Clinical Decision Model

Abstract

An embodiment of the invention provides a method for determining a patient-specific probability of disease. The method collects clinical parameters from a plurality of patients to create a training database. A fully unsupervised Bayesian Belief Network model is created using data from the training database; and, the fully unsupervised Bayesian Belief Network is validated. Clinical parameters are collected from an individual patient; and, such clinical parameters are input into the fully unsupervised Bayesian Belief Network model via a graphical user interface. The patient-specific probability of disease is output from the fully unsupervised Bayesian Belief Network model and sent to the graphical user interface for use by a clinician in pre-operative planning. The fully unsupervised Bayesian Belief Network model is updated using the clinical parameters from the individual patient and the patient-specific probability of disease.

1 . A method for determining a patient-specific probability of malignancy in a thyroid nodule, said method including:
 collecting clinical parameters from a plurality of patients to create a training database, the clinical parameters including fine needle aspiration biopsy results, ultrasound data, lymph node size, and imaging data;   creating a fully unsupervised Bayesian Belief Network model using data from the training database;   validating the fully unsupervised Bayesian Belief Network model;   collecting the clinical parameters for an individual patient;   receiving the clinical parameters for the individual patient into the fully unsupervised Bayesian Belief Network model;   outputting the patient-specific probability of malignancy in the thyroid nodule from the fully unsupervised Bayesian Belief Network model to a graphical user interface for use by a clinician; and   updating the fully unsupervised Bayesian Belief Network model using the clinical parameters for the individual patient and the patient-specific probability of malignancy in the thyroid nodule.   
     
     
         2 . The method according to  claim 1 , wherein the clinical parameters further include functional status of the thyroid nodule, number of cervical lymph nodes, serum thyrotropin level, pre-operative diagnosis, nuclear medicine rating, age, and ethnicity. 
     
     
         3 . The method according to  claim 1 , wherein the fine needle aspiration biopsy results include an inadequate score, indeterminate score, negative score, and positive score; wherein the ultrasound data include a complex cyst score, mixed score, simple cyst score, and solid score; and wherein the lymph node size includes a less than 18 centimeters score, 18-31 centimeters score, and greater than 31 centimeters score. 
     
     
         4 . The method according to  claim 1 , wherein the imaging data includes results from electrical impedance scanning. 
     
     
         5 . The method according to  claim 4 , wherein the results from the electrical impedance scanning include a definitely benign score, probably benign score, suspicious for cancer score, probably cancer score, and definitely cancer score. 
     
     
         6 . The method according to  claim 1 , wherein said creating of the fully unsupervised Bayesian Belief Network model includes creating the fully unsupervised Bayesian Belief Network model without human-developed decision support rules. 
     
     
         7 . The method according to  claim 1 , further including estimating an accuracy of the patient-specific probability of malignancy in the thyroid nodule, the accuracy including at least one of model sensitivity, model specificity, positive and negative predictive values, and overall accuracy. 
     
     
         8 . A system for determining a patient-specific probability of malignancy in a thyroid nodule, said system including:
 a training database having clinical parameters collected from a plurality of patients, the clinical parameters including fine needle aspiration biopsy results, ultrasound data, lymph node size, and imaging data; and   a fully unsupervised Bayesian Belief Network model created using data from said training database,   said fully unsupervised Bayesian Belief Network model receiving clinical parameters for an individual patient and outputting the patient-specific probability of malignancy in the thyroid nodule to a graphical user interface,   said fully unsupervised Bayesian Belief Network model being updated using the clinical parameters for the individual patient and the patient-specific probability of malignancy in the thyroid nodule.   
     
     
         9 . The system according to  claim 8 , wherein the clinical parameters further include functional status of the thyroid nodule, number of cervical lymph nodes, serum thyrotropin level, pre-operative diagnosis, nuclear medicine rating, age, and ethnicity. 
     
