Lung Nodule Assessment in Computed Tomography: Precision of Attenuation Measurement Based on Computer-Aided Volumetry

Purpose: To compare the reproducibility (r) of CT value measurement of pulmonary nodules using volumetry software (LungCare, LC) and manual ROIs (mROI). Materials and Methods: 54 artificial nodules in a chest phantom were scanned three times with CT. CT values were measured with LC and mROI. The intrascan-r was assessed with three measurements in the first scan, and the interscan-r with measurements in three consecutive scans (one observer). Intrascan-r und interobserver-r (two obs.) were assessed in the first scan and in contrast-enhanced CT of 51 nodules from 15 patients (kernels b50f and b80f). Intrascan-r and Interscan-r were described as the mean range and interobserver-r as the mean difference of CT values. The significance of differences was tested using t-test and sign test. Results: Reproducibility was significantly higher for volumetry-based measurements in both artificial and patient nodules (range 0.11 vs. 6.16 HU for intrascan-r, 2.22 vs. 7.03 HU for interscan-r, difference 0.11 vs. 18.42 HU for interobserver-r; patients: 1.78 vs. 13.19 HU (b50f-Kernel) and 1.88 vs. 27.4 HU (b80f-Kernel) for intrascan-r, 3.71 vs. 22.43 HU for interobserver-r). Absolute CT values differed significantly between convolution kernels (pat./mROI: 29.3 [b50f] and 151.9 HU [b80f] pat./LC: 5 [b50f] and 147 HU [b80f]). Conclusion: The reproducibility of volumetry-based measurements of CT values in pulmonary nodules is significantly higher and should therefore be recommended, e.g. in dynamic chest CT protocols. Reproducibility does not depend on absolute CT values.