Methods for predicting an individual's clinical treatment outcome from sampling a group of patient's biological profiles

• US 9,342,657 B2
• Filed: 03/24/2003
• Issued: 05/17/2016
• Est. Priority Date: 03/24/2003
• Status: Active Grant

• Current Assignee: Nien-Chih Wei
• Original Assignee: Nien-Chih Wei
• Inventors: Wei, Nien-Chih
• Primary Examiner(s): NEGIN, RUSSELL SCOTT
• Application Number:
  US10/394,137
• Publication Number:
  US 20040193019A1
• Time in Patent Office: 4,803 Days
• Field of Search: None
• US Class Current: 1/1
• CPC Class Codes: A61K 31/337 having four-membered rings, e.g. taxol
• CPC Class Codes: A61K 31/4745 condensed with ring systems having nitrogen as a ring hetero atom, e.g. phenantrolines yohimbine derivatives, vinblastine A61K31/475; ergoline derivatives A61K31/48
• CPC Class Codes: A61K 31/513 having oxo groups directly attached to the heterocyclic ring, e.g. cytosine
• CPC Class Codes: A61K 31/519 ortho- or peri-condensed with heterocyclic rings
• CPC Class Codes: A61K 31/555 containing heavy metals, e.g. hemin, hematin, melarsoprol
• CPC Class Codes: A61K 31/664 Amides of phosphorus acids
• CPC Class Codes: A61K 33/243 Platinum; Compounds thereof
• CPC Class Codes: G16B 20/00 ICT specially adapted for functional genomics or proteomics, e.g. genotype-phenotype associations
• CPC Class Codes: G16B 20/20 Allele or variant detection, e.g. single nucleotide polymorphism [SNP] detection
• CPC Class Codes: G16B 25/00 ICT specially adapted for hybridisation; ICT specially adapted for gene or protein expression
• CPC Class Codes: G16B 25/10 Gene or protein expression profiling; Expression-ratio estimation or normalisation
• CPC Class Codes: G16B 40/00 ICT specially adapted for biostatistics; ICT specially adapted for bioinformatics-related machine learning or data mining, e.g. knowledge discovery or pattern finding
• CPC Class Codes: G16B 40/20 Supervised data analysis
• CPC Class Codes: G16B 5/00 ICT specially adapted for modelling or simulations in systems biology, e.g. gene-regulatory networks, protein interaction networks or metabolic networks
• CPC Class Codes: G16H 50/50 for simulation or modelling of medical disorders
• CPC Class Codes: G16H 50/70 for mining of medical data, e.g. analysing previous cases of other patients

Abstract

Methods, systems, and computer program products that predict an individual'"'"'s treatment outcome from a sampling of a group of patients'"'"' biological profiles. Biological profile information is received from patients who had a medical condition and who received a treatment. Treatment outcome information regarding the patients who had the medical condition and who received the treatment is also received. A discriminant analysis-based pattern recognition process is then performed on the biological profile information and the treatment outcome information, thereby generating a model that correlates between the biological profile information and the treatment outcome information. The model can be used for, among other things, predicting treatment outcome for the new patient for the treatment.

Claims

1. A method for treating cancer, comprising obtaining the results of a correlation between patient biological profile information and the predictive information of post-treatment efficacy for a particular treatment selected from one or more treatments for a cancer, wherein the particular treatment comprises a drug treatment, wherein the correlation is obtained by:

i) identifying a group of patients who had the cancer and who underwent the particular treatment for the cancer'"'"'s metastasis;

ii) receiving multiple gene based pre-treatment biological profile information obtained through analysis of one or more of tissue and biological specimens obtained from the group of patients prior to the particular treatment;

iii) receiving post-treatment efficacy information from the group of patients indicative of an impact of the particular treatment on the cancer'"'"'s metastasis;

iv) performing a discriminant analysis-based pattern recognition process on the multiple gene based pre-treatment biological profile information and the post-treatment efficacy information, including identifying discriminating features of the multiple gene based pre-treatment biological profile information that correlate to different post treatment efficacies, and including generating a prediction model for the particular treatment that maps the multiple gene based pre-treatment biological profile information, using the discriminating features, to the corresponding different post treatment efficacies for the cancer'"'"'s metastasis;

