Description of Collaborative Activity: |
This project has its origin in an ASPE-funded PCOR project: Training Data for Machine Learning to Enhance Patient-Centered Outcomes Research Data Infrastructure. The project acquires, curates, and uses high-quality training data sets to predict tuberculosis (TB) drug resistance, in collaboration with NIAID and its TB Portals program. TB drug resistance has been concerning because it is difficult to diagnose in a patient during the early stages and it complicates and prolongs treatment considerably. The idea of the project is to develop new ways of detecting drug-resistance in chest radiographs by using deep learning in combination with large image repositories, clinical data, and genomic information of the bacteria for drug-sensitive and drug-resistant tuberculosis. Researchers use this data to train machine learning models that can detect signs of TB drug resistance.; |
