| [Objective] To evaluate the application of iterative reconstruction (iDose4and IMR) reducing radiation dose and controlling image quality in prospective electrocardiogram(ECG)-gated computed tomography enhancement scan of heart combined with weight adjusting tube voltage and current.[Materials and methods]1. General informationTen pigs were included,5male and5female, age4-16month, weight26kg-52kg. This study was approved by the hospital ethics committee of animal experiment and was carried out under their supervision.2. Examining methods2.1Anesthetization:Basal anesthesia and maintenance of anesthesia was adopted. Basal anesthesia: ketamine(10mg/kg) and diazepam(0.4mg/kg) by intramuscular injections; maintenance of anesthesia:propofol(10mg/kg) by intravenous drip via ear vein.2.2Scanning parameters:All pigs were underwent prospective electrocardiogram-gated computed tomography enhancement scan of heart adopted by256-section multi-slice spiral CT scanner (Brilliance iCT, Philips). Plain scan was performed before enhancement scan and value of hounsfield unit of plain scan of the ascending aorta root was measured and recorded. Collimation(96-128) X0.625mm; Rotation time0.27s/r; Thickness0.9mm; Reconstruction interval0.45mm. Non-ionic contrast medium(Iopamiro,370mg I/ml)60ml+normal saline30ml was injected via ear vein by automatic high pressure injector; Injection speed3.5ml/s. Bolus-tracking was used, with a region of interest (ROI) in the descending aorta at the level of the carina of trachea. When the CT value reached100HU, a7s post threshold delay was applied before the scan started. The tube voltage and current of conventional dose (Group A) was adjusted by weight for all pigs and the tube current was then reduced by30%(Group B),50%(Group C),70%(Group D).Continuous intravenous infusion of normal saline was performed after every CT scan in order to accelerate the excretion of the contrast agent (The urine collection bag should be changed in time). Keep the CT value of plain scan of the ascending aorta root the same as the first CT value of plain scan before enhancement scan everytime. The scan range was from superior aperture of thorax to the5cm beneath left diaphragm.3. Image postprocessingThe three reconstruction methods FBP, iDose4, IMR were performed for all the raw data and the four groups were accordingly divided into12subgroup:A1(Group A,FBP), A2(Group A,iDose4), A3(Group A,IMR), B1(Group B,FBP), B2(Group B,iDose4), B3(Group B,IMR), Cl(Group C,FBP), C2(Group C,iDose4), C3(Group C,IMR), D1(Group D, FBP), D2(Group D,iDose4), D3(Group D,IMR). The most clear images were uploaded to the external workstation (Cardiac Viewer, Extended Brilliance Workspace (version4.0); Philips Healthcare) to analyze and post-process.4. Image quality evaluation4.1Objective indicators:The objective indicators include image noise, Signal to Noise Ratio (SNR), Contrast to Noise Ratio (CNR).Measurements were performed by an experienced cardiovascular radiologist by blind method. The compared images were at the same level. The ROI, the area of which was50±1mm2, was drawed by copy-and-paste function to ensure the size and shape stable; Every ROI was measured for three times and the result was recorded as the mean of the three measurements. The CT number (SI) and standard deviation (SD) of the ascending aorta root (the opening of left main coronary artery level) and the center of left ventricular chamber was measured. The influence of the aortic valve shoud be avoided in the course of measurement. The SD of ROI stands for noise. In addition, erector spinae was regarded as ground tissue. The CT number (SI) and standard deviation (SD) of central area of bilateral erector spinae in the same level was measured (the result was recorded as the mean of bilateral erector spinae) and the vessels and fat should be avoided. The SNR and CNR was calculated by the following formula:SNRaorta=SIaorta/SDaorta, SNRheart=SIheart/SDheart, CNRaorta=(SIaorta-SImuscle)/SDmuscle, CNRheart=(SIheart-SImuscle)/SDmuscle.4.2Subjective indicators: The subjective indicators include the subjective image noise, the overall image quality, the coronary artery image quality.The subjective indicators of all images were evaluated by two experienced cardiovascular radiologist used by blind method. The two evaluators need to re-read images together and reach an agreement when there was disagreement between them.4.2.1Subjective image noise:5=minimal image noise,4=less than average noise,3=average image noise,2=above average noise,1=unacceptable image noise.4.2.2Overall image quality:5=excellent image quality, excellent anatomical clarity;4=good image quality, good anatomical