| ObjectiveTo study the advantage and limitation of virtual non-contrasted of suspected liver mass genarated from dual source computed tomography,and discuss its postential clinical application in upper abdomen.Materials and MethodsThis prospective study was approved by the institutional review board; all patients provided written informed consent.Fifty there person diagnosed to have liver mass by type-B ultrasonic underwent preoperative DE CT which included unenhanced(ture noncontrast, TNC), arteries, portalvein, and delayed phases.DE sequence was performed during portalvein phase 60-70s arfter bolus injecting Ultravist 370 mg I/ml(Bayer Schering Pharma, Berlin, Germany) 70ml or 1.5ml/kg from antecubital vein to generated iodine graphic. DE sequence parameters were 80 kV/468 mAs and 140 kV/96 mAs (effectivemAsm a ratio of 1:3 between tube currents). Collimation was 14±1.2 mm. Two sequens of imagine (80Kv,140kV) with slice thickness of 0.75mm and position imcreacement of 5mm were automaticly generated and then loded into workstation(syngo MMWP,version 2008A;Siemens Medical Solutions)to genetated VNC imagine by changing the mix ratio of iodine to 0% using liver VNC utility program;then change the same window level and width as TNC before saving. Mean CT numbers, standard error of the same ROI in both TNC and VNC teams were measured and recorded by the same investigator as well as catogories of artifact which namely steak, subtraction, rim, cupping artifact. Image quality and noise (1--none,5-severe) and acceptability for VNE and TNE images scored by two radiologists with clinical experience of 25 years and 7 years respectely were rated. Effective radiation doses for DE CT and TNE images were calculated. ROI included isodense liver parenchyma, spleen, pancreas, both kedneys at remal hilum level,aorta, erector spinae muscle, gallbladder, retroperitoneal fat.ROI should be selected at the same localization and similar area between 0.5-3cm2 by using side-by-side comparasion of TNC and VNC, and follow the principles as:to get out of the way of visible focus(such as low density,calculi or calcification, cholangiectasis), to avoid capacity and the rim of B bulb.Results14 person out of 53 were overwight(BMI surpass normal range),17 livers excluded from the 26-cm field of view detector with an average of 1.6cm(measurement took on the maximum cross section).13 cases were diagnosized as hepatic cell carcinoma,2 cholangiocellular carcinoma,2 hepatic metastasis,8 cavernous hemangioma of liver,3 Focal Nodular Hyperplasia,3 hepatic cirrhosis,9 fatty liver,6 hepatic cyst,1 intrahepatic bile duct stone,2 calcification endo-liver,9 no abnormality seen.Other morbib change were:5 renal cyst,3 cholecystolithiasis,8 renal calculi.Mean CT numbers (standard deviation) on TNE and VNE images, respectively, for isodense liver parenchyma were 58.5HU±5.6,58.2HU±6.7, P±.223; low density lesions(including fatty liver,43.8HU±2.9,44.0HU±3.2,P=.622, and without fatty liver,38.8HU±8.1,37.4HU±8.3,P=.000),39.8HU±7.6,38.7HU±8.0, P=.000; Difference between liver and spleen in fatty cases of the two team were 3.1HU±3.0,1.2HU±2.9, P=0.000; hydatid fluid,10.5HU±9.4,-0.4HU±6.1,left kidney, 35.6FHU±4.0,23.0HU±5.2,P=0.000right kidney,35.5HU±4.1,23.2HU±5.0,P=0.000, pancreas,49.1HU±4.9,45.8HU±5.9,P=0.000,spleen,48.4HU±3.4,45.7HU±3.5, P=0.000, aorta,47.4HU±3.9,38.6HU±4.7,P=0.000, gallbladder,20.3HU±7.6; 17.9HU±10.6, P=0.000, erector spinae muscle,52.5HU±4.0; 52.3HU±5.2, P=0.658,)Retroperitoneal fat,-98.1HU±16.3,-96.1HU±10.4, P=0.392. Mean CT numbers of both kidney on VNC images had no significant deviation (P=0.258).There was no significant deviation in Signal-noise ratio(SNR) of pancreas,gallbladder and isodense liver parenchyma on VNC and TNC images, P=0.314,0.207, 0.940,respectely. Mean effective dose for DE CT scans of the abdomen was 3.36 mSv±0.9 and that for nonenhanced scans was 5.08mSv±0.55, P=.000, correlation coefficient r=0.622, Paired Differences Mean was 1.73 mSv.2 rim,3 substraction,1 radiated artifacts on VNCimages; 2 steak,1 radiated artifact on TNC images.Both radiologists considered that the artifacts and the exclution from the 26-cm field of view detector had no affection on diagnosis.Since images can ne easily discriminated, radiologists were not not blinded as which image reprensents the TNC and VNC data set. Both adiologist 1 and 2 rated 41 of 53 VNC as excellent,4 and 5 as good,5 and 4 as fair,3 poor and no not interpretable. In all but three patients radiologists accepted VNE images as replacement for TNE images.Those who were rated as fair or poor sufferred with wasting disease such as hepatic carcinoma, metastatic tumor, which induced lack of fat in the abdominal cavity and accumulation of intestinal canal, and resulted in lack of contradistinction between different tissues.1 case was rated as good and fair respectely by the two readers because the upper part of porta hepatis was clear but ambiguous in the under part. Significant deviation was found in images quality between TNC and VNC(P=0.000, relevanted samples nonparameter test),which means that VNC images were inferior toTNC images. No missing focus on VNC images,1 kidney stone covered by excretion of iodin.The enhancing part of focus can be macroscopicly and exactly discriminated while using side-by-side comparison(VNC/Overlay map) generated by the Dual Energy software,and all the cases can reach a satisfactory diagnosis when VNC combined togrther with the three phase.ConclusionVNC images had lower spatial resolution than TNC images, accompanyed with rim and subtraction artifacts,insensitivety of fatty liver,and poor or fair images quality for those had little fat,staking with intestinal canal in abdominal cavity.though,VNC can reflact macroscopisly and accuracyly the enhancement of liver focus, which is helpful for diagnosis, that means VNC images can reach a good diagnosis level and save radiation dose at the same time.Also, potential applications for other focus in upper abdomen,such as gastroenteric tumor, cholecystopathy, nephropathy remain to be discussed.
|
PDF Full Text Request