GRACE, a gastric-specific pathology foundation model trained on multicenter HE-stained slides, outperforms pancancer models on 28 tasks and improves pathologist accuracy, speed, and agreement in a reader study while enabling case triaging.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CV 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
A Pathology Foundation Model for Gastric Cancer with Real-World Validation
GRACE, a gastric-specific pathology foundation model trained on multicenter HE-stained slides, outperforms pancancer models on 28 tasks and improves pathologist accuracy, speed, and agreement in a reader study while enabling case triaging.