     
         10 . The system according to  claim 8 , wherein the fine needle aspiration biopsy results include an inadequate score, indeterminate score, negative score, and positive score; wherein the ultrasound data include a complex cyst score, mixed score, simple cyst score, and solid score; and wherein the lymph node size includes a less than 18 centimeters score, 18-31 centimeters score, and greater than 31 centimeters score. 
     
     
         11 . The system according to  claim 8 , wherein the imaging data includes results from electrical impedance scanning. 
     
     
         12 . The system according to  claim 11 , wherein the results from the electrical impedance scanning include a definitely benign score, probably benign score, suspicious for cancer score, probably cancer score, and definitely cancer score. 
     
     
         13 . The system according to  claim 8 , wherein said fully unsupervised Bayesian Belief Network model lacks human-developed decision support rules. 
     
     
         14 . The system according to  claim 8 , wherein said graphic user interface further outputs an estimated accuracy of the patient-specific probability of malignancy in the thyroid nodule, the estimated accuracy including at least one of model sensitivity, model specificity, positive and negative predictive values, and overall accuracy. 
     
     
         15 . A computer program product for determining a patient-specific probability of malignancy in a thyroid nodule, said computer program product including:
 a computer readable storage medium;   first program instructions to collect clinical parameters from a plurality of patients to create a training database, the clinical parameters including fine needle aspiration biopsy results, ultrasound data, lymph node size, and imaging data;   second program instructions to create a fully unsupervised Bayesian Belief Network model using data from the training database;   third program instructions to validate the fully unsupervised Bayesian Belief Network model;   fourth program instructions to collect the clinical parameters for an individual patient;   fifth program instructions to input the clinical parameters for the individual patient into the fully unsupervised Bayesian Belief Network model via a graphical user interface;   sixth program instructions to output the patient-specific probability of malignancy in the thyroid nodule from the fully unsupervised Bayesian Belief Network model to the graphical user interface for use by a clinician; and   seventh program instructions to update the fully unsupervised Bayesian Belief Network model using the clinical parameters for the individual patient and the patient-specific probability of malignancy in the thyroid nodule,   the first program instructions, the second program instructions, the third program instructions, the fourth program instructions, the fifth program instructions, the sixth program instructions, and the seventh program instructions are stored on the computer readable storage medium.   
     
     
         16 . A method for determining a patient-specific probability of transplant glomerulopathy, said method including:
 collecting clinical parameters from a plurality of patients to create a training database, the clinical parameters including biomarker levels from biopsy tissue;   creating a fully unsupervised Bayesian Belief Network model using data from the training database;   validating the fully unsupervised Bayesian Belief Network model;   collecting the clinical parameters for an individual patient;   receiving the clinical parameters for the individual patient into the fully unsupervised Bayesian Belief Network model;   outputting the patient-specific probability of transplant glomerulopathy from the fully unsupervised Bayesian Belief Network model to a graphical user interface for use by a clinician; and   updating the fully unsupervised Bayesian Belief Network model using the clinical parameters for the individual patient and the patient-specific probability of transplant glomerulopathy.   
     
     
         17 . The method according to  claim 16 , wherein the biomarker levels include gene expression levels for an ICAM-1 gene, IL-10 gene, CCL-3 gene, CD-86 gene, CCL-2 gene, CXCL-11 gene, CD-80 gene, GNLY gene, and PRF-1 gene. 
     
     
         18 . The method according to  claim 17 , wherein the biomarker levels further include gene expression levels for a CD40LG gene, IFNG gene, CD-28 gene, CXCL-10 gene, CCR-5 gene, CD-40 gene, CTLA-4 gene, TNF gene, CXCL-9 gene, CX3CR-1 gene, FOXP-3 gene, EDN-1 gene, CD-4 gene, TBX-21 gene, FASLG gene, C-3 gene, CD3E gene, CXCR-3 gene, and CCL-5 gene. 
     
     
         19 . The method according to  claim 16 , wherein the biomarker levels include gene expression levels for the VCAM1 gene, MMP9 gene, Banff C4d gene, MMP7 gene, and LAMC2 gene. 
     
