Methods and software for self-gating a set of images

• US 20030016851A1
• Filed: 05/30/2002
• Published: 01/23/2003
• Est. Priority Date: 07/17/2001
• Status: Active Grant

• Current Assignee: Merative US LP
• Original Assignee: AccuImage Diagnostics Corp. (International Business Machines Corporation)
• Inventors: Kaufman, Leon, Griess, Freiderike
• Granted Patent: US 7,142,703 B2
• Time in Patent Office: Days
• US Class Current: 382/131
• CPC Class Codes: A61B 5/352 Detecting R peaks, e.g. for synchronising diagnostic apparatus; Estimating R-R interval
• CPC Class Codes: A61B 5/7239 using differentiation including higher order derivatives
• CPC Class Codes: A61B 5/7257 using Fourier transforms
• CPC Class Codes: A61B 5/7289 Retrospective gating, i.e. associating measured signals or images with a physiological event after the actual measurement or image acquisition, e.g. by simultaneously recording an additional physiological signal during the measurement or image acquisition
• CPC Class Codes: A61B 6/03 Computed tomography [CT] echo-tomography A61B8/14
• CPC Class Codes: A61B 6/032 Transmission computed tomography [CT]
• CPC Class Codes: A61B 6/503 for diagnosis of the heart
• CPC Class Codes: A61B 6/504 for diagnosis of blood vessels, e.g. by angiography
• CPC Class Codes: A61B 6/5264 due to motion
• CPC Class Codes: A61B 8/5276 due to motion

Abstract

Methods and software for self-gating a set of images. In exemplary embodiments, a fundamental heart frequency of the patient can be measured without the use of an ECG signal. In one method, the fundamental heart frequency can be determined by analyzing the size of the heart in the images. In another method, the fundamental heart frequency can be determined by applying a Fourier Transform. The measured fundamental heart frequency can thereafter be used to select slice images from the image scan for creation of a sagittal or coronal projection image. In exemplary embodiments, the resultant projection image can be used for coronary calcium detection and scoring.

Claims

1. A method of self-gating a set of images, the method comprising:

acquiring a set of overlapping slice images of a patient'"'"'s heart;

generating a projection with the set of slice images;

marking a region of the projection;

analyzing the marked region to calculate a heart frequency and phase of the patient'"'"'s heart motion;

selecting groups of slice images from the set of slice images, based on their relative position in the calculated heart motion frequency and phase; and

generating a plurality of groups of slices that correspond to different phases of heart motion.

2. The method of claim 1 comprising measuring a size of the heart in a marked region of the set of slice images.

3. The method of claim 2 wherein measuring comprises applying a derivative filter to measure a local intensity signal derived from the marked region of the slice images.

4. The method of claim 1 comprising:

obtaining a projection image from each set of selected slice images, each representing a phase of the heart motion;

displaying the group of projections of the patient'"'"'s heart; and

highlighting the projection image of the patient'"'"'s heart in which the marked region of the heart has its largest size.

5. The method of claim 4 wherein highlighting comprises displaying the projection image of the patient'"'"'s heart in which the patient'"'"'s marked region of the heart has its largest size as a larger image.

6. The method of claim 4 wherein highlighting comprises marking by a ranking number the projection of the patient'"'"'s heart in which the patient'"'"'s marked area of the heart has its largest size.

7. The method of claim 1 comprising:

selecting a set of slices of the patient'"'"'s heart on the basis of a preferred projection of the set of slices; and

calcium scoring the selected set of slices of the patient'"'"'s heart.

8. The method of claim 1 comprising:

selecting a set of slices of the patient'"'"'s heart on the basis of a preferred projection of the set of slices; and

3-D rendering the set of slices of the patient'"'"'s heart.

9. The method of claim 1 wherein analyzing comprises:

summing an intensity value along a slice direction of a region of the projection of each slice set that was marked; and

Fourier transforming the intensity value to generate Fourier components in frequency space.

10. The method of claim 9 comprising computing a frequency power spectrum from the Fourier components.

11. The method of claim 10 comprising finding a maximum value of the frequency power spectrum of the heart motion.

12. The method of claim 11 comprising smoothing the frequency power spectrum of the heart motion.

13. The method of claim 11 comprising taking a middle point of a half-height interval of the maximum value of the frequency power spectrum of the heart motion.

