Application value of deep learning algorithm combined with low-kV and low-dose contrast agent scan in coronary CT angiography in overweight patients
WANG Faquan
WANG Zhongming
SHANG Wei
ZHU Xin
Abstract：Objective To explore the application value of deep learning algorithm(DLA)combined with low-kV and low-dose contrast agent scan(dual-low scan)in coronary CT angiography(CCTA)in overweight patients.Methods A total of 100 overweight patients[body mass index(BMI)≥26 kg/m2]with suspected or confirmed coronary artery disease were prospectively enrolled in Jiaozuo People's Hospital of Henan Province from July to Oct.2024.They were divided into control group(n=50,tube voltage:100 kVp,contrast agent dose:0.7 mL/kg)and experimental group(n=50,tube voltage:80 kVp,contrast agent dose:0.35 mL/kg)using random digital table method.Image reconstruction was performed with 50%adaptive statistical iterative reconstruction(50%ASIR-V)and deep learning image reconstruction-high(DLIR-H)algorithm for 2 groups,respectively,with all other parameters kept consistent.CT values and noise were measured at the proximal,middle and distal segments of the right coronary artery,left anterior descending artery and left circumflex artery,followed by the calculation of signal-to-noise ratio(SNR)and contrast-to-noise ratio(CNR).The subjective image quality was evaluated by 2 radiologists using the 5-point Likert scale in 2 groups.The data conforming to normal distribution were analyzed with independent-samples t-test,otherwise the Mann-Whitney U test was used.Kappa test was adopted for consistency analysis of subjective scores.Results There was no statistically significant difference in the baseline data between 2 groups.The mean effective radiation dose and contrast agent dose were 3.70±0.92 mSv and 28.86±4.09 mL,respectively,in experimental group,which were reduced by approximately 26.6%and 48.8%compared with those in control group(5.04±1.58 mSv and 56.41±7.63 mL,respectively;t=-5.157 and-22.517,respectively,P<0.001).However,the image noise of CCTA images was lower,while both SNR and CNR were higher in experimental group than those control group(P<0.05).The experimental group had significantly better subjective scores in noise level,vascular margin and overall image quality than control group(P<0.01),whereas there was no statistically significant difference in lumen contrast score between 2 groups(P=0.075).Conclusion For overweight patients,80kVp combined with DLA improves significantly CCTA image quality,and reduces radiation dose and contrast agent dose.
Keywords：coronary computed tomography angiographylow-kV and low-dose contrast agent scanoverweightdeep learning algorithm
Publication Date：2025-12-20
Online Publishing Date：2026-03-25(First online date of this platform, not the publication date of the document)
Pages：6( 1507-1511,1517 )