Virtual histological staining of tissue from a single label-free autofluorescence image using deep learning

Yair Rivenson, Hongda Wang, Zhensong Wei, Kevin de Haan, W. Dean Wallace, Aydogan Ozcan

Keywords: Deep learning, Histology, Digital pathology

TL;DR: We demonstrate virtual staining of unlabeled tissues using single autofluorescent images.

Abstract: Histochemical staining of tissue samples is required for the diagnosis of many diseases, including cancer; however, the staining process is often time consuming, slow, costly, and does not support tissue preservation for advanced molecular analysis of the sample. Recently, we presented a deep learning framework that can perform virtual histochemical staining, in silico, of unlabeled tissue sections using a single autofluorescence image, emanating from the tissues endogenous fluorophores. We validated the success of this technique through a direct comparison between the virtually and histochemically stainedslides, as well as by a blind study performed by a panel of board-certified pathologists.

Code Of Conduct: I have read and accept the code of conduct.

Remove If Rejected: Remove submission from public view if paper is rejected.

Link: https://www.nature.com/articles/s41551-019-0362-y