Effect of Nodule Characteristics on Variability of Semiautomated Volume Measurements in Pulmonary Nodules Detected in a Lung Cancer Screening Program

PURPOSE: To retrospectively assess volume measurement variability in solid pulmonary nodules (volume, 15-500 mm(3)) detected at lung cancer screening and to quantify the independent effects of nodule morphology, size, and location. MATERIALS AND METHODS: This retrospective study was a substudy of the screening program that was approved by the Dutch Ministry of Health, and all participants provided written informed consent. Two independent readers used semiautomated software to measure the volume of pulmonary nodules detected in 6774 participants aged 50-75 years (5917 men). Nodules were classified according to their location (purely intraparenchymal, pleural based, juxtavascular, or fissure attached), morphology (smooth, polylobulated, spiculated, or irregular), and size (50 mm(3)). The level of agreement was expressed by using the absolute values of the relative volume differences (RVDs). Multivariate logistic regression analysis was performed, and odds ratios (ORs) were computed to quantify the independent effects of morphology, location, and size on RVD categories. RESULTS: Altogether, 4225 nodules in 2239 participants were included. Complete agreement in volume was obtained for 3646 (86%) of the nodules. Disagreement was small (absolute value of RVD or= 5% but or= 15%) for 174 (4%). Multivariate analysis showed that the ORs of volume disagreement were 15.7, 3.1, and 1.9 for irregular, spiculated, and polylobulated nodules, respectively; 3.5, 2.6, and 2.1 for juxtavascular, pleural-based, and fissure-attached nodules, respectively; and 1.3 for large nodules compared with smooth, purely intraparenchymal, and small reference nodules. CONCLUSION: Nodule morphology, location, and size influence volume measurement variability, particularly for juxtavascular and irregular nodules.