System and a method thereof for automatic detection of epileptogenic focus in pharmacoresistant epilepsy

A method for detection of epileptogenic focus in subjects with pharmacoresistant epilepsy, comprising steps of:

• acquiring PET and MRI data from a medical hardware system comprising a PET detector and an MRI scanner;
• preprocessing said PET and MRI data on said medical hardware system with system processing tools;
• transferring said pre-processed PET and MRI data and associated metadata to a storage and communication module integrated with the medical hardware system;
• storing said pre-processed PET and MRI data to a computation module;
• accessing said pre-processed PET and MRI data including a FLAIR MRI image and a TI-weighted MRI image from said storage and communication module by a computation module;
• co-registering rigidly the pre-processed PET and MRI data including the FLAIR MRI image to the T1-weighted MRI image obtaining co-registered images;
• flipping said co-registered images left to right while maintaining anterior-posterior orientation obtaining flipped T1-weighted MRI, flipped FLAIR MRI and flipped PET images;
• performing multi spectral segmentation of the flipped T1-weighted MRI and the flipped FLAIR MRI images and the not flipped T1-weighted MRI and the not flipped FLAIR MRI images using the not flipped T1-weighted MRI and the not flipped FLAIR MRI images to create grey and white matter probability maps with the computing tool;
• creating an average symmetrical template using said computing tool for image registration through nonlinear co-registration of the grey and white matter of the flipped and not flipped MRI images;
• warping the flipped and not flipped PET and T1-weighted MRI images to the average symmetrical template using the corresponding flow fields of the not flipped PET and T1-weighted MRI images by flow field 1 and the flipped PET and T1-weighted MRI images by flow field 2 images thereby obtaining warped images corrected for interhemispheric structural asymmetry;
• warping inversely the warped flipped and warped not flipped PET and T1-weighted MRI images using flow field 1 causing warping of the warped not flipped PET and T1-weighted MRI images back to the native space and warping of the warped flipped PET and T1-weighted MRI images to anatomically overlap the grey and white matter of the not flipped PET and T1-weighted MRI images;
• smoothing said inversely warped PET images using a smoothing tool to improve the signal-to-noise ratio;
• calculating an asymmetry index image after inverse warping as per expression:
• AI=(flipped−not flipped)/max (flipped, not flipped) wherein the difference between the inversely warped flipped and not flipped PET images corresponds to the interhemispheric metabolic asymmetry;
• multiplying the grey matter tissue probability map to said AI image, restricting the analysis to the grey matter;
• smoothing-the grey matter-restricted AI images a second time;
• thresholding and clustering said smoothed AI images by converting said smoothed AI images to z-score AI images;
• and
• thresholding said z-score AI images at z>4, followed by clustering the region of peak intensity;
• thereby localizing and detecting epileptogenic focus in the pharmacoresistant epilepsy subjects.