| Background Radiation plays an important role in the diagnosis and treatment of diseases,meanwhile, it could cause radiation injury on patients to some extent, for example,carcinogenesis. Because of higher density resolution and faster imaging speed, X-ray computed tomography(CT) is more and more widely used in clinic,which has brought great help to diagnosis and treatment of diseases. However,the radiation dose of CT scanning on the human body is still higher. The mean effective radiation dose for different parts in the single phase scanning is about 2.0m Sv~20m Sv and significantly higher than 1m Sv/a which is the public safety dose threshold recommended by the International Commission on Radiological Protection(ICRP). The radiation dose that the patient received in the whole abdomen intestinal CT examination is higher because of the larger scanning scope and multi scanning phase,it has already been one of the public concerns. Reduced-dose CT can be achieved by modifying the scanning parameters(tube current, tube potential, pitch, etc). But there is a limit to reduce the radiation dose by adjusting the scanning parameters,since the excessive reduction of radiation dose can induce too much CT image noise to meet the diagnostic needs. In recent years,in order to lower radiation dose of CT scanning, iterative reconstruction(IR)technology has been applied in clinical practice. The ability of IR to remove the image noise is stronger than that of the conventional filtered back projection(FBP) reconstruction. Therefore, the usage of IR allows to reduce more radiation dose of CT scanning. However, due to the lack of scientific and standardized methods to establish the scanning parameters of low-dose CT with iterative reconstruction, the scanning parameters of low-dose CT, at present, are not accurate. Some of the low-dose CT scanning parameters were on the high side, resulting in less radiation dose reduced and the image quality of low-dose CT was higher than that of the conventional CT image. Some low-dose CT scanning parameters were on the low side,resulting in excessive radiation dose reduced and the image quality of low-dose CT significantly decreased. Therefore, these scanning methods were difficult to be popularized in clinic, which has hindered the application of IR. How to set accurately the scanning parameters of low-dose CT with IR to ensure the quality of image and meet the diagnosis needs is a practical problem in clinical work. In previous studies, the water phantom was used to estimate the scanning parameters of low-dose CT, but the methods were not very accurate. Therefore,it is necessary to develop a more accurate and comprehensive approach to optimize the scanning parameters of low-dose CT on the basis of previous research.Objective The purpose of this study is as follows: First, to establish the combinational curves of tube current—image noise and tube current—CTDIvol at different tube potential by scanning the water phantom with different tube potential, tube current and reconstruction algorithm. Next, to verify the accuracy and reliability of setting the low-dose abdominal CT scanning parameters with IR by the curves. Finally, to develop a new method to optimize the scanning parameters of low-dose CT with IR and effectively reduce the radiation dose received by the patients during the CT examination.Methods Part 1 Phantom Study The water phantom was scanned at 14 different reference currents, from 700 m As to 50 m As with a decrement of 50 m As, and 3 different tube potential(80k V, 100 k V and 120 k V) in the single-source mode on a dual-source 128-MDCT scanner. Transverse images were reconstructed using FBP and sinogram affirmed iterative reconstruction 3(SAFIRE3) respectively. Images noise(standard deviation of CT attenuation,SD) of each group was measured by using a workstation after scanning. The average noise of 10 consecutive images in one group was used as the noise of this group. The volume computed tomography dose index( CTDIvol) of each scan was also recorded from the workstation. Image noise of different reconstruction algorithms with the same scanning parameters were compared with paired t test. The functions of reference current and images noise,and the functions of reference current and CTDIvol were fitted by SPSS19.0, then the combined fitted-curves were drawn by Excel. The low-dose abdominal CT scanning parameters were read from the curves, at which the noise of image reconstructed with SAFIRE3 was equal to that of the routine dose image reconstructed with FBP. Finally, the spatial resolution and density resolution of the CT images acquired by those parameters were examined by scanning the Catphan 500 phantom.Part 2 Clinical Research 240 patients who would undergo the enhanced total abdominal intestinal CT examination were assigned into group A(18.5kg/m2?BMI<23kg/m2) and group B(23kg/m2?BMI<27kg/m2), then assigned randomly into the control group(subgroup A1 and B1), low tube current group(subgroup A2 and B2) and low tube potential group(subgroup A3 and B3). The tube current,tube potential and the method of image reconstruction of each group were set according to the combined curves. The other scanning parameters were as same as that used in scanning water phantom in the phantom study. After scanning, the CT attenuation and image noise(standard deviation of the CT attenuation, SD) of liver,hepatic portal vein,abdominal aorta,erector inaemuscle,bladder,subcutaneous fat and background