Skip to content
DraconiticResearch
Search Sources API
226,742 studies31 sourcesUpdated 20 min ago
Peer-reviewedCC BY 4.0Frontiers in Cardiovascular Medicine

Artificial intelligence for intracardiac echocardiography interpretation and guidance during atrial fibrillation ablation

1 more authors
7 affiliations
  1. 1Department of Biomedical Engineering, Medical School of Tianjin University, Tianjin University
  2. 2Beijing Shuxi Technology Co., Ltd
  3. 3State Key Laboratory of Advanced Medical Materials and Devices, Medical School of Tianjin University
  4. 4Department of Cardiology, Beijing Friendship Hospital, Capital Medical University
  5. 5Department of Radiology, Beijing Friendship Hospital, Capital Medical University
  6. 6Department of Cardiology, Affiliated Hospital of Nantong University
  7. 7Department of Cardiology, The First Affiliated Hospital of Wenzhou Medical University

Abstract

Background: Intracardiac echocardiography (ICE) has becoming indispensable for ensuring safety and efficacy during atrial fibrillation (AF) ablation. However, ICE interpretation remains highly observer-dependent, presenting a steep learning curve in rapidly distinguishing complex, dynamic cardiac structures.

Objectives: This study aimed to develop and validate a multi-task deep learning system (AI-ICE) for real-time, automated interpretation of cardiac structures and ultrasound views during AF ablation.

Methods: The interpretation task is defined as two sub-tasks, one for detecting 12 anatomical structures and the other for classifying 5 standard views. A total of 373,091 ICE images from 350 patients undergoing AF ablation across three centers were collected. The dataset was split into training/internal validation cohorts (two centers) and an independent external testing cohort (one center, 35,335 frames). The AI-ICE model combined a ResUnet backbone with an LSTM module to capture spatiotemporal dependencies across consecutive frames, reflecting the streaming nature of real-time clinical ICE imaging. The two tasks were formulated as a multi-task coupled framework with shared representations, exploiting the intrinsic interdependence between the imaging view and the visible anatomy.

Results: For the internal validation, the average AUC for classifying ultrasound views and detecting structures reached 0.971 (95% CI: 0.947–0.992) and 0.974 (95% CI: 0.968–0.980). For the external validation, the AI-ICE system achieved an AUC of 0.984 (95% CI: 0.974–0.993) for ultrasound view classification and 0.955 (95% CI: 0.948–0.962) for anatomical structure objection. Compared with three cardiologist experts, it achieved higher accuracy (AUC: 0.941 vs. 0.916) and superior inter-rater agreement (Kappa coefficient: 0.791 vs. 0.716). Additionally, the system demonstrated high efficiency, with an inference time of 12 milliseconds per ICE image.

Conclusions: The multi-task AI-ICE system enables real-time, operator-independent ICE interpretation during AF ablation, establishing a foundation for automated intra-operative guidance.

