Prediction of Post-Ablation Outcome in Atrial Fibrillation Using Shape Parameterization and Partial Least Squares Regression

Abstract

Abstract.To analyze left atrial remodeling may reveal shape features related to post-ablation outcome in atrial fibrillation, which helps in identifying suitable candidates before ablation. In this article, we propose an application of diffeomorphometry and partial least squares regression to address this problem. We computed a template of left atrial shape in control group and then encoded the shapes in atrial fibrillation with a large set of parameters representing their diffeomorphic deformation. We applied a two-step partial least squares regression. The first step eliminates the influence of atrial volume in shape parameters. The second step links deformations directly to post-ablation recurrence and derives a few principle modes of deformation, which are unrelated to volume change but are involved in post-ablation recurrence. These modes contain information on ablation success due to shape differences, resulting from remodeling or influencing ablation procedure. Some details are consistent with the most complex area of ablation in clinical practice. Finally, we compared our method against the left atrial volume index by quantifying the risk of post-ablation recurrence within six months. Our results show that we get better prediction capabilities (area under receiver operating characteristic curves = 0.73) than left atrial dilation ( = 0.47), which outperforms the current state of the art.

A new machine learning approach for predicting likelihood of recurrence following ablation for atrial fibrillation from CT

“…The left atrial sphericity-a measure of how closely the atrium resembles a sphere-was introduced by Brisbal et al [6] to quantify remodeling of atrial morphology by MRI and to predict post-ablation recurrence. Jia et al [3], mapped diffeomorphic features of patients' atrium on pre-ablation images relative to a constructed atlas to predict AF recurrence using a partial least square fitting. In a recent study by Bieging et al [22], a particle based modeling (PBM) approach from cardiac magnetic resonance scans done in the DE-MRI-Guided Ablation vs.…”

“…Previous studies have identified left atrial measurements such as volume [2][3][4], anterior-posterior (AP) diameter [5], and atrial sphericity [6] as predictors of AF recurrence. However, there is a lack of consensus on the strength of these individual biomarkers in gauging the risk of AF recurrence.…”

Left atrial shape is independent predictor of arrhythmia recurrence after catheter ablation for atrial fibrillation: A shape statistics study

“…Thus, the prediction for each patient was built from a model learned from a series of shapes that did not comprise the patient’s own LA shape. To analyze shape features that are independent from atrial size and AF persistence, we separated the analysis into 2 sequential partial least squares regressions 16 : the first one to remove the shape variations that are linearly related to LA volume and AF persistence, and the second one to compute the shape patterns that are linearly related to postablation recurrence. This model (implemented in Matlab2017a) allowed us to provide each patient with an LA shape score normalized to the range from 0 to 1, according to the proximity between the studied patient shape and the shape patterns associated with recurrence.…”