Comparison of the characteristics between machine learning and deep learning algorithms for ablation site classification in a novel cloud-based system

Abstract

Background

CARTONET is a cloud-based system for the analysis of ablation procedures using the CARTO system. The current CARTONET R14 model employs deep learning, but its accuracy and positive predictive value (PPV) remain underevaluated.

Objective

This study aimed to compare the characteristics of the CARTONET system between the R12.1 and the R14 models.

Methods

Data from 396 atrial fibrillation ablation cases were analyzed. Using a CARTONET R14 model, the sensitivity and PPV of the automated anatomic location model were investigated. The distribution of potential reconnection sites and confidence level for each site were investigated. We also compared the difference in the data between the CARTONET R12.1, the previous CARTONET version, and the CARTONET R14 models.

Results

We analyzed the overall tags of 39,169 points and the gap prediction of 625 segments using the CARTONET R14 model. The sensitivity and PPV of the R14 model significantly improved compared with the R12.1 model (R12.1 vs R14: sensitivity, 71.2% vs 77.5% [P < .0001]; PPV, 85.6% vs 86.2% [P = .0184]). The incidence of reconnections was highly observed in the posterior area of the right pulmonary veins (98/238 [41.2%]) and left pulmonary veins (190/387 [49.1%]). In contrast, the possibility of reconnection was highest in the roof area for the right pulmonary veins (14% [5.5%–41%]) and left pulmonary veins (16% [8%–22%]).

Conclusion

The R14 model significantly improved sensitivity and PPV compared with the R12.1 model. The tendency for predicting potential reconnection sites was similar to that of the previous version, the R12 model.