WO2022099221A1

WO2022099221A1 - Compositions and methods for cancer diagnosis, prognosis and management

Info

Publication number: WO2022099221A1
Application number: PCT/US2021/059085
Authority: WO
Inventors: Heather H. JOHNSON
Original assignee: Johnson Heather H
Priority date: 2020-11-09
Filing date: 2021-11-12
Publication date: 2022-05-12

Abstract

The present invention includes compositions, methods and kits for cancer diagnosis, prognosis and management. In particular, the present invention provides compositions, methods and kits for screening and diagnosis of cancer, for diagnosis of clinically significant and insignificant cancer for risk stratification and need for immediate treatment, for monitoring cancer progression during active surveillance, for cancer treatment decision-making, for detection and prediction of cancer metastasis, for prediction of cancer recurrence, for prediction of treatment resistance and cancer remission after treatment, for measuring treatment efficacy, for monitoring treatment outcome, and for predicting cancer patient survival and survival time.

Description

COMPOSITIONS AND METHODS FOR CANCER DIAGNOSIS, PROGNOSIS AND MANAGEMENT

CROSS REFERENCE TO RELATED APPLICATIONS

[0001] This application claims the benefit of, and priority to, U.S. Provisional Patent Application 63111508, filed on November 09, 2020, and U.S. Provisional Patent Application 63274431, filed on November 01, 2021, the entire disclosures of which are hereby incorporated by reference.

FIELD OF THE INVENTION

[0002] The present invention relates to compositions and methods for cancer diagnosis and prognosis, cancer surveillance, measuring cancer treatment efficacy, monitoring treatment outcome, prediction of treatment resistance and cancer remission after treatment, and predicting cancer patient survival and survival time.

BACKGROUD OF THE INVENTION

[0003] Cancer is a leading cause of death worldwide. Many patients undergo surgery have recurrent cancer within a few years. Metastatic cancer is incurable. Currently there is no effective and safe treatment for metastatic cancer; many metastatic cancer patients develop treatment resistance. A lot of metastatic cancer patients die within a few years or a few months after diagnosis with deteriorated quality of life.

[0004] Several diagnostic methods are being used to detect and confirm the presence or absence of cancer in patients. Biopsy is a primary diagnostic tool in many types of cancers, yet biopsies tend to be invasive with some risks. A biopsy may not give accurate diagnosis due to error by diagnosing physician, sampling error, or insufficient sampling that misses tissues containing tumor, thus leading to false negative or false positive diagnosis. In addition, biopsy may cause pain, bleeding, infection, and damages of tissue or organ in the patient.

[0005] Endoscopy is another diagnostic method for certain types of cancer, such as stomach cancer, colorectal cancer. There are many types of scopes designed to view particular areas of the body. However, they are invasive and unpleasant for many patients.

[0006] Diagnostic imaging is used for cancer diagnosis for many types of cancer, in which an internal picture of the body and its structures are produced. Diagnostic imaging includes X-rays that can reveal abnormal areas indicating the presence of cancer, CAT scan (computerized axial tomography) using radiographic beams to create detailed computerized pictures taken with a specialized X-ray machine with more precise and clearer image of cancer tissue, Magnetic Resonance Imaging (MRI) using a powerful magnetic field to create detailed computer images of the body’s soft tissue, large blood vessels and major organs, Ultrasound using high-frequency sound waves to determine if a suspicious lump is solid or fluid. However, the imaging tests are expensive to conduct and require large machines and well-trained specialist for diagnosis. [0007] Detection of cancer biomarkers in patient body fluids and tissues, such as blood and urine, is a better tool for cancer diagnosis and prognosis. Cancer biomarkers are substances found in the blood, urine, stool, other bodily fluids, or tissues of patients. Cancer biomarkers can be used to diagnose cancer, determine and monitor cancer progression, predict patient response to certain cancer treatments, monitor patient treatment outcome, monitor cancer progression, predict cancer recurrence after treatment, diagnose and predict cancer metastasis, and predict patient survival time.

[0008] Currently, there are many cancer biomarkers in clinical use. For example, blood test for prostate specific antigen (PSA) is often used for prostate cancer diagnosis with higher than normal PSA levels indicating cancer. Y et most of these biomarkers lack high sensitivity and/or specificity, preventing them to be used as accurate and/or definitive cancer diagnostic method. In addition, few sensitive and specific biomarkers can be used to diagnose or predict metastatic cancer, thus preventing metastatic patients or patients who will develop metastatic cancer to be diagnosed early with more aggressive treatment to stop cancer spreading and save lives.

[0009] Many cancer treatments are effective only for cancers at certain stage or non-metastatic cancer, thus accurately assessing cancer staging and predicting cancer metastasis are important for successful treatment. Similarly, predicting and measuring treatment outcome, predicting cancer remission after treatment, monitoring cancer progression and predicting cancer recurrence after treatment, predicting the development of treatment resistance, and predicting patient survival are all pivotal in oncology treatment decisionmaking and cancer management. Currently few cancer biomarkers can be used for these tasks with high accuracy. Thus, it is imperative to identify and develop novel cancer biomarkers or novel biomarker panels with high sensitivity and specificity for cancer screening and diagnosis, patient stratification for treatment or surveillance, predicting cancer metastasis, predicting and monitoring patient response to treatments, determining cancer progression, predicting cancer recurrence and remission after treatment, predicting drug resistance, and predicting patient survival so more effective treatments can be used to save patients’ lives.

SUMMARY OF THE INVENTION

[0010] The present invention provides compositions and methods for cancer diagnosis and prognosis, treatment decision-making, managing cancer surveillance, measuring treatment efficacy, predicting treatment outcome, and predicting cancer patient survival.

[0011] In one aspect, the present invention provides a method to determine if a subject has cancer during cancer screening or diagnosis, or if a subject needs biopsy (e.g. a subject suspected of having prostate cancer, lung cancer, bladder cancer or kidney cancer). Generally, the method includes the following steps:

(1) obtaining a biological sample from a subject;

(2) using the biological sample to measure expression levels of a panel of genes, which comprises at least three or more genes selected from the group of genes including PIP5K1A, CCND1, GSTP1, CST3, CCNA1, LMTK2, MY06, HPN, CDK1, PSCA, PTEN, GOLM1, PMP22, EZH2, FGFR1, FN1, VEGFA, TMPRSS2, ANXA3, CRISP3, BIRC5, AMACR, HIF1A, KLK3, PCA3, KLK2, MSMB, FLT1, MMP9, AR, TERT, PGC, SPINK1, STAT3, STAT5, and TFF3,-

(3) determining a diagnostic score by (i) calculating the relative expression level of each gene as compared to one or more housekeeping genes, (ii) combining the calculated relative expression level of each gene with a predefined algorithm to determine a diagnostic score; and

(4) making a diagnosis by (a) comparing the diagnostic score with a predetermined cancer diagnostic score cutoff value, and (b) if the diagnostic score is higher than a predetermined cancer diagnostic score cutoff value, then the subject is diagnosed to have cancer or need biopsy, (c) if the diagnostic score is equal to or lower than a predetermined cancer diagnostic score cutoff value, then the subject is diagnosed to have no cancer or no need for biopsy.

