Methods and software for improving thresholding of coronary calcium scoring

• US 7,149,331 B1
• Filed: 09/03/2002
• Issued: 12/12/2006
• Est. Priority Date: 09/03/2002
• Status: Expired due to Fees

• Current Assignee: International Business Machines Corporation
• Original Assignee: Cedara Software, Corp (International Business Machines Corporation)
• Inventors: Greiss, Friederike, Kaufman, Leon
• Primary Examiner(s): Ahmed, Samir
• Assistant Examiner(s): Mistry, O'Neal R.
• Application Number:
  US10/234,984
• Time in Patent Office: 1,561 Days
• Field of Search: None
• US Class Current: 382/128
• CPC Class Codes: G06T 2207/30101 Blood vessel; Artery; Vein; Vascular
• CPC Class Codes: G06T 7/0012 Biomedical image inspection
• CPC Class Codes: G16H 50/30 for calculating health indices; for individual health risk assessment

Abstract

The present invention provides methods and software for improving thresholding of coronary calcium scoring. In exemplary embodiments, the present invention applies an automatic search algorithm to slice images of an image scan to generate an individualized signal threshold for each particular patient. In other embodiments, an individualized signal threshold is generated for each slice image of the image scan.

Claims

1. A method for improving coronary calcium scoring of an image scan comprising a plurality of slice images, the method comprising:

applying an automatic search algorithm to at least one slice image to calculate an individualized signal threshold for the at least one slice image;

applying the individualized signal threshold to the at least one slice image to identify calcium therein; and

selecting a region of interest in the at least one slice image in one or more areas that cannot have calcium, wherein the automatic search algorithm is applied to the region of interest to calculate the individualized signal threshold for the at least one slice image,

wherein applying the automatic search algorithm comprises;

selecting a lesion criteria, the lesion criteria comprising an acceptable number N of false positives that can be located in the region of interest;

applying a first signal threshold HU₁ to the region of interest to locate false positives that have signal higher than the first signal threshold HU₁;

counting the number of false positives located in the region of interest with the first signal threshold HU₁;

comparing N with the number of false positives located in the region of interest with the first signal threshold HU₁; and

adjusting the first signal threshold HU₁ until the number of false positives in the region of interest is less than or equal to N, wherein the individualized signal threshold for the at least one slice image is set substantially close to an adjusted signal threshold HU in which the number of false positives is less than or equal to N.

2. The method of claim 1 wherein the false positives comprise a grouping of pixels that have a minimum area, a minimum signal, and a minimum connectedness.

3. The method of claim 1 wherein the lesion criteria further comprises at least one of a minimum connectedness and/or a minimum area of pixels.

4. The method of claim 3 wherein selecting the lesion criteria comprises prompting an operator to provide at least one of the minimum connectedness, the minimum area of pixels, and/or a value representative of N.

5. The method of claim 4 wherein prompting the operator comprises having an operator provide each of the minimum connectedness, the minimum area of pixels, and the value representative of N before applying the automatic search algorithm.

6. The method of claim 3 wherein the minimum area of pixels is approximately 1 mm².

7. The method of claim 3 wherein the minimum connectedness comprises between approximately 1 pixel and 8 pixels.

8. The method of claim 1 wherein the value of N comprises zero or an integer.

9. The method of claim 1 wherein the value of N is a percent probability that a lesion is false.

10. The method of claim 1 wherein the first signal threshold HU₁ is between approximately 0 HU and 50 HU.

11. The method of claim 1 wherein adjusting the first signal threshold HU₁ comprises performing at least one of raising or lowering the signal threshold according to a predetermined program.

12. The method of claim 11 wherein the predetermined program comprises:

repetitively raising the first signal threshold HU₁ a first step until the number of located false positives becomes N or smaller;

repetitively lowering the raised first signal threshold a second step that is smaller than the first step until the number of located false positives exceeds N; and

repetitively raising the signal threshold by a third step that is smaller than the second step until the number of located false positives becomes N or smaller.

13. The method of claim 12 wherein the first step is greater than the second step, and the second step is greater than the third step.

14. The method of claim 12 wherein the first step is between approximately 10 HU and 50 HU, the second step is between approximately 5 HU and 15 HU, and the third step is approximately 1 HU and 5 HU.

15. The method of claim 12 wherein the first step is between 10 and 50 HU.

16. The method of claim 12 wherein the first step is between 1 and 10 HU.

17. The method of claim 12 wherein the predetermined program further comprises setting the signal threshold for the at least one slice image after a predetermined number n of steps,

wherein the signal threshold is selected as a HU value where one step increase of value n changes the number of false positives from a value larger than N, to N.

18. The method of claim 17 wherein the step n equals 1 HU.

19. The method of claim 17 where the step n equals 5 HU.

20. The method of claim 1 wherein adjusting the first signal threshold HU₁ comprises repeatedly raising or lowering the signal threshold by steps of 1 HU.

21. The method of claim 1 wherein adjusting the first signal threshold HU₁ comprises repeatedly raising or lowering the signal threshold by steps of 5 HU or less.

