Methods of diagnosing inflammatory bowel disease
View Patent ↗The present invention provides methods, systems, and code for accurately classifying whether a sample from an individual is associated with inflammatory bowel disease (IBD) or a clinical subtype thereof. In particular, the present invention is useful for classifying a sample from an individual as an IBD sample using a statistical algorithm and/or empirical data. The present invention is also useful for differentiating between a clinical subtype of IBD such as Crohn's disease (CD) and ulcerative colitis (UC) using a statistical algorithm and/or empirical data. Thus, the present invention provides an accurate diagnostic prediction of IBD or a clinical subtype thereof and prognostic information useful for guiding treatment decisions.
1. A method for classifying whether a sample from an individual is associated with inflammatory bowel disease (IBD), said method comprising:
(a) analyzing a sample obtained from said individual to determine the presence or level of an anti-neutrophil cytoplasmic antibody (ANCA), anti- Saccharomyces cerevisiae immunoglobulin A (ASCA-IgA), anti- Saccharomyces cerevisiae immunoglobulin G (ASCA-IgG), an anti-outer membrane protein C (anti-OmpC) antibody, an anti-flagellin antibody, and a perinuclear anti-neutrophil cytoplasmic antibody (pANCA) in said sample; and
(b) applying a combination of at least two learning statistical classifier systems in tandem to the presence or level of said ANCA, ASCA-IgA, ASCA-IgG, anti-OmpC antibody, anti-flagellin antibody, and pANCA determined in step (a) to classify said sample as an IBD sample or non-IBD sample with an overall accuracy of at least 70%, wherein said combination of at least two learning statistical classifier systems comprises a classification and regression tree (C&RT) or random forest and a neural network, and wherein said C&RT or random forest is first applied to the presence or level of said ANCA, ASCA-IgA, RSCA-IgG, anti-OmpC antibody, anti-flagellin antibody, and pANCA determined in step (a) to generate a prediction or probability value.
2. The method of claim 1 , wherein the presence or level of said ANCA, ASCA-IgA, ASCA-IgG, anti-OmpC antibody, or anti-flagellin antibody is determined by assaying said sample with an immunoassay.
3. The method of claim 2 , wherein said immunoassay is an enzyme-linked immunosorbent assay (ELISA).
4. The method of claim 1 , wherein the presence or level of said pANCA is determined by assaying said sample with an immunohistochemical assay.
5. The method of claim 4 , wherein said immunohistochemical assay is an immunoflourescence assay.
6. The method of claim 1 , wherein said sample is selected from the group consisting of serum, plasma, whole blood, and stool.
7. The method of claim 1 , wherein step (b) further comprises applying said neural network to said prediction or probability value and the presence or level of said ANCA, ASCA-IgA, ASCA-IgG, anti-OmpC antibody, anti-flagellin antibody, and pANCA determined in step (a) to classify said sample as an IBD sample or non-IBD sample.
8. The method of claim 1 , wherein said method further comprises sending the results from said classification to a clinician.
9. The method of claim 1 , wherein said method further provides a diagnosis in the form of a probability that said individual has IBD.
10. The method of claim 1 , wherein said method further comprises administering to said individual a therapeutically effective amount of a drug useful for treating one or more symptoms associated with IBD.
11. The method of claim 10 , wherein said drug is selected from the group consisting of aminosalicylates, corticosteroids, thiopurines, methotrexate, monoclonal antibodies, free bases thereof, pharmaceutically acceptable salts thereof, derivatives thereof, analogs thereof, and combinations thereof
12. A tangible computer-readable medium that stores code for controlling one or more processors to classify whether a sample from an individual is associated with IBD, said code including instructions to apply a statistical process comprising a combination of at least two learning statistical classifier systems in tandem to a data set indicating the presence or level of an anti-neutrophil cytoplasmic antibody (ANCA), anti- Saccharomyces cerevisiae immunoglobulin A (ASCA-IgA), anti- Saccharomyces cerevisiae immunoglobulin G (ASCA-IgG), an anti-outer membrane protein C (anti-OmpC) antibody, an anti-flagellin antibody, and a perinuclear anti-neutrophil cytoplasmic antibody (pANCA) in said sample to produce a statistically derived decision classifying said sample as an IBD sample or non-IBD sample based upon the presence or level of said ANCA, ASCA-IgA, ASCA-IgG, anti-OmpC antibody, anti-flagellin antibody, and pANCA to classify said sample as an IBD sample or non-IBD sample with an overall accuracy of at least 70%,
wherein said combination of at least two learning statistical classifier systems comprises a classification and regression tree (C&RT) or random forest and a neural network, and
wherein said C&RT or random forest is first applied to the presence or level of said ANCA, ASCA-IgA, ASCA-IgG, anti-OmpC antibody, anti-flagellin antibody, and pANCA to generate a prediction or probability value.
