Application of three dimensional artificial intelligence localization technology in CT chest scanning
10.3760/cma.j.cn112149-20210301-00168
- VernacularTitle:三维人工智能定位技术在CT胸部扫描中的应用
- Author:
Shangwen YANG
1
;
Xiaoqian ZHU
;
Xiaoyan XIN
;
Xin ZHANG
;
Mingjun WANG
;
Bing ZHANG
Author Information
1. 南京大学医学院附属鼓楼医院医学影像科，南京 210008
- Keywords:
Artificial intelligence;
Tomography, X-ray computed;
Thorax
- From:
Chinese Journal of Radiology
2022;56(1):50-54
- CountryChina
- Language:Chinese
-
Abstract:
Objective:To evaluate the clinical application of three dimensional artificial intelligence(3D-AI) localization technology in chest CT scan.Methods:A total of 100 patients who underwent chest CT for COVID-19 screening in Nanjing University Medical School Affiliated Drum Tower Hospital were collected from September 2020 to October 2020 were analyzed retrospectively. The patients were divided into manual positioning group ( n=50) and 3D-AI automatic positioning group ( n=50) with block randomization method. All patients were scanned with the same CT scanning protocol. The off-center distance, CT dose index (CTDI), dose length product (DLP) and CT examination time were measured and recorded. Quantitative image evaluation of mediastinal window images and qualitative image evaluation of chest window images were assessed by two radiologists. The off-center distance, CTDI, DLP, CT examination time and objective indexes of image quality of two groups were compared by independent sample t test. The quantitative image quality scores were compared with χ 2 test. Results:Compared with manual positioning group, the overall off-center distance of 3D-AI automatic positioning group was reduced by 42.86% [(15.4±9.7) vs. (8.8±7.2)mm, t=3.65, P<0.01], CTDI was reduced by 10.67%[(7.5±2.5) vs. (6.7±2.6)mGy, t=0.59, P=0.04], DLP was reduced by 13.33%[(270±95) vs. (234±86)mGy·cm, t=1.98, P=0.02], the average examination time was reduced by 29.91% [(214±26) vs. (150±14)s, t=15.79, P<0.01]. There were no significant differences in the background noise, signal to noise ratio of descending aorta and erecting spinal muscle, and subjective score between two groups ( P>0.05). Conclusion:The 3D-AI automatic positioning technology can greatly improve the accuracy of patient positioning and reduce the radiation dose for chest CT imaging, and improve work efficiency with qualified chest CT image quality.
