Hui Ju Chang
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Hui Ju Chang has a total of 29 co-authors. Here are the top co-authors by number of shared publications:
- Albert Joseph Rogers
Stanford University
3 shared publications - Prasanth Ganesan
Department of Medicine, Division of Cardiovascular Medicine, and Cardiovascular Institute, Stanford University
3 shared publications - Sanjiv M. Narayan
Department of Medicine, Stanford University, Palo Alto, California
3 shared publications - Brototo Deb
Department of Medicine, Cardiovascular Institute and Institute for Computational and Mathematical Engineering, Stanford University, Stanford, CA, USA
2 shared publications - Kelly A. Brennan
Department of Medicine, Division of Cardiovascular Medicine, and Cardiovascular Institute, Stanford University
2 shared publications Maxime Pedron
Division of Cardiology, Stanford Cardiovascular Institute, Stanford University, Palo Alto, CA, USA (P.G., M.P., R.F., A.J.R., B.D., H.J.C., S.R.-C., V.S., K.A.B., T.B., P.C., P.J.W., S.M.N.).
2 shared publications
Latest Publications
- 7 July 2026
- Published by Wolters Kluwer Health in Circulation: Arrhythmia and Electrophysiology
- Vol. 19 (7) , e014765
Identifying the Presence and Characteristics of Mid-Myocardial and Epicardial Fibrosis From Intracardiac Electrograms in Patients Undergoing Ventricular Arrhythmia Ablation Using a Transformer-Based Self-Supervised Classifier
BACKGROUND: Catheter ablation is an essential tool for ventricular arrhythmia management, yet sustained procedural success is hindered by the limited ability to identify nonendocardial arrhythmogenic substrates during the procedure. Although delayed enhancement cardiac magnetic resonance imaging is the reference standard for detecting myocardial fibrosis, barriers including cost, workflow complexity, and artifacts in patients with implantable devices limit its preprocedural use. We hypothesized that intracardiac electrograms provide sufficient information to infer scar beyond the endocardial surface and that this information can be harnessed by machine learning techniques. METHODS: This retrospective study included a total of 131 584 cardiac contact electrogram (EGM) signals collected from 46 patients undergoing ventricular arrhythmia ablation. A stratified patient-wise split was used to create the training/validation set (N=37) and the testing set (N=9), while ensuring a similar distribution of scar types. We developed a novel image-processing workflow to create scar labels using coregistered cardiac magnetic resonance imaging and electroanatomic mapping surface meshes. We developed a transformer-based self-supervised model, EGM2Scar-AI, alongside basic convolutional neural network models using either EGM waveforms or EGM-derived short-time Fourier transform spectrograms. RESULTS: The average task-specific area under the receiver operating characteristic curve of both the basic and short-time Fourier transform-based convolutional neural networks was 0.729 (0.725–0.732) and 0.729 (0.726–0.733), respectively, while EGM2Scar-AI performed significantly better with an average area under the receiver operating characteristic curve of 0.822 (0.819–0.825) across all 3 scar types. All models performed better on endocardial and mid-myocardial fibrosis identification, with a modest reduction in performance for epicardial fibrosis. Sensitivity improved significantly with the transformer-based architecture without appreciable changes in specificity. CONCLUSIONS: Our study demonstrates that routine intracardiac electrograms enable the identification of endocardial, mid-myocardial, and epicardial scar using a transformer-based self-supervised deep learning model. Ultimately, our model has the potential to provide magnetic resonance imaging-like fibrosis maps during EP procedures without the need for external imaging and to increase the data set size for patient-specific models in ventricular arrhythmia research.
