System and method for detection and display of diseases and abnormalities using confidence imaging

• US 7,599,542 B2
• Filed: 07/28/2005
• Issued: 10/06/2009
• Est. Priority Date: 04/08/2005
• Status: Active Grant

• Current Assignee: 361º Systems Incorporated
• Original Assignee: 361º Systems Incorporated
• Inventors: Carruthers, William Benjamin III, Brockway, John Philip
• Primary Examiner(s): Dang; Duy M
• Application Number:
  US11/191,768
• Publication Number:
  US 20060228015A1
• Time in Patent Office: 1,531 Days
• Field of Search: 382/128, 382132-133, 600424-427
• US Class Current: 382/132
• CPC Class Codes: G06T 2207/30068 Mammography; Breast
• CPC Class Codes: G06T 7/0012 Biomedical image inspection

Abstract

The present invention provides a method, apparatus, and computer instructions for improved detection and display of abnormal regions in mammals such as tumors, lesions, and other abnormalities (collectively referred to as abnormalities). In a preferred embodiment, a system and process is disclosed for classification of image points in the spatial dimensions and subsequent segmentation and classification of regions using morphological descriptors that operate in up to three spatial dimensions. Additionally, mapping of a set of classification images to color and opacity parameters is provided for display purposes. After image data is captured and readied for processing, each spatial point in the image is evaluated against predetermined intensity-time parameters. The resulting intensity-time confidence image is then processed to identify distinct regions within the image, and evaluate morphological characteristics of the identified regions using predetermined morphology templates/parameters. A confidence value is determined for each region, and this value is applied to the intensity-time value for each spatial point. The resulting output is a confidence image for the patient'"'"'s region of interest, that can be used to detect different abnormalities, and display them in a conveniently manipulatable manner so a medical service provider can better understand the abnormality and take more directed actions (e.g., refined procedures) to remedy it as appropriate. The computational system disclosed can thoroughly and automatically detect these temporal patterns, as well as morphological patterns, and not only marks them for visual inspection with a confidence level, but also identifies the type or kind of cancer with an assigned probability, giving accurate indications of the extent of the cancer.

Claims

1. A method for processing an image of a region of interest in a person or other mammal to detect abnormalities, comprising:

(a) receiving the image having spatial information of the region of interest taken at plural times by capturing the image via a magnetic resonance imaging (MRI) system and processing the image to enhance detectability of the features;

(b) processing the image using at least one predetermined feature template, morphology parameters, and intensity parameters varying over the plural times, to determine a confidence image, wherein each feature template represents at least one predetermined detectable feature for the region of interest and the confidence image represents a likelihood that the at least one feature is present in the region of interest; and

(c) outputting the confidence image for further processing and detection of abnormal features.

2. The method of claim 1, wherein step (b) comprises processing the image, at least one template and intensity parameters to determine a composite intensity value for each spatial point in the image.

3. The method of claim 2, comprising processing the composite intensity values to identify plural regions within the region of interest, each region sharing common predetermined characteristics.

4. The method of claim 3, wherein the step of identifying the regions comprises identifying continuous spatial points each having a composite intensity value substantially within predetermined minimum and maximum thresholds.

5. The method of claim 3, further comprising determining a region confidence value for each identified region based on at least one morphology characteristic, using predetermined morphology parameters.

6. The method of claim 5, wherein the region confidence value is determined based on a comparison of characteristics of each identified region to predetermined characteristics of at least one of the group of a measure of the regions curvature, eccentricity, filled volume, and solidity.

7. The method of claim 5, wherein the confidence image includes a value for each spatial point based on the composite intensity value and a region confidence value for each said spatial point.

8. The method of claim 2, further comprising processing the image and at least one intensity range filter to determine an intensity range confidence image, processing the image and at least one relative curve filter to determine a relative curve confidence image, and processing both the relative curve confidence image and the intensity range confidence image to determine the composite intensity values, the composite intensity values together forming the confidence image.

9. The method of claim 2, step (c) further comprising determining, based on the composite intensity values, each identified region having characteristics of a first abnormality.

10. The method of claim 9, further comprising displaying the confidence image as a 3-D manipulatible image with at least one region having characteristics of the first abnormality visually displayed with a similar color to assist with visual identification of at least one region having characteristics of the first abnormality.

11. The method of claim 9, further comprising comparing each identified region having characteristics of said first abnormality with previously determined information about the spatial location of said first abnormality in the region of interest, and iteratively adjusting at least one parameter for at least one of the predetermined feature templates, morphology parameters, and intensity parameters to determine if said adjusted parameter results in confidence image better identifying the spatial location of said first abnormality.

12. The method of claim 9, step (c) further comprising inputting a user selection of spatial locations within the image and in response to the user selection, and automatically determining at least one of a region or a parameter based on one or more parameters of the selected spatial locations.

13. The method of claim 12, wherein the step of automatically determining comprises at least one of:

(c1) automatically determining at least one new region of interest differing from each identified region, based on the user selection of spatial locations and further user selection of a feature template and confirmation of derived time-intensity parameters; and

(c2) automatically determining at least one of a modified boundary and dimension for an identified region, based on the user selection of spatial locations and further user selection of a feature template and confirmation of derived morphology parameters.

14. The method of claim 1, wherein the preprocessing comprises normalizing the image, and the composite intensity value for each spatial point in the image is one of the group of an average or a product of all normalized intensity values for each said spatial point at the plural times.

15. The method of claim 1, wherein the image comprises spatial information taken at least three predetermined times after administration of a chemical substance having detectible wash-in and wash-out intensity characteristics in abnormalities within the period of the predetermined times.