22. The method of claim 1 wherein if the first threshold HU₁ results in N false positives or less, then adjusting the first signal threshold comprises:

reducing the first signal threshold HU₁ by a first step until the number of false positives exceeds N; and

increasing the reduced first signal threshold HU₁ by a second step until the number of false positives is less than or equal to N,

wherein the first step is greater than the second step.

23. The method of claim 1 wherein adjusting the first signal threshold HU₁ comprises raising the first signal threshold by steps of 1 HU until N false positives are found.

24. The method of claim 1 wherein the region of interest that cannot have calcium comprises a blood pool, an aorta, or a heart muscle.

25. The method of claim 1 wherein the at least one slice image is obtained with a mechanical CT scanner, an electron beam CT scanner, or a helical CT scanner.

26. The method of claim 1 comprising using the identified amount of calcium from the at least one slice image to calculate a patient'"'"'s calcium volume.

27. The method of claim 1 comprising using the identified amount of calcium from the at least one slice image to calculate a patient'"'"'s Agatston score.

28. The method of claim 1 comprising calculating a mean HU and a standard deviation of the region of interest, wherein the first signal threshold HU₁ is equal to approximately the mean HU plus one standard deviation.

29. The method of claim 1 wherein applying the first signal threshold HU₁ comprises analyzing a size of a lesion relative to a size of the region of interest, wherein if the size of the lesion is larger than a predetermined percentage, then the signal threshold is increased.

30. A method of generating an individualized signal threshold for each slice image of an image scan, the method comprising:

selecting at least one region of interest in each of the plurality of slice images;

generating a signal threshold for each of the plurality of slice images by applying an automatic search algorithm to the selected region of interest in each of the slice images comprising steps of;

applying a first signal threshold HU₁ to the first region of interest;

counting the number of false positives located in the region of interest that are located with the first signal threshold HU₁;

comparing the number of counted false positives with the value of N; and

performing at least one of lowering or raising the first signal threshold HU₁ until a signal threshold is found that generates N false positives or lower, wherein the individualized signal threshold for the slice image is et substantially closes to an adjusted signal threshold in with N false positives are found; and

selecting lesion criteria, wherein the lesion criteria comprises a minimum area of the lesion, a minimum connectedness of the lesion, and a number N of acceptable false positives in the region of interest, wherein the signal threshold is generated based on the lesion criteria.

31. The method of claim 30 wherein the region or interest is operator selected, wherein the region of interest is an area of the slice image that does not contain calcium.

32. The method of claim 31 wherein the region of interest is a blood pool, an aorta, or a heart muscle.

33. The method of claim 30 wherein the image slices are obtained with a mechanical CT scanner, an electron beam CT scanner, or a helical CT scanner.

34. The method of claim 30 comprising using the identified amount of calcium from each slice image to calculate a patient'"'"'s calcium volume.

35. The method of claim 30 comprising using the identified amount of calcium from each slice image to calculate a patient'"'"'s Agatston score.

36. A method of using an operator-defined lesion criteria to generate an individualized signal threshold for individual slice images of an image scan, the method comprising:

selecting the lesion criteria for locating false positives in a region of interest in each slice image, wherein the lesion criteria includes an operator-provided acceptable number of false positives, represented by N; and

computer processing the slice image according to the lesion criteria to generate the individualized signal threshold for each slice images,

wherein the region of interest is examined with a first threshold and a number of false positives that meet the criteria of calcium lesions is counted until that number exceeds N, wherein computer processing comprises;

repetitively raising the first threshold by an incremental step S until the number of false positives that meet the criteria of calcium lesions counted no longer exceeds N;

repetitively lowering the first threshold by an incremental step size s, which is smaller than S, until the number of false positives that meet the criteria of calcium lesions exceeds N; and

repetitively raising the first threshold by a step z until the number of false positives counted no longer exceeds N,

wherein the signal threshold for the slice image is set substantially close to the value of the signal threshold value where the criteria of the calcium lesion counted no longer exceeds the value N.

37. The method of claim 36 wherein the region of interest is placed by the operator over a substantially homogeneous area in the slice image.

38. The method of claim 36 wherein the criteria further comprises at least one of a minimum connectedness and/or a minimum area.

39. The method of claim 38 wherein the region of interest is examined with a first threshold and the number of false positives that meet the criteria of calcium lesions is counted, wherein if the first threshold results in a number of false positives that is lower than or equal to the value N, computer processing comprises:

repetitively lowering the threshold by steps of 1 HU until the number of false positives that meet the criteria of calcium lesions counted is at N or exceeds the value N; and

setting the threshold substantially at the signal threshold value when the number of lesions is at or below N.

40. The method of claim 36 wherein the step S is between approximately 25 HU to 50 HU, step s is between approximately 10 HU and 25 HU, and step z is between approximately 1 and 10 HU.

41. The method of claim 36 where the step z is 1 HU.

42. The method of claim 36 where the step z is between 2 HU and 10 HU.

43. The method of claim 36 wherein substantially close is a value one HU step z below a signal threshold where the number of located false positives becomes N or smaller.

