Automated co-registration of prostate MRI data

An apparatus comprising:

• a non-transitory computer-readable storage medium storing a plurality of imaging series of magnetic resonance imaging (MRI) data for imaged tissue of a patient;
• and
• a processor coupled to the computer-readable storage medium;
• wherein the processor is configured to
• receive the imaging series of MRI data;
• identify a volume of interest (VOI) of each image of the imaging series of MRI data;
• compute, using a genetic algorithm applied to the VOIs, registration parameters for the VOIs of the plurality of imaging series through the maximization of mutual information of the VOIs;
• and
• register the VOIs of the plurality of imaging series using the computed registration parameters.

A non-transitory computer-readable storage medium storing instructions that, when executed, cause one or more processors to:

• receive images comprising magnetic resonance imaging (MRI) data for imaged tissue of a patient;
• correct, by the computing device, signal inhomogeneities in one or more volumes of interest (VOI) for the images;
• compute, using a genetic algorithm applied to the VOIs, registration parameters for the corrected VOIs through the maximization of mutual information of the corrected VOIs;
• and
• register, by the computing device, the VOIs using the computed registration parameters;
• apply a multiparametric model to the registered VOIs to generate a Composite Biomarker Score (CBS) map that maps the imaged tissue using a respective CBS value for one or more voxels of the imaged tissue;
• determine locations of the CBS map that correspond to CBS values that satisfy a threshold;
• generate a visual indication of the determined locations of the CBS map as regions of the imaged tissue predicted to include cancer;
• and
• output the visual indication.

A method comprising:

• receiving, by a computing device, a plurality of imaging series of magnetic resonance imaging (MM) data for imaged tissue of a patient;
• identifying, by the computing device, a volume of interest (VOI) of each image of the imaging series of the MM data;
• correcting, by the computing device, distortions that affect intensity values in each of the VOIs caused by signal inhomogeneities due to the presence of the endorectal coil (ERC) in each of the VOIs to generate corrected VOIs, wherein correcting the distortions comprises:
• modeling the ERC as two parallel wires perpendicular to an axial imaging plane of the imaged tissue;
• calculating a differential contribution of each wire segment to a magnetic field in the xy-plane for each of the VOIs;
• summing the differential contributions to obtain a coil sensitivity profile for each of the VOIs;
• registering the sensitivity profile for each of the VOIs to the respective imaging series using a minimum variation metric as a similarity metric;
• and
• dividing the intensity values of the imaging series by the respective registered coil sensitivity profile on a pixel-wise basis;
• computing, by the computing device, registration parameters for the corrected VOIs through a maximization of mutual information of the corrected VOIs;
• and
• registering, by the computing device, the corrected VOIs of the images of the plurality of imaging series using the computed registration parameters.