New study demonstrates that CAD4TB outperforms human radiologists in detecting TB

The study ‘Using artificial intelligence to read chest radiographs for tuberculosis detection: A multi-site evaluation of the diagnostic accuracy of three deep learning systems’ was carried out. This study demonstrates that CAD4TB was better than human readers in detecting bacteriologically confirmed TB, which could play an important role, especially in settings with a shortage of trained human readers. Furthermore, the study showed that the latest version of CAD4TB showed improved performance.

This study demonstrates that these DL systems have the potential to increase capacity and aid TB diagnosis, especially in settings with a shortage of trained human readers which have been noted as shortcomings in CXR use.