Computer-aided detection system utilizing temporal analysis as a precursor to spatial analysis

• US 7,620,227 B2
• Filed: 12/29/2005
• Issued: 11/17/2009
• Est. Priority Date: 12/29/2005
• Status: Active Grant

• Current Assignee: General Electric Company
• Original Assignee: General Electric Company
• Inventors: Gupta, Sandeep N., Avinash, Gopal B., Gering, David Thomas
• Primary Examiner(s): AZARIAN, SEYED H
• Application Number:
  US11/321,510
• Publication Number:
  US 20070165920A1
• Time in Patent Office: 1,419 Days
• Field of Search: 382/128, 382/129, 382/130, 382/131, 382/132, 382/133, 382/134, 382/154, 382/162, 382/168, 382/173, 382/181, 382/194, 382/199, 382/203, 382/224, 382/232, 382/240, 382/256, 382/274, 382/305, 382/312, 382/318, 600/410, 600/300, 600/408
• US Class Current: 382/128
• CPC Class Codes: A61B 2090/364 Correlation of different images or relation of image positions in respect to the body
• CPC Class Codes: A61B 5/055 involving electronic [EMR] or nuclear [NMR] magnetic resonance, e.g. magnetic resonance imaging
• CPC Class Codes: A61B 5/7264 Classification of physiological signals or data, e.g. using neural networks, statistical classifiers, expert systems or fuzzy systems neural networks per se G06N3/00; expert systems per se G06N5/00
• CPC Class Codes: A61B 90/36 Image-producing devices or illumination devices not otherwise provided for
• CPC Class Codes: G06T 2207/10076 4D tomography; Time-sequential 3D tomography
• CPC Class Codes: G06T 2207/20104 Interactive definition of region of interest [ROI]
• CPC Class Codes: G06T 2207/30068 Mammography; Breast
• CPC Class Codes: G06T 2207/30096 Tumor; Lesion
• CPC Class Codes: G06T 7/0012 Biomedical image inspection
• CPC Class Codes: G06T 7/11 Region-based segmentation
• CPC Class Codes: G06T 7/143 involving probabilistic approaches, e.g. Markov random field [MRF] modelling
• CPC Class Codes: G06V 10/25 Determination of region of interest [ROI] or a volume of interest [VOI]
• CPC Class Codes: G06V 2201/03 Recognition of patterns in medical or anatomical images

Abstract

A computer-aided detection system for evaluating tissue based on a series of timed images acquired both before and after a contrast agent is administered performs a temporal analysis of the tissue and then a spatial analysis in which the tissue is categorized. After the temporal and spatial analysis is performed, the results can be displayed. The displayed results can include both tissue characterization results, underlying curves used to determine the characterization, and confidence data regarding how good the characterization is expected to be. The confidence data can be provided, for example, by using variations in color schemes, displaying numerical confidence levels, or providing graphical features such as piecewise linear models.

Claims

1. A method for computer-aided detection in a four-dimensional series of medical images, the method comprising the following steps:

(a) accessing from a memory a series of three dimensional medical images of a region of interest over a period of time, at least one set in the series being taken before administration of a contrast agent and a plurality of sets taken after the administration of a contrast agent to produce a four dimensional image;

(b) performing a temporal analysis using a processing unit to characterize each voxel in the four dimensional image, wherein the performing a temporal analysis step includes the steps of accessing at least one stored curve representative of a known tissue type, comparing the intensity over time at the voxel to the at least one curve, and classifying tissue based on the probability that the tissue at the voxel is the type represented by the curve; and

(c) performing a spatial analysis influenced by the temporal characterizations using the processing unit to segment the voxels into tissue types.

2. The method as recited in claim 1, further comprising the steps of identifying voxels of interest based on the analysis in step (b) and using the identified voxels of interest in step (c) to perform the spatial analysis.

3. The method as recited in claim 1, further comprising the step of registering the images in the series of three-dimensional images.

4. The method as recited in claim 1, wherein step (b) comprises applying a curve-fitting analysis.

5. The method as recited in claim 1, wherein step (c) comprises determining a probability that each voxel is of the selected tissue type.

