Search

EP-4738264-A1 - WEIGHTED RECONSTRUCTION OF CT IMAGES OF THE HEART

EP4738264A1EP 4738264 A1EP4738264 A1EP 4738264A1EP-4738264-A1

Abstract

A mechanism for producing three-dimensional (3D) computed tomography (CT) images representing a target volume of a subject (e.g., a patient). Specifically, the proposed method applies to target volumes that are associated with a quasi-periodic motion, e.g., the heart. Said quasi-periodic motion is characterized by a regular, repetitive and/or rhythmic motion of the target volume that may not have a consistent frequency and/or period (e.g., where said frequency and/or period varies with time).

Inventors

  • KOEHLER, THOMAS
  • GRASS, MICHAEL
  • PROKSA, ROLAND

Assignees

  • Koninklijke Philips N.V.

Dates

Publication Date
20260506
Application Date
20241031

Claims (15)

  1. A computer-implemented method (200) of generating a three-dimensional, 3D, computed tomography, CT, image representing a target volume of a subject that is associated with a quasi-periodic motion, the computer-implemented method comprising: determining, for each of a plurality of sub-volumes of the target volume, a target phase of the quasi-periodic motion by processing one or more preliminary 3D images representing the target volume (210), wherein each target phase meets one or more predefined criteria associated with the corresponding sub-volume; obtaining CT projection data of the target volume (220), wherein the CT projection data comprises a plurality of projection values; and processing the CT projection data to produce the 3D CT image using a back projection technique (230), wherein the back projection technique comprises applying weightings to projection values of the CT projection data responsive to the target phases.
  2. The computer-implemented method of claim 1, wherein, for each of the plurality of sub-volumes, the target phase corresponds to a phase of the quasi-periodic motion when a velocity associated with said sub-volume is estimated to be a local minimum of the quasi-periodic motion and/or below a predetermined threshold.
  3. The computer-implemented method of any one of claims 1 or 2, wherein: the 3D CT image comprises a plurality of voxels each having a voxel value; and the back projection technique comprises, for each voxel of the 3D CT image, processing the projection values of the CT projection data representing the voxel of the 3D CT image to produce the voxel value (300), wherein a weighting applied to each projection value used to produce the voxel value is responsive to a target phase for the voxel, wherein the target phase for the voxel is responsive to the target phases for one or more sub-volumes in the vicinity of a region of the target volume represented by the voxel.
  4. The computer-implemented method of claim 3, wherein, for each voxel value, applying a weighting to a projection value used to produce the voxel value comprises multiplying said projection value by the applied weighting.
  5. The computer-implemented method of any one of claims 3 or 4, wherein: the computer-implemented method further comprises determining a target phase function responsive to the target phases for the plurality of sub-volumes (215), wherein the target phase function describes target phases as a function of position in the target volume; and the back projection technique comprises, for each voxel, determining the target phase for the voxel responsive to the target phase function.
  6. The computer-implemented method of any one of claims 3 to 5, wherein: the CT projection data is acquired over a plurality of projection angles; and for each voxel value, the weighting applied to each projection value used to produce the voxel value is responsive to the projection angle used to acquire said projection value.
  7. The computer-implemented method of claim 6, wherein: for each voxel value, the weighting applied to each projection value used to produce the voxel value is responsive to an angular weighting function (410, 510) that describes weightings as a function of projection angle, wherein the angular weighting function defines a preferred projection angle range for acquiring CT projection data of the target volume at a reference phase of the quasi-periodic motion; and the back projection technique comprises, for each voxel: modifying the angular weighting function responsive to the target phase for the voxel (305); and determining the weighting to apply to each projection value from the modified angular weighting function (420, 520).
  8. The computer-implemented method of claim 7, wherein modifying the angular weighting function responsive to the target phase for the voxel comprises: determining a phase difference between the target phase for the voxel and the reference phase; determining an angular shift responsive to the determined phase difference; and shifting the angular weighting function by an amount equal to the determined angular shift.
  9. The computer-implemented method of any one of claims 7 or 8, wherein the reference phase corresponds to either: an average of the target phases for the plurality of sub-volume; or the target phase for the sub-volume closest to a center of the target volume.
  10. The computer-implemented method of any one of claims 7 to 9, wherein: the reference phase is associated with one or more reference projection angles; the angular weighting function has a maximum weighting at each of the one or more reference projections angles; and the angular weighting function has a non-zero weighting at least a predefined deviation above and below each of the one or more reference projection angles.
  11. The computer-implemented method of any one of claims 1 to 10, wherein the one or more preliminary 3D images comprise a sequence of preliminary 3D images representing motion of the target region over at least one period of the quasi-periodic motion.
  12. The computer-implemented method of claim 11, wherein the target phases are determined responsive to a measure of difference between consecutive preliminary 3D images of the sequence of preliminary 3D images.
  13. The computer-implemented method of any one of claims 1 to 12, wherein the target volume comprises a heart of the subject.
  14. A computer program product comprising computer program code means which, when executed on a computing device having a processing system (120), cause the processing system to perform all of the steps of the computer-implemented method of any one of claims 1 to 13.
  15. A processing system (120) for generating a three-dimensional, 3D, computed tomography, CT, image representing a target volume of a subject that is associated with a quasi-periodic motion, the processing system being configured to: determine, for each of a plurality of sub-volumes of the target volume, a target phase of the quasi-periodic motion by processing one or more preliminary 3D images representing the target volume, wherein each target phase meets one or more criteria associated with the corresponding sub-volume; obtain CT projection data of the target volume, wherein the CT projection data comprises a plurality of projection values; and process the CT projection data to produce the 3D CT image using a back projection technique, wherein the back projection technique comprises applying weightings to projection values of the CT projection data responsive to the target phases.

