Image-assisted analysis of pathologist slide review as a means of informing optimal digital slide acquisition strategies

In digital pathology, the practice of scanning slides at a single resolution results in more data than is clinically useful. In contrast, using a light microscope, the pathologist adapts magnification according to specimen content. We developed a method of digitizing the analog microscope workflows of pathologists using a microscope camera and video mosaicking with the Scale Invariant Feature Transform (SIFT). Matching this digital record to the whole slide image generates a slide interaction map over time and magnification tiers. The results of the pilot studies suggest data size can be significantly reduced without diminishing the information content required for diagnosis.