Effects of Different Scanning Methods on Artificial Intelligence
Reconstruction and Analysis in CCTA
GENG Jigang1, LI Xiaoshi1, ZHU Yinhu1, JIN Dayong1, LI Xin1, ZHANG Weiqi2, QIN Yue1
1. Department of Radiology, Xi’an Daxing Hospital, Xi’an Shaanxi 710016, China;
2. Department of Radiology, Xijing Hospital Affiliated to Air Force Military Medical University, Xi’an Shaanxi 710032, China
Abstract:Objective To analyze and evaluate the effect of two different scanning modes on the evaluation of post-processing
results of coronary computed tomography angiography (CCTA) and artificial intelligence (AI). Methods A total of 100 patients
who underwent coronary CT angiography review in our hospital from April to December 2021 were selected, and the patients
were randomly divided into group A and group B, with 50 cases in each group. Group A and group B were then divided into two
subgroups according to heart rate: the heart rate was ≤75 times/min for the A1 (33 cases) and B1 (35 cases) subgroups, and the heart
rate was >75 times/min for the A2 (17 cases) and B2 (15 cases) subgroups. Group A used volume scanning mode, and group B used
helical scanning mode. All images were sent to the Shukun AI post-processing station for reconstruction and analysis, and subjective
quality scores were performed on the volume rendering (VR) images and curved projection reformation (CPR) images automatically
reconstructed by AI according to different subgroups. By measuring the coronary area of interest, including CT value, SD value, and
calculating the contrast signal to noise ratio (CNR) and signal to noise ratio (SNR) values, objective image quality measurement was
performed, and the radiation doses between the two group A and group B were compared. Results In the VR subjective image quality
score, there was no statistically significant difference between group A1 and group B1 (P>0.05), and the image quality of group A2
was slightly higher than that of group B2 (P<0.05). In CPR subjective image quality evaluation, there was no statistically significant
difference between group A1 and group B1 (P>0.05), the score of image quality of group A2 was lower than that of group B2, and
the difference was statistically significant (P<0.05). In the objective image quality comparison, there was no statistically significant
difference in the SNR value of the distal segment of the anterior descending artery in the low heart rate subgroup between group A and group
B (P=0.8091), and the SNR values of all other segments in group B1 were significantly higher than group A1 (P<0.05). While in the high
heart rate subgroup, the SNR value of the distal segment in group B2 was lower than that in group A2, and the SNR value of the proximal
and middle segments in group B2 was greater than that in group A2 (P<0.05). Conclusion When the heart rate is ≤75 times/min, the image quality of the post-processed images of the volume scanning AI was similar to that of the helical scanning. When the heart rate is
>75 times/min, the image quality of the helical scanning is better. But because the radiation dose of volumetric scanning is greatly
reduced, it is recommended to use volume scanning when the heart rate is low (≤75 times/min), and use helical scanning when the
heart rate is high (>75 times/min).
耿纪刚1,李笑石1,朱寅虎1,金大永1,李馨1,张巍琪2,秦越1. 不同扫描方式对CCTA人工智能重建与分析的影响[J]. 中国医疗设备, 2022, 37(8): 70-75.
GENG Jigang1, LI Xiaoshi1, ZHU Yinhu1, JIN Dayong1, LI Xin1, ZHANG Weiqi2, QIN Yue1. Effects of Different Scanning Methods on Artificial Intelligence
Reconstruction and Analysis in CCTA. China Medical Devices, 2022, 37(8): 70-75.