Clinical phenotypes among patients with normal cardiac perfusion using unsupervised learning: a retrospective observational study.

Cluster analysis Coronary artery disease Machine learning Myocardial perfusion

Journal

EBioMedicine
ISSN: 2352-3964
Titre abrégé: EBioMedicine
Pays: Netherlands
ID NLM: 101647039

Informations de publication

Date de publication:
01 Jan 2024
Historique:
received: 31 08 2023
revised: 24 11 2023
accepted: 05 12 2023
medline: 4 1 2024
pubmed: 4 1 2024
entrez: 3 1 2024
Statut: aheadofprint

Résumé

Myocardial perfusion imaging (MPI) is one of the most common cardiac scans and is used for diagnosis of coronary artery disease and assessment of cardiovascular risk. However, the large majority of MPI patients have normal results. We evaluated whether unsupervised machine learning could identify unique phenotypes among patients with normal scans and whether those phenotypes were associated with risk of death or myocardial infarction. Patients from a large international multicenter MPI registry (10 sites) with normal perfusion by expert visual interpretation were included in this cohort analysis. The training population included 9849 patients, and external testing population 12,528 patients. Unsupervised cluster analysis was performed, with separate training and external testing cohorts, to identify clusters, with four distinct phenotypes. We evaluated the clinical and imaging features of clusters and their associations with death or myocardial infarction. Patients in Clusters 1 and 2 almost exclusively underwent exercise stress, while patients in Clusters 3 and 4 mostly required pharmacologic stress. In external testing, the risk for Cluster 4 patients (20.2% of population, unadjusted hazard ratio [HR] 6.17, 95% confidence interval [CI] 4.64-8.20) was higher than the risk associated with pharmacologic stress (HR 3.03, 95% CI 2.53-3.63), or previous myocardial infarction (HR 1.82, 95% CI 1.40-2.36). Unsupervised learning identified four distinct phenotypes of patients with normal perfusion scans, with a significant proportion of patients at very high risk of myocardial infarction or death. Our results suggest a potential role for patient phenotyping to improve risk stratification of patients with normal imaging results. This work was supported by the National Heart, Lung, and Blood Institute at the National Institutes of Health [R35HL161195 to PS]. The REFINE SPECT database was supported by the National Heart, Lung, and Blood Institute at the National Institutes of Health [R01HL089765 to PS]. MCW was supported by the British Heart Foundation [FS/ICRF/20/26002].

Sections du résumé

BACKGROUND BACKGROUND
Myocardial perfusion imaging (MPI) is one of the most common cardiac scans and is used for diagnosis of coronary artery disease and assessment of cardiovascular risk. However, the large majority of MPI patients have normal results. We evaluated whether unsupervised machine learning could identify unique phenotypes among patients with normal scans and whether those phenotypes were associated with risk of death or myocardial infarction.
METHODS METHODS
Patients from a large international multicenter MPI registry (10 sites) with normal perfusion by expert visual interpretation were included in this cohort analysis. The training population included 9849 patients, and external testing population 12,528 patients. Unsupervised cluster analysis was performed, with separate training and external testing cohorts, to identify clusters, with four distinct phenotypes. We evaluated the clinical and imaging features of clusters and their associations with death or myocardial infarction.
FINDINGS RESULTS
Patients in Clusters 1 and 2 almost exclusively underwent exercise stress, while patients in Clusters 3 and 4 mostly required pharmacologic stress. In external testing, the risk for Cluster 4 patients (20.2% of population, unadjusted hazard ratio [HR] 6.17, 95% confidence interval [CI] 4.64-8.20) was higher than the risk associated with pharmacologic stress (HR 3.03, 95% CI 2.53-3.63), or previous myocardial infarction (HR 1.82, 95% CI 1.40-2.36).
INTERPRETATION CONCLUSIONS
Unsupervised learning identified four distinct phenotypes of patients with normal perfusion scans, with a significant proportion of patients at very high risk of myocardial infarction or death. Our results suggest a potential role for patient phenotyping to improve risk stratification of patients with normal imaging results.
FUNDING BACKGROUND
This work was supported by the National Heart, Lung, and Blood Institute at the National Institutes of Health [R35HL161195 to PS]. The REFINE SPECT database was supported by the National Heart, Lung, and Blood Institute at the National Institutes of Health [R01HL089765 to PS]. MCW was supported by the British Heart Foundation [FS/ICRF/20/26002].

