Value of artificial intelligence image analysis in reducing radiation dose of head and neck CT angiography

LUO Peilin; LI Wei; HONG Xuanyang; ZHONG Wen; DAI Weichuang

Abstract：Objective:To study the specific manifestations and application value of artificial intelligence image analy-sis technology in head and neck CT angiography with different radiation doses,explore appropriate low-radiation-dose scanning protocols.Methods:Ninety patients who were suspected of having cerebrovascular diseases from April 2022 to August 2023 were selected to undergo head and neck CT angiography.These patients were divided into three groups of 30 cases each according to the random number table method.The tube voltage of each group was 80 kV.The radiation dose(CTDI)in group A was 15 mGy per slice(CTDI),10 mGy per slice in group B,and 5 mGy per slice in group C.Both objective and subjective analyses were performed on the images.The objective analysis included measur-ing the CT values of major blood vessels,background noise(SD),signal-to-noise ratio(SNR),and contrast-to-noise ratio(CNR).The subjective analysis involved scoring the VR,MIP processing,and axial images after being pro-cessed by artificial intelligence image analysis technology.Results:There was no significant difference in the mean value of intravascular CT among the three groups(P＞0.05).There was no significant difference in SD and SNR between group A and group B(P＞0.05),but there was significant difference in SD and SNR between group C and group A and B(P＜0.01).There was statistical difference in SD and SNR among the three groups(P＜0.01),but there was no sta-tistical difference in CNR among the three groups(P＞0.05).There was no significant difference between the three groups in displaying the anatomical structure of the head and neck vessels and removing the skull base bone(P＞0.05).The vascular structure of each group could be displayed,and the subjective scores of image quality were excellent or a-bove.The radiation dose of group C was reduced by 66.6%and 33.3%respectively compared with that of group A and group B.Conclusion:Through artificial intelligence image analysis technology,it is possible to achieve a better three-dimensional display of the anatomical structure of head and neck blood vessels under extremely low radiation dose con-ditions.It is suitable for patients who need low-dose head and neck CT angiography or perfusion imaging.