Answers are grounded in this paper and cite their sources.
We developed a deep-learning system that detects colorectal cancer specimens with dMMR or MSI using H&E-stained slides; it detected tissues with dMMR with an AUROC of 0.96 in a large, international validation cohort. This system might be used for high-throughput, low-cost evaluation of colorectal tissue specimens.
No takes yet. Share an insight, caveat, or question.
Echle et al. (2020) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: