CARDIOVASCULAR JOURNAL OF AFRICA • SMC-PAFCIC Abstracts October 2022 49 AFRICA the CoA and its complexity by showing a very tortuous and kinked isthmus aorta. Percutaneous procedure was done under general anesthesia, using at first a right femoral artery approach which failed to cross the extremely tight CoA. A secondary arterial access via the radial artery had allowed anterograde catheterization and angiography of the ascending aorta. Several predilatations of the stenosis were performed using initially coronary balloons than vascular balloons with successively higher diameter sizes .A 0.035 super stiff guide wire was than exchanged and retrieved through femoral artery sheath using a lasso catheter. After careful measurement of all aortic segments, a covered CP stent was mounted and crimped on a balloon in balloon (BIB) catheter than advanced through a 14 French delivery sheath to the CoA area via femoral approach. Final angiographic inspection following stent deployment showed no immediate complication and well stent expansion. No residual peak-to-peak aortic gradient was noted. No further complications had occurred during hospitalization and the patient was discharged 2 days later. Conclusion: Stent repair of complex coaractation is a safe and feasible strategy facilited with noninvasive imaging techniques , good planning and teamwork attitude . Submission ID: 1632 A NEW FEASIBLE AND SAFE ACCESS POINT IN INTERVENTIONAL CARDIOLOGY : THE TRANS-PALMAR ACESS ALI KHORCHANI, HOUCINE ABDELHAFIDH, ZINE EL ABIDINE BEN ALI, FEKIH RIDHA, IMANE BOUALAOUI, TAHA LASSOUED, SOUAD FERJENI, SAMI MILOUCHI TUNISIA Background: The selection of the vascular access is a crucial moment during an interventional cardiology procedure. Radial approach is recommended access over the femoral access because of a lower rate of bleeding complication. Many distal access were tested but their reliability is not yet validated. A new feasible and safe access point in interventional cardiology: The trans-palmar acess Aim: To evaluate the feasibility and safety of the distal ulnar palmar access on coronary angiography and intervention. Results: In 21 out of 27 patients, distal palmar (DUP) artery catheterization was feasible with an average puncture time of 3.5 minutes. Haemostasis was obtained in all cases within 1 hour compressive bandage which was well tolerated. No major complication was reported. Conclusion: The DUP approach is a new innovative access with preliminary results suggesting its feasibility and safety. Larger studies are needed to confirm this results and clarify its reliability. Submission ID: 1633 AUTOMATIC CORONARY ANGIOGRAM ANALYSIS ALGORITHM: KEYFRAME EXTRACTION PHASE AIMAN GHRAB, HOUNAIDA MOALLA FOURATI, AMINE BAHLOUL, BASSEM BEN HAMED, RANIA HAMMAMI, LEILA ABID TUNISIA Introduction: Coronary angiography is the gold standard for the diagnosis of coronary artery disease (CAD). Treatment decision is based on the severity of the disease and the obstruction percentage. This evaluation, often made by visual assessment done by clinicians during catheterization procedure, is the subject of high interobserver variability. Objective: We aim to develop a fully automated coronary angiogram analysis method to have an accurate evaluation of CAD. To do so, keyframe extraction is the first and most important step. Methods: The full dataset was collected from the exams performed by a single catheterization laboratory during the period between January 2018 and December 2021. We used a sample of angiograms for the keyframe extraction step. Each one was annotated by two experienced physicians using a web application. Each frame, being a set of pixels, undergo processing based on predefined filters. Filter pipelines have also been used to seek better performance. The frames were tested sometimes with and sometimes without cropping. The surviving pixel were used to calculate an intensity score. The frame with highest score and the six neighboring frames were chosen as keyframes. Each processing method was compared to the manually annotated frames. The processing methods are detailed in table 1. Results: The full dataset consisted of 3159 angiographic study: a total of 37209 coronary angiogram was extracted. We used a sample of 45 angiogram to extract a total of 1434 frames. The manual annotation found 474 keyframe and 960 non-keyframe. By using the sato filter, MODERATED POSTER SESSION
RkJQdWJsaXNoZXIy NDIzNzc=