[0012] In some embodiments, the group of genes consists of PIP5K1A, CCND1, GSTPI, CST3, CCNA1, LMTK2, MY06, HPN, CDKI, PSCA, PTEN, GOIM1, PMP22, EZH2, FGFRI, FNI, VEGFA, TMPRSS2, ANXA3, CRISP 3, BIRC5, AMACR, HIF1A, KLK3 and PCA3.

[0013] In another embodiment, the group of genes consists of CCND1, HIF1A, FGFR1, BIRC5, AMACR, CRISP3, FNI, HPN, MY06, PSCA, PMP22, GO1M1, EMTK2, EZH2, GSTP1, PCA3, VEGFA, CST3, PTEN, PIP5KIA, CDKI, TMPRSS2, ANXA3 and CCNA1.

[0014] In another embodiment, the group of genes consists of PIP5K1A, CCND1, GSTP1, CST3, CCNA1, LMTK2, MY06, HPN, CDK1, PSCA, PTEN, GO1M1, PMP22, EZH2, FGFR1, FNI, VEGFA, TMPRSS2, ANXA3, CRISP 3, BIRC5, AMACR, HIF1A and KLK3.

[0015] In another embodiment, the group of genes consists of PTEN, PIP5K1A, CDK1, TMPRSS2, ANXA3, HIF1A, FGFRI, BIRC5, AMACR, CRISP3, PMP22, G0IM1, EZH2, GSTPI, PCA3, VEGFA, CST3, CCNA1, CCNDI, FNI, MY06, KLK3 and PSCA.

[0016] In another embodiment, the group of genes consists of PTEN, CDK1, TMPRSS2, HIF1A, FGFRI, BIRC5, CRISP 3, FNI, HPN, PSCA, PMP22, LMTK2, EZH2, GSTP , PCA3, VEGFA, ANXA3 and KLK3.

[0017] In one aspect, the present invention provides a method to determine if a subject has high risk, clinically significant cancer and needs immediate treatment or low risk, clinically insignificant cancer and needs active surveillance (e.g. a subject diagnosed as having prostate cancer, lung cancer, bladder cancer or kidney cancer). Generally, the method includes the following steps:

(1) obtaining a biological sample from a subject diagnosed as having cancer;

(2) using the biological sample to measure expression levels of a panel of genes, which comprises at least three or more genes selected from the group of genes including PIP5K1A, CCNDI, GSTPI, CST3, CCNA1, LMTK2, MY06, HPN, CDKI, PSCA, PTEN, GO1M1, PMP22, EZH2, FGFRI, FNI, VEGFA, TMPRSS2, ANXA3, CRISP3, BIRC5, AMACR, HIF1A, KLK3, PCA3, KLK2, MSMB, FLT1, MMP9, AR, TERT, PGC, SPINK1, STAT3, STAT5, and TFF3;

(3) determining a stratification score by (i) calculating the relative expression level of each gene as compared to one or more housekeeping genes, (ii) combining the calculated relative expression level of each gene with a predefined algorithm to determine a stratification score; and

(4) making a diagnosis by (a) comparing the stratification score with a predetermined high risk stratification score cutoff value, and (b) if the stratification score is higher than a predetermined high risk stratification score cutoff value, then the subject is diagnosed as having high risk, clinically significant cancer and needs immediate treatment, (c) if the stratification score is equal to or lower than a predetermined high risk stratification score cutoff value, then the subject is diagnosed as having low risk, clinically insignificant cancer and needs active surveillance.

[0018] In some embodiments, the present invention provides a method for monitoring cancer progression and/or performing active surveillance to determine if a subject has cancer progression and needs immediate treatment or has no cancer progression and will continue active surveillance without treatment.

[0019] In some embodiments, the group of genes consists of PIP5K1A, CCNDI, GSTPI, CST3, CCNA1, LMTK2, MY06, HPN, CDKI, PSCA, PTEN, GO1M1, PMP22, EZH2, FGFRI, FNI, VEGFA, TMPRSS2, ANXA3, CRISP 3, BIRC5, AMACR, HIF1A, KLK3 and PCA3. [0020] In another embodiment, the group of genes consists of CCND1, HIF1A, FGFR1, BIRC5, AMACR, CRISP3, FNI, HPN, MY06, PSCA, PMP22, GO1M1, EMTK2, EZH2, GSTPI, PCA3, VEGFA, CST3, PTEN, PIP5K1A, CDK1, TMPRSS2, ANXA3 and CCNA1.

[0021] In another embodiment, the group of genes consists of PIP5K1A, CCND1, GSTPI, CST3, CCNA1, EMTK2, MY06, HPN, CDK1, PSCA, PTEN, G0IM1, PMP22, EZH2, FGFR1, FNI, VEGFA, TMPRSS2, ANXA3, CRISP 3, BIRC5, AMACR, HIF1A and KLK3.

[0022] In another embodiment, the group of genes consists of PTEN, PIP5K1A, CDK1, TMPRSS2, ANXA3, HIF1A, FGFR1, BIRC5, AMACR, CRISP3, PMP22, G0IM1, EZH2, GSTPI, PCA3, VEGFA, CST3, CCNA1, CCND1, FNI, MY06, KLK3 and PSCA.

[0023] In another embodiment, the group of genes consists of PTEN, CDK1, TMPRSS2, HIF1A, FGFR1, BIRC5, CRISP 3, FNI, HPN, PSCA, PMP22, LMTK2, EZH2, GSTPI, PCA3, VEGFA, ANXA3 and KLK3.

[0024] In one aspect, the present invention provides a method to predict if a subject having cancer will have metastatic cancer in the future (e.g. a subject diagnosed as having prostate cancer, lung cancer, bladder cancer or kidney cancer). Generally, the method includes the following steps:

(1) obtaining a biological sample from a subject diagnosed as having cancer;

(2) using the biological sample to measure expression levels of a panel of genes, which comprises at least three or more genes selected from the group of genes including PIP5K1A, CCND1, GSTPI, CST3, CCNA1, EMTK2, MY06, HPN, CDK1, PSCA, PTEN, G0EM1, PMP22, EZH2, FGFR1, FNI, VEGFA, TMPRSS2, ANXA3, CRISP3, BIRC5, AMACR, HIF1A, KLK3, PCA3, KLK2, MSMB, FLTI, MMP9, AR, TERT, PGC, SPINK1, STAT3, STAT5, and TFF3;

(3) determining a metastasis score by (i) calculating the relative expression level of each gene as compared to one or more housekeeping genes, (ii) combining the calculated relative expression level of each gene with a predefined algorithm to determine a metastasis score; and

(4) making a prediction by (a) comparing the metastasis score with a predetermined metastatic cancer score cutoff value, and (b) if the metastasis score is higher than a predetermined metastatic cancer score cutoff value, then the subject is predicted to have metastatic cancer in the future, (c) if the metastasis score is equal to or lower than a predetermined metastatic cancer score cutoff value, then the subject is predicted to not have metastatic cancer in the future.

