http://chineseinput.net/에서 pinyin(병음)방식으로 중국어를 변환할 수 있습니다.
변환된 중국어를 복사하여 사용하시면 됩니다.
간 자기공명영상검사에서 VIBE 시퀀스의 고식적인 방법과딥러닝의 비교 및 유용성 평가
이충성,백인성,김경환,강보걸 대한자기공명기술학회 2023 대한자기공명기술학회지 Vol.33 No.3
Liver DCE imaging uses the VIBE sequence, but previous research into conventional and deep-learning methods using AI is lacking. We therefore used image evaluation and examination direction and validity . The ACR phantom study was repeated 30 times. In the low contrast resolution evaluation area, SNR and CNR were evaluated using Syngovia View&Go. In the spatial resolution evaluation area, the height of signal intensity and FWHM were evaluated through MATLAB. For the patient study, Matrix 352 was set up based on the phantom study and tested on 30 patients. SNR and CNR were evaluated at the liver parenchyma, hepatic portal vein, and descending aorta. Spatial resolution was evaluated at the borders of the hepatic portal vein and descending aorta. The data were analyzed using a two-way ANOVA and a post-hoc analysis was performed using the Duncan test. The results revealed significant differences in the SNR and CNR of the phantom study under Matrix 416, while spatial resolution could not be evaluated under the conventional method Matrix 352 or the deep-learning method Matrix 288. The results of the patient study showed significant differences in SNR, CNR, and spatial resolution. The deep learning method improved the image more than the classical method, and the acquisition time was reduced by an average of 4 seconds (22.4%). When Matrix 352 was applied in the deep learning method, there was a decrease in reproducibility and respiration artifacts due to a reduction in scan time. Accordingly, we recommend applying Matrix size in the deep learning method.