoses in the remaining cases.

Ko and colleagues acknowledged that the model is not exact. “Depending on [the] statistical threshold, a fraction of clinically stable nodules will be incorrectly flagged as having an abnormal increased growth rate. To reduce the frequency of false-positive findings, at the cost of increasing the false-negative rate, the threshold for differentiation can be increased.”

Additional studies are required and if they confirm these findings the model could provide a tool to differentiate stable nodules from ground-glass-containing nodules that require aggressive treatment, the authors concluded.