Serum Patterns Predictive of Breast Cancer

• US 20080312514A1
• Filed: 05/12/2006
• Published: 12/18/2008
• Est. Priority Date: 05/12/2005
• Status: Abandoned Application

• Current Assignee: Vermillion, Inc.
• Original Assignee: Vermillion, Inc.
• Inventors: Mansfield, Brian C.
• Application Number:
  US11/914,091
• Publication Number:
  US 20080312514A1
• Time in Patent Office: Days
• US Class Current: 600/300
• CPC Class Codes: G01N 33/57415 of breast

Abstract

Models for classifying a biological sample are developed from samples taken from a mammalian subject into one of at least two possible biological states related to breast cancer. Samples may be processed by mass spectral and other high-throughput analytical techniques.

Claims

1. A model for classifying a biological sample taken from a mammalian subject into one of at least two possible biological states related to breast cancer using a data stream that is obtained by performing a mass spectral analysis of the biological sample, the data stream including magnitude values for a range of mass-to-charge values, comprising:

at least one classifying hypervolume associated with one of the at least two biological states related to breast cancer and disposed within a vector space having n dimensions, each dimension corresponding to a different mass-to-charge value;

wherein n is at least three and at least a first of the dimensions corresponds to a mass-to-charge value in a range of m/z values selected from the m/z ranges consisting of between 200 to 300, 300 to 400, 400 to 500, 500 to 600, 600 to 700, and 700 to 900.

2. The model of claim 1, wherein n is at least 5.

3. The model of claim 1, wherein n is between 5 and 25.

4. The model of claim 1, wherein at least a second of the dimensions corresponds to a mass-to-charge value of between 500 and 1100.

5. The model of claim 1, wherein at least a second of the dimensions corresponds to a mass-to-charge value of between 500 and 900.

6. The model of claim 1, wherein at least a second of the dimensions corresponds to a mass-to-charge value of between 700 and 900.

7. The model of claim 1, the at least one classifying hypervolume being a first classifying hypervolume, further comprising:

a second classifying hypervolume disposed within the vector space;

the first classifying hypervolume being associated with a presence of breast cancer, the second classifying hypervolume being associated with an absence of breast cancer.

8. The model of claim 1, the at least one classifying hypervolume being a first classifying hypervolume, further comprising:

a second classifying hypervolume disposed within the vector space;

the first classifying hypervolume and the second classifying hypervolume being associated with a presence of breast cancer.

9. The model of claim 1, the at least one classifying hypervolume being a first classifying hypervolume, further comprising:

a second classifying hypervolume disposed within the vector space;

the first classifying hypervolume and the second classifying hypervolume being associated with an absence of breast cancer.

10. The model of claim 1, wherein the classifying hypervolume is associated with a presence of breast cancer.

11. The model of claim 10, wherein the classifying hypervolume is associated with a presence of in situ breast cancer.

12. The model of claim 10, wherein the classifying hypervolume is associated with a presence of invasive breast cancer.

13. The model of claim 10, wherein the classifying hypervolume is associated with a likelihood of metastasis of the invasive breast cancer.

14. The model of claim 1, wherein the classifying hypervolume is associated with an absence of breast cancer.

15. The model of claim 14, wherein the classifying hypervolume is associated with a benign breast condition.

16. The model of claim 14, wherein the benign breast condition is selected from the group consisting of hyperplasia, radial scar, calcification, and fibroadenoma.

17. The model of claim 14, wherein the classifying hypervolume is associated with a likelihood of a future occurrence of breast cancer

18. The model of claim 1, wherein the model has at least a 65% accuracy.

19. The model of claim 1, wherein the model has at least a 70% accuracy.

20. The model of claim 1, wherein the model has at least a 80% sensitivity.

21. The model of claim 1, wherein the model has at least a 80% specificity.

22. The model of claim 1, where in the hypervolume is a hypersphere.

62. The model of claim 1, wherein the first of the dimensions corresponds to a mass-to-charge value of between 520 and 590.

63. The model of claim 1, wherein the first of the dimensions corresponds to a mass-to-charge value of about 537.

64. The model of claim 1, wherein the first of the dimensions corresponds to a mass-to-charge value of about 579.

65. The model of claim 1, wherein the first of the dimensions corresponds to a mass-to-charge value of between 535 and 540.