v) receiving multiple gene based pre-treatment biological profile information obtained through analysis of the one or more of tissue and biological specimens obtained from a new patient prior to the new patient receiving the particular treatment for the cancer; and

vi) processing the new patient multiple gene based pre-treatment biological profile information according to the model, including using the discriminating features of the new patient multiple gene based pre-treatment biological profile information to predict an efficacy of the particular treatment for the new patient prior to the new patient receiving the particular treatment;

wherein the multiple gene based pre-treatment biologic profile information and the post-treatment efficacy information are received within a suitably programmed computer system, and

wherein the discriminant analysis-based pattern recognition process is performed within the suitably programmed computer system,

and treating the new patient with the particular treatment when the model predicts efficacy of the particular treatment for the new patient.

2. The method according to claim 1, further comprising generating the model to predict the efficacy of the particular treatment on a scale.

3. The method according to claim 1, further comprising:

identifying one or more additional groups of patients who had the cancer and who underwent one or more corresponding other particular treatments;

repeating, for each of the one or more additional groups of patients and the one or more corresponding other particular treatments, the receiving multiple gene based pre-treatment biological profile information, the receiving post-treatment efficacy outcome information, and the performing a discriminant analysis-based pattern recognition process, including generating one or more other corresponding prediction models for the one or more other particular treatments; and

repeating the receiving new patient multiple gene based pre-treatment biological profile information and the processing of the new patient multiple gene based pre-treatment biological profile information for each of the one or more other predictions models to determine an efficacy of the one or more other particular treatments for the new patient prior to the new patient receiving any of the one or more other particular treatments.

4. The method according to claim 1, wherein the multiple gene based pre-treatment biological profile information comprises DNA profile information.

5. The method according to claim 1, wherein the multiple gene based pre-treatment biological profile information comprises RNA profile information.

6. The method according to claim 1, wherein the multiple gene based pre-treatment biological profile information comprises multiple protein profile information.

7. The method according to claim 1, wherein the multiple gene based pre-treatment biological profile information is embedded on a microarray.

8. The method according to claim 7, wherein the multiple gene based pre-treatment biological profile information is generated from a genomic analysis.

9. The method according to claim 1, wherein the receiving pre-treatment biological profile information from the group of patients comprises receiving the pre-treatment biological profile information from fewer than 100 patients.

10. The method according to claim 1, wherein the receiving pre-treatment biological profile information from the group of patients comprises receiving the pre-treatment biological profile information from greater than 100 patients.

11. The method according to claim 1, wherein the performing a discriminant analysis-based pattern recognition process comprises reducing the multiple gene based pre-treatment biological profile information to a sub-set of genes that account for a substantial portion of diversity.

12. The method according to claim 11, further comprising embedding a representation of the sub-set of genes on a microarray chip as part of the model.

13. The method according to claim 12, wherein the receiving pre-treatment biological profile information from a new patient comprises receiving genetic profile information from the new patient regarding only the sub-set of genes.

14. The method according to claim 11, wherein the receiving pre-treatment biological profile information from a new patient comprises receiving genetic profile information from the new patient regarding only the sub-set of genes.

15. The method according to claim 1, wherein the processing of the new patient pre-treatment biological profile information to predict the efficacy of the particular treatment for the new patient is performed within a suitably programmed computer system.

16. The method according to claim 1, further comprising using the model to predict efficacy of a first line treatment for a cancer patient.

17. The method according to claim 1, further comprising using the model to predict efficacy of a single drug treatment.

18. The method according to claim 1, further comprising using the model to predict efficacy of a multiple combination drug treatment for the cancer.

19. The method according to claim 1, further comprising using the model to predict efficacy of a first line single drug treatment.

20. The method according to claim 1, further comprising using the model to predict efficacy of a first line multiple combination drug treatment for the cancer.

21. The method according to claim 1, further comprising using the model to predict efficacy of a later-line drug treatment for the cancer.

22. The method according to claim 1, further comprising:

receiving multiple gene based pre-treatment biological profile information obtained through analysis of the one or more of tissue and biological specimens obtained from a second new patient prior to the second new patient receiving the particular treatment for a second cancer; and

processing the second new patient multiple gene based pre-treatment biological profile information according to the model, generated with respect to the first cancer of i) and never trained by the second cancer, including using the discriminating features of the second new patient multiple gene based pre-treatment biological profile information to predict an efficacy of the particular treatment for the second new patient with respect to the second cancer.

