Dhamanpreet Kaur
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Previous Organizations / Institutes
The organizations / institutes from the scholar's other articles in the past ten years
Research Subjects
Dhamanpreet Kaur has a total of 129 co-authors. Here are the top co-authors by number of shared publications:
- Euan Angus Ashley
Stanford University School of Medicine
3 shared publications - William Hiesinger
Department of Cardiothoracic Surgery, Stanford University, Stanford, USA
3 shared publications - Albert Joseph Rogers
Stanford University
2 shared publications - Curtis P. Langlotz
Stanford University
2 shared publications - Cyril Zakka
Department of Cardiothoracic Surgery, Stanford University, Stanford, USA
2 shared publications Guson Kang
VA Palo Alto Health Care System, Cardiovascular Medicine, CA (G.K.).
2 shared publications
Latest Publications
- Open Access
- 1 September 2026
- Published by Wolters Kluwer Health in Circulation
- Vol. 154 (9) , 789-800
A Multitask Deep Learning Model for Pediatric Echocardiography Analysis
BACKGROUND: Congenital heart defects afflict ≈1% of all births worldwide. Although deep learning has shown significant promise in automating and improving adult echocardiography analysis, existing pediatric-based models are often limited to single tasks and specific echocardiographic views. To address this, we introduce EchoAI-Peds, a multitask deep learning model for pediatric echocardiography. Our model was developed using the most comprehensive set of pediatric labels to date and is designed to integrate information from multiple views simultaneously. METHODS: We trained a video-based vision transformer to simultaneously detect 28 congenital heart defects, structural and functional abnormalities, repairs, and interventions directly from complete pediatric echocardiography studies with multiple videos. Our model was developed using >700 000 videos derived from >12 000 studies performed at Stanford Medicine between 2014 and 2021. Specifically, our model was trained on 10 815 studies (median age, 8 [IQR 2–14] years; 45% female) and validated on 1294 studies (median age, 9 [IQR 2–15] years; 46% female). Model efficacy was tested on an internal held-out data set of 1336 studies from that same period. In addition, model generalizability was tested on a spatially and temporally distinct patient cohort at the Children’s Hospital of Philadelphia using 2121 studies performed between 2024 and 2025. RESULTS: Our model achieved macroaveraged area under the receiver operating characteristic curve (AUROC) values of 0.91 (95% CI, 0.90–0.92) on the internal test set (median age, 8 [IQR 2–14] years; 47% female) and AUROC values of 0.89 (95% CI, 0.88–0.90) on the external test set (median age, 6 [IQR 0.92–13] years; 44.4% female). Moreover, EchoAI-Peds significantly outperformed adult-based echocardiography foundation models trained on substantially larger data sets, including EchoCLIP (internal AUROC=0.58, 95% CI, 0.56–0.60; external AUROC=0.61, 95% CI, 0.59–0.62) and EchoPrime (internal AUROC=0.58, 95% CI, 0.57–0.61; external AUROC=0.60, 95% CI, 0.59–0.62). Finally, our model demonstrated robust performance across patient age, patient sex, and studies with varying numbers of videos. CONCLUSIONS: Our findings demonstrate the remarkable potential for multitask deep learning models to aid the interpretation of pediatric echocardiograms. In addition, our results underscore the need for models that are specifically tailored to pediatric populations.