     
         20 . The method according to  claim 20 , wherein the biomarker levels further include gene expression levels for a TNC gene, S100A4 gene, NPHS1 gene, NPHS2 gene, AFAP gene, PDGF8 gene, SERPINH1 gene, TIMP4 gene, TIMP3 gene, VIM gene, SERPINE1 gene, TIMP1 gene, FN1 gene, ANGPT2 gene, TGFB1 gene, ACTA2 gene, TIMP2 gene, COL4A2 gene, MMP2 gene, COL1A1 gene, COL3A1 gene, GREM1 — 2 gene, SPARC gene, IGF1 gene, SMAD3 gene, HSPG2 gene, FN1 gene, ANGPT2 gene, TGFB1 gene, ACTA2 gene, THBS1 gene, CTNNB1 gene, FGF2 gene, TJP1 gene, FAT gene, CDH1 gene, SMAD7 gene, CD2AP gene, CDH3 gene, CTGF gene, ACTN4 gene, SPP1 gene, AGRN gene, VEGF gene, and BMP7 gene. 
     
     
         21 . The method according to  claim 16 , wherein said creating of the fully unsupervised Bayesian Belief Network model includes creating the fully unsupervised Bayesian Belief Network model without human-developed decision support rules. 
     
     
         22 . The method according to  claim 16 , further including estimating an accuracy of the patient-specific probability of transplant glomerulopathy, the accuracy including at least one of model sensitivity, model specificity, positive and negative predictive values, and overall accuracy. 
     
     
         23 . A method for determining a patient-specific probability of impaired wound healing, said method including:
 collecting clinical parameters from a plurality of patients to create a training database, the clinical parameters including biomarker levels from at least one of serum, wound effluent and biopsy tissue, the biomarker levels including gene expression levels for an IP-10 gene, IL-6 gene, MCP-1 gene, IL-5 gene, and RANTES gene;   creating a fully unsupervised Bayesian Belief Network model using data from the training database;   validating the fully unsupervised Bayesian Belief Network model;   collecting the clinical parameters for an individual patient;   receiving the clinical parameters for the individual patient into the fully unsupervised Bayesian Belief Network model;   outputting the patient-specific probability of transplant glomerulopathy from the fully unsupervised Bayesian Belief Network model to a graphical user interface for use by a clinician; and   updating the fully unsupervised Bayesian Belief Network model using the clinical parameters for the individual patient and the patient-specific probability of transplant glomerulopathy.   
     
     
         24 . The method according to  claim 23 , wherein the biomarker levels include:
 a first gene expression level for the IP-10 gene taken from the serum on a first measurement day;   a second gene expression level for the IP-10 gene taken from the serum on a second measurement day;   a third gene expression level for the IP-10 gene taken from the serum on a third measurement day;   a fourth gene expression level for the IL-6 taken from the serum on the first measurement day;   a fifth gene expression level for the IL-6 taken from the serum on the third measurement day;   a sixth gene expression level for the MCP-1 gene taken from the wound effluent on the first measurement day;   a seventh gene expression level for the MCP-1 gene taken from the serum on the first measurement day;   a eighth gene expression level for the MCP-1 gene taken from the serum on the second measurement day;   a ninth gene expression level for the MCP-1 gene taken from the serum on a thirtieth measurement day;   a tenth gene expression level for the IL-5 gene taken from the wound effluent on the third measurement day;   a eleventh gene expression level for the IL-5 gene taken from the wound effluent on the thirtieth measurement day; and   a twelfth gene expression level for the RANTES gene taken from the wound effluent on the third measurement day.   
     
