Decentralized Distributed Multi-institutional PET Image Segmentation Using a Federated Deep Learning Framework

Clinical Nuclear Medicine, vol. 47, pp. 606–617

Abstract

PURPOSE: The generalizability and trustworthiness of deep learning (DL)-based algorithms depend on the size and heterogeneity of training datasets. However, because of patient privacy concerns and ethical and legal issues, sharing medical images between different centers is restricted. Our objective is to build a federated DL-based framework for PET image segmentation utilizing a multicentric dataset and to compare its performance with the centralized DL approach. METHODS: PET images from 405 head and neck cancer patients from 9 different centers formed the basis of this study. All tumors were segmented manually. PET images converted to SUV maps were resampled to isotropic voxels (3 × 3 × 3 mm3) and then normalized. PET image subvolumes (12 × 12 × 12 cm3) consisting of whole tumors and background were analyzed. Data from each center were divided into train/validation (80% of patients) and test sets (20% of patients). The modified R2U-Net was used as core DL model. A parallel federated DL model was developed and compared with the centralized approach where the data sets are pooled to one server. Segmentation metrics, including Dice similarity and Jaccard coefficients, percent relative errors (RE%) of SUVpeak, SUVmean, SUVmedian, SUVmax, metabolic tumor volume, and total lesion glycolysis were computed and compared with manual delineations. RESULTS: The performance of the centralized versus federated DL methods was nearly identical for segmentation metrics: Dice (0.84 ± 0.06 vs 0.84 ± 0.05) and Jaccard (0.73 ± 0.08 vs 0.73 ± 0.07). For quantitative PET parameters, we obtained comparable RE% for SUVmean (6.43% ± 4.72% vs 6.61% ± 5.42%), metabolic tumor volume (12.2% ± 16.2% vs 12.1% ± 15.89%), and total lesion glycolysis (6.93% ± 9.6% vs 7.07% ± 9.85%) and negligible RE% for SUVmax and SUVpeak. No significant differences in performance (P > 0.05) between the 2 frameworks (centralized vs federated) were observed. CONCLUSION: The developed federated DL model achieved comparable quantitative performance with respect to the centralized DL model. Federated DL models could provide robust and generalizable segmentation, while addressing patient privacy and legal and ethical issues in clinical data sharing.

Authors 15

1. Isaac Shiri corresponding

   University Hospital of Geneva

   Affiliation as printed

   Division of Nuclear Medicine and Molecular Imaging, Geneva University Hospital

2. Alireza Vafaei Sadr Aachen Institute of Pathology

   University of Geneva · Universitätsklinikum Aachen · RWTH Aachen University

   Affiliation as printed

   Department of Theoretical Physics and Center for Astroparticle Physics, University of Geneva, Geneva, Switzerland

   Institute of Pathology, RWTH Aachen University Hospital, Aachen, Germany

3. Mehdi Amini corresponding

   University Hospital of Geneva

   Affiliation as printed

   Division of Nuclear Medicine and Molecular Imaging, Geneva University Hospital

4. Yazdan Salimi corresponding

   University Hospital of Geneva

   Affiliation as printed

   Division of Nuclear Medicine and Molecular Imaging, Geneva University Hospital

5. Amirhossein Sanaat corresponding

   University Hospital of Geneva

   Affiliation as printed

   Division of Nuclear Medicine and Molecular Imaging, Geneva University Hospital

6. Azadeh Akhavanallaf corresponding

   University Hospital of Geneva

   Affiliation as printed

   Division of Nuclear Medicine and Molecular Imaging, Geneva University Hospital

7. Behrooz Razeghi
   Affiliation as printed

   Department of Computer Science

8. Sohrab Ferdowsi

   University of Geneva · HES-SO Genève

   Affiliation as printed

   HES-SO, University of Geneva, Geneva

9. Abdollah Saberi corresponding

   University Hospital of Geneva

   Affiliation as printed

   Division of Nuclear Medicine and Molecular Imaging, Geneva University Hospital

10. Hossein Arabi corresponding

    University Hospital of Geneva

    Affiliation as printed

    Division of Nuclear Medicine and Molecular Imaging, Geneva University Hospital

11. Minerva Becker

    University Hospital of Geneva

    Affiliation as printed

    Division of Radiology, Geneva University Hospital, Geneva, Switzerland

12. Slava Voloshynovskiy
    Affiliation as printed

    Department of Computer Science

13. Denız Gündüz corresponding

    Imperial College London

    Affiliation as printed

    Faculty of Engineering, Department of Electrical and Electronic Engineering, Imperial College London, London, United Kingdom

14. Arman Rahmim Aachen Institute of Pathology

    University of Geneva · BC Cancer Agency · University of British Columbia · Universitätsklinikum Aachen · RWTH Aachen University

    Affiliation as printed

    Department of Theoretical Physics and Center for Astroparticle Physics, University of Geneva, Geneva, Switzerland

    Institute of Pathology, RWTH Aachen University Hospital, Aachen, Germany

    Department of Integrative Oncology, BC Cancer Research Institute, Vancouver, BC, Canada

    Department of Radiology and Physics, University of British Columbia, Vancouver, BC

15. Habib Zaidi corresponding

    University of Geneva · University Medical Center Groningen · University of Groningen · University of Southern Denmark · University Hospital of Geneva

    Affiliation as printed

    Department of Nuclear Medicine and Molecular Imaging, University of Groningen, University Medical Center Groningen, Groningen, The Netherlands

    Department of Nuclear Medicine, University of Southern Denmark, Odense, Denmark

    Division of Nuclear Medicine and Molecular Imaging, Geneva University Hospital

    Geneva University Neurocenter, University of Geneva, Geneva, Switzerland