- 7 July 2026
- Published by Wolters Kluwer Health in Circulation: Arrhythmia and Electrophysiology
- Vol. 19 (7) , e014765
Identifying the Presence and Characteristics of Mid-Myocardial and Epicardial Fibrosis From Intracardiac Electrograms in Patients Undergoing Ventricular Arrhythmia Ablation Using a Transformer-Based Self-Supervised Classifier
BACKGROUND: Catheter ablation is an essential tool for ventricular arrhythmia management, yet sustained procedural success is hindered by the limited ability to identify nonendocardial arrhythmogenic substrates during the procedure. Although delayed enhancement cardiac magnetic resonance imaging is the reference standard for detecting myocardial fibrosis, barriers including cost, workflow complexity, and artifacts in patients with implantable devices limit its preprocedural use. We hypothesized that intracardiac electrograms provide sufficient information to infer scar beyond the endocardial surface and that this information can be harnessed by machine learning techniques. METHODS: This retrospective study included a total of 131 584 cardiac contact electrogram (EGM) signals collected from 46 patients undergoing ventricular arrhythmia ablation. A stratified patient-wise split was used to create the training/validation set (N=37) and the testing set (N=9), while ensuring a similar distribution of scar types. We developed a novel image-processing workflow to create scar labels using coregistered cardiac magnetic resonance imaging and electroanatomic mapping surface meshes. We developed a transformer-based self-supervised model, EGM2Scar-AI, alongside basic convolutional neural network models using either EGM waveforms or EGM-derived short-time Fourier transform spectrograms. RESULTS: The average task-specific area under the receiver operating characteristic curve of both the basic and short-time Fourier transform-based convolutional neural networks was 0.729 (0.725–0.732) and 0.729 (0.726–0.733), respectively, while EGM2Scar-AI performed significantly better with an average area under the receiver operating characteristic curve of 0.822 (0.819–0.825) across all 3 scar types. All models performed better on endocardial and mid-myocardial fibrosis identification, with a modest reduction in performance for epicardial fibrosis. Sensitivity improved significantly with the transformer-based architecture without appreciable changes in specificity. CONCLUSIONS: Our study demonstrates that routine intracardiac electrograms enable the identification of endocardial, mid-myocardial, and epicardial scar using a transformer-based self-supervised deep learning model. Ultimately, our model has the potential to provide magnetic resonance imaging-like fibrosis maps during EP procedures without the need for external imaging and to increase the data set size for patient-specific models in ventricular arrhythmia research.
- 1 February 2025
- Published by Wolters Kluwer Health in Circulation: Arrhythmia and Electrophysiology
- Vol. 18 (2) , e012860
Comparing Phenotypes for Acute and Long-Term Response to Atrial Fibrillation Ablation Using Machine Learning
BACKGROUND: It is difficult to identify patients with atrial fibrillation (AF) most likely to respond to ablation. While any arrhythmia patient may recur after acutely successful ablation, AF is unusual in that patients may have long-term arrhythmia freedom despite a lack of acute success. We hypothesized that acute and chronic AF ablation outcomes may reflect distinct physiology and used machine learning of multimodal data to identify their phenotypes. METHODS: We studied 561 consecutive patients in the Stanford AF ablation registry (66±10 years, 28% women, 67% nonparoxysmal), from whom we extracted 72 data features of electrograms, electrocardiogram, cardiac structure, lifestyle, and clinical variables. We compared 6 machine learning models to predict acute and long-term end points after ablation and used Shapley explainability analysis to contrast phenotypes. We validated our results in an independent external population of n=77 patients with AF. RESULTS: The 1-year success rate was 69.5%, and the acute termination rate was 49.6%, which correlated poorly on a patient-by-patient basis (φ coefficient=0.08). The best model for acute termination (area under the curve=0.86, Random Forest) was more predictive than for long-term outcomes (area under the curve=0.67, logistic regression; P <0.001). Phenotypes for long-term success reflected clinical and lifestyle features, while phenotypes for AF termination reflected electrical features. The need for AF induction predicted both phenotypes. The external validation cohort showed similar results (area under the curve=0.81 and 0.64, respectively) with similar phenotypes. CONCLUSIONS: Long-term and acute responses to AF ablation reflect distinct clinical and electrical physiology, respectively. This de-linking of phenotypes raises the question of whether long-term success operates through factors such as attenuated AF progression. There remains an urgent need to develop procedural predictors of long-term AF ablation success.
- 1 February 2025
- Published by Wolters Kluwer Health in Circulation: Arrhythmia and Electrophysiology
- Vol. 18 (2) , e012860
Comparing Phenotypes for Acute and Long-Term Response to Atrial Fibrillation Ablation Using Machine Learning
BACKGROUND: It is difficult to identify patients with atrial fibrillation (AF) most likely to respond to ablation. While any arrhythmia patient may recur after acutely successful ablation, AF is unusual in that patients may have long-term arrhythmia freedom despite a lack of acute success. We hypothesized that acute and chronic AF ablation outcomes may reflect distinct physiology and used machine learning of multimodal data to identify their phenotypes. METHODS: We studied 561 consecutive patients in the Stanford AF ablation registry (66±10 years, 28% women, 67% nonparoxysmal), from whom we extracted 72 data features of electrograms, electrocardiogram, cardiac structure, lifestyle, and clinical variables. We compared 6 machine learning models to predict acute and long-term end points after ablation and used Shapley explainability analysis to contrast phenotypes. We validated our results in an independent external population of n=77 patients with AF. RESULTS: The 1-year success rate was 69.5%, and the acute termination rate was 49.6%, which correlated poorly on a patient-by-patient basis (φ coefficient=0.08). The best model for acute termination (area under the curve=0.86, Random Forest) was more predictive than for long-term outcomes (area under the curve=0.67, logistic regression; P <0.001). Phenotypes for long-term success reflected clinical and lifestyle features, while phenotypes for AF termination reflected electrical features. The need for AF induction predicted both phenotypes. The external validation cohort showed similar results (area under the curve=0.81 and 0.64, respectively) with similar phenotypes. CONCLUSIONS: Long-term and acute responses to AF ablation reflect distinct clinical and electrical physiology, respectively. This de-linking of phenotypes raises the question of whether long-term success operates through factors such as attenuated AF progression. There remains an urgent need to develop procedural predictors of long-term AF ablation success.
