A recent article in CytoJournal on the Introduction of the Thin Prep Imaging System™ (TIS) illustrates a not-so-obvious use-case for VisualStrata™. During the evaluation of TIS cytotechnologists reviewed the image as well to see how well TIS (and the cytotechnologists!) were doing in interpreting images for detecting cervical cancer.
The detection results were promising. However, there is something still missing: The criteria used in interpreting the image. In other words what visual indicators led TIS and the ctyotechnologists to make a determination. More importantly, in what visual ways were the criteria different.
The image collaboration features of VisualStrata™ make it possible to automatically visually layer (e.g., pointers, lines, shapes) the images with the auto-detection criteria and at the same time provide the same image in an unobstructed way to the cytotechnolgogist. The result is a comparison of the visual detection crieteria between machine and human. These results would server not only to improve the automated and human detection processes, it would also provide valuable insight into the differences in detection.