66. The model of claim 1, wherein the first of the dimensions corresponds to a mass-to-charge value of between 575 and 580.

67. The model of claim 1, wherein the second of the dimensions corresponds to a mass-to-charge value of about 827.

68. The model of claim 1, wherein the second of the dimensions corresponds to a mass-to-charge value of between 820 and 830.

23. A method of classifying a biological sample taken from a subject into one of at least two possible biological states related to breast cancer by analyzing a data stream that is obtained by performing a mass spectral analysis of the biological sample, the data stream including magnitude values for a range of mass-to-charge values, comprising:

abstracting the data stream to produce a sample vector that characterizes the data stream in a vector space having n dimensions and containing a diagnostic hypervolume, the vector space having at least a first dimension, a second dimension, and a third dimension, the first dimension corresponding to a mass-to-charge value of between 500 and 600, the second dimension corresponding to a mass-to-charge value of between 700 and 900, the diagnostic hypervolume corresponding to one of the presence or absence of breast cancer; and

determining whether the sample vector rests within the diagnostic hypervolume.

24. The method of claim 23, wherein the hypervolume corresponds to the presence of breast cancer and further comprising:

if the sample vector rests within the diagnostic hypervolume, identifying the biological sample as indicating that the subject has breast cancer.

25. The method of claim 23, wherein the third dimension corresponds to a mass-to-charge value of between 500 and 1100.

26. The method of claim 23, wherein the third dimension corresponds to a mass-to-charge value of between 500 and 900.

27. The method of claim 23, the diagnostic hypervolume is a first diagnostic hypervolume, wherein the vector space contains a second diagnostic hypervolume, the first diagnostic hypervolume and the second diagnostic hypervolume corresponding to the presence of breast cancer.

28. The model of claim 23, the diagnostic hypervolume is a first diagnostic hypervolume corresponding to the presence of breast cancer, wherein the vector space contains a second diagnostic hypervolume, the second diagnostic hypervolume corresponding to an absence of breast cancer.

29. The method of claim 23, wherein the hypervolume is a hypersphere.

30. The method of claim 23, wherein the hypervolume corresponds to the presence of in situ breast cancer.

31. The method of claim 23, wherein the hypervolume corresponds to the presence of invasive breast cancer.

32. The method of claim 24, wherein the hypervolume corresponds to the absence of breast cancer and to the presence of a benign breast condition.

33. The model of claim 32, wherein the benign breast condition is selected from the group consisting of hyperplasia, radial scar, calcification, and fibroadenoma.

34. A model for classifying a biological sample taken from a mammalian subject into one of at least two possible biological states related to breast cancer using a data stream that is obtained by performing an mass spectral analysis of the biological sample, the data stream including magnitude values for a range of mass-to-charge values, comprising:

at least one classifying hypervolume disposed within an vector space having n-dimensions, each dimension corresponding to a different mass-to-charge value,

wherein n is greater than three, at least two of the dimensions correspond to mass-to-charge values in table A.

35. The model of claim 34, wherein at least three of the dimensions correspond to mass-to-charge values in table A.

36. The model of claim 34, wherein n is between 5 and 25.

37. The model of claim 34, the at least one classifying hypervolume being a first classifying hypervolume, further comprising:

a second classifying hypervolume disposed within the vector space,

the first classifying hypervolume being associated with a presence of breast cancer, the second classifying hypervolume being associated with an absence of breast cancer.

38. The model of claim 34, the at least one classifying hypervolume being a first classifying hypervolume, further comprising:

a second classifying hypervolume disposed within the vector space,

the first classifying hypervolume and the second classifying hypervolume being associated with a presence of breast cancer.

39. The model of claim 34, the at least one classifying hypervolume being a first classifying hypervolume, further comprising:

a second classifying hypervolume disposed within the vector space,

the first classifying hypervolume and the second classifying hypervolume being associated with an absence of breast cancer.

40. The model of claim 34, wherein the classifying hypervolume is associated with a presence of breast cancer.

41. The model of claim 40, wherein the classifying hypervolume is associated with a presence of in situ breast cancer.

42. The model of claim 40, wherein the classifying hypervolume is associated with a presence of invasive breast cancer.

43. The model of claim 42, wherein the classifying hypervolume is associated with a likelihood of metastasis of the invasive breast cancer.