air were measured on the workstation. CTDIvol and product of dose and length(DLP) of each scanning were recorded. Signal to noise ratio(SNR) of each part,contrast to noise ratio(CNR) of liver,hepatic portal vein and aorta abdominalis were calculated subsequently. Effective radiation dose(ED) of each scanning was calculated too. Two radiologists who didn't knew the scanning parameters and reconstruction method blindly evaluated the subjective image quality of each group. One-way ANOVA and nonparametric tests were used for statistical analysis. Interrater correlation of subjective image quality scores was assessed by using kappa statistic.Results Part 1 Phantom Study Compared with the noise of the image reconstructed with FBP, the noise of image reconstructed with SAFIRE3 were significantly reduced when their scanning parameters were same. The reduction rate of 80 k V, 100 k V and 120 k V group was 14.1%~25.9%, 22.4%~27.8% and 25.3%~28.7% respectively. The reduction rate increased gradually with the decrease of FBP image noise. The image noise and the reference tube current, from 50 m As to 600 m As, followed a power equation. The CTDIvol had linear correlation with the reference tube current under the experimental conditions. The determination coefficients were more than 0.99. The combinational curves objectively, accurately and comprehensively reflected the effect of different tube voltage, tube current and reconstruction method on image noise and CTDIvol. The scanning parameters and reconstruction algorithm of low-dose CT at which the image noise was same as the routine dose CT image(120k V/400 m As/FBP,subgroup A1)were 120 k V/210 m As/ SAFIRE3(subgroup A2) and 100 k V/410 m As/SAFIRE3(subgroup A3). The other scanning parameters and reconstruction algorithm of low-dose CT at which the image noise was same as the routine dose CT image(120k V/500 m As/FBP,subgroup B1) were 120 k V/260 m As/SAFIRE3( subgroup B2) and 100 k V/500 m As/SAFIRE3(subgroup B3). The spatial resolution of each group was 5 Lp/cm, and the density resolution of each group was 4mm when the contrast was 1%,meeting the needs of diagnosis.Part 2 Clinical Research 1) There was no significant difference in objective image noise(SD) of each location between subgroup A1, A2 and A3(P>0.05),which was consistent with the expected result. 2) Compared with in subgroup A1 or A2, the CT attenuation of liver, abdominal aorta and portal vein significantly increased(P<0.05), and that of subcutaneous fat significantly decreased( P<0.05) and that of other location were comparable(P>0.05) in subgroup A3. There was no significant difference in CT attenuation of each location between subgroup A1 and A2(P>0.05). 3) Compared with subgroup A1 or A2, the SNR of abdominal aorta and portal vein significantly increased(P<0.05) and of other location were comparable(P>0.05) in subgroup A3. There were no significant difference in SNR of each location between subgroup A1 and A2(P>0.05). 4) Compared with subgroup A1 or A2, the CNR of abdominal aorta and portal vein significantly increased(P<0.05) and the CNR of liver slightly increased(P>0.05) in subgroup A3. There were no significant difference in CNR of the three locations between subgroup A1 and A2(P>0.05). 5) There was no significant difference in the subjective image quality score between subgroup A1, A2 and A3(P>0.05). 6) Compared with subgroup A1, CTDIvol, DLP and ED of subgroup A2(mean, CTDIvol, 7.7m Gy vs. 16.5m Gy, DLP, 352.5m Gycm vs. 749.7m Gycm, ED, 5.3m Sv vs. 11.2m Sv)and subgroup A3(mean, CTDIvol, 9.3m Gy vs. 16.5m Gy, DLP, 424.8m Gycm vs. 749.7m Gycm, ED, 6.4m Sv vs. 11.2m Sv)significantly decreased(P<0.05),and CTDIvol, DLP and ED of subgroup A2 were significantly decreased than that of subgroup A3(P<0.05),which was consistent with the results expected from the combined curves. 7) The difference of image noise,CT attenuation,SNR, CNR, subjective image scores, CTDIvol, DLP and ED between subgroup B1, B2 and B3 were similar with that between subgroup A1, A2 and A3. 8) Interrater agreement for subjective image quality scores of subgroup A1, A2, A3, B1, B2 and B3 was excellent(kappa=0.806~0.898).Conclusion On the basis of previous studies,the functional relationship between the tube current and the image noise, and the functional relationship between the tube current and CTDIvol were fitted by statistical methods for the first time in this study. The tube current—FBP image noise curve, the tube current—SAFIRE3 image noise curve and the tube current—CTDIvol curve were drawn in the same coordinate systema,then the curves of 3 levels of tube voltage were combined together to form the combinational curves,which demonstrate the effect of scanning and reconstruction methods on image noise and radiation dose intuitively and comprehensively. This method could not only improve the accuracy of low-dose CT scanning parameters,but also compare the radiation dose of lower tube current scanning with that of lower tube potential scanning. Clinical studies has showed that it will be more convenient,accurate and quicker to read the optimal low-dose abdominal CT scanning parameters with iterative reconstruction at which the images noise is comparable with routine abdominal CT images noise by this combinational curves. This method will promote the application of iterative reconstruction in clinical practice,and effectively reduce the radiation dose received by the patients in abdominal CT examination. |
PDF Full Text Request