Atlas and methods for segmentation and alignment of anatomical data

• US 7,324,842 B2
• Filed: 01/22/2002
• Issued: 01/29/2008
• Est. Priority Date: 01/22/2002
• Status: Active Grant

• Current Assignee: Cortechs Labs, Inc.
• Original Assignee: Cortechs Labs, Inc.
• Inventors: Dale, Anders, Gicquel, Se
• Primary Examiner(s): Mercader; Eleni Mantis
• Assistant Examiner(s): Roy; Baisakhi
• Application Number:
  US10/055,256
• Publication Number:
  US 20030139659A1
• Time in Patent Office: 2,198 Days
• Field of Search: 600/407, 600/410, 600/411, 600/420, 600/427, 600/431, 600/473, 600/475, 702/183, 702/189, 378/4, 378/20, 378/15, 378/19, 324/309, 324/307, 382/128, 382/131
• US Class Current: 600/407
• CPC Class Codes: A61B 5/055 involving electronic [EMR] or nuclear [NMR] magnetic resonance, e.g. magnetic resonance imaging

Abstract

The present invention provides an atlas comprising values representative of magnetic resonance properties of a magnetic resonance (MR) scan and optionally, prior probability data relating to tissue type. Further embodiments of the invention involve a system including an MR scanner and the atlas for use in alignment of an MR scan, such as a localizer scan, to obtain a specific geometry of the data acquired during a subsequent scan. Also, a system includes an MR scanner and the atlas for automatic segmentation of an MR scan. Methods of making and using the atlas and system are also provided.

Claims

1. A method for obtaining information about a subject, comprising the steps of:

providing a prior probability MRI based atlas having magnetic resonance data including tissue type prior probability derived from at least one other subject;

processing with a processor information received from a magnetic resonance scanner pertaining to said subject, wherein said information includes data from a magnetic resonance scan;

reading said atlas; and

determining alignment of said magnetic resonance scan to obtain a specific geometry of a subsequent magnetic resonance scan.

2. The method of claim 1, further comprising the steps of:

communicating alignment data from said processor to said scanner; and

automatically aligning said magnetic resonance scan to obtain said specific geometry of a subsequent magnetic resonance scan by the use of said alignment data.

3. The method of claim 1, wherein said step of providing a prior probability MRI based atlas having magnetic property values derived from at least one other subject involves data derived from a plurality of subjects.

4. A method of using a processor for obtaining information about a subject, comprising the steps of:

providing magnetic property values from a magnetic resonance scanner

pertaining to said subject;

providing a prior probability MRI based atlas having at least two magnetic property values for at least one corresponding voxel derived from at least one other subject; and

labeling by said processor of tissue types of a tissue corresponding to said magnetic resonance property values pertaining to said subject by the use of said atlas having said magnetic resonance values derived from at least one other subject.

5. The method of claim 4, wherein said step of providing a prior probability MRI based atlas having magnetic property values derived from at least one other subject involves data derived from a plurality of other subjects.

6. A method for creating an a prior probability MRI based atlas, comprising the steps of:

providing a first magnetic resonance modality volume pertaining to a subject and divided into voxels;

correcting distortion of said first magnetic resonance modality volume;

recording a magnetic property value in a node of said atlas corresponding to a voxel of said first magnetic modality volume;

providing a second magnetic resonance modality volume pertaining to a second subject and divided into voxels;

correcting distortion of said second magnetic resonance modality volume; and

updating said magnetic property data in said node of said atlas corresponding to a voxel of said second magnetic resonance modality volume.

7. The method of claim 6, wherein the step of correcting involves the correction of distortion caused by at least one of the group consisting of chemical shift, magnetic susceptibility, per-acquisition motion, gradients non-linearity, main magnetic field non-homogeneity, eddy currents, and Maxwell effects.

8. The method of claim 6, further comprising the steps of:

providing a plurality of magnetic resonance modality volumes pertaining to a plurality of subjects and each of said plurality of magnetic resonance modality volumes divided into voxels;

correcting distortion of each of said plurality of magnetic resonance modality volumes; and

updating said magnetic property value in said node of said atlas corresponding to a voxel of each of said plurality of magnetic resonance modality volumes;

wherein said magnetic property value in said node of said atlas includes statistical data.