23. The method according to claim 22, wherein the first cancer of i) comprises colon cancer and the second cancer comprises ovarian cancer.

24. The method according to claim 22, wherein the first cancer of i) comprises a first type of tumor and the second cancer comprises a second type of tumor related to the first type of tumor.

25. The method according to claim 24, wherein the first type of tumor is primary tumor and the second type of tumor is a metastatic tumor.

26. The method according to claim 1, wherein the multiple gene based pre-treatment biological profile information comprises microarray expression profile information.

27. The method according to claim 26, wherein the microarray expression profile data is obtained from one or more of the following sources:

nucleic acid expression profiles;

protein expression profiles; and

single nucleotide polymorphisms.

28. The method according to claim 1, further comprising using the model for a deterministic treatment administration tool for personalized medical treatment.

29. The method according to claim 1, wherein the treatment efficacy information comprises objective features.

30. The method according to claim 1, wherein the pre-treatment biological profile information and the treatment efficacy information is received from patients who benefitted from a clinical trial of the treatment.

31. The method according to claim 30, further comprising submitting an application for approval of the particular treatment for use with new patients for whom the model outputs a favorable predicted treatment efficacy.

32. The method according to claim 1, wherein the performing a discriminant analysis-based pattern recognition process of part iv comprises:

dividing the biological profile information and the treatment efficacy information into a training sub-set and a testing sub-set;

performing the discriminant analysis-based pattern recognition process on the training sub-set; and

testing the resultant model with the testing sub-set.

33. The method according to claim 1, wherein the performing a discriminant analysis-based pattern recognition process comprises:

dividing the biological profile information and the post-treatment efficacy information into a training sub-set of biological profile information and post-treatment efficacy information and a testing sub-set of biological profile information and post-treatment efficacy information;

performing a plurality of different types of discriminant analysis-based pattern recognition processes on the training sub-set of biological profile information and treatment efficacy information, including generating a plurality of models that correlate between the training set of biological profile information and the training set of post-treatment efficacy information;

testing the plurality of models with the testing sub-set of biological profile information and treatment efficacy information; and

selecting the model that most accurately predicts the testing sub-set of treatment efficacy information from the testing sub-set of biological profile information.

34. The method according to claim 33, further comprising:

selecting new testing sub-sets of biological profile information and treatment efficacy information;

repeating the performing a plurality of different types of discriminant analysis-based pattern recognition processes for the new sub-sets of biological profile information; and

averaging the results for each of the plurality of models;

wherein the selecting the model comprises selecting the model that most accurately predicts, on average, the testing sub-set of treatment efficacy information from the testing sub-set of biological profile information.

35. The method according to claim 1, wherein the multiple gene based pre-treatment biological profile information is obtained from targeted tissue samples.

36. The method according to claim 1, further comprising selecting a mode of treatment for the new patient based on the predicted treatment efficacy.

37. The method according to claim 1, wherein the performing a discriminant analysis-based pattern recognition process is performed without biological pathway information.

38. The method according to claim 1, further comprising:

submitting an application for approval of the particular treatment for use with new patients for whom the model outputs a favorable predicted treatment efficacy.

39. The method according to claim 1, further comprising:

analyzing the one or more of tissue and biological specimens of the group of patients to obtain the multiple gene based pre-treatment biological profile information regarding the group of patients.

40. The method according to claim 1, wherein the multiple gene based pre-treatment biological profile information comprises one or more of the following:

genetic profile information;

DNA profile information;

RNA profile information; and

multiple protein profile information.

41. The method of claim 1, wherein the efficacy information includes cell response information associated with tissue to which the particular treatment is targeted.

42. The method of claim 41, wherein the efficacy information includes tissue scan information.

43. The method of claim 41, wherein the efficacy information includes blood information.

44. The method of claim 1, further comprising:

classifying the multiple gene based pre-treatment biological profile information of the group of patients with respect to sensitivity of the patients to the treatment based on the efficacy information;

wherein the performing of the discriminant analysis-based pattern recognition process includes performing the discriminant analysis-based pattern recognition process on the multiple gene based pre-treatment biological profile information and classifications of sensitivity of the patients to the treatment, including,

identifying discriminating features of the multiple gene based pre-treatment biological profile information that correlate to different classifications of sensitivity, and

generating a prediction model for the particular treatment that maps the multiple gene based pre-treatment biological profile information, using the discriminating features, to the corresponding different classifications of sensitivity; and

wherein the processing of the new patient pre-treatment biological profile information includes using the discriminating features of the new patient multiple gene based pre-treatment biological profile information to identify a classification of sensitivity to the particular treatment for the new patient.

