In the final round of the 18th Challenge Cup National Extra-curricular Academic Science and Technology Competition for University Students just concluded, the project "Application of Deep Learning in Digital Subtraction Angiography (DSA): a new potential method to reduce radiation dose" recommended by the School of Statistics and Data Science won a national second prize, which is a breakthrough in the School history.
The Challenge Cup competition was co-sponsored by the Central Committee of the Communist Youth League, China Association for Science and Technology, the Ministry of Education, the Chinese Academy of Social Sciences, the All-China Federation of Students and the People's Government of Guizhou Province, and hosted by Guizhou University, attracting over 2.5 million students from around 2,000 universities, who submitted over 400,000 works.
The team included five students, Zhijiang Tang, Xuantai Wu, Qiyu Zhao, Chen Song and Zongyuan Li, and was supervised by Jun Xu, Ph.D. and Qiang Xiong, M.D.
This study aims to artificially reduce the number of DSA shots and radiation exposure of patients by synthesizing low-frame rate DSA images into high-frame rate ones through a deep learning system. DSA is a method of obtaining blood flow images through continuous X-rays films (more than 100). A DSA procedure causes about 3mSv of radiation to the human body, which is equal to the annual exposure under natural conditions. The method proposed in the study can significantly reduce the radiation exposure while providing detailed blood vessel information for doctors as foundations for follow-up diagnoses and treatments.