[0025] In some embodiments, the method can be used for measuring metastatic cancer treatment efficacy during or after treatment by comparing a metastasis score obtained from a subject during or after metastatic cancer treatment, and if the metastasis score is higher than a predetermined metastatic cancer score cutoff value, then the subject is determined to still have metastatic cancer; or if the metastasis score is equal to or lower than a predetermined metastatic cancer score cutoff value, then the subject is determined to have no metastatic cancer.

[0026] In some embodiments, the group of genes consists of PIP5K1A, CCND1, GSTPI, CST3, CCNA1, LMTK2, MY06, HPN, CDK1, PSCA, PTEN, GOIM1, PMP22, EZH2, FGFR1, FNI, VEGFA, TMPRSS2, ANXA3, CRISP 3, BIRC5, AMACR, HIF1A, KLK3 and PCA3.

[0027] In another embodiment, the group of genes consists of CCND1, HIF1A, FGFR1, BIRC5, AMACR, CRISP3, FNI, HPN, MY06, PSCA, PMP22, G0EM1, EMTK2, EZH2, GSTPI, PCA3, VEGFA, CST3, PTEN, PIP5K1A, CDK1, TMPRSS2, ANXA3 and CCNA1. [0028] In another embodiment, the group of genes consists of PIP5K1A, CCND1, GSTPI, CST3, CCNA1, IMTK2, MY06, HPN, CDK1, PSCA, PTEN, GOIMI, PMP22, EZH2, FGFR1, FNI, VEGFA, TMPRSS2, ANXA3, CRISP 3, BIRC5, AMACR, HIFIA and KLK3.

[0029] In another embodiment, the group of genes consists of PTEN, PIP5K1A, CDK1, TMPRSS2, ANXA3, HIFIA, FGFR1, BIRC5, AMACR, CRISP3, PMP22, GOIMI, EZH2, GSTP1, PCA3, VEGFA, CST3, CCNAI, CCND1, FNI, MY06, KLK3 and PSCA.

[0030] In another embodiment, the group of genes consists of PTEN, CDK1, TMPRSS2, HIF1A, FGFR1, BIRC5, CRISP 3, FNI, HPN, PSCA, PMP22, IMTK2, EZH2, GSTPI, PCA3, VEGFA, ANXA3 and KLK3.

[0031] In one aspect, the present invention provides a method to predict the development of treatmentresistant cancer (e.g. a subject diagnosed as having prostate cancer, lung cancer, bladder cancer or kidney cancer). Generally, the method includes the following steps:

(1) obtaining a biological sample from a subject before cancer treatment;

(2) using the biological sample to measure expression levels of a panel of genes, which comprises at least three or more genes selected from the group of genes including PIP5K1A, CCND1, GSTPI, CST3, CCNA1, IMTK2, MY06, HPN, CDKI, PSCA, PTEN, GOIMI, PMP22, EZH2, FGFR1, FNI, VEGFA, TMPRSS2, ANXA3, CRISP3, BIRC5, AMACR, HIFIA, KLK3, PCA3, KLK2, MSMB, FLT1, MMP9, AR, TERT, PGC, SPINKI, STAT3, STAT5, and TFF3;

(3) determining a treatment-resistance score by (i) calculating the relative expression level of each gene as compared to one or more housekeeping genes, (ii) combining the calculated relative expression level of each gene with a predefined algorithm to determine a treatment-resistance score; and

(4) making a prediction by (a) comparing the treatment-resistance score with a predetermined treatmentresistance score cutoff value, and (b) if the treatment-resistance score is higher than a predetermined treatment-resistance score cutoff value, then the subject is predicted to have treatment-resistance in the future, (c) if the treatment-resistance score is equal to or lower than a predetermined treatment-resistance score cutoff value, then the subject is predicted to have no treatment-resistance in the future.

[0032] In some embodiments, the treatment-resistant cancer is castration-resistant prostate cancer.

[0033] In another embodiment, the treatment-resistant cancer is metastatic castration-resistant prostate cancer.

[0034] In some embodiments, the group of genes consists of PIP5K1A, CCND1, GSTPI, CST3, CCNA1, IMTK2, MY06, HPN, CDKI, PSCA, PTEN, GOIMI, PMP22, EZH2, FGFR1, FNI, VEGFA, TMPRSS2, ANXA3, CRISP 3, BIRC5, AMACR, HIFIA, KLK3 an PCA3.

[0035] In another embodiment, the group of genes consists of CCND1, HIF1A, FGFR1, BIRC5, AMACR, CRISP3, FNI, HPN, MY06, PSCA, PMP22, GOIMI, IMTK2, EZH2, GSTPI, PCA3, VEGFA, CST3, PTEN, PIP5KIA, CDKI, TMPRSS2, ANXA3 and CCNAI.

[0036] In another embodiment, the group of genes consists of PIP5K1A, CCND1, GSTPI, CST3, CCNA1, IMTK2, MY06, HPN, CDKI, PSCA, PTEN, GOIMI, PMP22, EZH2, FGFRI, FNI, VEGFA, TMPRSS2, ANXA3, CRISP 3, BIRC5, AMACR, HIFIA and KLK3.

[0037] In another embodiment, the group of genes consists of PTEN, PIP5KIA, CDKI, TMPRSS2, ANXA3, HIFIA, FGFRI, BIRC5, AMACR, CRISP3, PMP22, GOIMI, EZH2, GSTPI, PCA3, VEGFA, CST3, CCNAI, CCND1, FNI, MY06, KLK3 and PSCA. [0038] In another embodiment, the group of genes consists of PTEN, CDK1, TMPRSS2, HIF1A, FGFR1, BIRC5, CRISP 3, FN1, HPN, PSCA, PMP22, EMTK2, EZH2, GSTP1, PCA3, VEGFA, ANXA3 and KLK3.

