The Influence of CT Reconstruction Methods on the Accuracy of Monitoring Lung Nodule Diameters at Different Dose Levels

Abstract

Purpose: This study aims to investigate the effect of filtered back projection (FBP) and sinogram-affirmed iterative reconstruction (SAFIRE) on the accuracy of lung nodule diameter measurements at different dose levels. Method: 48 CT images were acquired (at tube-current time product of 10, 20, 30 and 40 mAs) using an anthropomorphic phantom Lungman N1 ©, containing simulated spherical lung nodules of +100 Hounsfield Units of 5, 8 and 12mm diameter. Images were reconstructed with FBP and SAFIRE strengths 1, 3, and 5. Twelve participants, with radiographic experience, performed nodule diameter measurements for all images. Nodule edge sharpness was calculated for all images by measuring the angle of profile edge slope.Contrast to Noise Ratio (CNR) values were obtained from pixel values in regions of interest (ROIs) in the lung nodule and background air. Measurement accuracy was assessed by calculating the absolute error percentage (AEP) between participant’s measurements and actual nodule size.Results: There is no significant difference in nodule diameter measurement between mAs values and reconstruction algorithms (p-value 0,009 - 0,969). AEP showed no significant difference (p-value 0,041-0,969) for any of the reconstruction algorithms. Discussion: Previous research using SAFIRE suggests a decrease of mAs while maintaining image quality. Furthermore, SAFIRE has the ability to increase CNR and decrease image artefacts. However, the findings in this study suggest that accuracy of lung nodule measurement does not improve with an increase of CNR values nor the line profiles of edge sharpness. Conclusion: Our study suggests that image dose levels can be reduced without compromising nodule diameter measurement accuracy, regardless of reconstruction method.