clarity, all structures clearly interpretable;3=average image quality, fair anatomical clarity, the clinical diagnosis met;2=poor image quality, incomplete demonstration of anatomical structures, poor anatomical detail;1=very poor image quality, no useful information obtained.4.2.3Coronary artery image quality:The image quality of the nine coronary artery segments (left main coronary artery (LM), proximal, middle and distal left anterior descending coronary artery(LAD), proximal and distal left circumflex artery (LCX), and proximal, middle and distal right coronary artery (RCA)) was evaluated according to a five-grade scale: 5=much clear visualization without any artefacts;4=clear visualization with some artefacts;3=a little blurring visualization with moderate artefacts;2=blurring visualization with many artefacts, not evaluated;1=serious artefacts, most coronary segments not visualized. For grades3or higher, the coronary arteries were considered as detected or sufficient for complete diagnosis. Coronary artery diagnostic rate is defined as a ratio for grades3or higher and coronary artery fineness rate is defined as a ratio for grades4or higher. Four of the nine segments, including the LM, proximal LAD, proximal LCX and proximal RCA, were re-classified as proximal coronary segments while the other five (middle and distal LAD, distal LCX, middle and distal RCA) were re-classified as distal coronary segments.5. Radiation doseThe volume CT dose index (CTDIvol), dose-length product (DLP) of every CT scan was recorded. Effective dose (ED) was calculated by the formula:ED=DLP*k, k=0.014.6. Statistical analysisAll the statistic analyses were fulfilled by the software SPSS13.0.Objective indicators:Comparing of the image noise, SNR, CNR which was underwent the three reconstruction methods in each group:If it was normal distributed and homogeneous of variance, One-Way ANOVA was adopted and Bonferroni test was used for two groups comparision; If normal distributed but heterogeneous of variance, Welch test was adopted; If abnormal distributed, Kruskal-Wallis nonparametric test was adopted. T-test of independent sampler was used to analysis any two objective indicators. P<0.05was regarded as statistical significance.Subjective indicators:Comparing of the subjective image noise, the overall image quality, the coronary artery image quality scores which were underwent the three reconstruction methods in each group was analyzed by Kruskal-Wallis nonparametric test. Mann-Whitney nonparametric test was used to analysis any two subjective indicators’scores. Chi-square test was adopted to compare any diagnostic rates or fineness rates. P<0.05was considered as statistical significance.[Results]1. Influence of low dose CT scan on image quality by the conventional FBP reconstruction1.1Effect on objective indicators:(1) Image noise:A1<B1<C1<D1(P=0.000);(2) Image SNR:A1>B1>C1>D1(P=0.000);(3) Image CNR:A1>B1>C1>D1(P=0.000).1.2Effect on subjective indicators:(1) Subjective image noise score:A1>B1>C1>D1(P=0.000).(2) Overall image quality:A1(3.80±0.42)> B1(3.60±0.52)> C1(3.00±0.67)> D1(2.00±0.67)(P=0.000).(3) Coronary artery image quality score:1) Proximal coronary artery:A1(4.20±0.56)> B1(3.90±0.71)> C1(2.93±0.73)>D1(2.00±0.68)(P=0.000).2) Distant coronary artery:A1(3.34±0.75)> B1(2.94±0.77)> C1(2.10±0.79)> D1(1.22±0.42)(P=0.000).(4) Coronary diagnostic rate:1) Proximal coronary artery:A1(100%)> B1(95%)> C1(70%)> D1(22.5%)(P=0.000).2) Distant coronary artery:A1(88%)> B1(72%)> C1(36%)> D1(0%)(P=0.000).(5) Coronary fineness rate:1) Proximal coronary artery:A1(92.5%)> B1(80%)> C1(22.5%)> D1(0%)(P=0.000).2) Distant coronary artery:A1(48%)> B1(24%)> C1(0%)=D1(0%)(P=0.000).2. Influence of iterative reconstruction (iDose4and IMR) on objective indicators of image quality2.1Image noise: (1) For each group:FBP>iDose4>IMR (P<0.05); Multiple comparisons in each group:Differences were all statistically significant (P<0.05).(2)1) Root of the ascending aorta:The image noise of Group D2was higher than Group A1(P>0.05), difference was not statistically significant between Group C2and Group A1(P>0.05); the image noise of Group D3was lower than Group A1(P=0.000).2) Left ventricular chamber:Difference was not statistically significant between Group D2and Group A1(P=0.105); the image noise of Group D3was lower than Group A1(P=0.000).2.2Image SNR:(1) For each group:FBP<iDose4<IMR (P<0.05); Multiple comparisons in each group:Differences were all statistically significant (P<0.05).