冠状动脉VR成像人工重建与人工智能重建的对比研究

Comparative study of artificial reconstruction and artificial intelligence reconstruction in coronary VR imaging

目的探讨在冠状动脉容积再现(Volume Rendering,VR)成像中人工重建与人工智能重建(Artificial Intelligence,AI)的对比研究。方法使用东芝320排Aquilion ONE CT机对2021年4月17日—4月22日在我院进行冠脉CT检查的患者102例[男56例,女46例,平均年龄(60.28±10.37)岁]扫描所得的图像分别进行VR成像的人工重建(A组)和人工智能重建(B组),对两组图像的总评分、分析组评分及非分析组评分两两对比研究。结果将102例患者,其中男56例,女46例,年龄37~84岁,平均(60.28±10.37)岁,进行人工重建和人工智能重建两组所得的各项分值进行对比,人工组的总评分为(52.22±7.79)%,其中分析组评分为(69.84±9.26)%,非分析组评分为(19.41±14.57)%;人工智能组的总评分为(88.42±5.98)%,其中分析组评分为(92.16±5.79)%,非分析组评分为(79.81±14.21)%。结果显示人工智能组在整体血管包括分析组血管及非分析组血管的数目、长度的显示均优于人工组。结论AI检出血管的分支数以及质量较人工重建高,并且操作时间短,提高了操作技师工作效率。

Objective To explore the comparative study of artificial reconstruction and artificial intelligence(AI) in coronary volume rendering(VR). Methods The images of 102 patients [56 males and 46 females, with an average age of(60.28±10.37) years] who underwent coronary CT examination in our hospital from April 17 to April 22, 2021 were scanned by Toshiba 320 row Aquilion ONE CT machine for VR imaging artificial reconstruction(group A) and artificial intelligence reconstruction(group B), respectively. The total score, analysis group score and non-analysis group score of the two groups were compared. Results 102 patients, including 56 males and 46 females, aged 37 to 84 years old, the average age was(60.28±10.37), were compared with the scores of the two groups of artificial reconstruction and artificial intelligence reconstruction. The total score of the artificial group was(52.22±7.79)%, of which the score of the analysis group was(69.84±9.26)%, and the score of the non-analysis group was(19.41±14.57)%. The total score of the AI group was(88.42±5.98)%, including(92.16±5.79)% in the analysis group and(79.81±14.21)% in the non-analysis group. The results showed that the number and length of vessels in the AI group, including those in the analysis group and those in the non-analysis group, were better than those in the artificial reconstruction group. Conclusion The number and quality of branches detected by AI is higher than that of artificial reconstruction, and the operation time is shorter, which improves the work efficiency of the operation technician.