Discussion On The Clinical Application Of AI-Assisted Head And Neck CTA In Diagnosis Of Intracranial Aneurysms

Objective:To explore the application value of AI-assisted head and neck CTA in the diagnosis of intracranial aneurysms,in order to further improve the efficiency and accuracy of the detection of intracranial aneurysms and provide more reliable imaging basis for the clinical diagnosis and treatment of patients.Methods:（1）The clinical data of 96 patients with suspected intracranial aneurysms who met the inclusion criteria in the Affiliated Hospital of Youjiang Medical University for Nationalities from May 2021 to May 2022 were retrospectively collected.（2）The CTA data of all included patients were analyzed by AI post-processing technology（AI group）and manual post-processing technology（manual group）,and the original images were post-processed.The subtraction effect and 3D-VR image display effect of the two techniques were evaluated,and the radiation dose and post-processing time of the two methods were calculated.（3）The detection and diameter measurement results of intracranial aneurysms in the two groups were recorded.Taking DSA as the gold standard,the sensitivity and accuracy of the diagnosis of intracranial aneurysms in the two groups were calculated and compared with the aneurysm measurement results of DSA.（4）SPSS22.0 statistical software was used to analyze the above data,and t test was used to compare the measurement data,P<0.05 was considered statistically significant.Non-normal distribution data were expressed as interquartile range M(P₂₅-P₇₅),Wilcoxon rank sum test was used for comparison,and P<0.05 was considered statistically significant.Enumeration data were analyzed by pairedχ²test,and the difference was considered statistically significant when P<0.05.Results:Of the 96 patients,71 were diagnosed and confirmed by DSA with a total of 92aneurysms.（1）Post-processing time:The post-processing time of the manual group was 380（342-434）s,and that of the AI group was 3.05（2.79-3.22）s.The post-processing time of the AI group was significantly shorter than that of the manual group（P<0.01）.（2）Radiation dose:The DLP and CTDI in the manual group were 929.04（876.43,985.28）m Gy·cm and20.46（19.67,20.77）m Gy,respectively,while those in the AI group were 470.91（440.24,496.46）m Gy·cm and 10.29（9.89,10.29）m Gy·cm,respectively.10.43)m Gy,the radiation dose of AI group was significantly lower than that of manual group（P<0.01）.（3）3D-VR image display effect:The 3D-VR images processed by the AI group could only be rotated in the top,bottom,left and right directions,which could not meet the observation requirements of the three-dimensional morphology and spatial relationship of the vascular lesion site in the head and neck,while the manual group could perform multi-directional rotation according to the needs,which could better present the complex geometric features and spatial relationship of the lesion site.（4）Subtraction effect:The AI group had poor subtraction effect on aneurysms located in the middle cerebral artery,which had statistical significance compared with the manual group（P<0.05）,and the coincidence rate of subtraction effect on aneurysms in other parts was higher（P>0.05）.（5）Detection efficiency:The sensitivity,specificity and accuracy of diagnosing intracranial aneurysms were 88.73%,80.65%and 91.66%in the AI group and 91.55%,80.00%and 96.87%in the manual group,respectively.There was no significant difference in the diagnosis of intracranial aneurysms at the patient level between AI group,manual group and DSA（P>0.05）.In terms of the number and location of aneurysms,the number of misdiagnosis and missed diagnosis in the AI group was higher than that in the manual group.（6）Quantitative analysis:Compared with the measured values of DSA,the measured values of tumor neck width in the AI group had statistical significance（P<0.05）,while the measured values in the manual group had no statistical significance（P>0.05）.There were statistically significant differences between the manual group and the AI group in the tumor<3 mm measurements（P<0.05）,but there were no statistically significant differences in the tumor length diameter 3～5 mm and>5 mm measurements（P>0.05）.Conclusion:（1）The subtraction effect of AI technology on intracranial aneurysm is more time-saving,labor-saving and can reduce radiation dose.If it is necessary to observe the complex morphology of aneurysms,manual subtraction technique should be selected.（2）AI-assisted head and neck CTA has a high value in diagnosis of intracranial aneurysms,and AI technology can be used as a intracranial aneurysm-screening method for further promotion and application.At the same time,this technique has a certain false positive rate and false negative rate,so it need to be flexibly combined with traditional manual diagnosis to achieve the best correct diagnosis.