English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

Artificial intelligence (AI) models are increasingly used in the medicaldomain. However, as medical data is highly sensitive, special precautions toensure its protection are required. The gold standard for privacy preservationis the introduction of differential privacy (DP) to model training. Prior workindicates that DP has negative implications on model accuracy and fairness,which are unacceptable in medicine and represent a main barrier to thewidespread use of privacy-preserving techniques. In this work, we evaluated theeffect of privacy-preserving training of AI models regarding accuracy andfairness 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 thetask of classifying the presence of pancreatic ductal adenocarcinoma (PDAC).Both were retrospectively collected and manually labeled by experiencedradiologists. We then compared non-private deep convolutional neural networks(CNNs) and privacy-preserving (DP) models with respect to privacy-utilitytrade-offs measured as area under the receiver-operator-characteristic curve(AUROC), and privacy-fairness trade-offs, measured as Pearson's r orStatistical Parity Difference. We found that, while the privacy-preservingtrainings yielded lower accuracy, they did largely not amplify discriminationagainst age, sex or co-morbidity. Our study shows that -- under the challengingrealistic circumstances of a real-life clinical dataset -- theprivacy-preserving training of diagnostic deep learning models is possible withexcellent diagnostic accuracy and fairness.

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