Boosting Digital Pathology with Machine Learning
Drug safety assessments are an essential part of treatment evaluation prior to any clinical application. Every year, hundreds of pre-clinical studies are performed to discard dangerous compounds and save precious development time. The safety assessment process involves evaluating whether the drug under development can cause toxic reactions on animal tissues, through the analysis of thousands of animal tissue slides. The process is extremely time intensive, and requires repeated expert validations.
Integrating ML models to automate parts of the screening therefore holds potential for significant time gains, and could enable experts to focus on more complex cases.
Taking advantage of recent advances in computer vision, our data engineers developed a solution which could help Roche automation the lesion detection process.
Watch the video below for a presentation of the project by our COO, Matteo Togninalli and Vanessa Schumbacher, Group Head – Tissue Biomarker and Digital Pathology, at Roche.
Learn about how the solution works, and it can help provide better patient care.