Researchers in the United States studied the use of the Virchow foundational model for computational analysis of pathological reports to predict biomarkers and identify cells in various forms of cancer. This model was trained on a large dataset of whole-slide images and demonstrated high accuracy in detecting both common and rare cancers. The study highlighted the utility of foundational models in pathology, showing that the Virchow model was able to perform comparably or even better than clinical-grade models despite being trained on datasets with fewer tissue labels. This suggests the potential for foundational models to improve cancer diagnosis and treatment outcomes.
Source link
AI model Virchow outperforms clinical methods in cancer detection