Identifiants

DOI: 10.1016/j.ebiom.2023.104930 PMID: 38168587
pubmed: 38168587
pii: S2352-3964(23)00496-6
doi: 10.1016/j.ebiom.2023.104930
pii:
doi:

Types de publication

Journal Article

Langues

eng

Sous-ensembles de citation

IM

Pagination

104930

Informations de copyright

Copyright © 2023. Published by Elsevier B.V.

Déclaration de conflit d'intérêts

Declaration of interests Dr. Robert Miller has received consulting and research support from Pfizer. Drs Berman and Slomka participate in software royalties for QPS software at Cedars-Sinai Medical Center. Dr Williams serves as the President-Elect of the British Society of Cardiovascular Imaging and is on the Board of Directors for the Society of Cardiovascular Computed Tomography; she has received consulting support from FEOPS and has given lectures for Canon Medical Systems, Siemens Healthineers and Novartis. Dr. Pieszko has served as a consultant for Medicalgorithmics S.A. Dr. Slomka has received consulting fees from Synektik. Drs. Berman, Sharir, Kaufmann, and Edward Miller have served as consultants for GE Healthcare. Dr. Dorbala has received honoraria from Novo Nordisk and Pfizer; her institution has received grant support from Attralus, Pfizer, GE Healthcare, Siemans, and Phillips. Dr. DiCarli has received institutional research grant support from Gilead Sciences and Amgen and consulting honoraria from Sanofi, Valo Health and MedTrace. Dr. Ruddy has received research grant support from GE Healthcare and Pfizer. Dr. Edward Miller has served as a consultant for ROIVANT; has received grant support from Anylam, Pfizer and Siemens, and has participated on the study advisory board of BioBridge. Dr. Sinusas serves a leadership role on the Society of Nuclear Medicine and Molecular Imaging Cardiovascular Council. Dr. Einstein receives royalties from Wolters Kluwer UpToDate and the American Society of Nuclear Cardiology/Society of Nuclear Medicine and Molecular Imaging, consulting fees from W.L Gore & Associates, support through patents with Columbia Technology Ventures, and has given lectures for Ionetix. Dr. Einstein's institution has received research support from GE Healthcare, Roche Medical Systems, W. L. Gore & Associates, Eidos Therapeutics, Attralus, Pfizer, Neovasc, Intellia Therapeutics, Ionis Pharmaceuticals, Canon Medical Systems, the International Atomic Energy Agency, National Council on Radiation Protection and Measurements, and the United States Regulatory Commission. The remaining authors have nothing to disclose.

Auteurs

Robert J H Miller (RJH)

Departments of Medicine (Division of Artificial Intelligence in Medicine), Biomedical Sciences, and Imaging, Cedars-Sinai Medical Center, Los Angeles, CA, USA; Department of Cardiac Sciences, University of Calgary and Libin Cardiovascular Institute, Calgary, AB, Canada.

Bryan P Bednarski (BP)

Departments of Medicine (Division of Artificial Intelligence in Medicine), Biomedical Sciences, and Imaging, Cedars-Sinai Medical Center, Los Angeles, CA, USA.

Konrad Pieszko (K)

Departments of Medicine (Division of Artificial Intelligence in Medicine), Biomedical Sciences, and Imaging, Cedars-Sinai Medical Center, Los Angeles, CA, USA.

Jacek Kwiecinski (J)

Departments of Medicine (Division of Artificial Intelligence in Medicine), Biomedical Sciences, and Imaging, Cedars-Sinai Medical Center, Los Angeles, CA, USA; Department of Interventional Cardiology and Angiology, Institute of Cardiology, Warsaw, Poland.

