Synthetic visualization and quantification of perfusion-related tissue viability

• US 8,532,353 B2
• Filed: 11/23/2011
• Issued: 09/10/2013
• Est. Priority Date: 11/23/2011
• Status: Active Grant

• Current Assignee: Canon Medical Systems Corporation (Canon Inc.)
• Original Assignee: Vital Images Incorporated (Canon Inc.)
• Inventors: Salazar-Ferrer, Pascal, Jackson, Bennett W., Atev, Stefan E.
• Primary Examiner(s): Seth, Manav
• Application Number:
  US13/304,103
• Publication Number:
  US 20130131507A1
• Time in Patent Office: 657 Days
• Field of Search: None
• US Class Current: 382/128
• CPC Class Codes: A61B 6/032 Transmission computed tomography [CT]
• CPC Class Codes: A61B 6/463 characterised by displaying multiple images or images and diagnostic data on one display
• CPC Class Codes: A61B 6/486 involving generating temporal series of image data
• CPC Class Codes: A61B 6/507 for determination of haemodynamic parameters, e.g. perfusion CT
• CPC Class Codes: A61B 6/5217 extracting a diagnostic or physiological parameter from medical diagnostic data
• CPC Class Codes: G01R 33/56366 Perfusion imaging
• CPC Class Codes: G06T 2200/04 involving 3D image data
• CPC Class Codes: G06T 2207/10081 Computed x-ray tomography [CT]
• CPC Class Codes: G06T 2207/10088 Magnetic resonance imaging [MRI]
• CPC Class Codes: G06T 2207/30016 Brain
• CPC Class Codes: G06T 7/0014 using an image reference approach
• CPC Class Codes: G16H 50/30 for calculating health indices; for individual health risk assessment

Abstract

A computing system and method for generating, displaying and manipulating synthetic 2D and 3D brain tissue viability images and associated metrics from multiple perfusion maps derived from CT or other imaging system input image datasets. Stored linguistic tissue classification rules characterizing infarct, ischemic and/or normal or other tissue classes as a function of one or more perfusion parameters are used. The perfusion maps are normalized to normal brain tissue regions. Elements of the input image dataset are classified into one of the tissue classes as a function of the normalized perfusion maps and the classification rules. The classification includes ranking each of the plurality of tissue classes for elements of the input image dataset, and assigning one of the classes to elements of the dataset as a function of the ranks.

Claims

1. A method for operating a computing system to generate a brain tissue viability map from an input image dataset acquired from a patient'"'"'s brain by a medical imaging system, comprising:

computing, for each of one or more different perfusion parameters, an imaged perfusion map of perfusion values as a function of the input image dataset;

identifying a normal brain reference region as a function of at least one of the one or more imaged perfusion maps;

identifying, for each perfusion parameter, a normal perfusion value as a function of perfusion values in the normal brain reference region;

computing, for each perfusion parameter, a normalized perfusion map as a function of the imaged perfusion map and the normal perfusion value;

storing linguistic tissue classification rules characterizing each of a plurality of tissue viability classes as a function of one or more of the one or more perfusion parameters;

classifying elements of the input image dataset into one of the plurality of tissue viability classes as a function of the one or more normalized perfusion maps and the tissue classification rules; and

generating a brain tissue viability map as a function of the classified elements of the input image dataset.

2. The method of claim 1 wherein the one or more different perfusion parameters are from a set of perfusion parameters including regional cerebral blood flow (rCBF), cerebral blood volume (CBV), mean transit time (MTT), time-to-peak (TTP), perfusion weighted imaging (PWI) and permeability.

3. The method of claim 2 wherein the plurality of tissue viability classes are from a set classes including infarct tissue, ischemic tissue, oligemic tissue and normal perfused tissue.

4. The method of claim 1 wherein identifying the normal brain reference region includes:

comparing left side brain perfusion values to right side brain perfusion values of at least one of the one or more imaged perfusion maps to determine a side of the brain with a region of interest; and

identifying a side of the brain contralateral to the side with the region of interest as the normal brain reference region.

5. The method of claim 4 wherein identifying the normal brain reference region further includes identifying a middle cerebral artery (MCA) region.

