AI-enhanced Imaging Analysis to Predict Post-EVAR

Aortic Aneurysm

Abdominal aortic aneurysm (AAA), defined as a focal dilatation of the aorta, is a life-threatening condition and has become a serious public health challenge worldwide, associated with high rates of mortality and hospitalization. Endovascular aortic repair (EVAR) requires long-term surveillance by a multidisciplinary team to detect and treat post-operative complications.This international, multicentre, retrospective, observational study aims to use Artificial Intelligence-enhanced imaging analysis for the prediction of post-operative outcomes following EVAR. The primary aim is to identify predictive factors and build prediction models of post-operative outcomes following eligible EVAR. The secondary objective is to investigate morphological aortic changes and remodelling following EVAR . Screening for eligible patients will be done by a vascular specialist at each participating centers. Only subjects fulfilling all the inclusion and none of the exclusion criteria will be accepted in the study. Inclusion criteria from clinical characteristics may encompass the following: - patients who underwent elective EVAR for non-ruptured AAA for the first time, - patients who underwent non-complex EVAR procedure with commercially available endografts with CE mark, - patients with non-syndromic AAA or with non-infective AAA, - patients aged more than 18 years old. Inclusion criteria from CTA characteristics may encompass the following: - patients for whom one pre-operative and at least one post-operative CTA (within 1-year) can be collected in DICOM format. Detailed criteria for eligibility include the following: - Pre-operative CTA should be performed maximum 6 months before the intervention, - At least one post-operative CTA should be performed between 1 month and 12 months following the intervention, - CTA needs to be contrast-enhanced with at least injection at arterial and/or portal time, - CTA should have a spacing no more than 1 mm.Primary outcome will be EVAR-related complications including endoleaks, limb-occlusion and stent-graft migration. Secondary outcomes will include in-hospital length of stay post-intervention, re-interventions (with indication for re-intervention), and deaths (all-cause and aneurysm-related deaths). Outcomes will be investigated within 1 year and at the last follow-up of patients. Morphological changes following EVAR will be analyzed and compared based on pre-operative and post-operative CTA images. Images will be collected and analysed, using the PRAEVAorta 2 software (Nurea). Anatomic features measured include the measurement of aneurysm volume diameters and diameters (including AAA maximum diameter), measurement of intraluminal thrombus and calcifications, proximal and distal aneurysm neck lengths and diameters, suprarenal and infrarenal angulation, distal iliac landing zone length and diameter, and iliac tortuosity index. Based on the measurements, aneurysm sac will be classified into three categories: stable/ shrinkage/ expansion.The study will allow to identify predictive factors and develop predictive models of outcomes following eligible EVAR. By leveraging two approaches (traditional statistical analysis and AI-driven algorithms) and providing AI-enhanced detailed imaging analysis of the morphological characteristics of AAA, this study will allow a better understanding of pre-operative features predictive of the patients' outcomes.

No related papers found yet.

France