P1443Accuracy of three algorithms in predicting accessory pathway localization in a pediatric population with ventricular pre-excitation

Abstract

Abstract BACKGROUND Prediction of accessory pathway location is extremely important before scheduling ablation, even more in paediatric patients. Given the absence in the literature of conclusive and recent studies that established the best algorithm to predict location in these patients, especially after the introduction of 3D mapping systems, we designed a study to compare three previously published algorithms. AIM The aim of this study is to assess and compare the accuracy in predicting accessory pathway location of the algorithm by Arruda, Boersma and Chiang. METHODS This is a multicenter, observational, retrospective clinical study based on the analysis of the resting 12-leads ECGs of children (aged 0-18 years) with ventricular pre-excitation pattern. The study lasts from January 2013 to June 2019. We analyzed the accuracy (defined as percentage of true positive values) of predicted location by each algorithm, which could include one or more than one of 13 regions around mitral and tricuspid annuli. RESULTS Study population was composed by 120 patients, mean age was 12.7 +/- 3.6 years (height 155.6 +/-19.3 cm and weight 48.3 +/- 17.1 kg). The algorithm by Boersma has the highest percentage of predicted pathway locations that are found to be concordant with the site of successful AP ablation (see table). When we corrected for pre-test probability, the algorithm by Arruda makes the possibility of one specific location 8 times higher, the one by Boersma 6.4 times higher and the one by Chiang 6.9 times higher than by chance. CONCLUSIONS The algorithm by Boersma showed the highest accuracy in accessory pathway location, on the other hand the one by Arruda seems to have the highest corrected accuracy in accessory pathway among children. table 1 Arruda Boersma Chiang Number of locations 13 9 13 Total accuracy 0,617 0,717 0,533 Corrected accuracy per locations 8,0158 6,4494 6,929 Abstract Figure. picture 1