[0039] In one aspect, the present invention provides a method to predict if a subject will have cancer recurrence after treatment (e.g. a subject diagnosed as having prostate cancer, lung cancer, bladder cancer or kidney cancer). Generally, the method includes the following steps:

(1) obtaining a biological sample from a subject before having cancer treatment;

(2) using the biological sample to measure expression levels of a panel of genes, which comprises at least three or more genes selected from the group of genes including PIP5K1A, CCND1, GSTPI, CST3, CCNA1, LMTK2, MY06, HPN, CDK1, PSCA, PTEN, G0LM1, PMP22, EZH2, FGFR1, FN1, VEGFA, TMPRSS2, ANXA3, CRISP3, BIRC5, AMACR, HIF1A, KLK3, PCA3, KLK2, MSMB, FLT1, MMP9, AR, TERT, PGC, SPINK1, STAT3, STAT5, and 7 3;

(3) determining a recurrence score by (i) calculating the relative expression level of each gene as compared to one or more housekeeping genes, (ii) combining the calculated relative expression level of each gene with a predefined algorithm to determine a recurrence score; and

(4) making cancer recurrence prediction by (a) comparing the recurrence score with a predetermined cancer recurrence score cutoff value, and (b) if the recurrence score is higher than a predetermined cancer recurrence score cutoff value, then the subject is predicted to have cancer recurrence after treatment in the future, (c) if the recurrence score is equal to or lower than a predetermined cancer recurrence score cutoff value, then the subject is predicted to have no cancer recurrence after treatment in the future.

[0040] In some embodiments, the group of genes consists of PIP5K1A, CCND1, GSTPI, CST3, CCNA1, LMTK2, MY06, HPN, CDK1, PSCA, PTEN, G0IM1, PMP22, EZH2, FGFR1, FN1, VEGFA, TMPRSS2, ANXA3, CRISP 3, BIRC5, AMACR, HIF1A, KLK3 an PCA3.

[0041] In another embodiment, the group of genes consists of CCND1, HIF1A, FGFR1, BIRC5, AMACR, CRISP3, FN1, HPN, MY06, PSCA, PMP22, G0LM1, LMTK2, EZH2, GSTPI, PCA3, VEGFA, CST3, PTEN, PIP5K1A, CDKI, TMPRSS2, ANXA3 and CCNA1.

[0042] In another embodiment, the group of genes consists of PIP5K1A, CCND1, GSTPI, CST3, CCNA1, LMTK2, MY06, HPN, CDKI, PSCA, PTEN, G0LM1, PMP22, EZH2, FGFR1, FN1, VEGFA, TMPRSS2, ANXA3, CRISP 3, BIRC5, AMACR, HIF1A and KLK3.

[0043] In another embodiment, the group of genes consists of PTEN, PIP5K1A, CDK1, TMPRSS2, ANXA3, HIF1A, FGFR1, BIRC5, AMACR, CRISP3, PMP22, G0IM1, EZH2, GSTPI, PCA3, VEGFA, CST3, CCNA1, CCND1, FN1, MY06, KLK3 and PSCA.

[0044] In another embodiment, the group of genes consists of PTEN, CDK1, TMPRSS2, HIF1A, FGFR1, BIRC5, CRISP 3, FN1, HPN, PSCA, PMP22, EMTK2, EZH2, GSTPI, PCA3, VEGFA, ANXA3 and KLK3.

[0045] In one aspect, the present invention provides a method to determine treatment efficacy by detecting cancer/residual cancer during or after treatment or detecting cancer recurrence after treatment (e.g. a subject diagnosed as having prostate cancer, lung cancer, bladder cancer or kidney cancer). Generally, the method includes the following steps:

(1) obtaining a biological sample from a subject during or after cancer treatment;

(2) using the biological sample to measure expression levels of a panel of genes, which comprises at least three or more genes selected from the group of genes including PIP5K1A, CCND1, GSTPI, CST3, CCNA1, LMTK2, MY06, HPN, CDKI, PSCA, PTEN, G0LM1, PMP22, EZH2, FGFR1, FNI, VEGFA, TMPRSS2, ANXA3, CRISP3, BIRC5, AMACR, HIF1A, KLK3, PCA3, KLK2, MSMB, FLT1, MMP9, AR, TERT, PGC, SPINK1, STAT3, STAT5, and TFF3; (3) determining a cancer diagnostic score by (i) calculating the relative expression level of each gene as compared to one or more housekeeping genes, (ii) combining the calculated relative expression level of each gene with a predefined algorithm to determine a cancer diagnostic score; and

(4) making a determination by (a) comparing the cancer diagnostic score with a predetermined cancer diagnostic score cutoff value, and (b) if the cancer diagnostic score is higher than a predetermined cancer diagnostic score cutoff value, then the subject is determined to have cancer/residual cancer during or after treatment or cancer recurrence after treatment, (c) if the cancer diagnostic score is equal to or lower than a predetermined cancer diagnostic score cutoff value, then the subject is determined to have no cancer/residual cancer during or after treatment or no cancer recurrence after treatment.

[0046] In some embodiments, the group of genes consists of PIP5K1A, CCND1, GSTPI, CST3, CCNA1, EMTK2, MY06, HPN, CDK1, PSCA, PTEN, GO1M1, PMP22, EZH2, FGFR1, FNI, VEGFA, TMPRSS2, ANXA3, CRISP 3, BIRC5, AMACR, HIF1A, KLK3 an PCA3.

[0047] In another embodiment, the group of genes consists of CCND1, HIF1A, FGFR1, BIRC5, AMACR, CRISP3, FNI, HPN, MY06, PSCA, PMP22, GO1M1, EMTK2, EZH2, GSTP1, PCA3, VEGFA, CST3, PTEN, PIP5K1A, CDKI, TMPRSS2, ANXA3 and CCNA1.

[0048] In another embodiment, the group of genes consists of PIP5K1A, CCND1, GSTPI, CST3, CCNA1, LMTK2, MY06, HPN, CDKI, PSCA, PTEN, G0LM1, PMP22, EZH2, FGFR1, FNI, VEGFA, TMPRSS2, ANXA3, CRISP 3, BIRC5, AMACR, HIF1A and KLK3.

[0049] In another embodiment, the group of genes consists of PTEN, PIP5K1A, CDK1, TMPRSS2, ANXA3, HIF1A, FGFR1, BIRC5, AMACR, CRISP3, PMP22, G0IM1, EZH2, GSTPI, PCA3, VEGFA, CST3, CCNA1, CCND1, FNI, MY06, KLK3 and PSCA.

[0050] In another embodiment, the group of genes consists of PTEN, CDK1, TMPRSS2, HIF1A, FGFR1, BIRC5, CRISP 3, FNI, HPN, PSCA, PMP22, LMTK2, EZH2, GSTPI, PCA3, VEGFA, ANXA3 and KLK3.

[0051] In one aspect, the present invention provides a method to predict cancer remission after treatment (e.g. a subject diagnosed as having prostate cancer, lung cancer, bladder cancer or kidney cancer). Generally, the method includes the following steps:

(1) obtaining a biological sample from a subject before cancer treatment;

(3) determining a remission score by (i) calculating the relative expression level of each gene as compared to one or more housekeeping genes, (ii) combining the calculated relative expression level of each gene with a predefined algorithm to determine a remission score; and

(4) making a prediction by (a) comparing the remission score with a predetermined remission score cutoff value and partial remission score cutoff value, and (b) if the remission score is higher than a predetermined remission score cutoff value, then the subject is predicted to have cancer remission after treatment in the future, (c) if the remission score is equal to or lower than a predetermined remission score cutoff value but higher than a predetermined partial remission score cutoff value, then the subject is predicted to have partial remission after treatment in the future, (d) if the remission score is equal to or lower than a predetermined partial remission score cutoff value, then the subject is predicted to not have cancer remission after treatment in the future. [0052] In some embodiments, the group of genes consists of PIP5K1A, CCND1, GSTPI, CST3, CCNA1, IMTK2, MY06, HPN, CDK1, PSCA, PTEN, G0IM1, PMP22, EZH2, FGFR1, FNI, VEGFA, TMPRSS2, ANXA3, CRISP 3, BIRC5, AMACR, HIFIA, KLK3 an PCA3.

