The role of self-supervised pretraining in differentially private medical image analysis

2026-01-27

AI Summary

This study addresses the significant drop in diagnostic performance often caused by differential privacy (DP) in medical imaging. It presents the first systematic evaluation, under full-model DP settings, of how general supervised (ImageNet), general self-supervised (DINOv3), and domain-specific supervised (MIMIC-CXR) pretraining affect chest X-ray classification. Using DP-SGD to train ConvNeXt models on over 800,000 chest radiographs and five external datasets, the experiments demonstrate that DINOv3 initialization outperforms ImageNet, while domain-specific supervised pretraining achieves performance closest to the non-private baseline. Moreover, it substantially enhances cross-institutional generalization, demographic fairness, and robustness to variations in data scale and model capacity, underscoring the critical role of initialization strategies in balancing privacy and utility.

Technology Category

• Machine Learning: Privacy
• Computer Vision: Diffusion Models for Vision
• Search and Optimization: Learning to Search

Application Category

• Security and Privacy: Data transparency and provenance
• Responsible Web: Data and user privacy-enhancing technologies for the Web
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking

Abstract

Differential privacy (DP) provides formal protection for sensitive data but typically incurs substantial losses in diagnostic performance. Model initialization has emerged as a critical factor in mitigating this degradation, yet the role of modern self-supervised learning under full-model DP remains poorly understood. Here, we present a large-scale evaluation of initialization strategies for differentially private medical image analysis, using chest radiograph classification as a representative benchmark with more than 800,000 images. Using state-of-the-art ConvNeXt models trained with DP-SGD across realistic privacy regimes, we compare non-domain-specific supervised ImageNet initialization, non-domain-specific self-supervised DINOv3 initialization, and domain-specific supervised pretraining on MIMIC-CXR, the largest publicly available chest radiograph dataset. Evaluations are conducted across five external datasets spanning diverse institutions and acquisition settings. We show that DINOv3 initialization consistently improves diagnostic utility relative to ImageNet initialization under DP, but remains inferior to domain-specific supervised pretraining, which achieves performance closest to non-private baselines. We further demonstrate that initialization choice strongly influences demographic fairness, cross-dataset generalization, and robustness to data scale and model capacity under privacy constraints. The results establish initialization strategy as a central determinant of utility, fairness, and generalization in differentially private medical imaging.

Problem

• differential privacy
• medical image analysis
• self-supervised pretraining
• model initialization
• diagnostic performance

Innovation

• differential privacy
• self-supervised learning
• medical image analysis
• model initialization
• fairness