44. The model of claim 34, wherein the classifying hypervolume is associated with an absence of breast cancer.

45. The model of claim 44, wherein the classifying hypervolume is associated with a benign breast condition.

46. The model of claim 45, wherein the benign breast condition is selected from the group consisting of hyperplasia, radial scar, calcification, and fibroadenoma.

47. The model of claim 44, wherein the classifying hypervolume is associated with a likelihood of a future occurrence of breast cancer.

48. The model of claim 34, wherein the model has at least a 65% accuracy.

49. The model of claim 34, wherein the model has at least a 70% accuracy.

50. The model of claim 34, wherein the model has at least a 80% sensitivity.

51. The model of claim 34, wherein the model has at least a 80% specificity.

52. A model for classifying a biological sample taken from a mammalian subject using a data stream that is obtained by performing a mass spectral analysis of the biological sample, the data stream including magnitude values for a range of mass-to-charge values, comprising:

at least one classifying hypervolume disposed within a vector space having n dimensions, each dimension corresponding to a different mass-to-charge value,

wherein n is at least three, at least a first of the dimensions corresponds to a mass-to-charge value of between 500 and 600, at least a second of the dimensions corresponds to a mass-to-charge value of between 600 and 700.

53. The model of claim 52, wherein n is at least 5.

54. The model of claim 52, wherein n is between 5 and 25.

55. The model of claim 52, wherein the model has at least a 65% accuracy.

56. The model of claim 52, wherein the model has at least a 70% accuracy.

57. A model for classifying a biological sample taken from a mammalian subject using a data stream that is obtained by performing a mass spectral analysis of the biological sample, comprising:

at least two classifying hypervolumes disposed within a vector space having at least three dimensions, one of the at least two classifying hypervolumes being associated with a presence of a disease, another of the at least two classifying hypervolumes being associated with an absence of the disease,

the model having at least a 65% accuracy.

58. The model of claim 57, wherein the vector space has at least 5 dimensions.

59. The model of claim 57, wherein the disease is breast cancer.

60. The model of claim 57, wherein the data stream includes magnitude values for a range of mass-to-charge values, a first of the at least three dimensions corresponds to a mass-to-charge value of between 500 and 600, and a second of the at least three dimensions corresponds to a mass-to-charge value of between 600 and 700.

61. The model of claim 57, wherein the data stream includes magnitude values for a range of mass-to-charge values, at least two of the at least three dimensions correspond to mass-to-charge values in table 1.

69. A model for classifying a biological sample taken from a mammalian subject into one of at least two possible biological states using a data stream that is obtained by performing a mass spectral analysis of the biological sample, the data stream including magnitude values for a range of mass-to-charge values, comprising:

at least one classifying hypervolume associated with the presence of ductal carcinoma in situ and disposed within a vector space having n dimensions, each dimension corresponding to a different mass-to-charge value.

70. The model of claim 69, wherein n is at least three, at least a first of the dimensions corresponds to a mass-to-charge value of between 900 and 905, and at least a second of the dimensions corresponds to a mass-to-charge value of between 610 and 620.

71. The model of claim 69, the at least one classifying hypervolume being a first classifying hypervolume, further comprising:

a second classifying hypervolume associated with the presence of lobular carcinoma in situ and disposed within the vector space.

72. A model for classifying a biological sample taken from a mammalian subject into one of at least two possible biological states using a data stream that is obtained by performing a mass spectral analysis of the biological sample, the data stream including magnitude values for a range of mass-to-charge values, comprising:

at least one classifying hypervolume associated with the presence of lobular carcinoma in situ and disposed within a vector space having n dimensions, each dimension corresponding to a different mass-to-charge value.

73. The model of claim 72, wherein n is at least three, at least a first of the dimensions corresponds to a mass-to-charge value of between 1050 and 1060, and at least a second of the dimensions corresponds to a mass-to-charge value of between 610 and 620.

74. A model for classifying a biological sample taken from a mammalian subject into one of at least two possible biological states associated with breast pathology using a data stream that is obtained by performing a mass spectral analysis of the biological sample, the data stream including magnitude values for a range of mass-to-charge values, comprising:

at least one classifying hypervolume disposed within a vector space having n dimensions, each dimension corresponding to a different mass-to-charge value.