Web-Based Algorithms Applied to Low-Dose CTs Could Automate Detection of COVID-19

Scientists are using machine learning in the fight against the coronavirus pandemic by creating algorithms that could be used to ensure a more precise classification of COVID-19 in the future.

The Technical University of Munich (TUM Munich, Germany) is starting a new research project that will focus on using computer tomography (CT) and machine learning to classify the COVID-19 pulmonary disease which is a novel viral pulmonary inflammation. The aim of the project “The Early Detection and Classification of COVID-19 Pneumonia by Means of Computer Tomography and Machine Learning” is to apply machine learning methods to low-dose CTs of COVID-19 patients in order to perform individual, automated detection, quantification and risk evaluation of the disease.

With low-dose CT of the lungs, not only can infections be detected, but doctors can also see the extent to which the lungs are affected, something which they cannot do with a standard COVID-19 laboratory test. Low-dose CTs require only a small amount of radiation. At the end of the project, the TUM researchers hope to develop web-based algorithms that could be rolled out to and used in hospitals. The Bavarian Research Foundation (BFS) is funding the project which will also involve Siemens Healthineers AG.

Related Links:
Technical University of Munich