Application Abstract: Aortic dissection is a high-risk aortic disease with high mortality.Thoracic endovascular aortic repair(TEVAR)has become the first choice and research focus for the treatment of type B aortic dissection due to its microtrauma and good remodeling effect.Preoperative diagnostic planning and postoperative treatment of aortic dissection require objective quantification of structural parameters,accurate positioning of the tearing intima and the location of the tear.Segmentation of aortic dissection in CT images is of great significance for preoperative diagnosis,surgical planning and postoperative repair.The lesion is adjacent to the three important branches of the aortic arch,and the coating is likely to block these branches and cause serious complications.Therefore,the selection of DSA imaging Angle in TEVAR surgery has a crucial impact on the placement of stents.Key technologies addressed in this paper include:1)A 3D MPRNet convolutional neural network architecture based on multi-plane reconstruction was proposed to achieve accurate and continuous segmentation of aorta and torn intima.2)The correlation measure of linear combination is established,and the precise registration of CTA and DSA data is realized through the accelerated optimization algorithm of conjugate direction.3)The problem of image drift and dynamic real-time matching during surgical operation was studied,and the position and direction Angle of the projection of the doctor and preoperative angiography data relative to intraoperative DSA as well as the internal parameters of the imaging system were constructed to optimize the optimal imaging Angle of c-arm.The projection relationship between DSA and CTA images was studied,and the real 3d coordinates of branch artery openings and stents were obtained to realize the positioning of the distal end of stents in the 3d model.4)The problem of image drift and dynamic real-time matching during surgical operation was studied,and the position and direction Angle of the projection of the doctor and preoperative angiography data relative to intraoperative DSA as well as the internal parameters of the imaging system were constructed to optimize the optimal imaging Angle of c-arm.The projection relationship between DSA and CTA images was studied,and the real 3d coordinates of branch artery openings and stents were obtained to realize the positioning of the distal end of stents in the 3d model.
