SPARK in digital pathology: what it does and why it matters to pathologists
A clear explanation of the SPARK study in Nature Medicine: what it does, what it takes as input, what it outputs, and why it matters to…
From a static image to a disease trajectory: what should pathologists ask of AI-based endpoints?
A professional reading of Beyond the Slide’s argument for moving from static histologic measurements to longitudinal, interpretable…
Roche’s acquisition of PathAI: what changes inside pathology laboratories?
A pathology-focused reading of Roche’s acquisition of PathAI, valued at up to $1.05 billion, and what it may mean for digital workflows and…
PRET from HKUST: a serious test of few-data AI in digital pathology
HKUST’s PRET system uses in-context learning to analyze whole-slide images from a small number of examples, with strong results and a clear…
PathAI and A.C.Camargo collaborate in Brazil: a practical test of AI in oncology pathology
A professional reading of PathAI's collaboration with A.C.Camargo in Brazil, and what it means for pathologists working with oncology…
Haiku: a trimodal model linking H&E, spatial proteomics, and clinical data
Haiku links H&E, mIF, and clinical text in a shared embedding space, with results in retrieval, survival prediction, and biomarker inference…
Virtual imaging of lymph nodes in breast cancer: where does the pathologist stand?
A study of D-FFOCT with deep learning for detecting lymph-node metastases in breast cancer, and what it means in practical terms for pathologists.
AI and HER2 assessment in biliary tract cancer: real value or just another number?
A new report discusses the use of AI to standardize HER2 assessment in biliary tract cancer, with a practical reading of what this means for…
Was pathology’s gold standard really solid?
A critical look at histology as an imperfect measurement model, and why digital pathology and AI need more precise endpoints.
A digital reading of pre-regulatory PD-L1 validation: where does the pathologist’s decision still matter?
A recent study of a deep-learning assistant for TPS scoring in PD-L1 testing for lung cancer, with a practical focus on the 1% and 50%…