14. The method of claim 1 comprising verifying that the marked region contains enough information to compute the frequency and phase.

15. The method of claim 14 where verifying comprises computing that at least 3 seconds of data were included in the marked region.

16. The method of claim 14 where verifying comprises computing that the marked region of the heart does not extend further than half of the field of view.

17. The method of claim 1 comprising limiting the frequency to a range between approximately {fraction (1/2000)} ms and {fraction (1/500)} ms.

18. The method of claim 11 comprising limiting the frequency to a range between approximately {fraction (1/2000)} ms and {fraction (1/500)} ms.

19. The method of claim 1 comprising ranking the groups of slices.

20. The method of claim 19 comprising displaying the projection images of the selected slice sets in order based on their ranking.

21. The method of claim 19 wherein ranking comprises applying at least one quality measure to each of the groups of slices, wherein each of the groups of slices is used to generate a projection image.

22. The method of claim 21 wherein applying the quality measure comprises:

in a marked region of the projection image of the slices of the heart, summing pixel intensity values over a certain threshold value for each of the lines representing a slice in the projection images of the slices in each of the groups;

normalizing the pixel intensity values by a total number of pixels in the projection images to provide an average intensity value for the slices in each of the groups; and

comparing the average intensity values of each of groups and ranking the groups based on the average intensity value.

23. The method of claim 21 wherein applying the quality measure comprises:

in a marked region of the projection image of the heart, summing pixel intensity values for each of the lines representing a slice in the projection images of the slices in each of the groups; and

comparing the summed intensity values of each of groups and ranking the groups based on the summed intensity value.

24. The method of claim 21 wherein applying the quality measure comprises:

in a marked region of the projection image of the slices of the heart counting the number of pixels in each of the lines representing a slice in the slices in each of the groups that is above a pixel intensity threshold;

normalizing the counted number of pixels that are above the pixel intensity by dividing the counted number of pixels by a total number of pixels in the slices in the group;

comparing the normalized number of each group of slice images; and

ranking the groups of slice images using the comparison of normalized numbers.

25. The method of claim 21 wherein applying the quality measure comprises:

in a marked region of the projection image of the slices of the heart counting the number of pixels in each of the lines representing a slice in the slices in each of the groups that is above a pixel intensity threshold;

comparing the counted number of each group of slice images; and

ranking the groups of slice images using the comparison of counted numbers.

26. The method of claim 1 wherein the set of images are CT images.

27. The method of claim 1 wherein marking comprises selecting a region along a border of the patient'"'"'s heart.

28. The method of claim 1 comprising filling in any gaps in the groups of slice images.

29. The method of claim 28 comprising re-sampling the selected slice images to provide approximately equally spaced slice images in each group of slice images.

30. The method of claim 28 wherein filling in any gaps comprises applying a Fourier interpolation to generate a set of needed images.

31. The method of claim 28 wherein filling in any gaps comprises applying a nearest neighbor interpolation to generate a set of needed images.

32. The method of claim 28 wherein filling in any gaps comprises applying a linear interpolation to generate a set of needed images.

33. The method of claim 28 wherein filling in any gaps comprises applying a high order interpolation to generate a set of needed images.

35. The method of claim 34 wherein acquiring is carried out substantially continuously over an acquisition time period.

36. The method of claim 34 comprising measuring coronary calcium in the selected subset of slice images of the heart.

37. The method of claim 34 comprising 3-D rendering the selected subset of slice images of the heart.

38. The method of claim 34 wherein selecting is carried out through a pairwise comparison wherein the slice set is depopulated until a substantially contiguous and substantially non overlapping set of slice images is obtained.

39. The method of claim 34 wherein acquiring is carried out without an ECG.

40. The method of claim 34 wherein measuring comprises drawing an outline around a region around a border of the heart in at least one projection of the overlapping slice images and wherein selecting comprises choosing the slice image for which the image of the heart in the region is largest.

41. The method of claim 34 wherein marking comprises drawing, in at least one slice image, an outline around a portion that is less than a complete image of the heart.

42. The method of claim 34 wherein marking comprises drawing in at least one projection of the overlapping slices of images of the heart an outline around a plurality of different portions of the heart.

