Multi-institutional PET/CT image segmentation using federated deep transformer learning

Isaac Shiri; Behrooz Razeghi; Alireza Vafaei Sadr; Mehdi Amini; Yazdan Salimi; Sohrab Ferdowsi; Peter Boor; Deniz Gündüz; Slava Voloshynovskiy; Habib Zaidi

doi:10.1016/j.cmpb.2023.107706

Multi-institutional PET/CT image segmentation using federated deep transformer learning

Comput Methods Programs Biomed. 2023 Oct:240:107706. doi: 10.1016/j.cmpb.2023.107706. Epub 2023 Jul 12.

Authors

Affiliations

¹ Division of Nuclear Medicine and Molecular Imaging, Geneva University Hospital, Geneva, Switzerland.
² Department of Computer Science, University of Geneva, Geneva, Switzerland.
³ Institute of Pathology, RWTH Aachen University Hospital, Aachen, Germany; Department of Public Health Sciences, College of Medicine, The Pennsylvania State University, Hershey, PA 17033, USA.
⁴ Institute of Pathology, RWTH Aachen University Hospital, Aachen, Germany.
⁵ Department of Electrical and Electronic Engineering, Imperial College London, UK.
⁶ Division of Nuclear Medicine and Molecular Imaging, Geneva University Hospital, Geneva, Switzerland; Geneva University Neurocenter, University of Geneva, Geneva, Switzerland; Department of Nuclear Medicine and Molecular Imaging, University of Groningen, Groningen, The Netherlands; Department of Nuclear Medicine, University of Southern Denmark, Odense, Denmark. Electronic address: habib.zaidi@hcuge.ch.

PMID: 37506602
DOI: 10.1016/j.cmpb.2023.107706

Abstract

Background and objective: Generalizable and trustworthy deep learning models for PET/CT image segmentation necessitates large diverse multi-institutional datasets. However, legal, ethical, and patient privacy issues challenge sharing of datasets between different centers. To overcome these challenges, we developed a federated learning (FL) framework for multi-institutional PET/CT image segmentation.

Methods: A dataset consisting of 328 FL (HN) cancer patients who underwent clinical PET/CT examinations gathered from six different centers was enrolled. A pure transformer network was implemented as fully core segmentation algorithms using dual channel PET/CT images. We evaluated different frameworks (single center-based, centralized baseline, as well as seven different FL algorithms) using 68 PET/CT images (20% of each center data). In particular, the implemented FL algorithms include clipping with the quantile estimator (ClQu), zeroing with the quantile estimator (ZeQu), federated averaging (FedAvg), lossy compression (LoCo), robust aggregation (RoAg), secure aggregation (SeAg), and Gaussian differentially private FedAvg with adaptive quantile clipping (GDP-AQuCl).

Results: The Dice coefficient was 0.80±0.11 for both centralized and SeAg FL algorithms. All FL approaches achieved centralized learning model performance with no statistically significant differences. Among the FL algorithms, SeAg and GDP-AQuCl performed better than the other techniques. However, there was no statistically significant difference. All algorithms, except the center-based approach, resulted in relative errors less than 5% for SUV_max and SUV_mean for all FL and centralized methods. Centralized and FL algorithms significantly outperformed the single center-based baseline.

Conclusions: The developed FL-based (with centralized method performance) algorithms exhibited promising performance for HN tumor segmentation from PET/CT images.

Keywords: Deep transformers; Federated learning; PET/CT; Privacy; Segmentation.

Publication types

Multicenter Study

MeSH terms

Algorithms
Deep Learning*
Humans
Image Processing, Computer-Assisted / methods
Neoplasms* / diagnostic imaging
Positron Emission Tomography Computed Tomography / methods