Systems and methods for treating, diagnosing and predicting the occurrence of a medical condition

• US 7,467,119 B2
• Filed: 03/14/2005
• Issued: 12/16/2008
• Est. Priority Date: 07/21/2003
• Status: Expired due to Term

• Current Assignee: David Sans
• Original Assignee: Aureon Biosciences, Inc.
• Inventors: Verbel, David A., Saidi, Olivier, Teverovskiy, Mikhail
• Primary Examiner(s): Holmes; Michael B
• Application Number:
  US11/080,360
• Publication Number:
  US 20050262031A1
• Time in Patent Office: 1,373 Days
• Field of Search: 706/21
• US Class Current: 706/21
• CPC Class Codes: G16H 50/20 for computer-aided diagnosis, e.g. based on medical expert systems
• CPC Class Codes: G16H 50/70 for mining of medical data, e.g. analysing previous cases of other patients
• CPC Class Codes: G16Z 99/00 Subject matter not provided for in other main groups of this subclass

Abstract

Methods and systems are provided that use clinical information, molecular information and computer-generated morphometric information in a predictive model for predicting the occurrence (e.g., recurrence) of a medical condition, for example, cancer.

Claims

1. A method of evaluating a risk of occurrence of a medical condition in a patient, the method comprising:

receiving a patient dataset for the patient; and

evaluating the patient dataset with a model predictive of the medical condition to produce a value indicative of the risk of occurrence of the medical condition in the patient, wherein the model is based on one or more clinical feature(s), one or more molecular feature(s), and one or more computer-generated morphometric feature(s) generated from one or more tissue image(s).

2. The method of claim 1, further comprising generating with a computer one or more morphometric feature(s) from a tissue image for the patient for inclusion in the patient dataset, wherein said generating the morphometric feature(s) comprises:

segmenting the tissue image into one or more objects;

classifying the one or more objects into one or more object classes; and

determining the morphometric features by taking one or more measurements for the one or more object classes.

3. The method of claim 2, wherein said classifying the one or more objects into one or more object classes comprises classifying each of one or more of the objects into a class from the group of classes consisting of stroma, cytoplasm, epithelial nuclei, stroma nuclei, lumen, red blood cells, tissue artifacts, and tissue background.

4. The method of claim 2, wherein said taking one or more measurements pertaining to the one or more object classes comprises taking one or more measurements of one or more spectral properties and/or one or more shape properties of the one or more object classes.

5. The method of claim 1, wherein the model is predictive of prostate cancer recurrence and one or more of the morphometric feature(s) are from the group of morphometric features consisting of a morphometric feature of a red blood cell, a morphometric feature of epithelial nuclei, a morphometric feature of stroma, a morphometric feature of lumen, a morphometric feature of cytoplasm, and a morphometric feature of tissue background.

6. The method of claim 5, wherein one or more of the clinical feature(s) are from the group of clinical features consisting of biopsy Gleason score, race, UICC stage, ploidy result, DRE result, lymph node involvement, dominant biopsy Gleason grade, percent ploidy in S phase, post-operative Gleason score, TNM stage, dominant post-operative Gleason grade, age, seminal vesicle involvement, pre-operative PSA, percent ploidy fraction. surgical margin involvement, and extracapsular involvement.

7. The method of claim 6, wherein one or more of the molecular feature(s) are from the group of molecular features consisting of AR-tumor, AR-gland, CD34-tumor/PIN, Ki67-tumor 2, CD45-PIN 3, CD34-tumor/stroma, Ki67-tumor 3, p27-tumor, C14-PIN, CD34-tumor, PSA-gland, PSMA-PIN, CD34-PIN/stroma, and CD45-tumor 3.

8. The method of claim 7, wherein the predictive model comprises a concordance index of at least about 0.88.

9. The method of claim 7, wherein the predictive model comprises a p value less than about 0.0001 for a log-rank test.

10. The method of claim 1, wherein the model is predictive of prostate cancer recurrence and one or more of the morphometric feature(s) are from the group of morphometric features consisting of a morphometric feature of a red blood cell, a morphometric feature of epithelial nuclei, a morphometric feature of stroma, a morphometric feature of lumen, and a morphometric feature of cytoplasm.

11. The method of claim 10, wherein one or more of the clinical feature(s) are from the group of clinical features consisting of TNM clinical stage, surgical margins, and lymph nodes.