     
         25 . The method according to  claim 24 , wherein the biomarker levels include:
 a lower first gene expression level, a middle first gene expression level, and an upper first gene expression level, the middle first gene expression level having a lower limit of 2427 mean fluorescent intensity (MFI) and an upper limit of 4040 MFI;   a lower second gene expression level, a middle second gene expression level, and an upper second gene expression level, the middle second gene expression level having a lower limit of 1371 MFI and an upper limit of 4108 MFI;   a lower third gene expression level, a middle third gene expression level, and an upper third gene expression level, the middle third gene expression level having a lower limit of 2164 MFI and an upper limit of 3509 MFI;   a lower fourth gene expression level, a middle fourth gene expression level, and an upper fourth gene expression level, the middle fourth gene expression level having a lower limit of 174 MFI and an upper limit of 955 MFI;   a lower fifth gene expression level, a middle fifth gene expression level, and an upper fifth gene expression level;   a lower sixth gene expression level, a middle sixth gene expression level, and an upper sixth gene expression level, the middle sixth gene expression level having a lower limit of 352 MFI and an upper limit of 709 MFI;   a lower seventh gene expression level, a middle seventh gene expression level, and an upper seventh gene expression level, the middle seventh gene expression level having a lower limit of 40.5 MFI and an upper limit of 199 MFI;   a lower eighth gene expression level, a middle eighth gene expression level, and an upper eighth gene expression level, the middle eighth gene expression level having a lower limit of 50.3 MFI and an upper limit of 352 MFI;   a lower ninth gene expression level, a middle ninth gene expression level, and an upper ninth gene expression level, the middle ninth gene expression level having a lower limit of 35.9 MFI and an upper limit of 88 MFI;   a lower tenth gene expression level, a middle tenth gene expression level, and an upper tenth gene expression level, the middle tenth gene expression level having a lower limit of 2.2 MFI and an upper limit of 8.7 MFI;   a lower eleventh gene expression level, a middle eleventh gene expression level, and an upper eleventh gene expression level, the middle eleventh gene expression level having a lower limit of 10.5 MFI and an upper limit of 50 MFI; and   a lower twelfth gene expression level, a middle twelfth gene expression level, and an upper twelfth gene expression level, the middle twelfth gene expression level having a lower limit of 122 MFI and an upper limit of 7538 MFI.   
     
     
         26 . The method according to  claim 24 , wherein:
 the twelfth gene expression level is usable to estimate the fourth gene expression level;   the fourth gene expression level is usable to estimate the seventh gene expression level;   the seventh gene expression level is usable to estimate the fifth gene expression level;   the first gene expression level is usable to estimate the eighth gene expression level;   the eighth gene expression level is usable to estimate the second gene expression level;   the second gene expression level is usable to estimate the third gene expression level;   the tenth gene expression level is usable to estimate the sixth gene expression level;   the sixth gene expression level is usable to estimate the second gene expression level;   the ninth gene expression level is usable to estimate the eighth gene expression level; and   the eleventh gene expression level is usable to estimate the first gene expression level.   
     
     
         27 . The method according to  claim 23 , wherein the biomarker levels further include gene expression levels for an IL-1α gene, IL-1β gene, IL-2 gene, IL-3 gene, IL-4 gene, IL-7 gene, IL-8 gene, IL-10 gene, IL-12(p40) gene, IL-12(p70) gene, IL-13 gene, IL-15 gene, Eotaxin gene, IFN-γ gene, GM-CSF gene, MIP-1α gene, and TNFα gene. 
     