(2)1) Root of the ascending aorta:The image SNR of Group D2was lower than Group A1(P<0.05), difference was not statistically significant between Group C2and Group Al (P>0.05); the image SNR of Group D3was higher than Group Al (P=0.000).2) Left ventricular chamber:the image SNR of Group D2was lower than Group Al (P<0.05), difference was not statistically significant between Group C2and Group A1(P>0.05); the image SNR of Group D3was higher than Group A1(P=0.000).2.3Image CNR:(1) For each group:FBP<iDose4<IMR (P<0.05); Multiple comparisons in each group:Differences were all statistically significant (P<0.05).(2)1) Root of the ascending aorta:Difference was not statistically significant between Group D2and Group Al (P>0.05); the image CNR of Group D3was higher than Group A1(P=0.000).2) Left ventricular chamber:Difference was not statistically significant between Group D2and Group Al (P>0.05); the image CNR of Group D3was higher than Group A1(P=0.000).3. Influence of iterative reconstruction (iDose4and IMR) on subjective indicators of image quality3.1Subjective image noise score:For each group:FBP<iDose4<IMR (P<0.05). 3.2Overall image quality:(1) For each group:FBP<iDose4<IMR (P<0.05).(2) The overall image quality score of Group D2was lower than Group A1(P>0.05), the overall image quality score of Group C2was about the same as Group Al (P>0.05); Group D3was higher than Group A1(P<0.05).3.3Coronary artery image quality score:(1) Proximal and distant coronary artery for each group:FBP<iDose4<IMR (P<0.05).(2)1) Proximal coronary artery:The quality score of Group D2was lower than Group Al (P=0.000), the quality score of Group C2was lower than Group A1(P<0.05), the quality score of Group B2was higher than Group A1(P<0.05); the quality score of Group D3was about the same as Group A1(P>0.05).2) Distant coronary artery:The quality score of Group D2was lower than Group A1(P=0.000), the quality score of Group C2was lower than Group A1(P>0.05), the quality score of Group B2was higher than Group A1(P>0.05); the quality score of Group D3was about the same as Group A1(P>0.05).3.4Coronary artery diagnostic rate:(1) Proximal and distant coronary artery for group D:FBP<iDose4<IMR (P<0.05); Multiple comparisons in group D:Differences were all statistically significant (P<0.05).(2)1) Proximal coronary artery:The diagnostic rate of Group D2was lower than Group A1(P<0.05), the diagnostic rate of Group C2was about the same as Group A1(P>0.05); the diagnostic rate of Group D3was about the same as Group A1(P>0.05).2) Distant coronary artery:The diagnostic rate of Group D2was lower than Group A1(P=0.000), the diagnostic rate of Group C2was about the same as Group A1(P>0.05); the diagnostic rate of Group D3was about the same as Group Al (P>0.05).3.5Coronary artery fineness rate:(1) The difference among coronary artery fineness rates by the three reconstruction methods in each group are all statistically significant (P<0.05), both for proximal and distant coronary artery.(2) Proximal coronary artery:the fineness rate of Group D3was higher than Group D2(P=0.000); Distant coronary artery:the fineness rate of Group D3was higher than Group D2(P=0.000).4. Radiation doseGroup A:CTDIvol (mGy)/DLP(mGy·cm)/ED(mSv) respectively were (16.20±3.79)/(261.75±73.96)/(3.66±1.04);Group B:CTDIvol (mGy)/DLP(mGy·cm)/ED(mSv) respectively were (11.20±2.49)/(184.74±47.18)/(2.59±0.66);Group C:CTDIvol (mGy)/DLP(mGy·cm)/ED(mSv) respectively were (8.04±1.86)/(137.50±40.67)/(1.93±0.57);Group D:CTDIvol (mGy)/DLP(mGy·cm)/ED(mSv) respectively were (4.85±1.15)/(81.90±20.68)/(1.15±0.29);The CTDIvol, DLP, ED of Group B, C, D were all reduced successively by30%,50%,70%compared to Group A.(Conclusion]1. The space of reducing radiation dose using the conventional reconstruction method FBP, reconstructed by which will lead to obvious increase of image noise, remarkable decrease of SNR and CNR, substantial reduction of image quality due to its sensitivity to noise is limited. The radiation dose of heart CT scanning using FBP reconstruction can only be reduced by30%at most; when reduced by50%, part of the image can’t be diagnosed.2. Iterative reconstruction technique (iDose4and IMR) both can memorably reducing image noise, enhancing image SNR and CNR, improving image quality and the effect of noise reduction, image quality improvement with the use of IMR is more remarkable than iDose4,especially in the distant coronary arteries. The radiation dose of prospective ECG-gated CT enhancement scanning of heart combined with weight adjusting tube voltage and current can be reduced by50%to make sure that the image quality meets clinical diagnosis with the use of iDose4reconstruction; when with the use of IMR reconstruction, the radiation dose can at least reduced by70%to make sure the acquisition of the same image quality as the routine dose. |
PDF Full Text Request