PO-02-073 FEASIBILITY OF ARTIFICIAL INTELLIGENCE 3D ALGORITHM INTEGRATION WITH INTRACARDIAC ECHOCARDIOGRAPHY FOR LA IMAGING

Intrachardiac echography (ICE) is an useful tool for LA anatomical rendering during AF mapping and ablation, but it operator-depandent. The advent of the CARTOSOUND FAM Module, artificial intelligence (AI) imaging algorithm, has potential to overcome this ICE limitation. This new technology based on Machine Learning (ML) algorithm automate 3D volume reconstruction while acquiring 2D ULS clips using SOUNDSTAR Catheters positioned in RA and/or RVOT. To evaluate feasibility reproducibility Module comparison standard-of-care CT images. In 28 patients, shells body adjacent structures (LAA, LSPV, LIPV, RSPV, RIPV) were reconstructed AI algorithm. selected frames within gated end-expiration phase max volume. results evaluated by comparing images three independent raters. For structures: 3 anchors scale (Poor contour= 0, average contour=1, good contour=2). body: 5 (1=dissatisfied , 2= less satisfied, 3=somewhat satisfied 4= very 5=extremely satisfied). generated both anatomies along with contours that overlaid corresponding ultrasound frames. calculation time was 65 sec. Comparisons between two techniques revealed fair agreement (Mean rating scores: LSPV-2.0; LIPV-1.9; RSPV-1.8; RIPV-1.8; LAA-1.8 overall body-4.3). addition, no differences raters found these structures, meaning high agreements reported across (Figure). Registration error: Average error - 2.7mm, STD 2.1mm respectively, Min 0mm, Max 11.97mm. A newer version resulted improved at 25 integration Catheter demonstrated comparable gold-standard images, as rated different operators. comparability poses future additional clinical implications.