6. The method of claim 5 wherein identifying the normal brain reference region includes identifying the normal brain reference region as a function of a perfusion map demonstrating a high sensitivity to abnormal brain tissue.

7. The method of claim 1 wherein identifying a normal perfusion value includes computing a statistical representation of the perfusion values in the normal brain reference region.

8. The method of claim 7 wherein computing a statistical representation includes computing a central tendency and optionally a dispersion of the perfusion values in the normal brain reference region.

9. The method of claim 8 wherein computing the central tendency and optionally a dispersion includes computing a mean and optionally a standard deviation of the perfusion values in the normal brain reference region.

10. The method of claim 7 wherein computing a normalized perfusion map includes computing a normalized perfusion map of elements representing statistical distances between elements of the imaged perfusion map and the normal perfusion value as a function of the statistical representation of the normal perfusion value.

11. The method of claim 7 wherein computing a normalized perfusion map includes computing a normalized perfusion map of elements representing distances between the elements of the imaged perfusion map and the normal perfusion value.

12. The method of claim 1 wherein:

storing tissue classification rules includes storing rules characterizing each of the tissue viability classes as a function of perfusion parameters; and

classifying elements of the input image data set includes;

ranking each of the plurality of tissue classes for voxel elements of the input image dataset; and

assigning one of the plurality of tissue classes to the voxel elements of the input dataset as a function of the ranks.

13. The method of claim 12 wherein storing tissue classification rules includes storing rules characterizing tissue viability classes as a function of relative values of perfusion parameters.

14. The method of claim 12 and further including computing a confidence index map representing a confidence value of elements of the brain tissue viability map as a function of the rankings.

15. The method of claim 14 and further including displaying a tissue viability image and confidence index as a function of the tissue viability map and the confidence index map.

16. The method of claim 1 and further including computing a confidence index map representing a confidence value of elements of the brain tissue viability map.

17. The method of claim 16 and further including displaying a tissue viability image and confidence index as a function of the tissue viability map and the confidence index map.

18. The method of claim 1 and further including smoothing the tissue viability map.

19. The method of claim 18 and further including displaying a tissue viability image as a function of the smoothened tissue viability map.

20. The method of claim 18 wherein smoothing the tissue viability map includes:

comparing classifications of adjacent elements of the tissue viability map; and

reclassifying elements of the tissue viability map as a function of the comparison.

21. The method of claim 1 wherein classifying elements of the input image dataset further includes tuning the classifications of the elements as a function of a user tuning input.

22. The method of claim 21 wherein classifying elements of the input image dataset includes weighting perfusion parameters as a function of the user tuning input.

23. The method of claim 21 wherein tuning the classifications of the elements includes tuning classifications between normal classifications and a group including infarct and ischemic classifications as a function of a user normal/abnormal weight control input.

24. The method of claim 21 wherein tuning the classifications of the elements includes tuning classifications between infarct classifications and ischemic classifications as a function of a user abnormal tissue core/penumbra weight control input.

25. The method of claim 21 and further including displaying a tissue viability image as a function of the tissue viability map representative of tuned classifications.

26. The method of claim 1 wherein:

storing tissue classification rules includes storing rules based on relative perfusion parameter values; and

classifying elements of the input image dataset includes fuzzy logic classifying the elements as a function of the one or more normalized perfusion maps and the tissue classification rules.

27. The method of claim 1 and further including displaying a tissue viability image as a function of the brain tissue viability map.

28. The method of claim 1 and further including pre-segmenting regions of interest from the one or more imaged perfusion maps, wherein the regions of interest are selected from a set including ventricles and vessels.

29. The method of claim 1 wherein:

identifying a normal perfusion value as a function of perfusion values in the normal brain reference region includes identifying perfusion values for white matter and gray matter;

computing a normalized perfusion map includes computing normalized perfusion maps for white matter and gray matter;

storing linguistic tissue classification rules includes storing white matter-specific rules and gray matter-specific rules; and

classifying elements of the input image dataset includes classifying the elements as a function of the white matter and gray matter perfusion maps and the white matter-specific rules and gray matter-specific rules.

