The utility of automated volumetric growth analysis in a dedicated pulmonary nodule clinic

Abstract

Objective: Our objectives were 3-fold: to define the correlation between automated volumetric and 2-dimensional measurements of pulmonary nodule growth and prospectively; to determine whether volumetric growth analysis represents a useful addition to 2-dimensional measurements; and to evaluate growth rates over time of biopsy-proven lung cancers using automated volumetric software.

Methods: Nodule growth on consecutive 2-dimensional computed tomographic scans was measured, and a decision regarding nodule biopsy was made. Automated volumetric software was then used to determine nodule growth, growth rates obtained from the 2 techniques were correlated, and the decision to perform a biopsy was reassessed. Biopsy-proven lung cancer growth rates were then documented over time.

Results: Growth rates measured using volumetric software were highly correlated with 2-dimensional measurements (r = 0.69; P < .00001). This correlation was affected by nodule type (irregular [r = .63] versus smooth [r = 0.84]; P = .02) as well as the interval between scans (<100 days [r = .5] versus >100 days [r = 0.76]; P = .02). The addition of volumetric growth analysis changed the decision to perform a biopsy after only a minority (6.2%) of scan comparisons; however, lung cancer was diagnosed in 43% of these cases. Growth curves for individual cancers were highly variable, with 45% of tumors showing at least 1 period of shrinkage.

Conclusions: Automated volumetric software influences biopsy decisions in only a minority of cases in a dedicated pulmonary nodule clinic, but seems to be useful in detecting lung cancer in this minority. Radiographically determined nodule growth rates, in general, need to be questioned as the sole determinate of the need to perform a biopsy.