Objective To evaluate the performance of reconstruction networks with different sparse views under the condition of keeping the same number of model parameters.Methods The number of network channels and network layers were adjusted to make the parameter quantity of each network similar when keeping the structure of each image-domain network and dual-domain network unchanged.The reconstruction performance of each network at different sparsity levels was compared.The AAPM Low-Dose CT Grand Challenge datasets were used in the experiment,including 10 976 images for training,979 images for validation,and 4 256 images for testing.The performance of each model was evaluated visually in combination with objective metrics such as peak signal-to-noise ratio,structural similarity and root mean square error.Results Before adjusting the model parameters,the hybrid domain network Tensor-Net obtained the best visual evaluation and objective evaluation metrics.After parament adjustment,with a similar number of parameters,Tensor-Net outperformed the other models at various projection angles in image anatomical detail recovery,but its structural similarity was slightly lower than that of RED-CNN.The parameters of the hybrid domain model Dual-FBPConvNet were all worse than those of FBPConvNet.Conclusion The hybrid domain model is advantageous in sparse-view CT reconstruction,but it faces more serious overfitting problems.Using a larger image domain model can achieve results similar to those of hybrid domain model.

Objective To investigate the low-dose CT image denoising and generalization performance of the existing mainstream deep learning based denoising networks.Methods The public AAPM Mayo challenge dataset was used to train the denoising network using 3 image-domain methods(REDCNN,WGAN-VGG,CTformer)and 2 projection-image dual-domain methods(VVBP-UNet,CLEAR),separately.The denoising networks were evaluated quantitatively for peak signal-to-noise ratio(PSNR),structural similarity index,root mean square error,number of network parameters and floating point operations,and their generalization performance was analyzed on the AbdomenCT-1K Dataset.Results Image-domain denoising networks effectively suppressed low-dose CT image noise,with REDCNN demonstrating the best denoising performance and achieving a PSNR of 42.0988 dB.The dual-domain denoising networks were better at preserving tiny tissue structures while removing image noise,with VVBP-UNet performing the best and increasing PSNR to 42.150 9 dB.Conclusion The projection-image dual-domain method exhibits superior denoising and generalization performances than the image-domain method,despite requiring a relatively large amount of network parameters and computations.When computing resources are sufficient,the denoising results obtained by dual-domain method better fulfill the requirements for clinical diagnosis.

Objective To investigate the low-dose CT image denoising and generalization performance of the existing mainstream deep learning based denoising networks.Methods The public AAPM Mayo challenge dataset was used to train the denoising network using 3 image-domain methods(REDCNN,WGAN-VGG,CTformer)and 2 projection-image dual-domain methods(VVBP-UNet,CLEAR),separately.The denoising networks were evaluated quantitatively for peak signal-to-noise ratio(PSNR),structural similarity index,root mean square error,number of network parameters and floating point operations,and their generalization performance was analyzed on the AbdomenCT-1K Dataset.Results Image-domain denoising networks effectively suppressed low-dose CT image noise,with REDCNN demonstrating the best denoising performance and achieving a PSNR of 42.0988 dB.The dual-domain denoising networks were better at preserving tiny tissue structures while removing image noise,with VVBP-UNet performing the best and increasing PSNR to 42.150 9 dB.Conclusion The projection-image dual-domain method exhibits superior denoising and generalization performances than the image-domain method,despite requiring a relatively large amount of network parameters and computations.When computing resources are sufficient,the denoising results obtained by dual-domain method better fulfill the requirements for clinical diagnosis.