- 1 January 2025
- Published by Wolters Kluwer Health in Circulation: Arrhythmia and Electrophysiology
- Vol. 18 (1) , e013023
Engineering of Generative Artificial Intelligence and Natural Language Processing Models to Accurately Identify Arrhythmia Recurrence
BACKGROUND: Large language models (LLMs) such as Chat Generative Pre-trained Transformer (ChatGPT) excel at interpreting unstructured data from public sources, yet are limited when responding to queries on private repositories, such as electronic health records (EHRs). We hypothesized that prompt engineering could enhance the accuracy of LLMs for interpreting EHR data without requiring domain knowledge, thus expanding their utility for patients and personalized diagnostics. METHODS: We designed and systematically tested prompt engineering techniques to improve the ability of LLMs to interpret EHRs for nuanced diagnostic questions, referenced to a panel of medical experts. In 490 full-text EHR notes from 125 patients with prior life-threatening heart rhythm disorders, we asked GPT-4-turbo to identify recurrent arrhythmias distinct from prior events and tested 220 563 queries. To provide context, results were compared with rule-based natural language processing and Bidirectional Encoder Representations from Transformer-based language models. Experiments were repeated for 2 additional LLMs. RESULTS: In an independent hold-out set of 389 notes, GPT-4-turbo had a balanced accuracy of 64.3%±4.7% out-of-the-box at baseline. This increased when asking GPT-4-turbo to provide a rationale for its answers, a structured data output, and in-context exemplars, to a balanced accuracy of 91.4%±3.8% ( P P <0.05). Results were consistent for GPT-3.5-turbo and Jurassic-2 LLMs. CONCLUSIONS: The use of prompt engineering strategies enables LLMs to identify clinical end points from EHRs with an accuracy that surpassed natural language processing and approximated experts, yet without the need for expert knowledge. These approaches could be applied to LLM queries for other domains, to facilitate automated analysis of nuanced data sets with high accuracy by nonexperts.
- 1 January 2025
- Published by Wolters Kluwer Health in Circulation: Arrhythmia and Electrophysiology
- Vol. 18 (1) , e013023
Engineering of Generative Artificial Intelligence and Natural Language Processing Models to Accurately Identify Arrhythmia Recurrence
BACKGROUND: Large language models (LLMs) such as Chat Generative Pre-trained Transformer (ChatGPT) excel at interpreting unstructured data from public sources, yet are limited when responding to queries on private repositories, such as electronic health records (EHRs). We hypothesized that prompt engineering could enhance the accuracy of LLMs for interpreting EHR data without requiring domain knowledge, thus expanding their utility for patients and personalized diagnostics. METHODS: We designed and systematically tested prompt engineering techniques to improve the ability of LLMs to interpret EHRs for nuanced diagnostic questions, referenced to a panel of medical experts. In 490 full-text EHR notes from 125 patients with prior life-threatening heart rhythm disorders, we asked GPT-4-turbo to identify recurrent arrhythmias distinct from prior events and tested 220 563 queries. To provide context, results were compared with rule-based natural language processing and Bidirectional Encoder Representations from Transformer-based language models. Experiments were repeated for 2 additional LLMs. RESULTS: In an independent hold-out set of 389 notes, GPT-4-turbo had a balanced accuracy of 64.3%±4.7% out-of-the-box at baseline. This increased when asking GPT-4-turbo to provide a rationale for its answers, a structured data output, and in-context exemplars, to a balanced accuracy of 91.4%±3.8% ( P P <0.05). Results were consistent for GPT-3.5-turbo and Jurassic-2 LLMs. CONCLUSIONS: The use of prompt engineering strategies enables LLMs to identify clinical end points from EHRs with an accuracy that surpassed natural language processing and approximated experts, yet without the need for expert knowledge. These approaches could be applied to LLM queries for other domains, to facilitate automated analysis of nuanced data sets with high accuracy by nonexperts.