45. The method of claim 44, wherein the classifications of sensitivity include a sensitive classification and a resistive classification.

46. The method of claim 44, wherein the classifications of sensitivity include a sensitivity scale.

47. The method of claim 1, wherein the drug treatment is chemotherapy.

48. The method of claim 1, wherein the drug treatment is chemotherapy and the disease is colorectal cancer.

49. The method of claim 1, wherein the drug treatment is chemotherapy and the disease is ovarian cancer.

50. The method of claim 49, wherein the chemotherapy is platinum based chemotherapy.

51. A method for treating cervical cancer, comprising obtaining the results of a correlation between patient biological profile information and the predictive information of post-treatment efficacy for a radiotherapy treatment for cervical cancer, wherein the correlation is obtained by:

i) identifying a group of patients who had the cervical cancer and who underwent the radiotherapy treatment for the cervical cancer;

ii) receiving multiple gene based pre-treatment biological profile information obtained through analysis of one or more of tissue and biological specimens obtained by surgery from the group of patients prior to the radiotherapy treatment;

iii) receiving post-treatment efficacy information from the group of patients indicative of an impact of the radiotherapy treatment;

iv) performing a K-Nearest-Neighbor discriminant analysis-based pattern recognition process on the multiple gene based pre-treatment biological profile information and the post-treatment efficacy information, including identifying discriminating features of the multiple gene based pre-treatment biological profile information that correlate to different post treatment efficacies, and including generating a prediction model for the radiotherapy treatment that maps the multiple gene based pre-treatment biological profile information, using the discriminating features, to the corresponding different post treatment efficacies;

v) receiving multiple gene based pre-treatment biological profile information obtained through analysis of the one or more of tissue and biological specimens obtained from a new patient prior to the new patient receiving the radiotherapy treatment for the cervical cancer; and

vi) processing the new patient multiple gene based pre-treatment biological profile information according to the model, including using the discriminating features of the new patient multiple gene based pre-treatment biological profile information to predict an efficacy of the radiotherapy treatment for the new patient prior to the new patient receiving the radiotherapy treatment;

wherein the multiple gene based pre-treatment biologic profile information and the post-treatment efficacy information are received within a suitably programmed computer system, and wherein the discriminant analysis-based pattern recognition process is performed within the suitably programmed computer system,

and treating the new patient with the radiotherapy treatment when the model predicts efficacy of the radiotherapy treatment for the new patient.

52. A method for treating colorectal cancer, comprising obtaining the results of a correlation between patient biological profile information and the predictive information of post-treatment efficacy of the cancer'"'"'s metastasis for a chemotherapy treatment for colorectal cancer comprising a combination of 5FU, leucovorin and CPT11, wherein the correlation is obtained by:

i) identifying a group of patients who had the colorectal cancer and who underwent the chemotherapy treatment for the colorectal cancer'"'"'s metastasis;

ii) receiving multiple gene based pre-treatment biological profile information obtained through analysis of one or more of tissue and biological specimens obtained by surgery from the group of patients prior to the chemotherapy treatment;

iii) receiving post-treatment efficacy information from the group of patients indicative of an impact of the chemotherapy treatment on the cancer'"'"'s metastasis;

iv) performing a K-Nearest-Neighbor discriminant analysis-based pattern recognition process on the multiple gene based pre-treatment biological profile information and the post-treatment efficacy information, including identifying discriminating features of the multiple gene based pre-treatment biological profile information that correlate to different post treatment efficacies, and including generating a prediction model for the chemotherapy treatment that maps the multiple gene based pre-treatment biological profile information, using the discriminating features, to the corresponding different post treatment efficacies for the cancer'"'"'s metastasis;

v) receiving multiple gene based pre-treatment biological profile information obtained through analysis of the one or more of tissue and biological specimens obtained from a new patient prior to the new patient receiving the chemotherapy treatment for the colorectal cancer; and

vi) processing the new patient multiple gene based pre-treatment biological profile information according to the model, including using the discriminating features of the new patient multiple gene based pre-treatment biological profile information to predict an efficacy of the chemotherapy treatment for the new patient prior to the new patient receiving the chemotherapy treatment;

wherein the multiple gene based pre-treatment biologic profile information and the post-treatment efficacy information are received within a suitably programmed computer system, and wherein the discriminant analysis-based pattern recognition process is performed within the suitably programmed computer system,

and treating the new patient with the chemotherapy treatment when the model predicts efficacy of the chemotherapy treatment for the new patient.

