Enhancing the Utility of Privacy-Preserving Cancer Classification using Synthetic Data

TL;DR

This paper explores privacy-preserving deep learning for breast cancer detection, demonstrating that synthetic data and differential privacy techniques can enhance model utility while protecting patient privacy.

Contribution

It introduces a malignancy-conditioned GAN for synthetic data generation and evaluates its effectiveness combined with DP-SGD in breast cancer classification.

Findings

Synthetic data improves privacy-utility tradeoffs.

Pretraining on synthetic data boosts model performance.

DP-SGD fine-tuning further enhances accuracy under privacy constraints.

Abstract

Deep learning holds immense promise for aiding radiologists in breast cancer detection. However, achieving optimal model performance is hampered by limitations in availability and sharing of data commonly associated to patient privacy concerns. Such concerns are further exacerbated, as traditional deep learning models can inadvertently leak sensitive training information. This work addresses these challenges exploring and quantifying the utility of privacy-preserving deep learning techniques, concretely, (i) differentially private stochastic gradient descent (DP-SGD) and (ii) fully synthetic training data generated by our proposed malignancy-conditioned generative adversarial network. We assess these methods via downstream malignancy classification of mammography masses using a transformer model. Our experimental results depict that synthetic data augmentation can improve privacy-utility tradeoffs in differentially private model training. Further, model pretraining on synthetic data achieves remarkable performance, which can be further increased with DP-SGD fine-tuning across all privacy guarantees. With this first in-depth exploration of privacy-preserving deep learning in breast imaging, we address current and emerging clinical privacy requirements and pave the way towards the adoption of private high-utility deep diagnostic models. Our reproducible codebase is publicly available at https://github.com/RichardObi/mammo_dp.