De-identify Aperio SVS whole slide image files

TL;DR: try the stand-alone application for deidentifying Aperio SVS files

The problem

Whole slide images (WSIs) - digitized microscope slides - are a valuable source of data for research on human diseases. When microscope slides are digitized, the scanner often captures an image of the slide label, to keep track of which slide was digitized. For pathology specimens of real human patients, the slide label often shows the patient's name and/or other identifiers, making the resulting WSIs into protected health information.

This demonstration WSI from the OpenSlide example data website demonstrates the label and macro images:

As you can see, the WSI file contains not just the high resolution tissue image, but also the slide label and a macro view showing the entire slide, including label, from above. In this case the label does not have patient information, but a clincal slide would. Depending on the label type and illumination in the scanner, the macro view can also show a readable label.

The solution

By removing the label and macro images, the WSI file can be safely used for research in a deidentified manner. The svs-deidentifier is an open-source application built to efficiently deidentify WSIs, specifically Aperio .svs image files.

See it in action

WSIs selected for deidentification:

Deidentification in progress:

Completed:

The app can be used in two modes: 1) create a deidentified copy; 2) modify a WSI in place. In the images above, the app was used in "Copy" mode, and the resulting deidentified images were created in a new folder.

The result:

Opening the resulting file with a whole slide image viewer demonstrates that the "associated images" for label and macro have been removed, and the label data cannot be shown anymore. Safely deidentified!

Downloads