The Clinical Application Of Hybrid Iterative Reconstruction Technique On256iCT Of Liver Tumors With Low Radiation Dose

Object:To explore the clinical application value of the hybird iterative reconstruction technique (iDose4-4) on256iCT of liver tumor with low radiation dose. The effective dose was reduced equally by50%,through three methods of reducing the tube voltage, increasing the tube voltage,and reducing the tube current time. Each of46patients, knowed to have liver tumors, randomly underwent256-section multidetector CT with conventional dose (group A:120kV,250mAs) and50%low dose (group B:120KV,125mAs; group C:140KV,85mAs; group D:100KV,210mAs) in the portal venous phase of contrast enhancement. Standard convolution FBP was used to reconstruct conventional dose and50%low dose (group B2> C2> D2) imges sets,and iDose4-4was used to reconstruct50%low dose (group B1、C1、D1) imges sets. The imge quality of each group was compared.Material and methods:1. The inclusion and exclusion criteriaInclusion criteria:Age>18years, The body mass index (BMI) is normal (21-25kg/m2).46patients confirmed with liver tumors and needed to review were obtained. Exclusion criteria: Patients were considered ineligible if they were younger than18years, were pregnant or lactating (female subjects only), had severe hepatic insufficiency or other diseases affecting the liver hemodynamic, such as fatty liver,or had any contraindication to iodinated contrast material,such as previous history of anaphylactic reaction or renal failure (serum creatinine clearance rate<120umol/L) and cardiac insufficiency.The cases with apparent motion artifact, breath with strong amplitude difference,and the subcutaneous fat of the anterior abdominal wall is too thin to be measured also were ruled out.This prospective, single-center study was approved by the hospital ethics committee,and be carried out under their supervision. Informed patient consent was obtained.2. The Scanning ParametersPreparing for scanning:Oral500～800m1pure water (gastrointestinal tract contrast medium) in15minutes before checking, when you are going to be checked,200ml pure water will be given orally again. Trained the patient’s breathing and breathless,and try to keep it consistent across the whole process.Equipment and Methods:Multiple-phase CT was performed by using a256-section multidetector CT scanner (Philips Brilliance). Scan Type:Helical; Slice Thickness:1.5mm; Pitch:0.9; Collimation:128×0.625; Reslution:Standard; Rotation Time(s):0.75; FOV(mm):350; Reconstruction Matrix:512x512; Window width/level:250hu/50hu. All patients received less100mL of nonionic contrast medium(ipoamidol,1.5mL of iodine per milliliter) intravenously. A mechanical power injector was used to administer the contrast medium at4mL/sec. Washing with20mL of normal saline. Arterial phase scanning was started12seconds after injection of contrast medium. Each of these patients were randomly performed the enhanced CT scan with standard radiation dose(group A:100kV;250mAs),and50%low dose (group B:120kV;125mAs; group C:140kV,85mAs; group D:100kV,210mAs) in the Portal vein phase,fixed time delays of60seconds.The scanning position and range:Scaning in the condition of breathless,with a supine position; The range is from diaphragm to the edge of the edge of the liver or liver lesions.3. Restruction Methods:Standard convolution FBP was used to reconstruct conventional dose and50%low dose (group B2、C2、D2) imges sets,and iDose4-4was used to reconstruct50%low dose (group B1、C1、D1) imges sets. These image sets were uploaded to the PACS diagnostic workstation for observation and analysis.4. Quantitative Image AnalysisThe images of each set were presented in random order to each of the readers.The patients name, age,and sex,as well as the CT parameters and all hospital record numbers,were removed from the images. The imge quantity was evaluated according to a unified scoring criteria.Ambient room lighting was maintained at a low and constant level for the periods.5. Evalution Indicators:The objective image quality:The objective indicators include noise, SNR and contrast to noise ratio. Quantitative measurements were performed by an abdominal radiology fellow who had5years of experience in gastrointestinal and hepatobiliary. For each patient, all of image sets were displayed side by side with a preset soft-tissue window (window width,250HU; window level,50HU).Mean CT numbers of the liver (in Hounsfield units) were obtained by manually placing circular regions of interest(ROIs) in the hepatic parenchyma(mean pixel number,100; range,93.2-106.8). RIOliver was recorded as the mean of four ROIs placed in the anterior and posterior segments of the right hepatic lobe and the medial and lateral segments of the left hepatic lobe. The contrast-to-noise ratio (CNR) and SNR were calculated with the following formula:CNR=(RIOliver-RIO muscle)/SDbackground; SNR=RIOliver SDbackground,where RIOmuscle is the mean attenuation of the right side of the shaft sma, SDnoise and SDbackground are background noise and the image noise-defined by the standard deviatio drawn in a homogeneous region of the subcutaneous fat of the anterior abdominal wall and the right side of the shaft sma,respectively. For all measurements, the size, shape,and position of the ROIs were kept constant between the two image sets by applying a copy-and-paste function at the workstation. Areas of focal changes in parenchymal density, large vessels, and prominent artifacts,if any, were carefully avoided. Areas of focal changes in parenchymal density, large vessels, and prominent artifacts,if any, were carefully avoided. To ensure consistency,all measurements were repeated three times at different levels,and average values were calculated.6. The subjective image qualityThe subject image quality,lesion conspicuity,and visibility of small portal vein branch were graded by two abdominal radiology fellow who had5years of experience in gastrointestinal and hepatobiliary, respectively,using five-point scale. The diagnostic confidence was assessed by using a four-point scale. The