Andrew Su

This study uses spatial single-cell analysis to identify 10 cell phenotypes in stage III colorectal cancers and links specific spatial features to recurrence risk.

The paper identifies a 26-gene expression signature in high-risk gastric intestinal metaplasia (GIM) that localizes to intestinal stem-like cells and is associated with increased gastric cancer risk.

This paper introduces HEMnet, a deep learning model that transfers molecular labels from immunohistochemistry to H&E images to identify cancer cells without manual annotation.

SpaCell integrates spatial gene expression and tissue morphology to improve cell type identification and image prediction.