Privacy risks of whole-slide image sharing in digital pathology
In their recent publication, researchers from BBMRI.at partner Med Uni Graz propose a model for the assessment of privacy risks associated with the sharing of whole-slide images in the growing field of digital and AI-assisted pathology.
In their paper regarding security risks in digital pathology, experts from BBMRI.at partner Med Uni Graz introduce a hierarchical taxonomy of whole-slide images based on the potential of these images to be linked to each other. In addition, they provide a thorough mathematical model aimed at assessing the risk of security threats associated with the sharing of whole-slide images and related data, especially regarding identity disclosure attacks. Furthermore, these risks are demonstrated using real-life data sets, evaluating the newly established risk assessment model in a series of experiments.
The authors conclude their findings by setting up detailed guidelines for the release of whole-slide images as anonymous or personal data and state their recommendations for the minimization of privacy risks in the context of sharing whole-slide images for digital pathology applications.
Holub, P.; Müller, H.; Bíl, T.; Pireddu, L.; Plass, M.; Prasser, F.; Schlünder, I; Zatloukal, K.; Neunutil, R.; Brázdil, T. Privacy risks of whole-slide image sharing in digital pathology. Nat Commun 14, 2577 (2023)
