Retrospective assessment of interobserver agreement and accuracy in classifications and measurements in subsolid...

To compare interobserver agreements among multiple readers and accuracy for the assessment of solid components in subsolid nodules between the lung and mediastinal window settings.

Seventy-seven surgically resected nodules with solid components smaller than 8 mm were included in this study. In both lung and mediastinal windows, five readers independently assessed the presence and size of solid component. Bootstrapping was used to compare the interobserver agreement between the two window settings. Imaging-pathology correlation was performed to evaluate the accuracy.

There were no significant differences in the interobserver agreements between the two windows for both identification (lung windows, k = 0.51; mediastinal windows, k = 0.57) and measurements (lung windows, ICC = 0.70; mediastinal windows, ICC = 0.69) of solid components. The incidence of false negative results for the presence of invasive components and the median absolute difference between the solid component size and the invasive component size were significantly higher on mediastinal windows than on lung windows (P < 0.001 and P < 0.001, respectively).

The lung window setting had a comparable reproducibility but a higher accuracy than the mediastinal window setting for nodule classifications and solid component measurements in subsolid nodules.

• Reproducibility was similar between the two windows in nodule classifications. • Reproducibility was similar between the two windows in solid component measurements. • Accuracy for solid component assessment was higher on lung windows.