Description

FIELD OF THE INVENTION The present invention relates to the field of computed tomography (CT), and in particular to CT imaging of the heart. BACKGROUND OF THE INVENTION CT imaging of the heart can be vital for accurately and non-invasively diagnosing heart-related illnesses and conditions in patients. To produce high quality images of the heart, more sophisticated techniques are required than standard CT imaging, however, due to motion of the heart (i.e., heartbeats) occurring during acquisition of the CT data. Current techniques include so-called cardiac-gated CT where an electrocardiogram (ECG) is acquired simultaneous to the CT data. A CT image of the heart may then be retrospectively reconstructed by only using CT data corresponding to a specific phase of the heart (i.e., a specific moment during a heartbeat) as determined from the ECG. Typically, the specific phase used in cardiac-gated CT corresponds to a phase when the heart (as a whole) is exhibiting the least amount of motion (i.e., a phase of least motion). It is recognized, however, that different parts of the heart move differently and, as such, may be associated with different phases of least motion. Consequently, in cardiac-gated CT, parts of the heart with a phase of least motion different to the phase used for reconstruction suffer from motion artifacts. There is thus a desire for a CT reconstruction technique that takes into account the spatial variation of the phase of least motion. SUMMARY OF THE INVENTION The invention is defined by the claims. According to examples in accordance with an aspect of the invention, there is provided a computer-implemented method of generating a three-dimensional (3D) computed tomography (CT) image representing a target volume of a subject that is associated with a quasi-periodic motion. The computer-implemented method first comprises determining, for each of a plurality of sub-volumes of the target volume, a target phase of the quasi-periodic motion by processing one or more preliminary 3D images representing the target volume, wherein each target phase meets one or more predefined criteria associated with the corresponding sub-volume. The method further comprises obtaining CT projection data of the target volume, wherein the CT projection data comprises a plurality of projection values, and processing the CT projection data to produce the 3D CT image using a back projection technique, wherein the back projection technique comprises applying weightings to projection values of the CT projection data responsive to the target phases. The present disclosure provides a technique for reconstructing 3D CT images representing a part of the human body (i.e., a target volume) that exhibits regular (i.e., quasi-periodic) motion, such as the heart or lungs. More specifically, the proposed approach provides a back projection technique that comprises weighting projection values of CT projection data (of the target volume) according to desired (i.e., target) phases determined for different parts of the target volume. In particular, each target phase, determined for a sub-volume of the target volume, represents a desired or preferred moment during the quasi-periodic motion to acquire CT projection data of said sub-volume. Accordingly, each target phase is determined such that it meets one or more predefined criteria associated with the corresponding sub-volume. The proposed approach thus provides a retrospective reconstruction technique of preferably weighting the CT projection data responsive to the target phases. Particularly, when reconstructing a particular sub-volume (or portion of a sub-volume) of the target volume, CT projections values (representing the sub-volume) that were acquired at a phase close to the target phase (for the present sub-volume) may be weighted more heavily than those acquired at a phase further away from the target phase. The proposed technique thus takes into account the spatial variation of the target phases across the target volume. In some examples, for each of the plurality of sub-volumes, the target phase may correspond to a phase of the quasi-periodic motion when a velocity associated with said sub-volume is estimated to be a local minimum of the quasi-periodic motion and/or below a predetermined threshold. In this way, motion artifacts may be minimized (or at least reduced) in the resulting 3D CT image. The 3D CT image typically comprises a plurality of voxels each having a voxel value. Accordingly, the back projection technique may comprise, for each voxel of the 3D CT image, processing the projection values of the CT projection data representing the voxel of the 3D CT image to produce the voxel value. In particular, a weighting applied to each projection value used to produce the voxel value may be responsive to a target phase for the voxel, wherein the target phase for the voxel is responsive to the target phases for one or more sub-volumes in the vicinity of a region of the target v