- 1 September 2026
- Published by Wolters Kluwer Health in Circulation
- Vol. 154 (9) , 789-800
Open AccessA Multitask Deep Learning Model for Pediatric Echocardiography Analysis
BACKGROUND: Congenital heart defects afflict ≈1% of all births worldwide. Although deep learning has shown significant promise in automating and improving adult echocardiography analysis, existing pediatric-based models are often limited to single tasks and specific echocardiographic views. To address this, we introduce EchoAI-Peds, a multitask deep learning model for pediatric echocardiography. Our model was developed using the most comprehensive set of pediatric labels to date and is designed to integrate information from multiple views simultaneously. METHODS: We trained a video-based vision transformer to simultaneously detect 28 congenital heart defects, structural and functional abnormalities, repairs, and interventions directly from complete pediatric echocardiography studies with multiple videos. Our model was developed using >700 000 videos derived from >12 000 studies performed at Stanford Medicine between 2014 and 2021. Specifically, our model was trained on 10 815 studies (median age, 8 [IQR 2–14] years; 45% female) and validated on 1294 studies (median age, 9 [IQR 2–15] years; 46% female). Model efficacy was tested on an internal held-out data set of 1336 studies from that same period. In addition, model generalizability was tested on a spatially and temporally distinct patient cohort at the Children’s Hospital of Philadelphia using 2121 studies performed between 2024 and 2025. RESULTS: Our model achieved macroaveraged area under the receiver operating characteristic curve (AUROC) values of 0.91 (95% CI, 0.90–0.92) on the internal test set (median age, 8 [IQR 2–14] years; 47% female) and AUROC values of 0.89 (95% CI, 0.88–0.90) on the external test set (median age, 6 [IQR 0.92–13] years; 44.4% female). Moreover, EchoAI-Peds significantly outperformed adult-based echocardiography foundation models trained on substantially larger data sets, including EchoCLIP (internal AUROC=0.58, 95% CI, 0.56–0.60; external AUROC=0.61, 95% CI, 0.59–0.62) and EchoPrime (internal AUROC=0.58, 95% CI, 0.57–0.61; external AUROC=0.60, 95% CI, 0.59–0.62). Finally, our model demonstrated robust performance across patient age, patient sex, and studies with varying numbers of videos. CONCLUSIONS: Our findings demonstrate the remarkable potential for multitask deep learning models to aid the interpretation of pediatric echocardiograms. In addition, our results underscore the need for models that are specifically tailored to pediatric populations.
- Top Cited PapersOpen Access
- 25 March 2026
- Published by Springer Nature in Nature Biomedical Engineering
- p. 1-16
A generalizable deep learning system for cardiac MRI
Cardiac MRI allows for a comprehensive assessment of myocardial structure, function and tissue characteristics. Here we describe a foundational vision system for cardiac MRI, capable of representing the breadth of human cardiovascular disease and health. Our deep-learning model is trained via self-supervised contrastive learning, in which visual concepts in cine-sequence cardiac MRI scans are learned from the raw text of the accompanying radiology reports. We train and evaluate our model on data from four large academic clinical institutions in the United States. We additionally showcase the performance of our models on the UK BioBank and two additional publicly available external datasets. We explore emergent capabilities of our system and demonstrate remarkable performance across a range of tasks, including the problem of left-ventricular ejection fraction regression and the diagnosis of 39 different conditions such as cardiac amyloidosis and hypertrophic cardiomyopathy. We show that our deep-learning system is capable of not only contextualizing the staggering complexity of human cardiovascular disease but can be directed towards clinical problems of interest, yielding impressive, clinical-grade diagnostic accuracy with a fraction of the training data typically required for such tasks.
- 25 March 2026
- Published by Springer Nature in Nature Biomedical Engineering
- p. 1-16
Top Cited PapersOpen AccessA generalizable deep learning system for cardiac MRI
Cardiac MRI allows for a comprehensive assessment of myocardial structure, function and tissue characteristics. Here we describe a foundational vision system for cardiac MRI, capable of representing the breadth of human cardiovascular disease and health. Our deep-learning model is trained via self-supervised contrastive learning, in which visual concepts in cine-sequence cardiac MRI scans are learned from the raw text of the accompanying radiology reports. We train and evaluate our model on data from four large academic clinical institutions in the United States. We additionally showcase the performance of our models on the UK BioBank and two additional publicly available external datasets. We explore emergent capabilities of our system and demonstrate remarkable performance across a range of tasks, including the problem of left-ventricular ejection fraction regression and the diagnosis of 39 different conditions such as cardiac amyloidosis and hypertrophic cardiomyopathy. We show that our deep-learning system is capable of not only contextualizing the staggering complexity of human cardiovascular disease but can be directed towards clinical problems of interest, yielding impressive, clinical-grade diagnostic accuracy with a fraction of the training data typically required for such tasks.
- Open Access
- 23 January 2026
- Published by Springer Nature in npj Digital Medicine
- Vol. 9 (1)
Sex disparities in deep learning estimation of ejection fraction from cardiac magnetic resonance imaging
The advent of artificial intelligence in cardiovascular imaging holds immense potential for earlier diagnoses, precision medicine, and improved disease management. However, the presence of sex-based disparities and strategies to mitigate biases in deep learning models for cardiac imaging remain understudied. In this study, we analyzed algorithmic bias in a foundation model that was pretrained on cardiac magnetic resonance imaging and radiology reports from multiple institutes and finetuned to estimate ejection fraction (EF) on the UK Biobank dataset. The model performed significantly worse in EF estimation for females than males in the diagnosis of reduced EF. Algorithmic fairness did not improve despite masking of protected attributes in radiology reports and data resampling, although explicit input of sex in model finetuning may improve EF estimation in some cases. The underdiagnosis of reduced EF among females holds critical implications for the exacerbation of existing sex-based disparities in cardiovascular health. We advise caution in the development of models for cardiovascular imaging to avoid such pitfalls.
