Catheter RF ablation using segmentation-based 2D-3D registration

1. A method for registering a first two-dimensional image of a cardiocirculatory structure and a three-dimensional image of the cardiocirculatory structure, comprising:

acquiring the three-dimensional image including the cardiocirculatory structure using a first imaging modality;

projecting the acquired three-dimensional image into two-dimensions to produce a two-dimensional projection image of the cardiocirculatory structure;

segmenting a structure of interest either from the acquired three-dimensional image or from the two-dimensional projection image;

acquiring the first two-dimensional image of the cardiocirculatory structure using a second imaging modality;

segmenting the structure of interest from the acquired first two-dimensional image;

generating a first distance map based on the two-dimensional projection image in which each pixel of the two-dimensional projection image is labeled with a distance to a nearest image feature pixel of the two-dimensional projection image;

generating a second distance map based on the acquired first two-dimensional image in which each pixel of the acquired first two-dimensional image is labeled with a distance to a nearest image feature pixel of the acquired first two-dimensional image; and

performing registration of the acquired three-dimensional image and the acquired first two-dimensional image by minimizing a difference between the first and second distance maps, wherein each of the above steps are performed using a computer system.

2. The method of claim 1 , wherein a contrast agent is administered prior to the acquisition of the three-dimensional image and segmenting a structure of interest either from the acquired three-dimensional image or from the two-dimensional projection image includes segmenting a contrast-enhanced structure.

3. The method of claim 1 , wherein a contrast agent is administered prior to the acquisition of the first two-dimensional image and segmenting the structure of interest from the acquired first two-dimensional image includes segmenting a contrast-enhanced structure.

4. The method of claim 1 , wherein cardiovascular phase information, corresponding to a moment in which the three-dimensional image is acquired, is stored and used to ensure that the acquired first two-dimensional image used in registration represents a comparable cardiac phase with respect to the acquired three-dimensional image.

5. The method of claim 1 , wherein background information is removed from the first acquired two-dimensional image prior to the performance of registration.

6. The method of claim 1 , wherein catheter information is removed from the acquired first two-dimensional image prior to the performance of registration.

7. The method of claim 1 , wherein the first imaging modality is a CT scan or an MRI.

8. The method of claim 1 , wherein the second imaging modality is a fluoroscope and the acquired first two-dimensional image includes a sequence of image frames acquired in real-time.

9. The method of claim 1 , wherein a rough alignment of the two-dimensional projection image and the acquired first two-dimensional image is performed prior to the generation of the first and second distance maps.

10. The method of claim 9 , wherein the rough alignment includes performing rough registration based on an image of a spine within the two-dimensional projection image and the acquired first two-dimensional image.

11. The method of claim 1 , wherein in projecting the acquired three-dimensional image into two-dimensions, an angle of projection is selected based on an angle of acquisition for the acquired first two-dimensional image.

12. The method of claim 1 , wherein the structure of interest includes pulmonary arteries, an aorta, a left atrium, or pulmonary veins.

13. The method of claim 1 , wherein for each of the first and second distance maps, pixel distances are calculated using a Euclidean distance transformation.

14. The method of claim 1 , wherein minimizing the difference between the first and second distance maps includes qualification of a similarity between the two distance maps by a sum of absolute difference or normalized cross correlation.

15. A method for providing real-time visualization of an anatomical structure during an electrophysiology intervention, comprising:

acquiring a three-dimensional image including the anatomical structure using a single-frame imaging modality;

projecting the acquired three-dimensional image into two-dimensions to produce a two-dimensional projection image of the anatomical structure;

segmenting a structure of interest either from the acquired three-dimensional image or from the two-dimensional projection image;

acquiring a sequence of two-dimensional image frames of the anatomical structure using a real-time imaging modality;

segmenting the structure of interest from a selected frame of the acquired sequence of two-dimensional image frames;

generating a first distance map based on the two-dimensional projection image in which each pixel of the two-dimensional projection image is labeled with a distance to a nearest image feature pixel of the two-dimensional projection image;

generating a second distance map based on the selected frame of the acquired sequence of two-dimensional image frames in which each pixel of the selected frame of the acquired sequence of two-dimensional image frames is labeled with a distance to a nearest image feature pixel of the selected frame of the acquired sequence of two-dimensional image frames;

performing registration of the acquired three-dimensional image and the sequence of two-dimensional image frames by minimizing a difference between the first and second distance maps; and

continuously superimposing the acquired three-dimensional image over each frame of the sequence of two-dimensional image frames and displaying the superposition for use in guiding electrophysiology intervention,

wherein each of the above steps are performed using a computer system.

16. The method of claim 15 , wherein cardiovascular phase information corresponding to a moment in which the three-dimensional image is acquired is stored and used to select the selected frame of the acquired sequence of two-dimensional image frames.

17. The method of claim 15 , wherein background information is removed from the selected frame of the acquired sequence of two-dimensional image frames prior to the performance of registration.

18. A computer system comprising:

a processor; and

a non-transitory, tangible, program storage medium, readable by the computer system, embodying a program of instructions executable by the processor to perform method steps for registering a two-dimensional image of a cardiocirculatory structure and a three-dimensional image of the cardiocirculatory structure, the method comprising:

acquiring the three-dimensional image including the cardiocirculatory structure using a first imaging modality;

projecting the acquired three-dimensional image into two-dimensions to produce a two-dimensional projection image of the cardiocirculatory structure;

acquiring a sequence of two-dimensional image frames of the cardiocirculatory structure using a fluoroscope;

selecting an image frame from the sequence of two-dimensional image frames by gating the sequence of two-dimensional image frames according to cardiac phase information corresponding to the acquisition of the three-dimensional image;

removing background information from the selected image frame of the acquired sequence of two-dimensional image frames;

generating a first distance map based on the two-dimensional projection image in which each pixel of the two-dimensional projection image is labeled with a distance to a nearest image feature pixel of the two-dimensional projection image;

generating a second distance map based on the background-removed selected image frame of the acquired sequence of two-dimensional image frames in which each pixel of the background-removed selected image frame of the acquired sequence of two-dimensional image frames is labeled with a distance to a nearest image feature pixel of the background-removed selected image frame of the acquired sequence of two-dimensional image frames; and

performing registration of the three-dimensional image and the acquired sequence of two-dimensional image frames by minimizing a difference between the first and second distance maps.

19. The computer system of claim 18 , wherein the first imaging modality is a CT scan or an MRI.

20. The computer system of claim 18 , wherein a rough alignment of the two-dimensional projection image and the background-removed selected image frame of the acquired sequence of two-dimensional image frames is performed prior to the generation of the first and second distance maps.