44. The method of claim 36 wherein the region of interest is examined with a first threshold and a number of false positives that meet the criteria of calcium lesions is counted, wherein if the first threshold results in a number of areas that are at or below N, computer processing comprises:

repetitively lowering the first threshold by an incremental step S until the number of false positives counted exceeds the value N;

repetitively raising the threshold by an incremental step size s, which is smaller than S, until the number of false positives is below the value N,

wherein the signal threshold for the slice image is set substantially close to the value of the signal threshold value where the false positives counted no longer exceeds N.

45. The method of claim 44 wherein the step s is between approximately 1 HU and 10 HU.

46. The method of claim 44 wherein the step s is between approximately 10 HU and 25 HU.

47. The method of claim 36 wherein the region of interest is examined with a first threshold and the number of false positives that meet the criteria of calcium lesions is counted until the number of false positives exceeds a value N;

and,

repetitively raising the threshold by steps of 1 HU until the number of false positives counted no longer exceeds the value N; and

setting the signal threshold for the slice image substantially close to the value of signal threshold where the number of counted false positives is at or below N.

48. The method of claim 36 wherein computer processing comprises:

applying a first signal threshold HU₁ to the region of interest; and

counting the number of false positives located in the region of interest,

wherein if the number of false positives exceeds N, the method comprises;

repetitively raising the first signal threshold by steps S until the number of false positives becomes N or smaller and storing the raised signal threshold HU_S;

returning to the HU value of the threshold just prior to the one where HU_s was reached;

repetitively raising the first signal threshold by steps s, that is smaller than steps S, until the number of false positives becomes N or smaller and storing the raised signal threshold HU_s;

returning to the HU value of the threshold just prior to the one where HU_s;

was reached;

repetitively raising the first signal threshold a steps z that is smaller than steps s, until the number of false positives becomes N or smaller and storing the raised signal threshold HU_z; and

setting the threshold at a signal level that is substantially the one where the number of located false positives becomes N.

49. The method of claim 48 where step S is between approximately 25 HU to 50 HU, step s is between approximately 10 HU and 25 HU, and step z is 1 HU.

50. The method of claim 48 where step S is between approximately 25 HU to 50 HU, step s is between approximately 10 HU and 25 HU, and step z is between 2 and 10 HU.

51. The method of claim 48 where substantially close is a value one step z below the value where the number of located false positives becomes N.

52. The method of claim 36 wherein computer processing comprises:

applying a first signal threshold HU₁ to the region of interest;

counting the number of false positives located in the region of interest;

wherein if the number of false positives is less than N the method comprises;

repetitively lowering the first signal threshold by steps S until the number of false positives becomes N or larger and storing the raised signal threshold HU_S;

returning to the HU value of the threshold just prior to the one where HU_S was reached;

repetitively lowering the first signal threshold by steps s, that is smaller than steps S, until the number of false positives becomes N or larger and storing the raised signal threshold HU_s;

returning to the HU value of the threshold just prior to the one where HU_s;

was reached;

repetitively lowering the first signal threshold by steps z that is smaller than steps s, until the number of false positives becomes N or larger and storing the raised signal threshold HU_z; and

setting the threshold at a signal level that is substantially the one where the number of located false positives becomes N.

53. The method of claim 52 where step S is between approximately 25 HU to 50 HU, step s is between approximately 10 HU and 25 HU, and step z is 1 HU.

54. The method of claim 52 where step S is between approximately 25 HU to 50 HU, step s is between approximately 10 HU and 25 HU, and step z is between 2 and 10 HU.

55. The method of claim 52 where substantially close is a value one step z below the value where the number of located false positives becomes N.

56. A method of generating a signal threshold for each slice image of an image scan, the method comprising:

selecting a slice image from among a plurality of slice images comprising the image scan;

selecting at least one region of interest in the selected slice image;

generating a signal threshold for the selected slice image by applying an automatic search algorithm to the selected region of interest comprising steps of;

applying a first signal threshold HU₁ to the first region of interest;

counting the number of false positives located in the region of interest that are located with the first signal threshold HU₁;

comparing the number of counted false positives with the value of N; and

performing at least one of lowering or raising the first signal threshold HU₁ until a signal threshold is found that generates N false positives or lower, wherein the individualized signal threshold for the slice image is et substantially closes to an adjusted signal threshold in with N false positives are found; and

selecting lesion criteria, wherein the lesion criteria comprises a minimum area of the lesion, a minimum connectedness of the lesion, and a number N of acceptable false positives in the region of interest, wherein the signal threshold is generated based on the lesion criteria; and

applying the signal threshold to the plurality of slice images to identify calcium in the slice images.

57. The method of claim 56 wherein the selected slice image is a slice image near a middle of the plurality of slice images.

58. The method of claim 56 wherein the selected slice image is one of a highest slice image from the plurality of slice images, a lowest slice image from the plurality of slice images, or a median slice image from the plurality of slice images.