[0053] In another embodiment, the group of genes consists of CCND1, HIF1A, FGFR1, BIRC5, AMACR, CRISP3, FNI, HPN, MY06, PSCA, PMP22, G0IM1, IMTK2, EZH2, GSTP1, PCA3, VEGFA, CST3, PTEN, PIP5KIA, CDKI, TMPRSS2, ANXA3 and CCNAI.

[0054] In another embodiment, the group of genes consists of PIP5K1A, CCND1, GSTPI, CST3, CCNA1, IMTK2, MY06, HPN, CDKI, PSCA, PTEN, GOIMI, PMP22, EZH2, FGFR1, FNI, VEGFA, TMPRSS2, ANXA3, CRISP 3, BIRC5, AMACR, HIFIA and KLK3.

[0055] In another embodiment, the group of genes consists of PTEN, PIP5K1A, CDK1, TMPRSS2, ANXA3, HIFIA, FGFR1, BIRC5, AMACR, CRISP3, PMP22, GOIMI, EZH2, GSTPI, PCA3, VEGFA, CST3, CCNAI, CCND1, FNI, MY06, KLK3 and PSCA.

[0056] In another embodiment, the group of genes consists of PTEN, CDK1, TMPRSS2, HIF1A, FGFR1, BIRC5, CRISP 3, FNI, HPN, PSCA, PMP22, IMTK2, EZH2, GSTPI, PCA3, VEGFA, ANXA3 and KLK3.

[0057] In one aspect, the present invention provides a method to predict survival of a subject diagnosed as having cancer (e.g. a subject diagnosed as having prostate cancer, lung cancer, bladder cancer or kidney cancer). Generally, the method includes the following steps:

(1) providing a biological sample from a subject diagnosed as having cancer;

(2) using the biological sample to measure expression levels of a panel of genes comprising at least three or more genes selected from the group of genes consists of PIP5K1A, CCND1, GSTPI, CST3, CCNA1, IMTK2, MY06, HPN, CDK1, PSCA, PTEN, GOIMI, PMP22, EZH2, FGFR1, FNI, VEGFA, TMPRSS2, ANXA3, CRISP3, BIRC5, AMACR, HIFIA, KLK3, PCA3, KLK2, MSMB, FLTI, MMP9, AR, TERT, PGC, SPINKI, STAT3, STAT5, and TFF3;

(3) determining a survival score by (i) calculating the relative expression level of each gene as compared to one or more housekeeping genes, (ii) combining the calculated relative expression level of each gene with a predefined algorithm to determine a survival score;

(4) making a prediction by (a) comparing the survival score with a predetermined survival score cutoff value, and (b) if the survival score is higher than a predetermined survival score cutoff value, then the subject is predicted to survive cancer in the future, (c) if the survival score is equal to or lower than a predetermined survival score cutoff value, then the subject is predicted to not survive or die of cancer in the future.

[0058] In some embodiments, the group of genes consists of PIP5K1A, CCND1, GSTPI, CST3, CCNA1, IMTK2, MY06, HPN, CDK1, PSCA, PTEN, GOIMI, PMP22, EZH2, FGFR1, FNI, VEGFA, TMPRSS2, ANXA3, CRISP 3, BIRC5, AMACR, HIFIA, KLK3 an PCA3.

[0059] In another embodiment, the group of genes consists of CCND1, HIF1A, FGFR1, BIRC5, AMACR, CRISP3, FNI, HPN, MY06, PSCA, PMP22, GOIMI, IMTK2, EZH2, GSTPI, PCA3, VEGFA, CST3, PTEN, PIP5KIA, CDKI, TMPRSS2, ANXA3 and CCNAI.

[0060] In another embodiment, the group of genes consists of PIP5K1A, CCND1, GSTPI, CST3, CCNA1, IMTK2, MY06, HPN, CDKI, PSCA, PTEN, GOIMI, PMP22, EZH2, FGFRI, FNI, VEGFA, TMPRSS2, ANXA3, CRISP 3, BIRC5, AMACR, HIFIA and KLK3. [0061] In another embodiment, the group of genes consists of PTEN, PIP5K1A, CDK1, TMPRSS2, ANXA3, HIF1A, FGFR1, BIRC5, AMACR, CRISPS, PMP22, GO1M1, EZH2, GSTP1, PCA3, VEGFA, CST3, CCNAI, CCND1, FNI, MY06, KLK3 and PSCA.

[0062] In another embodiment, the group of genes consists of PTEN, CDK1, TMPRSS2, HIF1A, FGFR1, BIRC5, CRISPS, FNI, HPN, PSCA, PMP22, LMTK2, EZH2, GSTPI, PCA3, VEGFA, ANXA3 and KLK3.

[0063] In one aspect, the present invention provides a method to predict survival time of a subject diagnosed as having cancer (e.g. a subject diagnosed as having prostate cancer, lung cancer, bladder cancer or kidney cancer). Generally, the method includes the following steps:

(1) providing a biological sample from a subject diagnosed as having cancer;

(2) using the biological sample to measure expression levels of a panel of genes comprising at least three or more genes selected from the group of genes consists of PIP5K1A, CCND1, GSTPI, CST3, CCNA1, LMTK2, MY06, HPN, CDKI, PSCA, PTEN, GOLM1, PMP22, EZH2, FGFRI, FNI, VEGFA, TMPRSS2, ANXA3, CRISP3, BIRC5, AMACR, HIF1A, KLK3, PCA3, KLK2, MSMB, FLT1, MMP9, AR, TERT, PGC, SPINK1, STATS, STAT5, and 7 3;

(3) determining a 5-year survival score by (i) calculating the relative expression level of each gene as compared to one or more housekeeping genes, (ii) combining the calculated relative expression level of each gene with a predefined algorithm to determine a 5-year survival score;

(4) determining a 10-year cancer survival score by (i) calculating relative expression level of each gene as compared to one or more housekeeping genes, (ii) combining the calculated relative expression level of each gene with a predefined algorithm to determine a 10-year survival score;

(5) determining a 20-year cancer survival score by (i) calculating relative expression level of each gene as compared to one or more housekeeping genes, (ii) combining the calculated relative expression level of each gene with a predefined algorithm to determine a 20-year survival score; and

(6) making a prediction: (a) if the 5-year survival score is equal to or lower than a predetermined 5-year survival score cutoff value, then the subject is predicted to have less than 5-year survival time, (b) if the 5- year survival score is higher than a predetermined 5 -year survival score cutoff value but the 10-year survival score is equal to or lower than a predetermined 10-year survival score cutoff value, then the subject is predicted to have 5-10 year survival time, (c) if the 10-year survival score is higher than a predetermined 10-year survival score cutoff value but the 20-year survival score is equal to or lower than a predetermined 20-year survival score cutoff value, then the subject is predicted to have 10-20 year survival time, (d) if the 20-year survival score is higher than a predetermined 20-year survival score cutoff value, then the subject is predicted to have more than 20-year survival time.