Objective To evaluate the performance of reconstruction networks with different sparse views under the condition of keeping the same number of model parameters.Methods The number of network channels and network layers were adjusted to make the parameter quantity of each network similar when keeping the structure of each image-domain network and dual-domain network unchanged.The reconstruction performance of each network at different sparsity levels was compared.The AAPM Low-Dose CT Grand Challenge datasets were used in the experiment,including 10 976 images for training,979 images for validation,and 4 256 images for testing.The performance of each model was evaluated visually in combination with objective metrics such as peak signal-to-noise ratio,structural similarity and root mean square error.Results Before adjusting the model parameters,the hybrid domain network Tensor-Net obtained the best visual evaluation and objective evaluation metrics.After parament adjustment,with a similar number of parameters,Tensor-Net outperformed the other models at various projection angles in image anatomical detail recovery,but its structural similarity was slightly lower than that of RED-CNN.The parameters of the hybrid domain model Dual-FBPConvNet were all worse than those of FBPConvNet.Conclusion The hybrid domain model is advantageous in sparse-view CT reconstruction,but it faces more serious overfitting problems.Using a larger image domain model can achieve results similar to those of hybrid domain model.

Objective: To evaluate the effect of height (HT), total body weight (TBW), body mass index (BMI), lean body weight (LBW), body surface area (BSA) and blood volume (BV) on aortic and liver contrast enhancement during upper abdominal contrast-enhanced CT scans. Methods: One hundred and thirteen enrolled patients underwent upper abdominal multiphase contrast-enhanced CT scans. The enhancement (ΔHU) of aorta in hepatic arterial phase and liver parenchyma in portal venous phase were measured and calculated. The ΔHU values difference of aorta and liver parenchyma in subgroups between males and females, TBW<60 kg and TBW≥60 kg, BMI<25 kg/m² and BMI≥25 kg/m² were compared. To evaluate the effect of the patient′s body parameters on aortic and hepatic enhancement, we performed simple linear regression analyses between the change in CT numbers per gram of iodine (ΔHU/gI) at aorta and liver and each of the following: HT, TBW, BMI, LBW, BSA, and BV. Pearson and t test were used. Results: The mean ΔHU values of aorta and liver were significantly lower at males than that of at females (P<0.05). They were higher at TBW<60 kg patients than at TBW≥60 kg patients(P<0.05), and the mean ΔHU values of the liver at BMI<25 kg/m² patients were significantly higher than at BMI≥25 kg/m² patients(P<0.05). The proportion of the ΔHU values at liver less than 50 HU was higher at males (18.3%, 11/60) than at females (7.5%, 4/53). The most obvious negative correlation coefficients were found between the ΔHU/gI of aorta and LBW at hepatic arterial phase (r=-0.559, P<0.01), and between the ΔHU/gI of liver and BSA at portal venous phase (r=-0.680, P<0.01). Conclusion LBW or BSA could be the alternative body index to TBW for the calculation of personalized iodine dose protocol at aortic and liver enhanced CT scans.

Objective To explore the optimal body size index for the calculation of iodine contrast agent dose required for multiphase liver enhanced CT scans based on the total body weight (TBW),lean body weight (LBW)and body surface area (BSA).Methods Two hundred and twenty enrolled patients were randomly divided into three groups,TBW-group (n＝75),LBW-group (n＝72)and BSA-group (n＝73),and administrated iodine doses were 600 mg I/TBW(kg),780 mg I/LBW(kg)and 22 g I/BSA(m2 ),respectively.All patients had taken upper abdominal plain scans and triple-phase enhanced CT scans.The enhanced values (ΔHU)of the aorta at hepatic arterial phase (HAP),the portal vein and liver parenchyma at portal venous phase (PVP)were compared.The correlation coefficients of adjusted maximal hepatic enhancement(aMHE)with TBW,LBW and BSA in three groups were evaluated,respectively.Results There were no statistical differences in the ΔHU values of the aorta at HAP and the portal vein and liver parenchyma at PVP in the three groups respectively.The smallest variances of the aorta at HAP,the portal vein and liver parenchyma at PVP were found in the LBW group. The aMHE showed mildly positive correlation with TBW (r＝0.230)with a P value of 0.047,but it was consistent with LBW (r＝0.158)and BSA (r＝－0.1 54)with corresponding P values of 0.1 85 and 0.1 9 2 ,respectively.Conclusion Compared with TBW and BSA,iodine contrast agent dose calculated based on the patient’s LBW can improve the patient-to-patient uniformities on aorta,portal vein and liver enhancement during the liver multiphase enhanced CT scans.The LBW is the best body index for the calculation of iodine dose on liver enhanced CT scans.