53. A method for treating ovarian cancer, comprising obtaining the results of a correlation between patient biological profile information and the predictive information of post-treatment efficacy of the cancer'"'"'s metastasis for a chemotherapy treatment for ovarian cancer comprising cisplatin or carboplatin, plus cyclophosphamide or taxol, wherein the correlation is obtained by:

i) identifying a group of patients who had the ovarian cancer and who underwent the chemotherapy treatment for the ovarian cancer'"'"'s metastasis;

ii) receiving multiple gene based pre-treatment biological profile information obtained through analysis of one or more of tissue and biological specimens obtained by surgery from the group of patients prior to the chemotherapy treatment;

iii) receiving post-treatment efficacy information from the group of patients indicative of an impact of the chemotherapy treatment on the cancer'"'"'s metastasis;

iv) performing a K-Nearest-Neighbor discriminant analysis-based pattern recognition process on the multiple gene based pre-treatment biological profile information and the post-treatment efficacy information, including identifying discriminating features of the multiple gene based pre-treatment biological profile information that correlate to different post treatment efficacies, and including generating a prediction model for the chemotherapy treatment that maps the multiple gene based pre-treatment biological profile information, using the discriminating features, to the corresponding different post treatment efficacies for the cancer'"'"'s metastasis;

v) receiving multiple gene based pre-treatment biological profile information obtained through analysis of the one or more of tissue and biological specimens obtained from a new patient prior to the new patient receiving the chemotherapy treatment for the ovarian cancer; and

vi) processing the new patient multiple gene based pre-treatment biological profile information according to the model, including using the discriminating features of the new patient multiple gene based pre-treatment biological profile information to predict an efficacy of the chemotherapy treatment for the new patient prior to the new patient receiving the chemotherapy treatment;

wherein the multiple gene based pre-treatment biologic profile information and the post-treatment efficacy information are received within a suitably programmed computer system, and wherein the discriminant analysis-based pattern recognition process is performed within the suitably programmed computer system,

and treating the new patient with the chemotherapy treatment when the model predicts efficacy of the chemotherapy treatment for the new patient.

54. A method for treating cancer, comprising obtaining the results of a correlation between patient biological profile information and the predictive information of post-treatment efficacy for a particular treatment selected from one or more treatments for a cancer, wherein the particular treatment comprises a radiotherapy treatment, wherein the correlation is obtained by:

i) identifying a group of patients who had the cancer and who underwent the particular treatment for the cancer;

ii) receiving multiple gene based pre-treatment biological profile information obtained through analysis of one or more of tissue and biological specimens obtained from the group of patients prior to the particular treatment;

iii) receiving post-treatment efficacy information from the group of patients indicative of an impact of the particular treatment;

iv) performing a discriminant analysis-based pattern recognition process on the multiple gene based pre-treatment biological profile information and the post-treatment efficacy information, including identifying discriminating features of the multiple gene based pre-treatment biological profile information that correlate to different post treatment efficacies, and including generating a prediction model for the particular treatment that maps the multiple gene based pre-treatment biological profile information, using the discriminating features, to the corresponding different post treatment efficacies;

v) receiving multiple gene based pre-treatment biological profile information obtained through analysis of the one or more of tissue and biological specimens obtained from a new patient prior to the new patient receiving the particular treatment for the cancer; and

vi) processing the new patient multiple gene based pre-treatment biological profile information according to the model, including using the discriminating features of the new patient multiple gene based pre-treatment biological profile information to predict an efficacy of the particular treatment for the new patient prior to the new patient receiving the particular treatment;

wherein the multiple gene based pre-treatment biologic profile information and the post-treatment efficacy information are received within a suitably programmed computer system, and

wherein the discriminant analysis-based pattern recognition process is performed within the suitably programmed computer system,

and treating the new patient with the particular treatment when the model predicts efficacy of the particular treatment for the new patient.