[0064] In some embodiments, the group of genes consists of PIP5K1A, CCND1, GSTPI, CST3, CCNA1, LMTK2, MY06, HPN, CDKI, PSCA, PTEN, GOIM1, PMP22, EZH2, FGFRI, FNI, VEGFA, TMPRSS2, ANXA3, CRISP 3, BIRC5, AMACR, HIF1A, KLK3 an PCA3.

[0065] In another embodiment, the group of genes consists of CCND1, HIF1A, FGFRI, BIRC5, AMACR, CRISP3, FNI, HPN, MY06, PSCA, PMP22, GO1M1, LMTK2, EZH2, GSTPI, PCA3, VEGFA, CST3, PTEN, PIP5KIA, CDKI, TMPRSS2, ANXA3 and CCNAI.

[0066] In another embodiment, the group of genes consists of PIP5K1A, CCND1, GSTPI, CSTS, CCNAI, LMTK2, MY06, HPN, CDKI, PSCA, PTEN, GO1M1, PMP22, EZH2, FGFRI, FNI, VEGFA, TMPRSS2, ANXA3, CRISPS, BIRC5, AMACR, HIF1A and KLK3.

[0067] In another embodiment, the group of genes consists of PTEN, PIP5K1A, CDKI, TMPRSS2, ANXA3, HIF1A, FGFRI, BIRC5, AMACR, CRISPS, PMP22, G0IM1, EZH2, GSTPI, PCA3, VEGFA, CSTS, CCNAI, CCNDI, FNI, MY06, KLK3 and PSCA. [0068] In another embodiment, the group of genes consists of PTEN, CDK1, TMPRSS2, HIF1A, FGFR1, BIRC5, CRISP 3, FN1, HPN, PSCA, PMP22, LMTK2, EZH2, GSTP1, PCA3, VEGFA, ANXA3 and KLK3.

[0069] In one aspect, the expression levels of a panel of genes are expression levels of mRNA, DNA methylation, protein, peptide, or their combination obtained from a biological sample of, but not limited to, blood, urine, ascites, other body fluids, tissue or cell from a subject.

[0070] In some embodiments, a kit is provided to make such measurement.

[0071] In some embodiments, an algorithm is used to make a diagnosis or prognosis by using expression levels of a panel of genes.

[0072] In another embodiment, a computer program is provided to make data analysis and diagnosis or prognosis, by taking the following steps: (1) receiving gene expression data on test genes in a panel; (2) determining an expression test score by (a) calculating the relative expression level of each gene in said panel as compared to one or more housekeeping genes, (b) combining the calculated relative expression level of each gene with a predefined algorithm to determine a score; (3) comparing the calculated expression test score to a predetermined diagnostic or prognostic score cutoff value (e.g. cancer diagnostic score cutoff value) to make a diagnosis or prognosis and display the result of diagnosis or prognosis.

INCORPORATION BY REFERENCE

[0073] All publications, patents, and patent applications mentioned in this specification are herein incorporated by reference to the same extent as if each individual publication, patent, or patent application was specifically and individually indicated to be incorporated by reference.

BRIEF DESCRIPTION OF THE DRAWINGS

[0074] The novel features and aspects of the invention are set forth with particularity in the appended claims. Full understanding of the features and advantages of the present invention will be obtained by reference to the following detailed description that sets forth illustrative embodiments, in which the principles of the invention are utilized with respect to the following drawings.

[0075] Table 1 Sensitivity, specificity, positive predictive value (PPV) and negative predictive value (NPV) of a 5-Gene Panel comprising of GSTP1, EMTK2, HPN, G0EM1 and PMP22 for prostate cancer diagnosis using prostate tissue specimens collected from 88 patients diagnosed with prostate cancer by pathological analysis of prostate tissue and 56 patients diagnosed with benign prostate by pathological analysis of prostate tissue. P value is shown.

[0076] Table 2 Sensitivity, specificity, positive predictive value (PPV) and negative predictive value (NPV) of an 8-Gene Panel comprising of MY06, EMTK2, PCA3, GSTP1, HPN, CCND1, FN1 and PMP22 for cancer risk stratification using prostate tissue specimens collected from 72 patients diagnosed with high risk, aggressive prostate cancer by pathological analysis of prostate tissue and 15 patients diagnosed with low risk, indolent prostate cancer by pathological analysis of prostate tissue. P value is shown.

[0077] Table 3 Sensitivity, specificity, positive predictive value (PPV) and negative predictive value (NPV) of a 25-Gene Panel comprising of PIP5K1A, CCND1, GSTP1, CST3, CCNA1, EMTK2, MY06, HPN, CDK1, PSCA, PTEN, GOLM1, PMP22, EZH2, FGFR1, FN1, VEGFA, TMPRSS2, ANXA3, CRISP3, BIRC5, AMACR, HIF1A, KLK3 and PCA3 for prostate cancer diagnosis using urine samples collected from 520 patients diagnosed with prostate cancer by pathological analysis and 94 patients diagnosed with benign prostate by pathological analysis. P value is shown.

[0078] Table 4 Sensitivity, specificity, positive predictive value (PPV) and negative predictive value (NPV) of a 25-Gene Panel comprising of PIP5K1A, CCND1, GSTP1, CST3, CCNA1, IMTK2, MY06, HPN, CDK1, PSCA, PTEN, G0LM1, PMP22, EZH2, FGFR1, FN1, VEGFA, TMPRSS2, ANXA3, CRISP3, BIRC5, AMACR, HIF1A, KLK3 and PCA3 for prostate cancer diagnosis using urine samples collected from 207 patients diagnosed with prostate cancer by pathological analysis and 189 patients diagnosed with benign prostate by pathological analysis. P value is shown.

[0079] Table 5 Sensitivity, specificity, positive predictive value (PPV) and negative predictive value (NPV) of a 14-Gene Panel comprising MPMP22. G0IM1, IMTK2, EZH2, GSTP1, PCA3, VEGFA, CST3, PTEN, PIP5K1A, CDK1, TMPRSS2, ANXA3 and CCND1 for prostate cancer diagnosis using urine samples collected from 202 patients diagnosed with prostate cancer by pathological analysis and 191 patients diagnosed with benign prostate by pathological analysis. P value is shown.

