In a digitized pathology workflow, a tissue section is mounted on a glass slide and scanned to generate a whole-slide image (WSI), a process which can introduce human or digital errors that hinder assessment (Fig. 1A). For example, air bubbles under a coverslip or out-of-focus regions due to scanning errors may limit a pathologist’s ability to fully evaluate the WSI.
Due to the large volume of cases in a digitized anatomic pathology (AP) laboratory and the time required for specimen reprocessing, an automated method to screen WSI quality prior to pathologist review is desirable.
To address this need, we developed a machine learning tool, termed ArtifactDetect*, to identify common types of artifacts and their root causes on WSIs.
Conference
USCAP 2025
View Poster