     
         28 . The method according to  claim 23 , wherein the clinical parameters further include RNA transcripts and translation products of genes selected from the group consisting: ACTA2, ACVR1, ADM, ALCAM, ANGPT 1, ANGPT 2, ANGPT 4, BAX, BCL2, BCL2L, 18S, 18S, CAV2, CCL1, CCL11, CCL17, CCL19, CCL 2, CCL 20, CCL22, CCL25, CCL27, CCL28, CCL3, COL3A1, COL4A1, COL4A3, CSSF1, CSF2, CSF3, CTGF, CX3CL1, CXCL1, CXCL10, CXCL11, CXCL12, FGF10, FGF11, FGF12, FGF13, FGF17, FGF2, FGF3, FGF5, FGF7, FGF8, FGF9, FIGF, IFNG, IGF1, IGF2, IGFBP1, IGFBP2, IGFBP3, IGFBP3, IGFBP4, IGFBP5, IGFBP6, IGFBP7, IL10, IL11, IL6, IL7, IL8, IL9, ITGA5, ITGAL, ITGAM, ITGB2, KDR, KITLG, LBP, LTA, MMP7, MMP8, MMP9, MPO, NCAM2, NFKB1, NFKB2, NOS2A, OSMR, PDGFA, PDGFB, PECAM1, SMAD6, SMAD7, SOCS1, SOCS3, SOCS5, STAT3, TEK, TGFA, TGFB1, TGFB2, TGFB3, TGFBR1, DCL2L2, BMP1, BMP15, BMP5, BMP3, BMP4, BMP5, BMP6, BMP7, BMP8A, BMP8B, CALCA, CALCB, CAV1, CCL4, CCL4L1, CCL4L2, CCL5, CCL7, CD14, CD4, CD40, CD40LG, CD83, CD8A, CD8B, COL18A1, COL1A1, CXCL13, CXCL2, CXCL5, CXCL9, ECGF1, EDN1, EGF, EGR1, EPO, FADD, FAS, FGF1, FLT1, FN1, GAPDH, GDF3, GDF5, MSTN, GDF9, HGF, HMGB1, IAPP, ICAM2, IFNB1, IL12A, IL13, IL15, IL16, IL17A, IL18, IL1A, 1L1B, 1L2, IL3, IL4, IL5, MAPK14, MET, MMP1, MMP10, MMP11, MMP12, MMP13, MMP14, MMP15, MMP2, MMP24. MMP3, PF4, PLA2G4A, PTGS1, PTGS2, SELE, SELP, SERPINE1, SLPI, SMAD1, SMAD2. SMAD3, SMAD4, TIE1, TIMP1, TIMP2, TIMP3, TNC, TNF, TNFSF10, VCAM1, VEGFB, VEGFC, XCL1, XCL2. 
     
     
         29 . The method according to  claim 23 , wherein the clinical parameters further include Injury Severity Score (ISS), Acute Physiology and Chronic Health Evaluation II (APACHE-II) scores, wound size, and associated vascular injury. 
     
     
         30 . A method for determining a patient-specific probability of disease, the disease including at least one of malignancy in a thyroid nodule, transplant glomerulopathy, impaired wound healing, and breast cancer, said method including:
 collecting clinical parameters from a plurality of patients to create a training database, the clinical parameters including at least one of fine needle aspiration biopsy results, ultrasound data, lymph node size, imaging data, Gail model cutoff, mammogram results, MRI results, breast size, personal history of breast disease, and biomarker levels from at least one of serum, wound effluent and biopsy tissue, the biomarker levels including gene expression levels for an IP-10 gene, IL-6 gene, MCP-1 gene, IL-5 gene, and RANTES gene;   creating a fully unsupervised Bayesian Belief Network model using data from the training database;   validating the fully unsupervised Bayesian Belief Network model;   collecting the clinical parameters for an individual patient;   receiving the clinical parameters for the individual patient into the fully unsupervised Bayesian Belief Network model;   outputting the patient-specific probability of disease from the fully unsupervised Bayesian Belief Network model to a graphical user interface for use by a clinician; and   updating the fully unsupervised Bayesian Belief Network model using the clinical parameters for the individual patient and the patient-specific probability of disease.   
     
     
         31 . The method according to  claim 30 , wherein the imaging data includes results from electrical impedance scanning,
 wherein the results from the electrical impedance scanning include a definitely benign score, probably benign score, suspicious for cancer score, probably cancer score, and definitely cancer score,   wherein the fine needle aspiration biopsy results include an inadequate score, indeterminate score, negative score, and positive score,   wherein the ultrasound data include a complex cyst score, mixed score, simple cyst score, and solid score, and   wherein the lymph node size includes a less than 18 centimeters score, 18-31 centimeters score, and greater than 31 centimeters score.   
     
