Feasibility Of Individualized Low Dose Chest Enhanced CT Imaging:Phantom And Clinical Study

Purpose:（1）To investigate the effect of different imaging parameters on the image quality and radiation dose of chest-enhanced CT by applying the PH-75 dual-energy CT quality control phantoms,and to investigate the optimal parameters for individualized low-radiation-dose chest-enhanced CT imaging.（2）To investigate the effect of the combination of Auto-prescription and ODM_full on image quality and radiation dose of chest enhanced CT and explore the feasibility of individualized low dose chest enhanced CT.Materials and methods:（1）phantom study:3 concentrations of iodine rods were placed in PH-75 dual-energy CT quality control molds（2 phantoms:TR-I,TR-J）:12mg I/ml,8 mg I/ml,and 4 mg I/ml.Two phantoms simulated the standard and larger size of patients,and the different material rods were used to simulate the degree of vascular strengthening during the arterial phase（CT value is about 300HU）,the degree of vascular strengthening during the venous phase（CT value is about 200HU）,the degree of strengthening of lung masses or solid nodules（CT is about 100HU）,and the remaining 5holes are placed with water equivalent rods（CT is about 0HU）.The following four scanning modes were used to repeat the scanning of three times,the voltage of group A tube was 120 k Vp,and the tube current was used in automatic tube current modulation mode（Smart-m A）;The tube voltage of group B was 120 k Vp,and thefull field of organ dose modulation(ODM_full);Group C tube voltage adopted automatic preset technology（Auto-prescription）,and the tube current was Smart-m A mode;The voltage of group D tube adopted Auto-prescription,and the tube current was ODM_full mode;Each group reconstructed adaptive iterative reconstruction algorithms（ASIR-V）40%～80%,with an interval of 20%,and each group reconstructed 3 subgroup images,denoted as A_1-3～D_1-3.The image data was transmitted to the AW 4.7 workstation,the CT value and SD value of iodine rods of three concentrations and water equivalent rod（background）were measured on the axial image,and SNR and CNR were calculated,the measurement was repeated 3 times and averaged,and the image quality was scored on a 5-point scale and kappa consistency test was performed by two observers,and the CTDI_vol and DLP values of each group were recorded.The image quality and radiation dose of each group were analyzed,and the subjective and objective scores of the images and the radiation dose were combined to select the best imaging mode.（2）Clinical study:The best imaging mode obtained in（1）(Auto-prescription/ODM_full,ASIR-V60%)was used for clinical research,and 55 patients with clinically confirmed pulmonary lesions（masses or solid nodules）were prospectively collected,and each patient underwent two chest enhanced CT scans in our hospital according to clinical needs,and the first enhanced CT parameters were the same as those in group A₁（120k Vp/Smart-m A,ASIR-V40%）,It is recorded as standard dose CT（SDCT）,and the second enhanced CT examination is recorded as low-dose CT（LDCT）using the D₂ group parameters in（1）.The image data was transmitted to the AW 4.7 workstation,and the CT value and SD value of the descending aorta and erector spinal muscle（background）at the tracheal carina level of the two sets of axial images of the arterial and venous phase were measured,and the SNR and CNR were calculated.The image quality was scored on a 5-point scale by two observers,and the imaging performance of the two groups was recorded,and the kappa test was used to check the consistency of the scores of the two observers,and the recommended value of the LDCT tube voltage system was recorded.The obtained image data is transmitted to u AI Research Portal V1.1 of UII,and the results of automatic segmentation of left lung,right lung and trachea volume were measured and recorded respectively.The paired sample t test or Wilcoxon rank test were used to compare the subjective and objective