Computer Aided Diagnosis

Name: Nico Karssemeijer
Organisation: Diagnostic Image Analysis Group & Institute for Computing and Information Sciences, Radboudumc
Abstract:

New developments in machine learning enable the development of medical image interpretation systems that match the sensitivity and specificity of human readers, and that may even outperform them in some applications. This makes computer aided diagnosis a promising technology to improve the quality of healthcare and to decrease cost. In particular screening applications are suitable for automation.

Computer aided detection (CAD) systems are already widely used in breast cancer screening to help radiologists with the reading of mammograms. However, currently every mammogram is still read by two radiologists to reduce errors. Because the quality of CAD systems is increasing it is expected that in the near future CAD systems will replace one of the readers. Other important applications are CAD in lung cancer and tuberculosis screening, and digital pathology.

Picture of the speaker