EXTRACTION OF HIDDEN INFORMATION FROM MEDICAL IMAGES
The invention consists in a new diagnostic method for processing images to make apparent the information hidden in the images. For example, the method allows the better characterization of the coronary artery wall morphology by re-analyzing angiography images. The method , taking the advantage of artificial intelligence approaches, is based on an image processing algorithm able to re-analyze the network of pixels of an image to provide additional information not visible in the original image.
Presently, diagnostic and therapeutic decision processes are supported by intravascular analysis by ultrasound (named IVUS) wherein the insertion of an ultrasound probe in the artery allows the straightforward geometry examination of the walls of the coronary artery. Such a detecting method has the drawback of being invasive, dangerous for the patient and expensive thus limiting its use in the coronary artery characterization. Furthermore, despite the IVUS technique is widely used it does not fully solve the problem of the lack of information about the walls of the whole coronary artery tree. The new method which allows to detect both the arterial lumen and the coronary wall morphology without further diagnostic procedure, comprises two different softwares: a prorotype software able to select the algorithm useful to catch the hidden images
- Diagnostic test to be used before or in place of coronary IVUS to detect or exclude the presence of plaques invisible at the coronary CT angiography because not protruding into the arterial lumen
- Tests performed on routine images show a strict correspondence between arterial wall detected by IVUS and images reprocessed by the invention
- Does NOT require an invasive procedures which could be dangerous or cause endothelial damages;
- Able to detect hidden images of very small arteries unable to be investigated by echocardiographic intravascular probe
- Allows the evaluation of the coronary artery lumen and/or the wall silhouette and thus the thickness of the coronary artery walls even to unskilled operators.