16. The method of claim 15, wherein the chemical substance comprises gadolinium.

17. An information processing system for detecting abnormalities in a region of interest in a person or other mammal, the system comprising an input capable of receiving an image file comprising spatial information of the region of interest taken at plural times, a processor, and an image processing application, the image processing application comprising plural instructions and the processor is operably configured to execute said plural instructions, the plural instructions comprising:

(a) confidence image determination instructions, for processing the image file using at least one predetermined feature template, morphology parameters, and intensity parameters varying over the plural times, to determine a confidence image, wherein each feature template represents at least one predetermined detectable feature for the region of interest and the confidence image represents a likelihood that the at least one feature is present in the region of interest; and

(b) detection instructions, for outputting the confidence image for further processing and detection of abnormal features.

18. The system of claim 17, wherein the confidence image determination instructions further comprise instructions operable for processing the image, at least one template and intensity parameters to determine a composite intensity value for each spatial point in the image.

19. The system of claim 18, wherein the confidence image determination instructions further comprise instructions operable for (a) processing the composite intensity values to identify plural regions within the region of interest, each region sharing common predetermined characteristics, and substantially continuous spatial points each having a composite intensity value substantially within predetermined minimum and maximum thresholds.

20. The system of claim 19, the confidence image determination instructions further comprise instructions operable for determining a region confidence value for each identified region based on at least one morphology characteristic, using predetermined morphology parameters, wherein the region confidence value is determined based on a comparison of characteristics of each identified region to predetermined characteristics of at least one of the group of a measure of the regions curvature, eccentricity, filled volume, and solidity, and wherein the confidence image includes a value for each spatial point determined based on the composite intensity value and a region confidence value for each said spatial point.

21. The system of claim 18, wherein the confidence image determination instructions further comprise instructions operable for processing the image and at least one intensity range filter to determine a intensity range confidence image, processing the image and at least one relative curve filter to determine a relative curve confidence image, and processing both the relative curve confidence image and the intensity range confidence image to determine the composite intensity values, the composite intensity values together forming the confidence image.

22. The system of claim 18, wherein the detection instructions further comprise instructions operable for determining, based on the composite confidence values, each identified region having characteristics of a first abnormality, and for displaying the confidence image as a 3-D manipulatible image with all regions having characteristics of the first abnormality visually displayed with a similar color to assist with visual identification of each said region.

23. The system of claim 22, wherein the detection instructions further comprise instructions operable for comparing each identified region having characteristics of said first abnormality with previously determined information about the spatial location of said first abnormality in the region of interest, and iteratively adjusting at least one parameter for at least one of the predetermined feature template, morphology parameters, and intensity parameters to determine if said adjusted parameter results in confidence image better identifying the spatial location of said first abnormality.

24. The system of claim 22, wherein the detection instructions further comprise instructions operable for inputting a user selection of spatial locations within the image and in response to the user selection, and automatically determining at least one of a region or a parameter based on one or more parameters of the selected spatial locations by:

(b1) automatically determining at least one new region of interest differing from said regions, based on the user selection of spatial locations and further user selection of a feature template and confirmation of derived time-intensity parameters; and

(b2) automatically determining at least one of a modified boundary and dimension for at least one of said regions, based on the user selection of spatial locations and further user selection of a feature template and confirmation of derived morphology parameters.

25. The system of claim 17, wherein the confidence image determination instructions further comprise instructions operable for capturing the image file via a magnetic resonance imaging (MRI) system and preprocessing the captured image to enhance detectability of the features, the image file comprising spatial information taken at least three predetermined times after the administration of a chemical having detectible wash-in and wash-out intensity characteristics in abnormalities within the period of the predetermined times, and the substance comprising gadolinium.

27. The method of claim 25, further comprising comparing each identified region having characteristics of said first abnormality with previously determined information about the spatial location of said first abnormality in the region of interest, and iteratively adjusting at least one parameter for at least one of the predetermined feature templates, morphology parameters, and intensity parameters to determine if said adjusted parameter results in confidence image better identifying the spatial location of said first abnormality.

28. The method of claim 25, further comprising, in response to the provider selection, automatically determining at least one of a region or a parameter based on one or more parameters of the selected spatial locations, wherein the step of automatically determining comprises at least one of:

(b1) automatically determining at least one new region of interest differing from each identified region, based on the user selection of spatial locations and further user selection of a feature template and confirmation of derived time-intensity parameters; and

(b2) automatically determining at least one of a modified boundary and dimension for an identified region, based on the user selection of spatial locations and further user selection of a feature template and confirmation of derived morphology parameters.

26. A method for determining regions of interest in a person or other mammal representing abnormalities from an MRI image comprising plural images taken at proximate times, the method comprising:

(a) processing the image using at least one predetermined feature template, morphology parameters, and intensity parameters varying over the plural times, to determine a confidence image comprising a composite intensity value for each spatial point in the image, and further processing the composite intensity values to identify one or more regions of interest within the image sharing common predetermined characteristics, wherein the confidence image represents a likelihood that the region of interest includes an abnormality;

(b) displaying at least one of a 2-D or 3-D display representation of the confidence image for a medical service provider, and in response to input from the provider based on the display, redetermining the region of interest, wherein the input includes selection via a user interface of at least one of (a) locations within the image but outside of a region of interest, and (b) at least one new parameter used in determining the location and volume of the region of interest.

(c) displaying a further display representation of the confidence image with a modified region of interest based on the user input; and

(d) repeating steps (c) and (d) as desired by the provider.