Accuracy of lung nodule volumetry in low-dose CT with iterative reconstruction: an anthropomorphic thoracic phantom study

Abstract

Objective: The purpose of this study was to assess accuracy of lung nodule volumetry in low-dose CT with application of iterative reconstruction (IR) according to nodule size, nodule density and CT tube currents, using artificial lung nodules within an anthropomorphic thoracic phantom. Methods: Eight artificial nodules (four diameters: 5, 8, 10 and 12mm; two CT densities: -630HU that represents ground-glass nodule and +100HU that represents solid nodule) were randomly placed inside a thoracic phantom. Scans were performed with tube current-time product to 10, 20, 30 and 50mAs. Images were reconstructed with IR and filtered back projection (FBP). We compared volume estimates to a reference standard and calculated the absolute percentage error (APE). Results: The APE of all nodules was significantly lower when IR was used than with FBP (7.5 +/- 4.7% compared with 9.0 +/- 6.9%; p<0.001). The effect of IR was more pronounced for smaller nodules (p<0.001). IR showed a significantly lower APE than FBP in ground-glass nodules (p<0.0001), and the difference was more pronounced at the lowest tube current (11.8 +/- 5.9% compared with 21.3 +/- 6.1%; p<0.0001). The effect of IR was most pronounced for ground-glass nodules in the lowest CT tube current. Conclusion: Lung nodule volumetry in low-dose CT by application of IR showed reliable accuracy in a phantom study. Lung nodule volumetry can be reliably applicable to all lung nodules including small, ground-glass nodules even in ultra-low-dose CT with application of IR. Advances in knowledge: IR significantly improved the accuracy of lung nodule volumetry compared with FBP particularly for ground-glass (-630HU) nodules. Volumetry in low-dose CT can be utilized in patient with lung nodule work-up, and IR has benefit for small, ground-glass lung nodules in low-dose CT.