Systems and methods for computer aided diagnosis and decision support in whole-body imaging

Abstract

A system for providing automatic diagnosis and decision support includes: a medical image database; generative learning and modeling modules that build distributional appearance models and spatial relational models of organs or structures using images from the medical image database; a statistical whole-body atlas that includes one or more distributional appearance models and spatial relational models of organs or structure, in one or more whole-body imaging modalities, built by the generative learning and modeling modules; and discriminative learning and modeling modules that build two-class or multi-class classifiers for performing at least one of organ, structure or disease detection or segmentation.

1 . A system for providing automatic diagnosis and decision support in whole-body imaging, comprising: 
 a medical image database;    generative learning and modeling modules that build distributional appearance models and spatial relational models of organs or structures using images from the medical image database;    a statistical whole-body atlas that includes one or more distributional appearance models and spatial relational models of organs or structure, in one or more whole-body imaging modalities, built by the generative learning and modeling modules; and    discriminative learning and modeling modules that build two-class or multi-class classifiers for performing at least one of organ, structure or disease detection or segmentation.    
   
   
       2 . The system of  claim 1 , wherein the generative learning and modeling modules extract statistics and find clusters using images from the medical image database.  
   
   
       3 . The method of  claim 2 , wherein the statistics comprise at least one of statistics on voxel intensities, statistics on global and local shape deformations, or statistics on joint articulations.  
   
   
       4 . The method of  claim 1 , wherein the statistical whole-body atlas comprises one or more three-dimensional canonical whole-body scans, each with associated statistics on voxel intensities, statistics on global and local shape deformations, and statistics on joint articulations.  
   
   
       5 . The system of  claim 1 , wherein the discriminative learning and modeling modules formulate organ detection and segmentation as discriminative learning and at least one of design or select discriminative features.  
   
   
       6 . The system of  claim 1 , further comprising software modules to output at least one of a location of one or more landmark points or a location of one or more organs, using images from the medical image database.  
   
   
       7 . The system of  claim 6 , wherein landmark points comprise at least one of an upper corner of a left lung, an upper corner of a right lung, a center of a left kidney or a center of a right kidney.  
   
   
       8 . A method for providing automatic diagnosis and decision support in whole-body imaging, comprising: 
 using whole-body imaging to obtain a first set of image data of a patient;    fitting a statistical whole-body atlas using the first set of image data, wherein the statistical whole-body atlas includes at least one of statistics on voxel intensities, statistics on global and local shape deformations, or statistics on joint articulations; and    using the statistical whole-body atlas to characterize pathological findings in terms of a diagnosis.    
   
   
       9 . The method of  claim 8 , wherein the first set of image data is obtained using one or more imaging modalities.  
   
   
       10 . The method of  claim 8 , wherein the first set of image data is obtained using whole-body positron emission tomography (PET) and one or more imaging modalities other than whole-body PET.  
   
   
       11 . The method of  claim 8 , wherein the statistical whole-body atlas comprises at least one of distributional appearance models or spatial relational models of organs or structures in one or more whole-body imaging modalities.  
   
   
       12 . The method of  claim 8 , wherein the first set of image data includes PET data, and wherein fitting the statistical whole-body atlas comprises automatically outlining selected regions of interest with pathological tracer uptakes while discounting physiological uptakes in the PET data.  
   
   
       13 . The method of  claim 8 , wherein the first set of image data further includes image data acquired by one or more imaging modalities other than PET, and wherein using the statistical whole-body atlas to characterize pathological findings in terms of a diagnosis comprises detecting anatomical or functional abnormalities of the whole body or body parts using the first set of image data.  
   
   
       14 . The method of  claim 8 , wherein using the statistical whole-body atlas to characterize pathological findings in terms of a diagnosis comprises automatically outlining selected regions of interest in the first set of image data and automatically extracting candidate features of interest from the selected regions of interest.  
   
   
       15 . The method of  claim 14 , wherein using the statistical whole-body atlas to characterize pathological findings in terms of a diagnosis further comprises automatically contouring and characterizing each of the candidate features of interest.  
   
   
       16 . The method of  claim 15 , wherein automatic contouring and characterizing is based on at least one of a standard uptake value or brain uptake normalized data.  
   
   
       17 . The method of  claim 8 , wherein the first set of image data comprises at least one of 2-D image data, 3-D image data or higher-dimensional image data.  
   
   
       18 . The method of  claim 8  further comprising: 
 using whole-body imaging to obtain a second set of image data of the patient; and    updating pathological findings based on the statistical whole-body atlas using the second set of image data.    
   
   
       19 . The method of  claim 18 , wherein updating pathological findings based on the statistical whole-body atlas comprises encoding anatomical or functional variations of the whole body or body parts using the second set of image data.  
   
   
       20 . The method of  claim 18  further comprising: 
 using whole-body imaging to obtain a third set of image data of the patient; and    updating pathological findings based on the statistical whole-body atlas using the third set of image data.    
   
   
       21 . The method of  claim 20 , wherein updating pathological findings based on the statistical whole-body atlas comprises encoding anatomical or functional variations of the whole body or body parts using the third set of image data.  
   
   
       22 . A method for providing automatic diagnosis and decision support in whole body imaging, comprising: 
 detecting and segmenting one or more selected regions of interest in whole-body images;    detecting one or more abnormalities by automatically interpreting whole-body images of the selected regions of interest for pathological findings; and    characterizing the pathological findings in terms of a diagnosis.    
   
   
       23 . The method of  claim 22 , wherein segmenting the selected regions of interest is accomplished using at least one of bounding boxes, centroid locations, bounding surfaces, or a bounding mask.  
   
   
       24 . The method of  claim 22 , further comprising detecting hotspots in the selected regions of interest.  
   
   
       25 . The method of  claim 24 , wherein the hotspots are segmented using organ-specific thresholds.  
   
   
       26 . The method of  claim 22 , wherein the whole-body images are obtained using at least one of whole-body PET, CT, MRI, SPECT, PET/CT, SPECT/CT, or PET/MRI.  
   
   
       27 . The method of  claim 22 , wherein, when a longitudinal data is available, performing clinical analysis of the longitudinal data and outputting changes in a clinically meaningful way.  
   
   
       28 . A program storage device readable by machine, tangibly embodying a program of instructions executable by the machine to perform method steps for providing automatic diagnosis and decision support, the method steps comprising: 
 using whole-body imaging to obtain a first set of image data of a patient;    fitting a statistical whole-body atlas using the first set of image data, wherein the statistical whole-body atlas includes at least one of statistics on voxel intensities, statistics on global and local shape deformations, or statistics on joint articulations; and    using the statistical whole-body atlas to characterize pathological findings in terms of a diagnosis.    
   
   
       29 . A program storage device readable by machine, tangibly embodying a program of instructions executable by the machine to perform method steps for providing automatic diagnosis and decision support, the method steps comprising: 
 detecting and segmenting one or more selected regions of interest in whole-body images;    detecting one or more abnormalities by automatically interpreting whole-body images of the selected regions of interest for pathological findings; and    characterizing the pathological findings in terms of a diagnosis.