- 23 January 2026
- Published by Springer Nature in npj Digital Medicine
- Vol. 9 (1)
Open AccessSex disparities in deep learning estimation of ejection fraction from cardiac magnetic resonance imaging
The advent of artificial intelligence in cardiovascular imaging holds immense potential for earlier diagnoses, precision medicine, and improved disease management. However, the presence of sex-based disparities and strategies to mitigate biases in deep learning models for cardiac imaging remain understudied. In this study, we analyzed algorithmic bias in a foundation model that was pretrained on cardiac magnetic resonance imaging and radiology reports from multiple institutes and finetuned to estimate ejection fraction (EF) on the UK Biobank dataset. The model performed significantly worse in EF estimation for females than males in the diagnosis of reduced EF. Algorithmic fairness did not improve despite masking of protected attributes in radiology reports and data resampling, although explicit input of sex in model finetuning may improve EF estimation in some cases. The underdiagnosis of reduced EF among females holds critical implications for the exacerbation of existing sex-based disparities in cardiovascular health. We advise caution in the development of models for cardiovascular imaging to avoid such pitfalls.
- Top Cited PapersOpen Access
- 7 March 2025
- Published by Springer Nature in npj Digital Medicine
- Vol. 8 (1) , 1-10
Red teaming ChatGPT in medicine to yield real-world insights on model behavior
Red teaming, the practice of adversarially exposing unexpected or undesired model behaviors, is critical towards improving equity and accuracy of large language models, but non-model creator-affiliated red teaming is scant in healthcare. We convened teams of clinicians, medical and engineering students, and technical professionals (80 participants total) to stress-test models with real-world clinical cases and categorize inappropriate responses along axes of safety, privacy, hallucinations/accuracy, and bias. Six medically-trained reviewers re-analyzed prompt-response pairs and added qualitative annotations. Of 376 unique prompts (1504 responses), 20.1% were inappropriate (GPT-3.5: 25.8%; GPT-4.0: 16%; GPT-4.0 with Internet: 17.8%). Subsequently, we show the utility of our benchmark by testing GPT-4o, a model released after our event (20.4% inappropriate). 21.5% of responses appropriate with GPT-3.5 were inappropriate in updated models. We share insights for constructing red teaming prompts, and present our benchmark for iterative model assessments.
- 7 March 2025
- Published by Springer Nature in npj Digital Medicine
- Vol. 8 (1) , 1-10
Top Cited PapersOpen AccessRed teaming ChatGPT in medicine to yield real-world insights on model behavior
Red teaming, the practice of adversarially exposing unexpected or undesired model behaviors, is critical towards improving equity and accuracy of large language models, but non-model creator-affiliated red teaming is scant in healthcare. We convened teams of clinicians, medical and engineering students, and technical professionals (80 participants total) to stress-test models with real-world clinical cases and categorize inappropriate responses along axes of safety, privacy, hallucinations/accuracy, and bias. Six medically-trained reviewers re-analyzed prompt-response pairs and added qualitative annotations. Of 376 unique prompts (1504 responses), 20.1% were inappropriate (GPT-3.5: 25.8%; GPT-4.0: 16%; GPT-4.0 with Internet: 17.8%). Subsequently, we show the utility of our benchmark by testing GPT-4o, a model released after our event (20.4% inappropriate). 21.5% of responses appropriate with GPT-3.5 were inappropriate in updated models. We share insights for constructing red teaming prompts, and present our benchmark for iterative model assessments.