[0080] Table 6 Sensitivity, specificity, positive predictive value (PPV) and negative predictive value (NPV) of a 25-Gene Panel comprising of PIP5K1A, CCND1, GSTP1, CST3, CCNA1, IMTK2, MY06, HPN, CDK1, PSCA, PTEN, G0LM1, PMP22, EZH2, FGFR1, FN1, VEGFA, TMPRSS2, ANXA3, CRISP3, BIRC5, AMACR, HIF1A, KLK3 and PCA3 for prostate cancer diagnosis using prostate tissue specimens collected from 55 patients diagnosed with prostate cancer by pathological analysis and 99 patients diagnosed with benign prostate by pathological analysis. P value is shown.

[0081] Table 7 Sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV) and odds ratio (OR) and AUC of a 25-Gene Panel comprising of PIP5K1A, CCND1, GSTP1, CST3, CCNA1, LMTK2, MY06, HPN, CDK1, PSCA, PTEN, G0LM1, PMP22, EZH2, FGFR1, FN1, VEGFA, TMPRSS2, ANXA3, CRISP3, BIRC5, AMACR, HIF1A, KLK3 and PCA3, PSA and their combination for prostate cancer diagnosis using urine samples collected from 193 patients diagnosed with prostate cancer by pathological analysis and 222 patients diagnosed with benign prostate by pathological analysis. P value is shown.

[0082] Table 8 Diagnosis of pre- and post-prostatectomy urine samples by a 25-Gene Panel comprising of PIP5K1A, CCND1, GSTP1, CST3, CCNA1, IMTK2, MY06, HPN, CDK1, PSCA, PTEN, G0IM1, PMP22, EZH2, FGFR1, FN1, VEGFA, TMPRSS2, ANXA3, CRISP 3, BIRC5, AMACR, HIF1A, KLK3 PCA3.

[0083] Table 9 Sensitivity, specificity, positive predictive value (PPV) and negative predictive value (NPV) of a 14-Gene Panel comprising MPMP22. G0LM1, LMTK2, EZH2, GSTP1, PCA3, VEGFA, CST3, PTEN, PIP5K1A, CDK1, TMPRSS2, ANXA3 and CCND1 for determining if a subject has higher risk or lower risk prostate cancer using urine samples collected from 149 patients diagnosed with higher risk prostate cancer by pathological analysis and 53 patients diagnosed with lower risk prostate cancer by pathological analysis in a prospective urine cohort. P value is shown.

[0084] Table 10 Sensitivity, specificity, positive predictive value and negative predictive value of a 25- Gene Panel comprising of PIP5K1A, CCND1, GSTP1, CST3, CCNA1, LMTK2, MY06, HPN, CDK1, PSCA, PTEN, G0LM1, PMP22, EZH2, FGFR1, FN1, VEGFA, TMPRSS2, ANXA3, CRISP3, BIRC5, AMACR, HIF1A, KLK3 and PCA3 for determining if a subject has clinically significant prostate cancer or clinically insignificant prostate cancer using urine samples collected from 162 patients diagnosed with clinically significant cancer by pathological analysis and 45 patients diagnosed with clinically insignificant cancer by pathological analysis in a prospective urine cohort. P value is shown.

[0085] Table 11 Sensitivity, specificity, positive predictive value and negative predictive value of a 24- Gene Panel comprising of CCND1, HIF1A, FGFR1, BIRC5, AMACR, CRISPS, FN1, HPN, MY06, PSCA, PMP22, G0IM1, IMTK2, EZH2, GSTP1, PCA3, VEGFA, CST3, PTEN, PIP5K1A, CDK1, TMPRSS2, ANXA3, and CCNA1 for determining if a subject has clinically significant prostate cancer or clinically insignificant prostate cancer using urine samples collected from 164 patients diagnosed with clinically significant cancer by pathological analysis and 43 patients diagnosed with clinically insignificant cancer by pathological analysis in a prospective urine cohort. P value is shown.

[0086] Table 12 Sensitivity, specificity, positive predictive value and negative predictive value of a 24- Gene Panel comprising of CCND1, HIF1A, FGFR1, BIRC5, AMACR, CRISPS, FN1, HPN, MY06, PSCA, PMP22, GOIM1, IMTK2, EZH2, GSTPI, PCA3, VEGFA, CST3, PTEN, PIP5KIA, CDK1, TMPRSS2, ANXA3, and CCNA1 for determining if a subject has clinically significant prostate cancer or clinically insignificant prostate cancer using urine samples collected from 272 patients diagnosed with clinically significant cancer by pathological analysis and 248 patients diagnosed with clinically insignificant cancer by pathological analysis in a retrospective urine cohort. P value is shown.

[0087] Table 13 Sensitivity, specificity, positive predictive value and negative predictive value of a 24- Gene Panel comprising of CCND1, HIF1A, FGFR1, BIRC5, AMACR, CRISPS, FN1, HPN, MY06, PSCA, PMP22, G0EM1, EMTK2, EZH2, GSTP1, PCA3, VEGFA, CST3, PTEN, PIP5K1A, CDK1, TMPRSS2, ANXA3, and CCNA1, cancer stage, Gleason score and their combination for determining if a subject has clinically significant prostate cancer or clinically insignificant prostate cancer using patient urine samples collected from 434 patients in a combined prospective and retrospective urine cohort. P value is shown.

[0088] Table 14 Sensitivity, specificity, positive predictive value and negative predictive value of a 24- Gene Panel comprising of CCND1, HIF1A, FGFR1, BIRC5, AMACR, CRISPS, FN1, HPN, MY06, PSCA, PMP22, GOIM1, IMTK2, EZH2, GSTP1, PCA3, VEGFA, CST3, PTEN, PIP5K1A, CDK1, TMPRSS2, ANXA3, and CCNA1 for determining if a subject has clinically significant prostate cancer or clinically insignificant prostate cancer using prostate tissue specimens collected from 45 patients diagnosed with clinically significant cancer by pathological analysis and 104 patients diagnosed with clinically insignificant cancer by pathological analysis in a prostate tissue specimen cohort. P value is shown.

[0089] Table 15 Sensitivity, specificity, positive predictive value and negative predictive value of an 18- Gene Panel comprising of PTEN, CDK1, TMPRSS2, HIF1A, FGFR1, BIRC5, CRISPS, FN1, HPN, PSCA, PMP22, IMTK2, EZH2, GSTPI, PCA3, VEGFA, ANXA3 and KLK3 for prediction of prostate cancer metastasis using prostate tissue specimens collected from 19 patients with metastatic prostate cancer and 131 patients without metastatic prostate cancer during follow-up in a prostate tissue cohort. P value is shown.

