99 - Application of perfusion-weighted imaging technology combined with artificial intelligence diagnostic system in acute ischemic stroke.

Autor: Huang, Dr. Yan, Li, Dr. Baosheng
Předmět:
Zdroj: Journal of Medical Imaging & Radiation Sciences; 2024 Supplement, Vol. 55 Issue 3, pN.PAG-N.PAG, 1p
Abstrakt: To explore the image quality and diagnostic efficacy of perfusion-weighted imaging technology combined with artificial. 52 patients with acute ischemic stroke underwent additional scanning ASL sequence scanning was selected as the observation group; another with DSC sequence was used as the control group. The images were uploaded to the Fast-Stroke AI diagnostic system for analysis. Two senior radiologists used the double-blind method to subjectively score the pseudo-color images processed by the two groups of sequence AI using the Likert scale and observed whether the two groups of pseudo-color images met the diagnostic requirements and whether there were differences in image clarity, image distortion, and artifacts between the two groups. Two weeks later, scored again and observed the consistency of the scores. Objective evaluation involves conducting inter-group tests based on two sets of CBF parameters calculated by AI. SPSS statistical software was used to statistically analyze the data. The measurement data were expressed as mean ± standard deviation, and a t-test was used for comparison between groups; the Kappa consistency test was performed on the subjective score results of two diagnostic physicians, with P < 0.05 as the difference with statistical significance. In subjective evaluation, the difference in scores between the two observers was not significant and there was strong to strong consistency (Kappa value ≥ 0.75). The subjective image quality of pseudo-color images in ASL and DSC is equivalent and there is no significant statistical difference (P>0.05). There was no significant statistical difference between the two groups of CBF parameters (P>0.05). Currently, ASL technology is developing rapidly, and AI diagnostic systems have good diagnostic performance for both ASL and DSC sequences. The combination of the two is worth promoting in acute cerebral infarction imaging. [ABSTRACT FROM AUTHOR]
Databáze: Supplemental Index