- 1 January 2024
- Published by Wolters Kluwer Health in Circulation: Heart Failure
- Vol. 17 (1) , e010879
Race, Sex, and Age Disparities in the Performance of ECG Deep Learning Models Predicting Heart Failure
BACKGROUND: Deep learning models may combat widening racial disparities in heart failure outcomes through early identification of individuals at high risk. However, demographic biases in the performance of these models have not been well-studied. METHODS: This retrospective analysis used 12-lead ECGs taken between 2008 and 2018 from 326 518 patient encounters referred for standard clinical indications to Stanford Hospital. The primary model was a convolutional neural network model trained to predict incident heart failure within 5 years. Biases were evaluated on the testing set (160 312 ECGs) using the area under the receiver operating characteristic curve, stratified across the protected attributes of race, ethnicity, age, and sex. RESULTS: There were 59 817 cases of incident heart failure observed within 5 years of ECG collection. The performance of the primary model declined with age. There were no significant differences observed between racial groups overall. However, the primary model performed significantly worse in Black patients aged 0 to 40 years compared with all other racial groups in this age group, with differences most pronounced among young Black women. Disparities in model performance did not improve with the integration of race, ethnicity, sex, and age into model architecture, by training separate models for each racial group, or by providing the model with a data set of equal racial representation. Using probability thresholds individualized for race, age, and sex offered substantial improvements in F1 scores. CONCLUSIONS: The biases found in this study warrant caution against perpetuating disparities through the development of machine learning tools for the prognosis and management of heart failure. Customizing the application of these models by using probability thresholds individualized by race, ethnicity, age, and sex may offer an avenue to mitigate existing algorithmic disparities.
- 1 January 2024
- Published by Wolters Kluwer Health in Circulation: Heart Failure
- Vol. 17 (1) , e010879
Race, Sex, and Age Disparities in the Performance of ECG Deep Learning Models Predicting Heart Failure
BACKGROUND: Deep learning models may combat widening racial disparities in heart failure outcomes through early identification of individuals at high risk. However, demographic biases in the performance of these models have not been well-studied. METHODS: This retrospective analysis used 12-lead ECGs taken between 2008 and 2018 from 326 518 patient encounters referred for standard clinical indications to Stanford Hospital. The primary model was a convolutional neural network model trained to predict incident heart failure within 5 years. Biases were evaluated on the testing set (160 312 ECGs) using the area under the receiver operating characteristic curve, stratified across the protected attributes of race, ethnicity, age, and sex. RESULTS: There were 59 817 cases of incident heart failure observed within 5 years of ECG collection. The performance of the primary model declined with age. There were no significant differences observed between racial groups overall. However, the primary model performed significantly worse in Black patients aged 0 to 40 years compared with all other racial groups in this age group, with differences most pronounced among young Black women. Disparities in model performance did not improve with the integration of race, ethnicity, sex, and age into model architecture, by training separate models for each racial group, or by providing the model with a data set of equal racial representation. Using probability thresholds individualized for race, age, and sex offered substantial improvements in F1 scores. CONCLUSIONS: The biases found in this study warrant caution against perpetuating disparities through the development of machine learning tools for the prognosis and management of heart failure. Customizing the application of these models by using probability thresholds individualized by race, ethnicity, age, and sex may offer an avenue to mitigate existing algorithmic disparities.
- Open Access
Race, Sex and Age Disparities in the Performance of ECG Deep Learning Models Predicting Heart Failure
Background Deep learning models may combat widening racial disparities in heart failure outcomes through early identification of individuals at high risk. However, demographic biases in the performance of these models have not been well studied. Methods This retrospective analysis used 12-lead ECGs taken between 2008 - 2018 from 290,252 patients referred for standard clinical indications to Stanford Hospital. The primary model was a convolutional neural network model trained to predict incident heart failure within 5 years. Biases were evaluated on the testing set (160,312 ECGs) using area under the receiver operating curve (AUC), stratified across the protected attributes of race, ethnicity, age, and sex. Results 50,956 incident cases of heart failure were observed within 5 years of ECG collection. The performance of the primary model declined with age. There were no significant differences observed between racial groups overall. However, the primary model performed significantly worse in Black patients aged 0 - 40 compared to all other racial groups in this age group, with differences most pronounced among young Black women. Disparities in model performance did not improve with integration of race, ethnicity, gender, and/or age into model architecture, by training separate models for each racial group, nor by providing the model with a dataset of equal racial representation. Using probability thresholds individualized for race, age, and gender offered substantial improvements in F1-scores. Conclusion The biases found in this study warrant caution against perpetuating disparities through the development of machine learning tools for the prognosis and management of heart failure. Customizing the application of these models by using probability thresholds individualized by race/ethnicity, age, and sex may offer an avenue to mitigate existing algorithmic disparities.