disagreement was resolved by the third senior radiologist.Interreader variability between the two readers for the subjective evaluation of image quality was assessed by using the unweighted k satitic.k Value of less0.20were considered to almost disagreement,0.20-0.40were considered to poor greement,0.41-0.60were considered to medium greement,while k Values of0.61-0.80and0.81-1.00indicated good and excellent agreement,respectively.7. The Radiation DoseThe radiation dose of patients was assessed with CT dose index(CTDI),the CT length product(DLP)and the effective radiation dose(ED).Record the value CTDI and DLP of the scanning process. Calculate the effective radiation dose,using the formula:ED=DLP*K (K=0.015mSv.mGy-1).8. Statistical AnalysisStatistic analyses were done by SPSS13.0software. the objective indicators, including noise, contrast to noise and SNR was analysed with the normality test. If it was meeting the nomal distribution and homogeneity of variance, the random-group analysis of variance was used,by using Bonferroni adjustment for two groups comparision. P<0.05was considered as statistical significance. while, the non-parameters Wilcoxon symbols test was used with the non-normal distribution of objective indicators and subjective quality indicators. P<0.05was considered as statistical significance. Compared between two groups, with P=0.0083indicating a statistically significant difference.Result1. The Effective DoseThe effective dose of group A was about twice as much as other groups in each of46patients.No statistically significant diference in the mean values of effective dose was seen between any two of groups B,C and D(P=0.9P9).The effective dose of group B was generally consistent with group C and group D.2. The influence of using50%-low dose and FBP restruction on image quality.(1) Objective Indicators①Objective Image Noise:the noise value of group A was lower than the other groups (P<0.001). the noise value of group D2is the highest,while the group C2is the lowest.②The CT value of Images:No statistically significant diference in the mean values of CT value between groupA and group B2was seen(P=0.260). Group A was higher than group C2(P<0.001), while lower than group D2(P<0.001).③The contrast to noise ratio of images:the CNR value of group A was high than the other groups (P<0.007). Group B2was higher than group C2(P<0.007), while lower than group D2(P<0.007)④The SNR of images:the SNR value of group A was high than the other groups (P<0.007). Group B2was higher than group C2(P=0.007), while lower than group D2(P=0.001)2) Subjective Indicators①The subjective quality score:the subjective quality score of group A was high than the other groups (P<0.001). Group C2and group B2were higher than group D2(P<0.0083)②Demonstrate of the fine branches of portal vein:No statistically significant diference in the median score between groupA and group D2was seen(P=0.038).Group A was higher than group B2and group D2(P<0.007)③Lesion Conspicuity:Assessment of lesion conspicuity,it is higher in the group A than that in the other groups (P<0.005). Group D2was higher than group B2and group C2(P<0.001)④Diagnostic confidence:Diagnostic confidence:No statistically significant diferencein the median score between group A and group D2was seen(P=0.059).Group A was higher than group B2and group D2 (P<0.001)3. The influence of using hybrid iterative restruction (iDose4-4) on image quality.(1) The effect of hybrid iterative restruction (iDose4-4) on image quality:Both the objective and subjective ratings of low-dose CT images were improved with hybrid iterative restruction (iDose4-4).(2) Comparison between groups:①Objective Image Noise:No statistically significant diference in the mean values of objective image noise between group D1and group A was seen (P=0.305).Group A was higher than group Bland group C1(P<0.001), while group B1higer than group C1(P=0.013)②The contrast to noise ratio of images:the CNR value of group D1was high than the other groups (P<0.001). Group A was higher than group C1(P<0.001), while lower than group B1(P<0.001)③The SNR of images:No statistically significant diference in the mean values of SNR between group C1and group A was seen (P=0.206). Group A was lower than group B1and D1(P<0.05),while group D1was higher than group B1(P=0.001).④The subjective quality score:Group C1was high than the other groups (P<0.0083). Group D1was lower than group B1(P=0.004), while group D1was generally equal to group A (P=0.695)⑤Demonstrate of the fine branches of portal vein:GroupD1was higher than the other groups (P<0.0083). Group C1was lower than group B1(P<0.001), while group B1was generally equal to group A(P=0.127).⑥Lesion Conspicuity:Group D1was generally equal to group A(P=0.394). Group B1was lower than group D1(P<0.001), while higher than group C (P<0.001). ⑦Diagnostic confidence:No statistically significant diferencein the median score between group A, group B1and group D1was seen (P=0.132, P=0.058).Group B1was lower than group D1(P<0.001), while higher than group C (P<0.001)Conclusion1. When the effective dose was reduced equally by50%through three methods of reducing the tube voltage(group D2), increasing the tube voltage(group C2),and reducing the tube current time (group B2), the image noise significantly increased, resulting in a lower image quality score. At the same time, the CNR and SNR decreased. Though the objective indicators decline in the reducing tube voltage(group D2), the image quality can be used for clinical diagnosis.2. The hybrid iterative restruction (iDose4-4) allows radiation dose reduction by50%,through through two methods of reducing the tube voltage(group D2), and reducing the tube current time (group B2), for liver tumor CT examinations whereas improving image quality to meet the diagnostic requirements. The way of decreasing the tube voltage (group D1) was better than that of decreasing tube current time (group B1) in improving CNR and SNR of the image, space resolution and density resolution.The image quality of Increasing the tube voltage (group C1) can not be used for clinical diagnosis.