Generative artificial intelligence enables the generation of bone scintigraphy images and improves generalization of deep learning models in data-constrained environments.

Journal: European journal of nuclear medicine and molecular imaging
Published Date: Jan 29, 2025

Abstract

PURPOSE: Advancements of deep learning in medical imaging are often constrained by the limited availability of large, annotated datasets, resulting in underperforming models when deployed under real-world conditions. This study investigated a generative artificial intelligence (AI) approach to create synthetic medical images taking the example of bone scintigraphy scans, to increase the data diversity of small-scale datasets for more effective model training and improved generalization.

Authors

  • David Haberl
    Division of Nuclear Medicine, Medical University of Vienna, Vienna, Austria.
  • Jing Ning
    Christian Doppler Laboratory for Applied Metabolomics, 1090 Vienna, Austria.
  • Kilian Kluge
    Department of Biomedical Imaging and Image-guided Therapy, Division of Nuclear Medicine, Medical University of Vienna, Spitalgasse 23, Vienna, 1090, Austria.
  • Katarina Kumpf
    IT4Science, IT Services & Strategic Information Management, Medical University of Vienna, Vienna, Austria.
  • Josef Yu
    Department of Biomedical Imaging and Image-guided Therapy, Division of Nuclear Medicine, Medical University of Vienna, Spitalgasse 23, Vienna, 1090, Austria.
  • Zewen Jiang
    Department of Biomedical Imaging and Image-guided Therapy, Division of Nuclear Medicine, Medical University of Vienna, Spitalgasse 23, Vienna, 1090, Austria.
  • Claudia Constantino
    Nuclear Medicine-Radiopharmacology, Champalimaud Clinical Centre, Champalimaud Foundation, Lisbon, Portugal.
  • Alice Monaci
    Department of Experimental and Clinical Biomedical Sciences, Nuclear Medicine Unit, University of Florence, Florence, Italy.
  • Maria Starace
    Department of Experimental and Clinical Biomedical Sciences, Nuclear Medicine Unit, University of Florence, Florence, Italy.
  • Alexander R Haug
    Department of Biomedical Imaging and Image-guided Therapy, Division of Nuclear Medicine, Medical University of Vienna, Spitalgasse 23, Vienna, 1090, Austria.
  • Raffaella Calabretta
    Department of Biomedical Imaging and Image-guided Therapy, Division of Nuclear Medicine, Medical University of Vienna, Spitalgasse 23, Vienna, 1090, Austria.
  • Luca Camoni
    ASST Spedali Civili of Brescia, Università degli Studi di Brescia, Brescia, Italy.
  • Francesco Bertagna
    Nuclear Medicine Unit ASST Spedali Civili Hospital, University of Brescia, Brescia, Italy.
  • Katharina Mascherbauer
    Department of Internal Medicine II, Division of Cardiology, Medical University of Vienna, Vienna, Austria.
  • Felix Hofer
    Department of Internal Medicine II, Division of Cardiology, Medical University of Vienna, Vienna, Austria.
  • Domenico Albano
    IRCCS Istituto Ortopedico Galeazzi, Milano, Italy; Sezione di Scienze Radiologiche, Dipartimento di Biomedicina, Neuroscienze e Diagnostica Avanzata, Università degli Studi di Palermo, Palermo, Italy.
  • Roberto Sciagra
    Department of Experimental and Clinical Biomedical Sciences, Nuclear Medicine Unit, University of Florence, Florence, Italy.
  • Francisco Oliveira
    Centro Clínico Champalimaud, Serviço de Medicina Nuclear, Champalimaud Foundation, Lisbon, Portugal.
  • Durval Costa
    Nuclear Medicine-Radiopharmacology, Champalimaud Clinical Centre, Champalimaud Foundation, Lisbon, Portugal.
  • Christian Nitsche
    Department of Internal Medicine II, Division of Cardiology, Medical University of Vienna, Vienna, Austria.
  • Marcus Hacker
    Division of Nuclear Medicine, Department of Biomedical Imaging and Image-Guided Therapy, Medical University of Vienna, Vienna, Austria.
  • Clemens P Spielvogel
    Division of Nuclear Medicine, Medical University of Vienna, Vienna, Austria.

Keywords

External Resources

Popular Topics
Recent Journals