Private, fair and accurate: Training large-scale, privacy-preserving AI models in medical imaging

Abstract

Artificial intelligence (AI) models are increasingly used in the medical domain. However, as medical data is highly sensitive, special precautions to ensure its protection are required. The gold standard for privacy preservation is the introduction of differential privacy (DP) to model training. Prior work indicates that DP has negative implications on model accuracy and fairness, which are unacceptable in medicine and represent a main barrier to the widespread use of privacy-preserving techniques. In this work, we evaluated the effect of privacy-preserving training of AI models regarding accuracy and fairness compared to non-private training. For this, we used two datasets: (1) A large dataset (N=193,311) of high quality clinical chest radiographs, and (2) a dataset (N=1,625) of 3D abdominal computed tomography (CT) images, with the task of classifying the presence of pancreatic ductal adenocarcinoma (PDAC). Both were retrospectively collected and manually labeled by experienced radiologists. We then compared non-private deep convolutional neural networks (CNNs) and privacy-preserving (DP) models with respect to privacy-utility trade-offs measured as area under the receiver-operator-characteristic curve (AUROC), and privacy-fairness trade-offs, measured as Pearson's r or Statistical Parity Difference. We found that, while the privacy-preserving trainings yielded lower accuracy, they did largely not amplify discrimination against age, sex or co-morbidity. Our study shows that -- under the challenging realistic circumstances of a real-life clinical dataset -- the privacy-preserving training of diagnostic deep learning models is possible with excellent diagnostic accuracy and fairness.

arXiv commentPublished in Communications Medicine. Nature Portfolio

References in corpus (8)

• Deep Learning with Differential Privacy
• Do Gradient Inversion Attacks Make Federated Learning Unsafe?
• Reinventing 2D Convolutions for 3D Images
• Towards Formalizing the GDPR's Notion of Singling Out
• Deep Learning-based Patient Re-identification Is able to Exploit the Biometric Nature of Medical Chest X-ray Data
• Collaborative Training of Medical Artificial Intelligence Models with non-uniform Labels
• Preserving privacy in domain transfer of medical AI models comes at no performance costs: The integral role of differential privacy
• Tight Auditing of Differentially Private Machine Learning

Cited by in corpus (6)

• Differential privacy for medical deep learning: methods, tradeoffs, and deployment implications
• The Impact of Speech Anonymization on Pathology and Its Limits
• Boosting multi-demographic federated learning for chest radiograph analysis using general-purpose self-supervised representations
• Differential privacy enables fair and accurate AI-based analysis of speech disorders while protecting patient data
• Effect of Reporting Mode and Clinical Experience on Radiologists' Gaze and Image Analysis Behavior in Chest Radiography
• Resolution scaling governs DINOv3 transfer performance in chest radiograph classification