Pathology departments worldwide are currently converting from traditional to digital microscopy, which produces 2-4 terabytes of new data per department every day. While the availability of large datasets has spurred the use of machine learning (ML) and artificial intelligence (AI) in clinical settings, pathologists have been slow to adopt these technologies.
The Inspiration Behind the Innovation
Current commercial and scientific AI solutions designed to aid in pathology-related image analysis are task specific. This makes them of limited use to pathologists, who are generalists and do not see value in a multitude of task-specific apps. In order to successfully bring an AI solution to market for pathologists, it’s imperative to involve them in the development of its ML models.
Pathology AI aims to address the future shortage of pathologists by simplifying and speeding up the diagnostic process using digital image analysis. In this project, machine vision observes the same area as the pathologist, accumulating “experience” based on the pathologist’s input.
The goal is for pathology AI to succeed in quantifying the size and number of cells to support diagnostic safety and efficiency, which will positively affect waiting times and safety in tissue sample diagnostics for patients across the country.
Thomas Lindkær Jensen: Senior Consultant; Pathology Department, Danish National Hospital, Copenhagen.
Rasmus Hartvig: Software Engineer