Resumen
In 2016 an interdisciplinary collaborative research team started an ambitious project to assist the diagnosis of tuberculosis (TB) in areas with over-burdened and poor healthcare infrastructure in the outskirts of Lima, capital of Peru. Currently we have implemented an integrated system which connects nurses, physicians and scientists, in a network of cooperation to identify and report suspicious cases of pulmonary TB. Our Electronic Mobile system (called ‘eRx’) is a Computer Automated Diagnosis (CAD) tool that can be used by local TB nurses allowing them to send X-rays along with additional prospect patient information, directly to be remotely evaluated by a physician. The physician accesses a web-based system using a computer or a tablet which allows him/her to evaluate the data coming from the distinct healthcare centers. The physician may also be assisted by a Beta version of an automatic diagnostic software which uses Deep Learning to identify the presence of lung abnormalities as well as the possible manifestations for TB. The goal of our research is to increase the speed at which a diagnosis can be made by providing a CAD tool to health care providers to improve quality and efficiency of TB diagnosis. To support future developments of the automatic diagnostic software, we rely on another web-based platform which contains over 10,000 X-rays images which are being annotated by a specialist who provides the regions where the TB manifestations can be found, which will be used later to train a Region-Based Convolution Neural Network. The software for the eRx system covers the obtaining of high-quality data, artificial intelligence support and mobile technologies directly in hands of the healthcare professionals, each software has already been validated individually, and will be evaluated together during a pilot study working directly with local healthcare centers, physicians and nurses. Our study has approval from the the IRB Boards at the University of Massachusetts Lowell IRB, and the School of Medicine at the Peruvian University Cayetano Heredia.
Idioma original | Inglés estadounidense |
---|---|
- | 100117 |
Publicación | Smart Health |
Volumen | 16 |
DOI | |
Estado | Indizado - may. 2020 |
Publicado de forma externa | Sí |
Nota bibliográfica
Publisher Copyright:© 2020 Elsevier Inc.