Healthcare
BioStack Raises $9.5 Million for Medical-AI Training Data
BioStack raised $9.5 million to create multimodal medical-AI datasets from more than 100 million de-identified patient records across four countries.
BioStack raised $9.5 million to turn hospital data into multimodal training datasets for medical artificial-intelligence development. The company works with hospitals in the United States, India, Brazil and Argentina and has access to more than 100 million de-identified patient records.
Its platform is designed to support development and evaluation of medical AI models. The data can also be used for clinical research, disease modeling, biomarker discovery, patient stratification and treatment-response analysis.
The company’s geographic footprint provides data from different patient populations and healthcare environments. De-identification is intended to allow the information to be used for model development while protecting patient identity.
The financing matters because data quality and diversity are central constraints in medical AI. Multimodal datasets spanning institutions and countries could improve model evaluation and help developers build tools that are more applicable across clinical settings.
Original source: Morningstar