Follicular lymphoma is a common type of cancer of the lymph system, affecting mainly older adults. We work to aid the grading process of the disease based on morphological characteristics in order to determine the appropriate treatment for a patient.
Follicular lymphoma cases are stratified into three grades (I, II, and III). A higher grade indicates a more aggressive disease, and therefore requires a more intense treatment plan. Therefore, in order to ensure that a patient gets the appropriate treatment, it is crucial for a pathologist to be able to correctly identify the grade of a follicular lymphoma image. We have developed systems that aid pathologists in the grading process. Currently, we are able to segment the follicular regions in an image, then we are able to further segment the follicular regions into their cytological components to detect centroblasts, and finally we can determine the grade of a follicular lymphoma image by the number of centroblasts detected in a follicle region. The overall process is outlined in the flowchart below, followed by more specific diagrams on the follicle detection and centroblast detection processes.
Figure 1 : Workflow
Figure 2 : Detection results with IHC-stained images
Figure 3 : Detection results with H and E-stained images
Figure 4 : Follicle detection results with H and E-stained images
Figure 5 : Centroblast Detection