43. The method of claim 34 where the selecting of the largest image of the portion of the heart is limited so that a substantially contiguous and substantially non overlapping set of images is obtained.

34. A method of gating a set of overlapping slice images without the use of a separate gating signal, the method comprising:

acquiring a set of overlapping slice images;

marking a portion of a heart in at least one of the slice images;

measuring a size of at least a portion of the heart in each of the marked slice images; and

selecting a subset of slice images from the set of overlapping slices, wherein the subset of slice images comprise slice images comprising the largest image of the marked portion of the patient'"'"'s heart.

44. A method of Fourier gating an image dataset, the method comprising:

obtaining a plurality of overlapping slice images of a patient'"'"'s heart;

calculating an intensity signal for the overlapping slice images;

Fourier transforming the intensity signal to find a fundamental frequency of the patient'"'"'s heart cycle;

analyzing the intensity signal with a derivative filter to locate slice images that were obtained during diastole of the patient'"'"'s heart cycle; and

selecting slices that correspond to the patient'"'"'s diastole.

45. The method of claim 44 wherein Fourier transforming further comprises:

computing a power spectrum from a Fourier series transformation;

smoothing the Fourier series with a Gaussian filter; and

computing a maximum frequency from the power spectrum, wherein the maximum frequency corresponds to the fundamental frequency.

46. The method of claim 45 comprising restricting a search of the fundamental frequency range to be between approximately {fraction (1/2000)} ms and {fraction (1/500)} ms.

47. The method of claim 44 wherein Fourier transforming further comprises:

computing a power spectrum from a Fourier series transformation; and

computing a maximum frequency from the power spectrum, wherein the maximum frequency corresponds to the fundamental frequency.

48. The method of claim 44 further comprising generating and displaying a plurality of projections of groups of slice images, wherein each of the groups of slice images correspond to the patient'"'"'s heart in different phases of the patient'"'"'s heart cycle.

49. The method of claim 48 comprising ranking the groups of slice images based on heart size in the projection.

50. The method of claim 49 wherein ranking comprises applying at least one quality measure to each of the groups of slices.

51. The method of claim 44 comprising resampling the selected slices to substantially equilize the spacing between the selected slices.

52. The method of claim 44 comprising filling in gaps between the selected slices through selection of a slice closest to a center of a nearest two slices spanning the gap.

53. The method of claim 44 comprising filling in gaps between the selected slices through linear interpolation of a nearest two slices spanning the gap.

54. The method of claim 44 comprising filling in gaps between the selected slices through high order interpolation of slices spanning the gap.

55. A method of Fourier gating an image dataset, the method comprising:

obtaining a plurality of overlapping slice images of a patient'"'"'s heart;

generating at least one of a coronal and sagittal projection with the set of slice images;

marking a region of the projection;

calculating an intensity signal along a direction of the slice images for the projection in the marked region;

Fourier transforming the intensity signal to find a fundamental frequency of a patient'"'"'s heart cycle;

analyzing the intensity signal with a derivative filter to locate slice images that were obtained during a diastole of the patient'"'"'s heart cycle;

using the intensity signal analysis to establish the phase of the fundamental frequency obtained from the Fourier transformation of the heart motion;

extending the selection process outside the marked region by obtaining the frequency of the heart motion from the Fourier transformation and the phase from the intensity signal; and

selecting slices that correspond to the patient'"'"'s diastole.

57. The method of claim 56 wherein further data processing is coronary calcium scoring.

58. The method of claim 56 wherein the further processing is 3-D volume rendering.

59. The method of claim 56 wherein presenting comprises ranking each data set by a size of the heart within the marked region and visually indicating to the operator the ranking in the presentation.

56. A method of Fourier gating an image dataset, the method comprising:

obtaining a plurality of overlapping slice images of a patient'"'"'s heart;

generating at least one of a coronal and sagittal projection with the set of slice images;

marking a region of the projection;

calculating an intensity signal along a direction of the slice for the projection of the overlapping slice images within the marked region;

Fourier transforming the intensity signal to find a fundamental frequency of the patient'"'"'s heart cycle;

obtaining a principal component of a Fourier spectrum within an allowed frequency window;

forming data sets of slices separated by a time interval that substantially corresponds to the principal component; and

presenting a projection image formed from the data sets for an operator to select a set for further processing.