12. The method of claim 11, wherein the one or more molecular feature(s) comprises AR staining index (tumor).

13. The method of claim 12, wherein the predictive model comprises a concordance index of at least about 0.87.

14. The method of claim 12, wherein the predictive model comprises a p value less than about 0.0001 for a log-rank test.

15. The method of claim 1, wherein the model is predictive of prostate cancer survival and one or more of the morphometric feature(s) are from the group of morphometric features consisting of a morphometric feature of a red blood cell, a morphometric feature of epithelial nuclei, and a morphometric feature of stroma.

16. The method of claim 15, wherein one or more of the clinical feature(s) are from the group of clinical features consisting of tnm and age.

17. The method of claim 16, wherein the one or more molecular feature(s) comprises psapsi.

18. The method of claim 17, wherein the predictive model comprises a concordance index of at least about 0.80.

19. The method of claim 17, wherein the predictive model comprises a p value less than about 0.0001 for a log-rank test.

20. The method of claim 1, wherein said evaluating the patient dataset with the predictive model comprises determining a diagnostic score for the patient.

21. The method of claim 1, wherein said evaluating the risk of occurrence of a medical condition comprises determining a likely time to occurrence of the medical condition in the patient.

22. The method of claim 1, wherein said evaluating the risk of occurrence of a medical condition comprises determining a likely responsiveness or unresponsiveness of the patient to a therapy.

23. The method of claim 1, further comprising outputting data indicating results of the evaluation.

24. The method of claim 23, wherein the results comprise results from the group of results consisting of a diagnostic score, information indicating one or more features of the patient dataset that were analyzed by the predictive model, information indicating an accuracy of the predictive model, or a combination thereof.

25. The method of claim 23, wherein said receiving a patient dataset for the patient comprises receiving data for the patient from a remote location and wherein said outputting data indicating results of the evaluation comprises transmitting the results to the remote location.

26. The method of claim 25, wherein said receiving and said transmitting comprise receiving and transmitting over one or more communications networks.

27. An apparatus for evaluating the risk of occurrence of a medical condition in a patient, the apparatus comprising:

a model predictive of the medical condition, wherein the model is based on one or more clinical feature(s), one or more molecular feature(s), and one or more computer-generated morphometric feature(s) generated from one or more tissue image(s), wherein the model is configured to;

receive a patient dataset for the patient; and

evaluate the patient dataset according to the model to produce a value indicative of the risk of occurrence of the medical condition in the patient.

28. The apparatus of claim 27, further comprising an image processing tool configured to generate one or more morphometric feature(s) from a tissue image for the patient for inclusion in the patient dataset, wherein said generating the morphometric feature(s) comprises:

segmenting the tissue image into one or more objects with the image processing tool;

classifying the one or more objects into one or more object classes by the image processing tool; and

determining the morphometric features by taking one or more measurements for the one or more object classes with the image processing tool.

29. The apparatus of claim 28, wherein said classifying the one or more objects into one or more object classes by the image processing tool comprises classifying by the image processing tool each of one or more of the objects into a class from the group of classes consisting of stroma, cytoplasm, epithelial nuclei, stroma nuclei, lumen, red blood cells, tissue artifacts, and tissue background.

30. The apparatus of claim 28, wherein said taking one or more measurements pertaining to the one or more object classes with the image processing tool comprises taking with the image processing tool one or more measurements of one or more spectral properties and/or one or more shape properties of the one or more object classes.

31. The apparatus of claim 27, wherein the model is predictive of prostate cancer recurrence and one or more of the morphometric feature(s) are from the group of morphometric features consisting of a morphometric feature of a red blood cell, a morphometric feature of epithelial nuclei, a morphometric feature of stroma, a morphometric feature of lumen, a morphometric feature of cytoplasm, and a morphometric feature of tissue background.

32. The apparatus of claim 31, wherein one or more of the clinical feature(s) are from the group of clinical features consisting of biopsy Gleason score, race, UICC stage, ploidy result, DRE result, lymph node involvement, dominant biopsy Gleason grade, percent ploidy in S phase, post-operative Gleason score, TNM stage, dominant post-operative Gleason grade, age, seminal vesicle involvement, pre-operative PSA, percent ploidy fraction, surgical margin involvement, and extracapsular involvement.

33. The apparatus of claim 32, wherein one or more of the molecular feature(s) are from the group of molecular features consisting of AR-tumor, AR-gland, CD34-tumor/PIN, Ki67-tumor 2, CD45-PIN 3, CD34-tumor/stroma, Ki67-tumor 3, p27-tumor, C14-PIN, CD34-tumor, PSA-gland, PSMA-PIN, CD34-PIN/stroma, and CD45-tumor 3.