     
         32 . The method according to  claim 30 , wherein said creating of the fully unsupervised Bayesian Belief Network model includes creating the fully unsupervised Bayesian Belief Network model without human-developed decision support rules. 
     
     
         33 . The method according to  claim 30 , further including estimating an accuracy of the patient-specific probability of disease, the accuracy including at least one of model sensitivity, model specificity, positive and negative predictive values, and overall accuracy. 
     
     
         34 . The method according to  claim 30 , wherein the clinical parameters further include functional status of the thyroid nodule, number of cervical lymph nodes, serum thyrotropin level, pre-operative diagnosis, nuclear medicine rating, age, and ethnicity. 
     
     
         35 . The method according to  claim 34 , wherein:
 the pre-operative diagnosis is usable to estimate the fine needle aspiration biopsy results,   the nuclear medicine rating is usable to estimate the lymph node size, imaging data, and the age,   the age is usable to estimate the nuclear medicine rating and imaging data,   the functional status of the thyroid nodule is usable to estimate the age,   the number of cervical lymph nodes is usable to estimate the imaging data, and   the ethnicity is usable to estimate the number of cervical lymph nodes.   
     
     
         36 . The method according to  claim 30 , wherein the biomarker levels include gene expression levels for an ICAM-1 gene, IL-10 gene, CCL-3 gene, CD-86 gene, CCL-2 gene, CXCL-11 gene, CD-80 gene, GNLY gene, PRF-1 gene, CD40LG gene, IFNG gene, CD-28 gene, CXCL-10 gene, CCR-5 gene, CD-40 gene, CTLA-4 gene, TNF gene, CXCL-9 gene, CX3CR-1 gene, FOXP-3 gene, EDN-1 gene, CD-4 gene, TBX-21 gene, FASLG gene, C-3 gene, CD3E gene, CXCR-3 gene, and CCL-5 gene. 
     
     
         37 . The method according to  claim 36 , wherein:
 the gene expression level for the CD-86 gene is usable to estimate the gene expression level for the CCL-3 gene,   the gene expression levels for the CCL-2 gene, the CD40LG gene and the CXCL11 gene are usable to estimate the gene expression level for the CD-86 gene,   the gene expression level for the PRF-1 gene is usable to estimate the gene expression level for the GNLY gene,   the gene expression levels for the GNLY gene and the CXCL-10 gene are usable to estimate the gene expression level for the CD-80 gene,   the gene expression level for the CD-80 gene is usable to estimate the gene expression level for the CXCL-11 gene,   the gene expression levels for the IFNG gene and the CD-28 gene are usable to estimate the gene expression level for the CD40LG gene,   the gene expression level for the CCR-5 gene is usable to estimate the gene expression level for the PRF-1 gene,   the gene expression levels for the CD-40 gene, the CD4 gene and the CD3E gene are usable to estimate the gene expression level for the CCR-5 gene,   the gene expression levels for the CTLA-4 gene, the TNF gene and the CX3CR1 gene are usable to estimate the gene expression level for the CD-40 gene,   the gene expression level for the CXCL9 gene is usable to estimate the gene expression level for the TNF gene,   the gene expression levels for the FOXP-3 gene and the EDN-1 gene are usable to estimate the gene expression level for the CX3CR1 gene,   the gene expression levels for the TBX-21 gene and the C-3 gene are usable to estimate the gene expression level for the CD-4 gene,   the gene expression level for the FASLG gene is usable to estimate the gene expression level for the TBX-21 gene,   the gene expression level for the CXCR-3 gene is usable to estimate the gene expression level for the CD3E gene, and   the gene expression level for the CCL-5 gene is usable to estimate the gene expression level for the CXCR-3 gene.   
     
