Mitigating bias in deep learning for diagnosis of coronary artery disease from myocardial perfusion SPECT images.

Journal: European journal of nuclear medicine and molecular imaging
PMID:

Abstract

PURPOSE: Artificial intelligence (AI) has high diagnostic accuracy for coronary artery disease (CAD) from myocardial perfusion imaging (MPI). However, when trained using high-risk populations (such as patients with correlating invasive testing), the disease probability can be overestimated due to selection bias. We evaluated different strategies for training AI models to improve the calibration (accurate estimate of disease probability), using external testing.

Authors

  • Robert J H Miller
    Department of Imaging and Medicine and the Smidt Heart Institute (E.E., S.C., H.G., S.C., R.J.H.M., P.J.S., B.K.T., D.S.B.), Cedars-Sinai Medical Center, Los Angeles, CA.
  • Ananya Singh
    Departments of Imaging, Medicine and Biomedical Sciences, Division of Artificial Intelligence in Medicine, Cedars-Sinai Medical Center, Los Angeles, CA, USA. Ananya.Singh@cshs.org.
  • Yuka Otaki
    Department of Imaging, Medicine, and Biomedical Sciences, Cedars-Sinai Medical Center, Los Angeles, California.
  • Balaji K Tamarappoo
  • Paul Kavanagh
    Division of Artificial Intelligence in Medicine, Department of Medicine, Imaging, and Biomedical Sciences, Cedars-Sinai Medical Center, Los Angeles, California.
  • Tejas Parekh
    Division of Artificial Intelligence in Medicine, Department of Medicine, Imaging, and Biomedical Sciences, Cedars-Sinai Medical Center, Los Angeles, California.
  • Lien-Hsin Hu
    Division of Nuclear Medicine, Department of Imaging, and Departments of Medicine and Biomedical Sciences, Cedars-Sinai Medical Center, Los Angeles, California.
  • Heidi Gransar
    Departments of Imaging, Medicine, and Biomedical Sciences, Cedars-Sinai Medical Center, Los Angeles, California.
  • Tali Sharir
    Department of Nuclear Cardiology, Assuta Medical Centers, Tel Aviv, Israel.
  • Andrew J Einstein
    Division of Cardiology, Department of Medicine, Columbia University Medical Center and New York-Presbyterian Hospital, New York, New York; Department of Radiology, Columbia University Medical Center and New York-Presbyterian Hospital, New York, New York.
  • Mathews B Fish
    Oregon Heart and Vascular Institute, Sacred Heart Medical Center, Springfield, Oregon.
  • Terrence D Ruddy
    Division of Cardiology, University of Ottawa Heart Institute, Ottawa, Canada.
  • Philipp A Kaufmann
    Cardiac Imaging, Department of Nuclear Medicine, University Hospital Zurich, Zurich, Switzerland; and.
  • Albert J Sinusas
    Section of Cardiovascular Medicine, Department of Internal Medicine, Yale University School of Medicine, New Haven, Connecticut.
  • Edward J Miller
    Section of Cardiovascular Medicine, Department of Internal Medicine, Yale School of Medicine, New Haven, CT, United States.
  • Timothy M Bateman
    Cardiovascular Imaging Technologies LLC, Kansas City, Missouri.
  • Sharmila Dorbala
    Department of Radiology, Division of Nuclear Medicine and Molecular Imaging, Brigham and Women's Hospital, Boston, Massachusetts.
  • Marcelo F Di Carli
    Cardiovascular Division, Department of Medicine, Brigham and Women's Hospital, Harvard Medical School, Boston, Massachusetts, USA.
  • Joanna X Liang
    Division of Nuclear Medicine, Department of Imaging, and Departments of Medicine and Biomedical Sciences, Cedars-Sinai Medical Center, Los Angeles, California.
  • Damini Dey
    Departments of Imaging and Medicine, and Cedars-Sinai Heart Institute, Cedars-Sinai Medical Center, 8700 Beverly Blvd, Taper A238, Los Angeles, CA, 90048, USA.
  • Daniel S Berman
    Departments of Imaging and Medicine, Cedars-Sinai Medical Center, Los Angeles, CA, USA.
  • Piotr J Slomka
    Department of Imaging, Medicine, and Biomedical Sciences, Cedars-Sinai Medical Center, Los Angeles, California Piotr.Slomka@cshs.org.