System and method for magnetic resonance imaging based respiratory motion correction for PET/MRI

1. A method for magnetic resonance imaging (MRI) based motion correction in position emission tomography (PET) images, comprising:

receiving a static 3D magnetic resonance (MR) image of a patient;

receiving PET image data of the patient;

receiving a series of 2D MR images of the patient acquired at a plurality of time points simultaneous to acquisition of the PET image data;

estimating a 3D+t motion field by registering the series of 2D MR images acquired at the plurality of time points to the static 3D MR image, wherein estimating a 3D+t motion field by registering the series of 2D MR images acquired at the plurality of time points to the static 3D MR image comprises:

estimating the 3D+t motion field using the registration of the series of 2D MR images and the static 3D MR image and a predictive respiratory motion model; and

generating a motion corrected PET image based on the estimated 3D+t motion field using motion corrected PET reconstruction.

2. The method of claim 1 , wherein estimating a 3D+t motion field by registering the series of 2D MR images acquired at the plurality of time points to the static 3D MR image comprises:

estimating a 3D+t motion field that minimizes an energy function including an image energy term and a model energy term.

3. The method of claim 2 , wherein the image energy term compares each 2D MR image to a corresponding slice of a transformed volume resulting from transforming the static 3D volume based on the 3D+t motion field at the respective one of the plurality of time points, and the model energy term compares the 3D+t motion field and a predicted motion field calculated using the predictive respiratory motion model.

4. The method of claim 3 , wherein the image energy term is a negative intensity-based similarity measure between each 2D MR image and the corresponding slice of the transformed volume.

5. The method of claim 3 , wherein the model energy term measures a difference between the 3D+t motion field at each of the plurality of time steps and the predictive respiratory motion model at a breathing phase corresponding to each of the plurality of time steps.

6. The method of claim 5 , wherein the breathing phase corresponding to each of the plurality of time steps is detected using a navigation signal.

7. The method of claim 3 , further comprising:

registering the static 3D image to an atlas used to train a standard respiratory motion model to obtain a warping function; and

applying an inverse of the warping function to the standard respiratory motion model to obtain a patient-specific respiratory motion model, wherein the model energy term compares the 3D+t motion field and the patient-specific respiratory motion model.

8. The method of claim 1 wherein the series of 2D MR images are slices in the transverse plane acquired in a sweep along a head-foot direction of the patient.

9. An apparatus for magnetic resonance imaging (MRI) based motion correction in position emission tomography (PET) images, comprising:

a processor; and

a memory storing computer program instructions, which when executed on the processor, cause the processor to perform operations comprising:

receiving a static 3D magnetic resonance (MR) image of a patient;

receiving PET image data of the patient;

receiving a series of 2D MR images of the patient acquired at a plurality of time points simultaneous to acquisition of the PET image data;

estimating a 3D+t motion field by registering the series of 2D MR images acquired at the plurality of time points to the static 3D MR image, wherein estimating a 3D+t motion field by registering the series of 2D MR images acquired at the plurality of time points to the static 3D MR image comprises:

estimating the 3D+t motion field using the registration of the series of 2D MR images and the static 3D MR image and a predictive respiratory motion model; and

generating a motion corrected PET image based on the estimated 3D+t motion field using motion corrected PET reconstruction.

10. The apparatus of claim 9 , wherein estimating a 3D+t motion field by registering the series of 2D MR images acquired at the plurality of time points to the static 3D MR image comprises:

estimating a 3D+t motion field that minimizes an energy function including an image energy term and a model energy term.

11. The apparatus of claim 10 , wherein the image energy term compares each 2D MR image to a corresponding slice of a transformed volume resulting from transforming the static 3D volume based on the 3D+t motion field at the respective one of the plurality of time points, and the model energy term compares the 3D+t motion field and a predicted motion field calculated using the predictive respiratory motion model.

12. The apparatus of claim 11 , wherein the image energy term is a negative intensity-based similarity measure between each 2D MR image and the corresponding slice of the transformed volume.

13. The apparatus of claim 11 , wherein the model energy term measures a difference between the 3D+t motion field at each of the plurality of time steps and the predictive respiratory motion model at a breathing phase corresponding to each of the plurality of time steps.

14. The apparatus of claim 11 , the operations further comprising:

registering the static 3D image to an atlas used to train a standard respiratory motion model to obtain a warping function; and

applying an inverse of the warping function to the standard respiratory motion model to obtain a patient-specific respiratory motion model, wherein the model energy term compares the 3D+t motion field and the patient-specific respiratory motion model.

15. A non-transitory computer readable medium storing computer program instructions for magnetic resonance imaging (MRI) based motion correction in position emission tomography (PET) images, the computer program instructions when executed on a processor cause the processor to perform operations comprising:

receiving a static 3D magnetic resonance (MR) image of a patient;

receiving PET image data of the patient;

receiving a series of 2D MR images of the patient acquired at a plurality of time points simultaneous to acquisition of the PET image data;

estimating a 3D+t motion field by registering the series of 2D MR images acquired at the plurality of time points to the static 3D MR image, wherein estimating a 3D+t motion field by registering the series of 2D MR images acquired at the plurality of time points to the static 3D MR image comprises:

estimating the 3D+t motion field using the registration of the series of 2D MR images and the static 3D MR image and a predictive respiratory motion model; and

generating a motion corrected PET image based on the estimated 3D+t motion field using motion corrected PET reconstruction.

16. The non-transitory computer readable medium of claim 15 , wherein estimating a 3D+t motion field by registering the series of 2D MR images acquired at the plurality of time points to the static 3D MR image comprises:

estimating a 3D+t motion field that minimizes an energy function including an image energy term and a model energy term.

17. The non-transitory computer readable medium of claim 16 , wherein the image energy term compares each 2D MR image to a corresponding slice of a transformed volume resulting from transforming the static 3D volume based on the 3D+t motion field at the respective one of the plurality of time points, and the model energy term compares the 3D+t motion field and a predicted motion field calculated using the predictive respiratory motion model.

18. The non-transitory computer readable medium of claim 17 , wherein the image energy term is a negative intensity-based similarity measure between each 2D MR image and the corresponding slice of the transformed volume.

19. The non-transitory computer readable medium of claim 17 , wherein the model energy term measures a difference between the 3D+t motion field at each of the plurality of time steps and the predictive respiratory motion model at a breathing phase corresponding to each of the plurality of time steps.

20. The non-transitory computer readable medium of claim 17 , the operations further comprising:

registering the static 3D image to an atlas used to train a standard respiratory motion model to obtain a warping function; and

applying an inverse of the warping function to the standard respiratory motion model to obtain a patient-specific respiratory motion model, wherein the model energy term compares the 3D+t motion field and the patient-specific respiratory motion model.