Multi-Parametric Classification Images of Cardiac LV RWMA in Nuclear Medicine

It is often difficult to extract information of interest from multiple radiological images as they may contain an overwhelming amount of information (and noise) that is only remotely or not at all related to the diagnostic information we are seeking. This problem led to the development of parametric images, representing 2-D spatial distributions of a single physical (or statistical) variable describing certain features of physiological functions (hence “functional images”). However, the current parametric images do not suggest directly how the input image data should be interpreted diagnostically. The processes and algorithms generating the current parametric images do not use information that is available outside the input image data of the studied single case (for example, diagnostically labeled image data of previous patients and the frequency of their occurrence in the patient population) and other diagnostically important information (results of diagnostic tests, age, sex, medical history, ... ).