34. The apparatus of claim 33, wherein the predictive model comprises a concordance index of at least about 0.88.

35. The apparatus of claim 33, wherein the predictive model comprises a p value less than about 0.0001 for a log-rank test.

36. The apparatus of claim 27, wherein the model is predictive of prostate cancer recurrence and one or more of the morphometric feature(s) are from the group of morphometric features consisting of a morphometric feature of a red blood cell, a morphometric feature of epithelial nuclei, a morphometric feature of stroma, a morphometric feature of lumen, and a morphometric feature of cytoplasm.

37. The apparatus of claim 36, wherein one or more of the clinical feature(s) are from the group of clinical features consisting of TNM clinical stage, surgical margins, and lymph nodes.

38. The apparatus of claim 37, wherein the one or more molecular feature(s) comprises AR staining index (tumor).

39. The apparatus of claim 38, wherein the predictive model comprises a concordance index of at least about 0.87.

40. The apparatus of claim 38, wherein the predictive model comprises a p value less than about 0.0001 for a log-rank test.

41. The apparatus of claim 27, wherein the model is predictive of prostate cancer survival and one or more of the morphometrie feature(s) are from the group of morphometric features consisting of a morphometric feature of a red blood cell, a morphometric feature of epithelial nuclei, and a morphometric feature of stroma.

42. The apparatus of claim 41, wherein one or more of the clinical feature(s) are from the group of clinical features consisting of tnm and age.

43. The apparatus of claim 42, wherein the one or more molecular feature(s) comprises psapsi.

44. The apparatus of claim 43, wherein the predictive model comprises a concordance index of at least about 0.80.

45. The apparatus of claim 43, wherein the predictive model comprises a p value less than about 0.0001 for a log-rank test.

46. The apparatus of claim 27, wherein the predictive model is configured to determine a diagnostic score for the patient.

47. The apparatus of claim 27, wherein the predictive model is configured to determine a likely time to occurrence of the medical condition in the patient.

48. The apparatus of claim 27, wherein the predictive model is configured to determine a likely responsiveness or unresponsiveness of the patient to a therapy.

49. The apparatus of claim 27, wherein the predictive model is further configured to output data indicating results of the evaluation.

50. The apparatus of claim 49, wherein the results comprise results selected from the group of results consisting of a diagnostic score, information indicating one or more features of the patient dataset that were analyzed by the predictive model, information indicating an accuracy of the predictive model, or a combination thereof.

51. The apparatus of claim 49, wherein said predictive model is configured to receive the patient dataset for the patient from a remote location and output the results for transmission to the remote location.

52. The apparatus of claim 51, wherein the patient dataset is received and the results are transmitted over one or more communications networks.

53. The apparatus of claim 27, wherein said predictive model comprises a neural network.

54. The apparatus of claim 27, wherein said predictive model comprises a support vector machine.

55. An apparatus for evaluating the risk of occurrence of a medical condition in a patient, the apparatus comprising:

a model predictive of the medical condition, wherein the model is based on one or more computer-generated morphometric feature(s) generated from one or more tissue image(s) and wherein the model is configured to;

receive a patient dataset for the patient; and

evaluate the patient dataset according to the model to produce a value indicative of the risk of occurrence of the medical condition in the patient.

56. The apparatus of claim 55, wherein the patient dataset comprises a patient dataset based on a liver tissue image for the patient and wherein the predictive model is configured to determine whether the liver tissue is normal or abnormal.

57. The apparatus of claim 55, wherein said model is based on the one or more computer-generated morphometric feature(s) and one or more clinical feature(s).

58. The apparatus of claim 57, wherein the patient dataset comprises a patient dataset based on a prostate tissue image for the patient and wherein the predictive model is configured to make a prediction with respect to prostate cancer recurrence for the patient.

59. The apparatus of claim 57, wherein the patient dataset comprises a patient dataset based on a prostate tissue image for the patient and wherein the predictive model is configured to make a prediction with respect to clinical failure for the patient.

60. A computer readable medium comprising computer executable instructions recorded thereon for performing the method comprising:

receiving a patient dataset for a patient; and

evaluating the patient dataset with a model predictive of a medical condition to produce a value indicative of the risk of occurrence of the medical condition in the patient, wherein the model is based on one or more clinical feature(s), one or more molecular feature(s), and one or more computer-generated morphometric feature(s) generated from one or more tissue image(s).