High-speed inter-modality image registration via iterative feature matching

• US 5,901,199 A
• Filed: 07/11/1997
• Issued: 05/04/1999
• Est. Priority Date: 07/11/1996
• Status: Expired due to Term

• Current Assignee: Board of Trustees of the Leland Stanford Junior University (Stanford Management Co.)
• Original Assignee: Board of Trustees of the Leland Stanford Junior University (Stanford Management Co.)
• Inventors: Cox, Richard S., Murphy, Martin J.
• Primary Examiner(s): Porta, David P.
• Assistant Examiner(s): Bruce, David Vernon
• Application Number:
  US08/890,776
• Time in Patent Office: 662 Days
• Field of Search: 378/65, 604/20, 604/19, 606/33
• US Class Current: 378/65
• CPC Class Codes: A61B 2090/364 Correlation of different images or relation of image positions in respect to the body
• CPC Class Codes: A61B 6/08 Auxiliary means for directing the radiation beam to a particular spot, e.g. using light beams
• CPC Class Codes: A61B 6/583 using calibration phantoms
• CPC Class Codes: A61N 2005/1061 using an x-ray imaging system having a separate imaging source
• CPC Class Codes: A61N 2005/1062 using virtual X-ray images, e.g. digitally reconstructed radiographs [DRR]
• CPC Class Codes: A61N 5/1049 for verifying the position of the patient with respect to the radiation beam
• CPC Class Codes: A61N 5/1065 Beam adjustment
• CPC Class Codes: A61N 5/1069 Target adjustment, e.g. moving the patient support

Abstract

This invention describes a method for aligning radiation therapy beams with a treatment target of a patient. Diagnostic computer tomography scanning is used to map the treatment target in the patient'"'"'s anatomy and to aid physicians in deciding where to aim the radiation therapy beams. Digitally reconstructed radiographs are generated from an intermediate 3-D image produced from the CT scans. These DRRs are compared with x-ray images taken of the treatment target'"'"'s position taken at the time of treatment. A transformation equation is used to compare the DRRs and the x-ray images. A minimization equation is used to determine the similarity between the two sets of images. If the difference between the two sets fall below a determined minimum value, then the position of the treatment target is deemed correct and the radiation therapy begins. If the difference does not fall below the determined minimum value, then either the radiation beams or the treatment target is adjusted and the process is repeated. This procedure may be used many times in the course of a single treatment session, depending on the movement of the patient.

Claims

1. A method of aligning a plurality of radiation therapy beams with a treatment target of a patient, comprising the steps of:

a) preprocessing, comprising the steps of;

1) conducting a computed tomography (CT) scan; and

2) assembling an intermediate 3-D image from said scan; and

b) treatment, comprising the steps of;

1) acquiring a set of at least two real images of position of said treatment target of a patient;

2) using said at least two real images to produce a first feature vector;

3) generating a set of at least two digitally reconstructed images (DRRs) from said intermediate 3-D image;

4) using said set of at least two digitally reconstructed images to produce a second feature vector;

5) calculating a difference between said first feature vector and said second feature vector; and

6) if said difference does not fall below a minimum allowable value, employing an algorithm to adjust positions of said radiation therapy beams such that said radiation therapy beams are aligned with said treatment target of said patient.

2. The method of claim 1 comprising repeating steps b) 1-6.

3. The method of claim 1, including the step of generating key pixels associated with anatomical edge features.

4. The method of claim 3, including the step of generating a mask for said at least two real images and said at least two digitally reconstructed images by highlighting key pixels.

5. The method of claim 3, wherein said feature vectors are determined by said key pixels.

6. The method of claim 1, using a preprocessing coordinate system and a treatment coordinate system.

7. The method of claim 6, using said preprocessing coordinate system wherein said treatment target of said patient is fixed.

8. The method of claim 6, using said treatment coordinate system wherein said treatment target of said patient is rotated and translated.

9. The method of claim 6 wherein said preprocessing coordinate system and said treatment coordinate system are related by a transformation equation.

10. The method of claim 1, allowing six degrees of freedom of position of said treatment target of said patient.

11. The method of claim 10, wherein said six degrees of freedom include translation measurements.

12. The method of claim 10, wherein said six degrees of freedom include rotational measurements.

13. The method of claim 1, wherein said difference between said vectors is determined by use of chi squared.

14. The method of claim 1, wherein said algorithm provides minimization of chi squared until convergence with said minimum allowable value.

15. The method of claim 14, including the step of adjusting patient anatomy in said intermediate 3-D image.

16. The method of claim 14, including the step of repositioning a camera model responsible for taking said at least two reconstructed images.

17. The method of claim 16, wherein repositioning of said camera model is guided by a lookup table.

18. The method of claim 17, wherein said lookup table provides the first derivative of said translation and rotation measurements.

19. The method of claim 14, wherein coordinates of repositioned said camera model are transformed into coordinates for alignment of said radiation therapy beams.

20. The method of claim 14, wherein coordinates of repositioned said camera are transformed into coordinates of alignment of said treatment target of said patient.

21. A method of aligning radiation therapy beams with a treatment target of a patient, comprising the steps of:

a) acquiring a plurality of current two-dimensional images indicative of a current position of said treatment target relative to a means for generating said therapy beams;

b) from a three-dimensional image comprising said treatment target, generating a corresponding plurality of reference two-dimensional images indicative of a reference position of said treatment target relative to said means for generating said therapy beams, wherein said treatment target is aligned with said means for generating said therapy beams in said reference position;

c) comparing current feature vectors defined by said current two-dimensional images with corresponding reference feature vectors defined by said reference two-dimensional images;

d) adjusting said current position if said current feature vectors do not match said reference feature vectors; and

e) applying said therapy beams if said current feature vectors match said reference feature vectors.