Slim and mobile Dual-Energy Technology Computed Tomography (Smart-DETeCT)
Slim and mobile Dual-Energy Technology Computed Tomography (Smart-DETeCT)

Slim and mobile Dual-Energy Technology Computed Tomography (Smart-DETeCT)

Novel type of compact CT scanner with artificial intelligence for rapid detection of lung diseases

Periode
-
Looptijd
12 months
Deel van call / Programma
/
Projectpartners
Maastricht University
Maastro
medPhoton

Austrian medical imaging company medPhoton develops a novel type of compact CT scanner that can be rapidly deployed for screening of covid-19 and other disease patterns. Specialists in quantitative CT imaging Maastro Clinic and University Maastricht partner with medPhoton to develop artificial intelligence (AI) software to rapidly analyze CT images for disease patterns.

The recent covid-19 pandemic has overwhelmed medical resources with more than 25 million of infections worldwide, and approaching one million of fatal cases (Sept 2020). The expected number of permanently affected patients with concomitant reduced economic participation in society, is enormous. Late diagnosis can result in serious disease and death. Early detection and intervention is essential to select the optimal treatment strategy. Besides the currently available biochemical tests, which take relatively long to yield results, a different method to detect lung disease is now being developed based on CT scanning.

Our approach is to develop a combination of a novel type of CT scanner that can be easily deployed worldwide. The scanner makes 2D and 3D images, which would be read out, not by a human operator, but by fast automated software. This software would be based on AI software, which we would develop. The software will be able to distinguish different covid-19 disease patterns in lungs. The system can be used for fast triage upon first contact with the patient, and also for follow-up during treatment and convalescence. The CT images will reveal much more relevant information for deciding on the treatment strategy, than the current simple biochemical tests. 

The deliverables of the project are (1) compact rapid cone beam CT scanner, capable of single and dual-energy CT scanning, (2) AI methods (based on Deep Learning) trained on large databases of available covid CT images to detect disease patterns automatically in 2D and 3D x-ray images. These methods were developed by MAASTRO-UM during the project. We showed that the 2D AI methods with the current quality of projection images needs to improve further to implement clinically. The 3D methods can be used successfully.  

Thank you! Your submission has been received!
Oops! Something went wrong while submitting the form.