Related

  1. Spontaneous Echo Contrast or Thrombus? Intracardiac Echocardiography and Isoproterenol Infusion Provide the Answer
    Can Sezer, Cem Çöteli, Gül Sinem Kılıç +3
    Anatolian Journal of CardiologyPeer-reviewedCC BY-NC 4.0Full textCardiovascular medicinePMC
  2. The “Floating Lotus” Sign on Intracardiac Echocardiography: Characterization and Predictors of a Novel Left Atrial Appendage Pseudo-Thrombus
    Liang Zhang, Yanyan Bai, Yanhua Zhu +4
    Reviews in cardiovascular medicinePeer-reviewedCC BYFull textCardiovascular medicineEurope PMC
  3. Acute Changes in Left Atrial Appendage Flow Velocity During Pulsed Field Ablation: An Intraprocedural Intracardiac Echocardiography Study
    Y. Pan, P. Osmancik, T. Knize +6
    medRxivPreprintAll rights reservedFull textCardiovascular medicinemedRxiv
  4. Contrast-Free Dual-Chamber Leadless Pacemaker Implantation Using Intracardiac Echocardiography and 3-Dimensional Electroanatomical Mapping
    Shuta Kikuchi, Yasuyuki Takada, Kazuhiro Satomi +1
    JACC Case ReportsPeer-reviewedCC BY-NC-ND 4.0Full text1 citedCardiovascular medicinePMC
  5. Impedance changes during intracardiac echocardiography-confirmed catheter contact in pulsed field ablation
    Kohki Kimura, Hirotoshi Nagae, Yuki Yamamoto +3
    Indian Pacing and Electrophysiology JournalPeer-reviewedCC BY-NC-ND 4.0Full text1 citedCardiovascular medicinePMC
BibTeX
@article{li2026artificial,
  title = {{Artificial intelligence for intracardiac echocardiography interpretation and guidance during atrial fibrillation ablation}},
  author = {Guang Li and Qingsong Wang and Qiyue Mu and Jinchen Li and Yi He and Fei Long and Jinglai Sun and Qi Lu and Xinlei Wu and Liping Sun and Haixia Xu and Chao Chen and Hui Yu},
  journal = {Frontiers in Cardiovascular Medicine},
  year = {2026},
  month = nov,
  doi = {10.3389/fcvm.2026.1992965},
  url = {https://www.frontiersin.org/articles/10.3389/fcvm.2026.1992965/full},
  note = {CC BY 4.0},
}
RIS
TY  - JOUR
TI  - Artificial intelligence for intracardiac echocardiography interpretation and guidance during atrial fibrillation ablation
AU  - Guang Li
AU  - Qingsong Wang
AU  - Qiyue Mu
AU  - Jinchen Li
AU  - Yi He
AU  - Fei Long
AU  - Jinglai Sun
AU  - Qi Lu
AU  - Xinlei Wu
AU  - Liping Sun
AU  - Haixia Xu
AU  - Chao Chen
AU  - Hui Yu
PY  - 2026
DA  - 2026/11/13
JO  - Frontiers in Cardiovascular Medicine
DO  - 10.3389/fcvm.2026.1992965
UR  - https://www.frontiersin.org/articles/10.3389/fcvm.2026.1992965/full
AB  - Background: Intracardiac echocardiography (ICE) has becoming indispensable for ensuring safety and efficacy during atrial fibrillation (AF) ablation. However, ICE interpretation remains highly observer-dependent, presenting a steep learning curve in rapidly distinguishing complex, dynamic cardiac structures.  Objectives: This study aimed to develop and validate a multi-task deep learning system (AI-ICE) for real-time, automated interpretation of cardiac structures and ultrasound views during AF ablation.  Methods: The interpretation task is defined as two sub-tasks, one for detecting 12 anatomical structures and the other for classifying 5 standard views. A total of 373,091 ICE images from 350 patients undergoing AF ablation across three centers were collected. The dataset was split into training/internal validation cohorts (two centers) and an independent external testing cohort (one center, 35,335 frames). The AI-ICE model combined a ResUnet backbone with an LSTM module to capture spatiotemporal dependencies across consecutive frames, reflecting the streaming nature of real-time clinical ICE imaging. The two tasks were formulated as a multi-task coupled framework with shared representations, exploiting the intrinsic interdependence between the imaging view and the visible anatomy.  Results: For the internal validation, the average AUC for classifying ultrasound views and detecting structures reached 0.971 (95% CI: 0.947–0.992) and 0.974 (95% CI: 0.968–0.980). For the external validation, the AI-ICE system achieved an AUC of 0.984 (95% CI: 0.974–0.993) for ultrasound view classification and 0.955 (95% CI: 0.948–0.962) for anatomical structure objection. Compared with three cardiologist experts, it achieved higher accuracy (AUC: 0.941 vs. 0.916) and superior inter-rater agreement (Kappa coefficient: 0.791 vs. 0.716). Additionally, the system demonstrated high efficiency, with an inference time of 12 milliseconds per ICE image.  Conclusions: The multi-task AI-ICE system enables real-time, operator-independent ICE interpretation during AF ablation, establishing a foundation for automated intra-operative guidance.
SN  - 2297-055X
ER  -