     
         38 . The method according to  claim 30 , wherein the biomarker levels include gene expression levels for the VCAM1 gene, MMP9 gene, Banff C4d gene, MMP7 gene, LAMC2 gene, TNC gene, S100A4 gene, NPHS1 gene, NPHS2 gene, AFAP gene, PDGF8 gene, SERPINH1 gene, TIMP4 gene, TIMP3 gene, VIM gene, SERPINE1 gene, TIMP1 gene, FN1 gene, ANGPT2 gene, TGFB1 gene, ACTA2 gene, TIMP2 gene, COL4A2 gene, MMP2 gene, COL1A1 gene, COL3A1 gene, GREM1 — 2 gene, SPARC gene, IGF1 gene, SMAD3 gene, HSPG2 gene, FN1 gene, ANGPT2 gene, TGFB1 gene, ACTA2 gene, THBS1 gene, CTNNB1 gene, FGF2 gene, TJP1 gene, FAT gene, CDH1 gene, SMAD7 gene, CD2AP gene, CDH3 gene, CTGF gene, ACTN4 gene, SPP1 gene, AGRN gene, VEGF gene, and BMP7 gene. 
     
     
         39 . The method according to  claim 38 , wherein:
 the gene expression levels for the TNC gene, the THBS-1 gene and the AFAP gene are usable to estimate the gene expression level for the LAMC-2 gene,   the gene expression levels for the NPHS-1 gene and the S100A4 gene are usable to estimate the gene expression level for the TNC gene,   the gene expression level for the NPHS-2 gene is usable to estimate the gene expression level for the NPHS-1 gene,   the gene expression level for the CTNNB-1 gene is usable to estimate the gene expression level for the THBS-1 gene,   the gene expression levels for the FGF-2 gene and the TJP-1 gene are usable to estimate the gene expression level for the CTNNB-1 gene,   the gene expression level for the CD-80 gene is usable to estimate the gene expression level for the CXCL-11 gene,   the gene expression levels for the FAT gene, the CDH-1 gene, the CD2AP gene, the CTGF gene and the ACTN-4 gene are usable to estimate the gene expression level for the TJP-1 gene,   the gene expression level for the SMAD-7 gene is usable to estimate the gene expression level for the CDH-1 gene,   the gene expression level for the CDH-3 gene is usable to estimate the gene expression level for the CD2AP gene,   the gene expression levels for the SPP-1 gene, the AGRN gene, the VEGF gene are usable to estimate the gene expression level for the ACTN-4 gene,   the gene expression level for the BMP-7 gene is usable to estimate the gene expression level for the VEGF gene,   the gene expression levels for the PDGFB gene, the TIMP-3 gene and the VIM gene are usable to estimate the gene expression level for the AFAP gene,   the gene expression level for the SERPINH-1 gene is usable to estimate the gene expression level for the PDGFB gene,   the gene expression level for the TIMP-4 gene is usable to estimate the gene expression level for the SERPINH-1 gene,   the gene expression levels for the SERPINE-1 gene, the TIMP-2 gene, the FN-1 gene, the TGFB-1 gene and the TIMP-1 gene are usable to estimate the gene expression level for the VIM gene,   the gene expression levels for the COL4A2 gene, the MMP-2 gene, the SMAD-3 gene and the HSPG-2 gene are usable to estimate the gene expression level for the TIMP-2 gene,   the gene expression level for the COL1A1 gene is usable to estimate the gene expression level for the MMP-2 gene,   the gene expression levels for the COL3A1 gene, the GREM-1-2 gene and the SPARC gene are usable to estimate the gene expression level for the COL1A1 gene,   the gene expression level for the IGF-1 gene is usable to estimate the gene expression level for the SPARC gene,   the gene expression level for the ANGPT-2 gene is usable to estimate the gene expression level for the FN-1 gene, and   the gene expression level for the ACTA-2 gene is usable to estimate the gene expression level for the TGFB-1 gene.   
     
     
         40 . The method according to  claim 30 , wherein said collecting of the clinical parameters includes collecting polymerase chain reaction data. 
     
     
         41 . The method according to  claim 30 , wherein the biopsy tissue includes tissue from a renal allograft biopsy. 
     
     
         42 . The method according to  claim 30 , further including outputting the patient-specific probability of disease to at least one of a web browser, a desktop user interface, and an electronic medical record. 
     