Objective To investigate the effect of tube voltage and iodine concentration of contrast medium (CM) on abdominal dynamic enhanced CT image quality.Methods Six miniature pigs underwent repeated upper abdomen dynamic contrast-enhanced CT scans in 4 scanning protocols with different concentration of CM and tube voltage,namely,protocol 1,CM with iodine concentration of 270 milligrams iodine per milliliter (mg/ml) and 80 kV tube voltage;protocol 2,270 mg/ml and 120 kV;protocol 3,370 mg/ml and 80 kV and protocol 4,370 mg/ml and 120 kV.The same iodine dose (600 mg/ml) and iodine delivery rate (IDR) (920 mg/s) were used in all protocols.The CM with iodine concentration of 270 mg/ml were injected at a flow rate of 3.4 ml/s,and 370 mg/ml CM injected at 2.5 ml/s.Image reconstruction was performed with iterative reconstruction (iDose4) in protocol 1 and 3,filtered back projection (FBP) was used in protocol 2 and 4.A subjective scoring system for image quality,image noise and sharpness was conducted by 2 radiologists independently.The measured values (peak of enhanced CT values,image noise of aorta,inferior vena cava,portal vein,hepatic vein and liver parenchyma) as well as the calculated values [their time-to-peak,signal-to-noise (SNR) and contrast-to-noise (CNR) ratios] were compared between among 4 protocols.The CT volume dose index (CDTIvol) and dose length product (DLP) were recorded from the CT console after each scanning.Factorial designed ANOVA was used for comparison of enhanced CT values of vessels and liver parenchyma,noise,SNR and CNR.The Kruskal-Wallis test was used for comparison of values among the 4 protocols,including the time-to-peak enhancement of vessels and liver parenchyma,the subjective scores of image quality indices.Result There was no significant differences in subjective scores of the image quality,image noise and image sharpness (P＞0.05).The scored were more than 3,and the images with 4 scanning protocols were all acceptable for diagnosis.There was no significant differences between protocol 1 and 3,protocol 2 and 4 in the peak enhancement CT values of aorta [(729±46) HU vs.(707±59)HU,(515±84)HU vs.(513±53)HU],inferior vena cava [(366±95)HU vs.(368±92)HU,(282±39)HU vs.(262 ± 67)HU],portal vein [(213± 18)HU vs.(201 ±29)HU,(180±21)HU vs.(176±27)HU],hepatic vein [(207±18)HU vs.(193±10)HU,(179±24)HU vs.(170±14)HU] and liver parenchyma [(128±7) HU vs.(127±4) HU,(135±5)HU vs.(135±6)HU] (P＞0.05).But the CT values of vessels (aorta,inferior vena cava,portal vein and hepatic vein) in protocol 1 and 3 were significantly higher than those in protocol 2 and 4 (P＜0.05),the CT values of liver parenchyma in protocol 1 and 3 were significantly lower than values in protocol 2 and 4 (P＜0.05).The image noises of vessels were higher in protocol 1 and 3 than noises in other protocols (P＜0.05),but there was no significant difference in liver parenchyma noise among protocols (P＞0.05).No significant differences were observed on the peak times,SNR and CNR in aorta,inferior vena cava,portal vein,hepatic vein and liver parenchyma among 4 protocols (P＞0.05).The CDTIvol and DLP were 199.67 mGy,1 597.4 mGy· cm respectively in protocol 1 and 3,585.12 mGy and 4 680.9 mGy· cm in protocol 2 and 4 (scanning with 120 kV).Conclusions CM with different iodinated concentration could achieve the same enhancement in the abdominal vessels and liver parenchyma by using the proper scan protocols,which have the same IDR and iodine dose per kilogram body weight.Higher vessel enhanced peak values were achieved when using the protocols with 80 kV tube voltage than 120 kV.By using a low dose protocol of 80 kV tube voltage with the iterative reconstruction algorithm,the quality of image can be warranted.