Open AccessRace, Sex and Age Disparities in the Performance of ECG Deep Learning Models Predicting Heart Failure
Background Deep learning models may combat widening racial disparities in heart failure outcomes through early identification of individuals at high risk. However, demographic biases in the performance of these models have not been well studied. Methods This retrospective analysis used 12-lead ECGs taken between 2008 - 2018 from 290,252 patients referred for standard clinical indications to Stanford Hospital. The primary model was a convolutional neural network model trained to predict incident heart failure within 5 years. Biases were evaluated on the testing set (160,312 ECGs) using area under the receiver operating curve (AUC), stratified across the protected attributes of race, ethnicity, age, and sex. Results 50,956 incident cases of heart failure were observed within 5 years of ECG collection. The performance of the primary model declined with age. There were no significant differences observed between racial groups overall. However, the primary model performed significantly worse in Black patients aged 0 - 40 compared to all other racial groups in this age group, with differences most pronounced among young Black women. Disparities in model performance did not improve with integration of race, ethnicity, gender, and/or age into model architecture, by training separate models for each racial group, nor by providing the model with a dataset of equal racial representation. Using probability thresholds individualized for race, age, and gender offered substantial improvements in F1-scores. Conclusion The biases found in this study warrant caution against perpetuating disparities through the development of machine learning tools for the prognosis and management of heart failure. Customizing the application of these models by using probability thresholds individualized by race/ethnicity, age, and sex may offer an avenue to mitigate existing algorithmic disparities.
- 18 March 2021
- Published by Oxford University Press (OUP) in Journal of the American Medical Informatics Association
- Vol. 28 (4) , 801-811
Application of Bayesian networks to generate synthetic health data
Objective: This study seeks to develop a fully automated method of generating synthetic data from a real dataset that could be employed by medical organizations to distribute health data to researchers, reducing the need for access to real data. We hypothesize the application of Bayesian networks will improve upon the predominant existing method, medBGAN, in handling the complexity and dimensionality of healthcare data. Materials and Methods: We employed Bayesian networks to learn probabilistic graphical structures and simulated synthetic patient records from the learned structure. We used the University of California Irvine (UCI) heart disease and diabetes datasets as well as the MIMIC-III diagnoses database. We evaluated our method through statistical tests, machine learning tasks, preservation of rare events, disclosure risk, and the ability of a machine learning classifier to discriminate between the real and synthetic data. Results: Our Bayesian network model outperformed or equaled medBGAN in all key metrics. Notable improvement was achieved in capturing rare variables and preserving association rules. Discussion: Bayesian networks generated data sufficiently similar to the original data with minimal risk of disclosure, while offering additional transparency, computational efficiency, and capacity to handle more data types in comparison to existing methods. We hope this method will allow healthcare organizations to efficiently disseminate synthetic health data to researchers, enabling them to generate hypotheses and develop analytical tools. Conclusion: We conclude the application of Bayesian networks is a promising option for generating realistic synthetic health data that preserves the features of the original data without compromising data privacy.
- 18 March 2021
- Published by Oxford University Press (OUP) in Journal of the American Medical Informatics Association
- Vol. 28 (4) , 801-811
Application of Bayesian networks to generate synthetic health data
Objective: This study seeks to develop a fully automated method of generating synthetic data from a real dataset that could be employed by medical organizations to distribute health data to researchers, reducing the need for access to real data. We hypothesize the application of Bayesian networks will improve upon the predominant existing method, medBGAN, in handling the complexity and dimensionality of healthcare data. Materials and Methods: We employed Bayesian networks to learn probabilistic graphical structures and simulated synthetic patient records from the learned structure. We used the University of California Irvine (UCI) heart disease and diabetes datasets as well as the MIMIC-III diagnoses database. We evaluated our method through statistical tests, machine learning tasks, preservation of rare events, disclosure risk, and the ability of a machine learning classifier to discriminate between the real and synthetic data. Results: Our Bayesian network model outperformed or equaled medBGAN in all key metrics. Notable improvement was achieved in capturing rare variables and preserving association rules. Discussion: Bayesian networks generated data sufficiently similar to the original data with minimal risk of disclosure, while offering additional transparency, computational efficiency, and capacity to handle more data types in comparison to existing methods. We hope this method will allow healthcare organizations to efficiently disseminate synthetic health data to researchers, enabling them to generate hypotheses and develop analytical tools. Conclusion: We conclude the application of Bayesian networks is a promising option for generating realistic synthetic health data that preserves the features of the original data without compromising data privacy.