     
         43 . The method according to  claim 30 , wherein said receiving of the clinical parameters for the individual patient further includes receiving the clinical parameters the individual patient from at least one of a database interface and a web interface. 
     
     
         44 . The method according to  claim 30 , wherein said collecting of the clinical parameters for the individual patient includes collecting an incomplete record wherein at least one of the clinical parameters is unknown, and wherein said outputting of the patient-specific probability of disease from the fully unsupervised Bayesian Belief Network model includes calculating the patient-specific probability of disease based on the incomplete record. 
     
     
         45 . The method according to  claim 30 , wherein said receiving of the clinical parameters for the individual patient into the fully unsupervised Bayesian Belief Network model includes receiving the clinical parameters for the individual patient into nodes of the fully unsupervised Bayesian Belief Network model, wherein each of the nodes include at least two bins, and wherein the bins in at least one of the nodes are sized such that data from the training database is evenly distributed between the bins. 
     
     
         46 . A method for determining a patient-specific probability of malignancy in at least one of a thyroid nodule and a breast, said method including:
 collecting clinical parameters from a plurality of patients to create a training database, the clinical parameters including at least four of fine needle aspiration biopsy results, ultrasound data, lymph node size, imaging data, Gail model cutoff, mammogram results, MRI results, breast size, and personal history of breast disease;   creating a fully unsupervised Bayesian Belief Network model using data from the training database;   validating the fully unsupervised Bayesian Belief Network model;   collecting the clinical parameters for an individual patient;   receiving the clinical parameters for the individual patient into the fully unsupervised Bayesian Belief Network model;   outputting the patient-specific probability of malignancy from the fully unsupervised Bayesian Belief Network model to a graphical user interface for use by a clinician; and   updating the fully unsupervised Bayesian Belief Network model using the clinical parameters for the individual patient and the patient-specific probability of malignancy.   
     
     
         47 . The method according to  claim 46 , wherein the imaging data includes results from electrical impedance scanning,
 wherein the results from the electrical impedance scanning include a definitely benign score, probably benign score, suspicious for cancer score, probably cancer score, and definitely cancer score,   wherein the fine needle aspiration biopsy results include an inadequate score, indeterminate score, negative score, and positive score,   wherein the ultrasound data include a complex cyst score, mixed score, simple cyst score, and solid score, and   wherein the lymph node size includes a less than 18 centimeters score, 18-31 centimeters score, and greater than 31 centimeters score.   
     
     
         48 . A method for determining a patient-specific probability of breast cancer, said method including:
 collecting clinical parameters from a plurality of patients to create a training database, the clinical parameters including Gail model cutoff, mammogram results, MRI results, breast size, and personal history of breast disease;   creating a fully unsupervised Bayesian Belief Network model using data from the training database;   validating the fully unsupervised Bayesian Belief Network model;   collecting the clinical parameters for an individual patient;   receiving the clinical parameters for the individual patient into the fully unsupervised Bayesian Belief Network model;   outputting the patient-specific probability of malignancy from the fully unsupervised Bayesian Belief Network model to a graphical user interface for use by a clinician; and   updating the fully unsupervised Bayesian Belief Network model using the clinical parameters for the individual patient and the patient-specific probability of malignancy.   
     
     
         49 . The method according to  claim 48 , wherein the Gail model cutoff includes one of a positive score and a negative score. 
     
     
         50 . The method according to  claim 48 , wherein the clinical parameters further include results from electrical impedance scanning. 
     
     
         51 . The method according to  claim 48 , wherein the clinical parameters further include ultrasound data and results of a clinical breast examination. 
     
     
         52 . The method according to  claim 48 , wherein said creating of the fully unsupervised Bayesian Belief Network model includes creating the fully unsupervised Bayesian Belief Network model without human-developed decision support rules. 
     
     
         53 . The method according to  claims 1 - 7  and  16 - 52 , further including updating the fully unsupervised Bayesian Belief Network model using a data pool, the data pool including clinical parameters from a plurality of patients.