30. A method for operating a computing system to generate and display a brain tissue viability map from an input image dataset acquired from a patient'"'"'s brain by a medical imaging system, comprising:

computing, for each of two or more different perfusion parameters, an imaged perfusion map of perfusion values as a function of the input image dataset;

identifying a normal brain reference region as a function of at least one of the imaged perfusion maps;

identifying, for each perfusion parameter, a normal perfusion value as a function of perfusion values in the normal brain reference region;

computing, for each perfusion parameter, a normalized perfusion map as a function of the imaged perfusion map and the normal perfusion value;

storing linguistic tissue classification rules characterizing each of a plurality of tissue viability classes as a function of one or more of the perfusion parameters;

classifying elements of the input image dataset into one of the plurality of tissue viability classes as a function of the normalized perfusion maps and the tissue classification rules;

including;

ranking each of the plurality of tissue classes for elements of the input image dataset; and

assigning one of the plurality of tissue classes to elements of the input image dataset as a function of the ranks;

generating a brain tissue viability map as a function of the classified elements of the input image dataset; and

displaying a tissue viability image as a function of the brain tissue viability map.

31. The method of claim 30 and further including:

providing a user tuning control; and

tuning the classifications of the elements as a function of a user input from the user tuning control when classifying the elements.

32. The method of claim 31 and further including:

computing a confidence index representing a confidence value of elements of the tissue viability map as a function of the rankings; and

displaying a tissue viability image and confidence index as a function of the tissue viability map and the confidence index map.

33. The method of claim 32 wherein:

identifying a normal perfusion value as a function of perfusion values in the normal brain reference region includes identifying perfusion values for white matter and gray matter;

computing a normalized perfusion map includes computing normalized perfusion maps for white matter and gray matter;

storing linguistic tissue classification rules includes storing white matter-specific rules and gray matter-specific rules; and

classifying elements of the input image dataset includes classifying the elements as a function of the white matter and gray matter perfusion maps and the white matter-specific rules and gray matter-specific rules.

34. A computing system for generating and displaying a brain tissue viability map from an input image dataset acquired from a patient'"'"'s brain by a medical imaging system, comprising:

a data store for storing information representative of;

a plurality of perfusion parameter mapping algorithms; and

linguistic tissue classification rules characterizing each of a plurality of tissue viability classes as a function of one or more perfusion parameters;

a processor coupled to the data store, wherein the processor;

computes, for each of two or more different perfusion parameters as a function of the perfusion parameter mapping algorithms, an imaged perfusion map of perfusion values as a function of the input image dataset;

identifies a normal brain reference region as a function of at least one of the imaged perfusion maps;

identifies, for each perfusion parameter, a normal perfusion value as a function of perfusion values in the normal brain reference region;

computes, for each perfusion parameter, a normalized perfusion map as a function of the imaged perfusion map and the normal perfusion value;

classifies elements of the input image dataset into one of the plurality of tissue viability classes as a function of the normalized perfusion maps and the tissue classification rules;

including;

ranking each of the plurality of tissue classes for elements of the input image dataset; and

assigning one of the plurality of tissue classes to elements of the input image dataset as a function of the ranks; and

generates a brain tissue viability map as a function of the classified elements of the input image dataset; and

a display to present a tissue viability image as a function of the brain tissue viability map.

35. The computing system of claim 34 wherein the processor:

causes a user tuning control to be presented on the display; and

tunes the classifications of the elements as a function of a user input from the user tuning control when classifying the elements.

36. The computing system of claim 35 wherein:

the processor computes a confidence index representing a confidence value of elements of the tissue viability map as a function of the rankings; and

the display presents a tissue viability image and confidence index as a function of the tissue viability map and the confidence index map.

37. The computing system of claim 36 wherein:

the data store stores tissue classification rules including white mater-specific rules and gray matter specific rules; and

the processor;

identifies normal perfusion values for white matter and gray matter;

computes normalized perfusion maps for white matter and gray matter; and

classifies elements as a function of the white matter perfusion maps and gray matter perfusion maps and the white matter-specific rules and gray matter-specific rules.