6. The method as recited in claim 5, wherein the probability is determined based on a method including at least one of a Bayesian maximum a posteriori analysis, a maximum likelihood analysis, an expectation maximization, a Markov model, a nearest-neighbor classifier, a fuzzy classifier, a classifier based on Parzen window estimation, and a Bayesian belief propagation network.

7. The method as recited in claim 1, wherein the medical images are acquired through at least one of an MRI, a CT, a PET, a SPECT, and an ultrasound scan.

8. The method as recited in claim 1, further comprising the step of providing a user interface and allowing a user to select a region of interest by selecting at least one of a temporal parameter and a spatial region.

9. The method as recited in claim 1, wherein step (c) further comprises the step of using a spatial attribute including at least one of an image gradient, a curvature, a size of segmented structures, a distance to a segmented boundary, a thickness of segmented structures, a bias field, a neighborhood coherence model, a registered anatomical model, and a fitted shape model to segment the voxels.

10. A method for computer-aided detection in a four-dimensional series of medical images, the method comprising the following steps:

(a) accessing from a memory a series of three dimensional medical images of a region of interest over a period of time, at least one set in the series being taken before administration of a contrast agent and a plurality of sets taken after the administration of a contrast agent to produce a four dimensional image;

(b) performing a temporal analysis using a processing unit to characterize each voxel in the four dimensional image;

(c) performing a spatial analysis influenced by the temporal characterizations using the processing unit to segment the voxels into tissue types, wherein the performing a spatial analysis step further comprises assigning a visual designator to each voxel based on at least one of the tissue type and a probability of belonging to a given tissue type.

11. The method as recited in claim 10, further comprising the step of assigning a color as the visual designator.

12. The method as recited in claim 11, further comprising the step of displaying the visual designator in at least one of a 2-D image, a 3-D surface rendering, a 3-D volume rendering, and a graph of intensity over time at each voxel.

13. The method as recited in claim 12, further comprising the steps of identifying voxels of interest in at least one of the analyses of step (b) and step (c) and using the identified voxels of interest to limit the display to at least one selected tissue type.

14. A method for computer-aided detection in a four-dimensional series of medical images, the method comprising the following steps:

(a) accessing from a memory a series of three dimensional medical images of a region of interest over a period of time, at least one set in the series being taken before administration of a contrast agent and a plurality of sets taken after the administration of a contrast agent to produce a four dimensional image;

(b) selecting a first scale space for processing the images and, for each of the images in the series;

(c) performing a temporal analysis using a processing unit to characterize each voxel in the four dimensional image

(d) performing a spatial analysis based on the temporal characterization using a processing unit to segment the voxels into at least one tissue type;

(e) selecting a new scale space and repeating steps (c-d) over a range of scale spaces, and

further comprising the steps of identifying voxels of interest based on the analysis in steps (c) and (d) to perform at least one of the steps of localizing the spatial analysis in step (d), adjusting the region of interest for the next scale space level, and limiting a display to at least one selected tissue type.

15. The method as recited in claim 14, further comprising the step of registering the images in the series of three-dimensional images.

16. The method as recited in claim 14, further comprising the steps of associating a visual designator based on at least one of a tissue type and a probability of belonging to a given tissue type and displaying the visual designator in at least one of a 2-D image, a 3-D surface rendering, a 3-D volume rendering, and a graph of intensity over time at each voxel.

17. A method of computer-aided detection in a four-dimensional series of medical images, the method comprising the following steps:

(a) accessing from a memory a series of three dimensional medical images of a region of interest over a period of time, at least one set in the series being taken before administration of a contrast agent and a plurality of sets taken after the administration of a contrast agent to produce a four dimensional image;

(b) selecting a first scale space for processing the images and, for each of the images in the series;

(c) performing a temporal analysis using a processing unit to characterize each voxel in the four dimensional image;

(d) performing a spatial analysis based on the temporal characterization using the processing unit to segment the voxels into at least one tissue type; and

(e) selecting a new scale space and repeating steps (c-d) over a range of scale spaces;

wherein the step of performing a temporal analysis comprises applying a curve-fitting analysis by accessing at least one stored curve representative of a known tissue type, comparing the intensity over time at the voxel to the at least one curve, and classifying tissue based on the probability that the tissue at the voxel is the type represented by the curve.