The Phantom And Clinic Study Of Impact Of Iterative Model Reconstruction On The CT Image Quality And Radiation Dose

Part one. The phantom study of knowledge based Iterative Model ReconstructionObject: To evaluation the CT image quality and dose reduction using knowledge based iterative model reconstruction(IMR) in a phantom study.Materials and Methods: Catphan 600 phantom was scanned by a 256- slice CT(Philips Brilliance 256 i CT) with 120 k V and varied m As from 11 to 500, raw data were reconstructed by filtered back projection(FBP), i Dose 4 and IMR 2(Routine, Soft and Sharp Plus). The volume CT dose index(CTDIvol) were recorded for dose evaluation. The evaluated image quality parameters were CT values, spatial and low-contrast resolution, image uniformity and noise.Results: 1. The CT values of PMP, LDPE and polystyrene were lower using IMR 2 than using FBP and i Dose 4(P < 0.05), the CT values of water were higher using IMR 2 than using FBP and i Dose 4(P < 0.05). 2. The differences of spatial resolution 6.48 ± 0.60、6.51 ± 0.46、8.29 ± 0.38、7.16 ± 0.32、9.49 ± 0.37, low-contrast resolution 0.70 ± 0.44、0.50 ± 0.31、0.33 ± 0.17、0.18 ± 0.10、1.20 ± 0.48 and noise 8.29 ± 4.91、5.87 ± 3.34、3.70 ± 1.24、2.11 ± 0.78、14.71 ± 5.13 for FBP, i Dose 4 and IMR 2(Routine、Soft and Sharp Plus) were statistically significant(P < 0.05). 3. There was no significant difference in image uniformity 2.80 ± 1.36、2.73 ± 1.24、2.21 ± 0.90、2.20 ± 0.78、2.01 ± 1.08 between different IR algorithms(P > 0.05). 4. The spatial and low-contrast resolution of LDCT-IMR is 8.0 lp/cm@10%MTF and 0.21 % @5mm@2.2m Gy,of ULDCT- IMR is 7.6 lp/cm@10%MTF、0.50% @5mm@0.8m Gy.Conclusion: Compared to FBP and i Dose 4 algorithms, IMR 2(Soft) can significantly reduce noise and improve low-contrast resolution, IMR(Sharp Plus) can significantly improve spatial resolution, IMR combine with LDCT/ULDCT can be used for lung cancer cancer.Part two. The effect of Iterative Model Reconstruction on clinical application of ultra low-dose non-enhanced chest CTObject: To evaluate the feasibility of using Iterative Model Reconstruction in ultra low-dose non-enhanced chest CT.Materials and Methods: 40 patients who underwent two consecutive non-enhanced scans including low-dose CT(120k V,30 m As) and ultra low-dose CT(120k V,10 m As) with raw data reconstruction using i Dose 4(Sharp(C)/Smooth(A)) and IMR(Sharp Plus/Soft) were enrolled. The effective dose, image noise, detection frequencies of lesion, subjective evaluation of normal lung and mediastinal structure were assessed.Results: 1 Effective dose was 65.42% less for the ultra low-dose protocol(P < 0.001). 2 Compared to LD-i Dose, the image noise were lower(P < 0.008) in LD-IMR and ULD-IMR. 3 Image quality of ULD-i Dose was lower than LD-i Dose,LD-IMR,ULD-IMR(P < 0.008). 4 Image quality of ULD-i Dose was significantly inferior for detection of GGO, emphysema.Conclusion: IMR allows ultra low-dose CT and affords diagnostic acceptable image quality.