301 Classsification Scheme For Type B Aortic Dissection And Creation Of An Artificial Intelligence Online Research Platform For Aortic Disease

Purpose:To evaluate the associations of new-establised 301 classification scheme with adverse events including thoracic aortic enlargenment,abdominal aortic enlargement,type la endoleak,distal stent graft-induced new entry(SINE)and composite endpoint after thoracic endovascular aortic repair(TEVAR)for type B aortic dissection(TBAD),and creat an artificial intelligence online research platform for aorta(AORTIST)to facilitate aortic diameter measurement,surgery planning,prognosis assessment of post-TEVAR adverse events,and management of patients'medical data.Methods:Between January 2011 and December 2016,201 patients who were treated with TEVAR for TBAD in ROBUST study were retrospectively analyzed.Patients'baseline characteristic,preoperative morphological features,post-TEVAR adverse events were collected based on electronic patient record or pre-and post-operative CTA scans.Based on the morphological features of thoracic false lumen branches(TFLBs),we derived 301 classification scheme,by which patients were classified into three groups:type ? dissection with true lumen descending closely along the thoracic vertebrae body(n=62),type ? dissection with false lumen coursing closely along the thoracic vertebrae body(n=24),and type ? dissection with a spiral configuration(n=115).Multivariable Cox proportional hazards regression was used to assess the independent relationship of 301 classification scheme and post-TEVAR adverse events.Based on Internet technology and airtificial interlligence,automatic aortic segmentation algorithm,image-based radiomics,and TEVAR expert system were used to establish AORTISTResults:The median follow-up period was 26.37 months(range,1.43-83.17)and the median imaging follow-up period was 12.43 months(range,0.10-83.07).Type ? and/or type ? dissection of 301 classification scheme was independently associated with the risk of post-TEVAR thoracic aortic enlargement(type ? vs ?:HR,4.55;95%CI,1.00-20.69;P =0.0497;type ? vs ?:HR,14.17;95%CI,2.74-73.18;P=0.0016),type la endoleak(type ? vs ?:HR,2.68;95%CI,0.87-8.26;P =0.0867;type ? vs ?:HR,8.46;95%CI,2.20-32.58;P =0.0019)and composite endpoint(type ? vs ?:HR,2.76;95%CI,1.11-6.90;P =0.0293;type ? vs ?:HR,2.49;95%CI,0.72-8.62;P=0.1486).Type ?&? dissection magnified the impact of distal oversizing ratio on the occurrence of distal SINE(type ?&?:HR,1.04;95%CI,1.01-1.08;P =0.0077;type ?:HR,0.94;95%CI,0.83-1.07;P =0.3768;P for interaction,0.0482).When the proximal oversizing ratio<0.9,the effect value of type ? dissection in 301 classification scheme was significantly larger than that of either type ?&? or type ? when the proximal oversizing ration ?0.9(P for interaction,0.0437).AORTIST was established preliminarily,which has multiple capabilities including automatic aortic measurements,automatic TEVAR planning,radiomics-based automatic prognosis of adverse events,and automatic management of patient's medical data.Conclusions:Type ? and type ? dissection in 301 classification scheme carry a higher risk of thoracic aortic enlargement,type la endoleak,and composite endpoint.AORTIST has the potential for facilitating TEVAR planning,prognosis of adverse events,and medical data management,but further optimization upgrading is needed.