The Effect of Radiation Dose and Tube Potential on Image Quality of CT: A Task-based Image Quality Assessment

Purpose: To compare the effect of radiation dose and tube potential on image quality of CT through the task-based image quality assessment parameters. Methods: We scanned Gammex 464 (the ACR quality assurance phantom) with GE Revolution Apex CT. Three radiation doses (5, 10, 20 mGy) and three tube potentials (80, 100, 120 kVp) were used to reconstruct nine sets of image. Bone and acrylic inserts from module 1 of the phantom was selected for the measurement of task-based transfer function (TTF, representing spatial resolution) and TTF_50% was recorded for each set of images. Module 3 was selected for the measurement of noise power spectrum (NPS, representing image noise) and noise value, spatial frequency (f-peak) and NPS peak value were recorded for each set of images. Detestability index (d' representing lesion detestability) was furtherly calculated based on TTF and NPS of images. The effect of radiation dose and tube potential on image quality was evaluated by One-way Anova analysis. Multiple comparisons for P value were corrected by FDR. Results: Compared with radiation dose, the effect of tube potential on TTF_50% was more obvious, but there was no significant difference between them in bone and acrylic substances. Noise and NPS peak significantly decreased with the increase of both radiation dose and tube potential but no statistical difference was found. Compared with tube potential, radiation dose showed greater impact on f-peak, but no statistical difference was found. d’ was significantly improved as radiation dose increased; while no statistical difference was found under different tube potentials. Conclusion: Image quality is predominantly influenced by radiation dose rather than tube potential. Image noise and lesion detestability is signifcantly improved as radiation dose elevates. Image quality could be comprehensively inflected by the task-based image quality assessment.