DLBCL-Morph: Morphological features computed using deep learning for an annotated digital DLBCL image set

Abstract Diffuse Large B-Cell Lymphoma (DLBCL) is the most common non-Hodgkin lymphoma. Though histologically DLBCL shows varying morphologies, no morphologic features have been consistently demonstrated to correlate with prognosis. We present a analysis of histology sections from 209 cases associated clinical and cytogenetic data. Duplicate tissue core were arranged in microarrays (TMAs), replicate stained H&E immunohistochemical stains for CD10, BCL6, MUM1, BCL2, MYC. The TMAs are accompanied by pathologist-annotated regions-of-interest (ROIs) that identify areas representative DLBCL. used deep learning model segment all tumor nuclei ROIs, computed several geometric each segmented nucleus. fit Cox proportional hazards demonstrate utility these predicting survival outcome, found it achieved C-index (95% CI) 0.635 (0.574,0.691). Our finding suggests prognostic importance, should be validated prospective studies.