OSU Department of Biomedical Informatics

Follicular Lymphoma

Manual grading of Follicular Lymphoma is time consuming and subject to variability. One solution is to develop a computerized system capable of histological grading of FL. Two systems were created, one that used H&E stained images, and another that used immunostained images. The H&E system consists of a number of image analysis steps, including feature measures extracted from cell spatial density, color feature clustering and morphological operations, concluding with identification of follicular boundaries. The true positive and false positive rates were 74.40%±13.48% and 14.20%±5.28%, respectively. The immunostained system offered a true positive rate of 97.08%±3.54% and a false positive rate of 19.24%±9.28%. Both systems are promising, and in the future may allow for efficient analysis of follicular lymphoma.



In this phase the boundaries of follicular lymphoma are identified.

Project Publications

[edit this page]