evaluation of the two groups,the imaging manifestations of the lesion,and the results of automatic segmentation of left lung,right lung,and trachea volume.Results:（1）Phantom study:the difference in radiation dose between groups was statistically significant（H=10.385 to 10.421,P<0.05）,and all radiation doses were lowest in group D.CTDI_vol was 2.22±0.01 m Gy,6.83±0.07 m Gy,respectively;DLP was 53.89±0.58 m Gy×cm,respectively.192.76±2.42 m Gy×cm.For the image quality of the 2 sizes of phantoms,when A₁ was used as the control group for a two-by-two comparison,（1）SNR_{12 mg I/ml}:the image SNR of B₁-B₂,C₁,and D₁ groups was not statistically different from A₁（P=0.083-0.997,P=0.072-0.999）.The SNR of the remaining groups was higher than A₁（F=38.970,F=16.52,P<0.05）;SNR8mg I/ml:the image SNR of A₂,B₁-B₂,C₁,and D₁ groups was not statistically different from A1（P=0.399-0.999,P=0.170-0.999）,and the SNR of the remaining groups was higher than A₁（F=11.910,F=10.820,P<0.05）;SNR_{4mg I/ml}:no statistical difference between image SNR of A₂,B₁-B₂,C₁,D₁ groups and A₁（P=0.096-0.809,P=0.085-0.999）,SNR of the remaining groups were higher than A₁（F=71.650,F=10.540,P<0.05）;（2）CNR_{12mg I/ml}:no statistical difference between CNR and A₁ in images of B₁,C₁ and D₁ groups（P=0.094-0.347,P=0.118-0.996）,CNR was higher than A₁ in the remaining groups（F=1160.000,F=105.000,P<0.05）;CNR_{8mg I/ml}:images of B₁,C₁ and D₁ groups CNR was not statistically different from A₁（P=0.128-0.898,P=0.698-0.822）,while CNR was higher than A₁ in the remaining groups（F=534.000,F=123.700,P<0.05）;CNR_{4mg I/ml:}CNR was not statistically different from A₁ in images of B1,C1 and D1groups（P=0.081-0.817,P=0.817）.0.817,P=0.052 to 0.976),and the CNR of the remaining groups was higher than A₁（F=64.850,F=45.830,P<0.05）;（3）subjective scores:2 observers had a good agreement on image quality（kappa=0.762 to 0.954,kappa=0.689 to 0.954,P<0.05）The subjective image scores of the A₂,B₁,C₁-C₂,and D₁-D₂ groups were not statistically different from A₁（P=1.527 to 0.999,P=0.391 to0.999,）;the remaining groups were lower than A₁（H=105.500,H=124.800,P<0.05,Figure 3）.Therefore,D₂ was the best imaging mode.（2）Clinical study:there was no statistical difference in age,height,weight,and BMI between the 2 groups（all P>0.05）.systemic tube voltage values in the LDCT group were80 k Vp（37 cases,67.30%）and 100 k Vp（18 cases,32.70%）,respectively.single-phase CTDI_vol was 3.23（2.89,4.40）m Gy in the LDCT group and single-phase in the SDCT group CTDI_vol was 6.15（4.80,7.97）m Gy（Z=6.509,P<0.001）;single-phase DLP was126.41（111.39,169.82）m Gy×cm in the LDCT group and 240.55（192.09,294.98）m Gy×cm in the SDCT group cm（Z=6.509,P<0.001）;the single-phase ED was 3.37（2.69,4.13）m Sv in the LDCT group and 1.767（1.56,2.38）m Sv in the SDCT group（Z=6.509,P<0.001）.the LDCT group had higher arterial and venous phase images of descending aortic CT The difference was statistically significant（t=12.524,t=12.032,P<0.001）;the difference between the background SD values of the LDCT group and the SDCT group was not statistically significant（P=0.094）.t=13.299,P<0.001);the consistency of image quality scores by the 2 observers was good（kappa=0.754,P<0.001）,and all images met the clinical diagnostic requirements.The scores of the images in both groups were 4.0（4.0,4.0）points,and the differences were not statistically significant（Z=-1.014,P>0.05）.automatic segmentation of the left lung parenchyma（t=-1.067,P=0.291）,right lung parenchyma（t=0.639,P=0.526）and trachea（t=0.108,P=0.914）in the arterial phase of the 2 groups based on AI software There was no statistical difference in the volume of the left lung parenchyma（t=1.045,P=0.302）,right lung parenchyma（t=-1.082,P=0.285）and bronchus（t=0.856,P=0.396）auto-segmented volumes based on AI software in the arterial phase in the 2 groups.Conclusion:Auto-prescription combined with ODM in chest CT enhanced imaging can achieve individualized low dose scanning with satisfactory image quality and diagnostic requirements.