13. A system for classifying whether a sample from an individual is associated with IBD, said system comprising:
(a) a data acquisition device configured to produce a data set indicating the presence or level of an anti-neutrophil cytoplasmic antibody (ANCA), anti- Saccharomyces cerevisiae immunoglobulin A (ASCA-IgA), anti- Saccharomyces cerevisiae immunoglobulin G (ASCA-IgG), an anti-outer membrane protein C (anti-OmpC) antibody, an anti-flagellin antibody, and a perinuclear anti-neutrophil cytoplasmic antibody (pANCA) in said sample;
(b) a data processing device configured to process the data set by applying a statistical process comprising a combination of at least two learning statistical classifier systems in tandem to the data set to produce a statistically derived decision classifying said sample as an IBD sample or non-IBD sample based upon the presence or level of said ANCA, ASCA-IgA, ASCA-IgG, anti-OmpC antibody, anti-flagellin antibody, and pANCA to classify said sample as an IBD sample or non-IBD sample with an overall accuracy of at least 70%, wherein said combination of at least two learning statistical classifier systems comprises a classification and regression tree (C&RT) or random forest and a neural network, and wherein said C&RT or random forest is first applied to the data set to generate a prediction or probability value; and
(c) a display device configured to display the statistically derived decision.
14. The method of claim 1 , wherein step (a) further comprises analyzing said sample to determine the presence or level of an anti-I2 antibody.
15. The method of claim 1 , wherein said anti-flagellin antibody is an anti-Cbir-1 antibody.
16. The method of claim 1 , wherein said IBD sample is classified as a Crohn's disease (CD) sample or ulcerative colitis (UC) sample.
17. The method of claim 1 , wherein said combination of at least two learning statistical classifier systems classifies said sample as an IBD sample or non-IBD sample with a sensitivity of at least 70%.
18. The method of claim 1 , wherein said combination of at least two learning statistical classifier systems classifies said sample as an IBD sample or non-IBD sample with a specificity of at least 80%.
19. The method of claim 1 , wherein said combination of at least two learning statistical classifier systems classifies said sample as an IBD sample or non-IBD sample with a positive predictive value (PPV) of at least 80%.
20. The method of claim 1 , wherein said combination of at least two learning statistical classifier systems classifies said sample as an IBD sample or non-IBD sample with a negative predictive value (NPV) of at least 70%.
21. The method of claim 1 , wherein the presence or level of said ANCA is determined by assaying the binding between said ANCA and fixed neutrophils.
22. The method of claim 1 , wherein the presence or level of said ASCA-IgA or ASCA-IgG is determined by assaying the binding between said ASCA-IgA or ASCA-IgG and an antigen selected from the group consisting of yeast cell wall mannan, a purified oligomannoside antigen, a synthetic oligomannoside antigen, and combinations thereof
23. The method of claim 22 wherein said antigen is yeast cell wall phosphopeptidomannan (PPM).
24. The method of claim 23 , wherein said yeast cell wall PPM is S. uvarum PPM.
25. The method of claim 1 , wherein the presence or level of said anti-OmpC antibody is determined by assaying the binding between said anti-OmpC antibody and an OmpC protein or an immunoreactive fragment of said OmpC protein.
26. The method of claim 1 , wherein the presence or level of said anti-flagellin antibody is determined by assaying the binding between said anti-flagellin antibody and a flagellin protein or an immunoreactive fragment of said flagellin protein.
27. The method of claim 26 , wherein said flagellin protein is selected from the group consisting of Cbir- 1 flagellin, flagellin X, flagellin A, flagellin B, immunoreactive fragments thereof, and combinations thereof
28. The method of claim 26 , wherein said immunoflourescence assay comprises determining the presence or absence of said pANCA in DNAse-treated neutrophils.
29. The method of claim 14 , wherein the presence or level of said anti-I2 antibody is determined by assaying the binding between said anti-I2 antibody and an I2 protein or an immunoreactive fragment of said I2 protein.
30. The computer-readable medium of claim 12 , wherein said IBD sample is classified as a Crohn's disease (CD) sample or ulcerative colitis (UC) sample.
31. The system of claim 13 , wherein said IBD sample is classified as a Crohn's disease (CD) sample or ulcerative colitis (UC) sample.
32. The method of claim 12 , wherein said neural network is then applied to said prediction or probability value and the presence or level of said ANCA, ASCA-IgA, ASCA-IgG, anti-OmpC antibody, anti-flagellin antibody, and pANCA to classify said sample as an IBD sample or non-IBD sample.
33. The method of claim 13 , wherein said neural network is then applied to said prediction or probability value and the presence or level of said ANCA, ASCA-IgA, ASCA-IgG, anti-OmpC antibody, anti-flagellin antibody, and pANCA to classify said sample as an IBD sample or non-IBD sample.