9. A method for creating a prior probability MRI based atlas, comprising the steps of:

providing a first magnetic resonance modality volume pertaining to a subject and divided into voxels;

providing a labeled volume indicating tissue types of tissue corresponding to said voxels;

correcting distortion of said first magnetic resonance modality volume;

extracting magnetic property distribution parameters for each tissue type identified at each voxel;

recording magnetic property data corresponding to each tissue type in a node of said atlas corresponding to a voxel of said first magnetic resonance modality volume;

providing a plurality of magnetic resonance modality volumes pertaining to a plurality of subjects and each of said plurality of magnetic resonance modality volumes divided into voxels;

providing a plurality of labeled volumes corresponding to said plurality of magnetic resonance modality volumes and indicating tissue types of tissue corresponding to said voxels;

correcting distortion of each of said plurality of magnetic resonance modality volume; and

updating said magnetic property data in said node of said atlas corresponding to a voxel of each of said plurality of magnetic resonance modality volumes.

10. A method for obtaining information about a subject, comprising the steps of:

providing a prior probability MRI based atlas having magnetic resonance data including more than one magnetic property value prior probability derived from at least one other subject;

processing with a processor information received from a magnetic resonance scanner pertaining to said subject, wherein said information includes data from a magnetiC resonance scan;

reading said atlas; and

determining alignment of said magnetic resonance scan to obtain a specific geometry of a subsequent magnetic resonance scan.

11. The method of claim 10 further comprising the steps of:

communicating alignment data from said processor to said scanner; and

automatically aligning said magnetic resonance scan to obtain said specific geometry of a subsequent magnetic resonance scan by the use of said alignment data.

12. The method of claim 10, wherein said step of providing a prior probability MRI based atlas having magnetic property values derived from at least one other subject involves data derived from a plurality of subjects.

13. A method of aligning and/or segmenting an MR scan, the method comprising:

scanning a subject with an MR scanner to produce a magnetic resonance image;

interfacing the magnetic resonance image with a prior probability MRI based atlas, which includes nodes which each include;

tissue type probability information, magnetic property values, and location information to determine, from the scan, the probability of tissue type by location in the magnetic resonance image; and

aligning the MR scan and/or segmenting the MR scan.

14. A method of aligning and/or segmenting an MR scan, the method comprising:

providing a prior probability MRI based atlas from at least three MR scans of other subjects which define node locations in anatomic atlas space which are independent of MR imaging platform voxel properties whereas voxels are dependent upon MR imaging instrument instructions and nodes are independent of MR imaging platform instructions, each node including probability information for at least four independent properties;

location in three dimensional atomic atlas space, prior probability of specific tissue types at this nodal location, statistics of at least one magnetic resonance property assigned to specific tissue types at this nodal location, and prior probability of specific tissue types of neighboring nodes for each tissue type at this nodal location, wherein magnetic resonance property values include one or more parameters such as T₁ or T₂ for the identical nodal location in anatomic space;

scanning a subject with an MR scanner to produce a magnetic resonance image representative of that subject comprised of voxel volume elements constructed in accordance with MR imaging instrument instructions;

interfacing the magnetic resonance image with the a priori nodal information from the atlas; and

aligning the MR scan and/or segmenting the MR scan to maximize the post probability of its voxel components matching both position in anatomic space and at least one MR tissue characteristic of the nodes determined a priori in the atlas.

15. The method of claim 14 in which two MR scans are obtained at different points in time to develop data about changes in anatomic distribution of organ tissue components andlor changes in their MR properties independent of imaging instrument instructions.

16. A method of creating an a priori nodal atlas as a reference for the consistent alignment and/or segmentation of subsequent MR scans comprising:

assigning to each node of the atlas information about location in three dimensional anatomic space, as well as information about magnetic resonance properties of each location in anatomic space for specific tissue types and prior probability for specific tissue types and prior probability of specific tissue types of neighboring nodes for each tissue type at the nodal location.

17. A method of using an a priori nodal atlas as a reference for alignment and/or segmentation of subsequent MR scans comprising:

reading data from an MR scan;

reading an a priori nodal atlas wherein, each node of the atlas is based on at least 3 MRI scans and wherein the atlas contains information about location in three-dimensional anatomic space, as well as information about magnetic resonance properties of each location in anatomic space for specific tissue types, prior probability for specific tissue types, and prior probability for specific tissue types of neighboring nodes for each tissue type at the nodal location; and

aligning and/or segmenting subsequent scans by operating on voxel based images created in accordance with instructions of an imaging instrument and applying transformations which maximize the probability of matching each voxel to a corresponding node in the atlas.

18. A system for obtaining information about a subject, the system comprising:

a processor adapted to receive information obtained from a magnetic resonance scan and read a prior probability MRI based atlas to determine a tissue type

the prior probability MRI based atlas comprising;

a plurality of nodes corresponding to a plurality of voxels representing spatial locations of a subject, each of said nodes configured to store at least one magnetic property value as determined by magnetic resonance imaging at least one tissue type prior probability value corresponding to a tissue type of a voxel, thereby determining the tissue type.