Michelle C Williams (MC)

Departments of Medicine (Division of Artificial Intelligence in Medicine), Biomedical Sciences, and Imaging, Cedars-Sinai Medical Center, Los Angeles, CA, USA; British Heart Foundation Centre for Cardiovascular Science, University of Edinburgh, Edinburgh, United Kingdom.

Aakash Shanbhag (A)

Departments of Medicine (Division of Artificial Intelligence in Medicine), Biomedical Sciences, and Imaging, Cedars-Sinai Medical Center, Los Angeles, CA, USA; Signal and Image Processing Institute, Ming Hsieh Department of Electrical and Computer Engineering, University of Southern California, Los Angeles, CA, USA.

Joanna X Liang (JX)

Departments of Medicine (Division of Artificial Intelligence in Medicine), Biomedical Sciences, and Imaging, Cedars-Sinai Medical Center, Los Angeles, CA, USA.

Cathleen Huang (C)

Departments of Medicine (Division of Artificial Intelligence in Medicine), Biomedical Sciences, and Imaging, Cedars-Sinai Medical Center, Los Angeles, CA, USA.

Tali Sharir (T)

Department of Nuclear Cardiology, Assuta Medical Centers, Tel Aviv, Israel; Israel and Ben Gurion University of the Negev, Beer Sheba, Israel.

M Timothy Hauser (MT)

Department of Nuclear Cardiology, Oklahoma Heart Hospital, Oklahoma City, OK, USA.

Sharmila Dorbala (S)

Division of Nuclear Medicine and Molecular Imaging, Department of Radiology, Brigham and Women's Hospital, Boston, MA, USA.

Marcelo F Di Carli (MF)

Division of Nuclear Medicine and Molecular Imaging, Department of Radiology, Brigham and Women's Hospital, Boston, MA, USA.

Mathews B Fish (MB)

Oregon Heart and Vascular Institute, Sacred Heart Medical Center, Springfield, OR, USA.

Terrence D Ruddy (TD)

Division of Cardiology, University of Ottawa Heart Institute, Ottawa, ON, Canada.

Timothy M Bateman (TM)

Cardiovascular Imaging Technologies LLC, Kansas City, MO, USA.

Andrew J Einstein (AJ)

Division of Cardiology, Department of Medicine and Department of Radiology, Columbia University Irving Medical Center and New York-Presbyterian Hospital, New York, NY, USA.

Philipp A Kaufmann (PA)

Department of Nuclear Medicine, Cardiac Imaging, University Hospital Zurich, Zurich, Switzerland.

Edward J Miller (EJ)

Section of Cardiovascular Medicine, Department of Internal Medicine, Yale University School of Medicine, New Haven, CT, USA.

Albert J Sinusas (AJ)

Section of Cardiovascular Medicine, Department of Internal Medicine, Yale University School of Medicine, New Haven, CT, USA.

Wanda Acampa (W)

Department of Advanced Biomedical Sciences, University of Naples "Federico II", Naples, Italy.

Donghee Han (D)

Departments of Medicine (Division of Artificial Intelligence in Medicine), Biomedical Sciences, and Imaging, Cedars-Sinai Medical Center, Los Angeles, CA, USA.

Damini Dey (D)

Departments of Medicine (Division of Artificial Intelligence in Medicine), Biomedical Sciences, and Imaging, Cedars-Sinai Medical Center, Los Angeles, CA, USA.

Daniel S Berman (DS)

Departments of Medicine (Division of Artificial Intelligence in Medicine), Biomedical Sciences, and Imaging, Cedars-Sinai Medical Center, Los Angeles, CA, USA.

Piotr J Slomka (PJ)

Departments of Medicine (Division of Artificial Intelligence in Medicine), Biomedical Sciences, and Imaging, Cedars-Sinai Medical Center, Los Angeles, CA, USA. Electronic address: Piotr.Slomka@cshs.org.

Classifications MeSH