[0090] Table 16 Univariate and multivariate cox regression analysis of an 18-Gene Panel comprising of PTEN, CDK1, TMPRSS2, HIF1A, FGFR1, BIRC5, CRISPS, FN1, HPN, PSCA, PMP22, IMTK2, EZH2, GSTPI, PCA3, VEGFA, ANXA3 and KLK3, Gleason score and PSA for prediction of prostate cancer metastasis using urine samples collected from 59 patients diagnosed with metastatic prostate cancer by bone scan and 148 patients diagnosed without metastatic prostate cancer by bone scan during follow-up in a prospective urine cohort. Hazard ratio (HR) is shown.

[0091] Table 17 Sensitivity, specificity, positive predictive value and negative predictive value of an 18- Gene Panel comprising of PTEN, CDK1, TMPRSS2, HIF1A, FGFR1, BIRC5, CRISPS, FN1, HPN, PSCA, PMP22, IMTK2, EZH2, GSTPI, PCA3, VEGFA, ANXA3 and KLK3 for prediction of prostate cancer metastasis using urine samples collected from 59 patients diagnosed with metastatic prostate cancer by bone scan and 148 patients diagnosed without metastatic prostate cancer by bone scan during follow-up in a prospective urine cohort. P value is shown.

[0092] Table 18 Sensitivity, specificity, positive predictive value and negative predictive value of an 18- Gene Panel comprising of PTEN, CDK1, TMPRSS2, HIF1A, FGFR1, BIRC5, CRISPS, FN1, HPN, PSCA, PMP22, LMTK2, EZH2, GSTPI, PCA3, VEGFA, ANXA3 and KLK3 for prediction of prostate cancer metastasis using urine samples collected from 8 patients diagnosed with metastatic prostate cancer by bone scan and 512 patients diagnosed without metastatic prostate cancer by bone scan during follow-up in a retrospective urine cohort. P value is shown.

[0093] Table 19 Sensitivity, specificity, positive predictive value and negative predictive value of an 18- Gene Panel comprising of PTEN, CDK1, TMPRSS2, HIF1A, FGFR1, BIRC5, CRISPS, FN1, HPN, PSCA, PMP22, LMTK2, EZH2, GSTPI, PCA3, VEGFA, ANXA3 and KLK3, PSA, Gleason score and their combination for prediction of prostate cancer metastasis using urine samples collected from 59 patients diagnosed with metastatic prostate cancer by bone scan and 148 patients diagnosed without metastatic prostate cancer by bone scan during follow-up in a prospective urine cohort. P value is shown.

[0094] Table 20 Sensitivity, specificity, positive predictive value and negative predictive value of a 23- Gene Panel comprising of PTEN, PIP5K1A, CDK1, TMPRSS2, ANXA3, HIF1A, FGFR1, BIRC5, AMACR, CRISPS, PMP22, GOLPH2, EZH2, GSTP1, PCA3, VEGFA, CST3, CCNA1, CCND1, FN1, MY06, KLK3, and PSCA for prediction of prostate cancer metastasis using urine samples collected from 67 patients diagnosed with metastatic prostate cancer by bone scan and 660 patients diagnosed without metastatic prostate cancer by bone scan during follow-up in a combined prospective and retrospective urine cohort. P value is shown.

[0095] Table 21 Sensitivity, specificity, positive predictive value and negative predictive value of a 24- Gene Panel comprising of CCND1, HIF1A, FGFR1, BIRC5, AMACR, CRISPS, FN1, HPN, MY06, PSCA, PMP22, GOIM1, IMTK2, EZH2, GSTPI, PCA3, VEGFA, CST3, PTEN, PIP5K1A, CDKI, IMPRSS2, ANXA3, and CCNA1 for prediction of prostate cancer metastasis using urine samples collected from 67 patients diagnosed with metastatic prostate cancer by bone scan and 660 patients diagnosed without metastatic prostate cancer by bone scan during follow-up in a combined prospective and retrospective urine cohort. P value is shown.

[0096] Table 22 Univariate and multivariate cox regression analysis of an 18-Gene Panel comprising of PTEN, CDKI, TMPRSS2, HIF1A, FGFR1, BIRC5, CRISPS, FN1, HPN, PSCA, PMP22, IMTK2, EZH2, GSTPI, PCA3, VEGFA, ANXA3 and KLK3, Gleason score and PSA for prediction of metastatic castrationresistant prostate cancer using urine samples collected from 14 patients diagnosed with metastatic castration-resistant prostate cancer and 59 patients diagnosed without metastatic castration-resistant prostate cancer during follow-up in a prospective mCRPC PCa Cohort. Hazard ratio (HR) is shown.

[0097] Table 23 Sensitivity, specificity, positive predictive value and negative predictive value of an 18- Gene Panel comprising of PTEN, CDK1, TMPRSS2, HIF1A, FGFR1, BIRC5, CRISPS, FN1, HPN, PSCA, PMP22, IMTK2, EZH2, GSTPI, PCA3, VEGFA, ANXA3 and KLK3, Gleason score, PSA, and their combination for prediction of metastatic castration-resistant prostate cancer using urine samples collected from 14 patients diagnosed with metastatic castration-resistant prostate cancer and 59 patients diagnosed without metastatic castration-resistant prostate cancer during follow-up in a prospective mCRPC PCa Cohort. P value is shown. [0098] Table 24 Univariate and multivariate cox regression analysis of an 18-Gene Panel comprising of PTEN, CDK1, TMPRSS2, HIF1A, FGFR1, BIRC5, CRISPS, FNI, HPN, PSCA, PMP22, IMTK2, EZH2, GSTPI, PCA3, VEGFA, ANXA3 and KLK3, Gleason score and PSA for prediction of metastatic castrationresistant prostate cancer using urine samples collected from 14 patients diagnosed with metastatic castration-resistant prostate cancer and 25 patients diagnosed without metastatic castration-resistant prostate cancer during follow-up in a prospective mCRPC MET Cohort. Hazard ratio (HR) is shown.

[0099] Table 25 Sensitivity, specificity, positive predictive value and negative predictive value of an 18- Gene Panel comprising of PTEN, CDK1, TMPRSS2, HIF1A, FGFR1, BIRC5, CRISPS, FNI, HPN, PSCA, PMP22, IMTK2, EZH2, GSTPI, PCA3, VEGFA, ANXA3 and KLK3, Gleason score, PSA, and their combination for prediction of metastatic castration-resistant prostate cancer using urine samples collected from 14 patients diagnosed with metastatic castration-resistant prostate cancer and 25 patients diagnosed without metastatic castration-resistant prostate cancer during follow-up in a prospective mCRPC MET Cohort. P value is shown.

[0100] Table 26 Sensitivity, specificity, positive predictive value (PPV) and negative predictive value (NPV) of a 23-Gene Panel comprising of PTEN, PIP5K1A, CDK1, TMPRSS2, ANXA3, HIF1A, FGFR1, BIRC5, AMACR, CRISP3, PMP22, GOLPH2, EZH2, GSTPI, PCA3, VEGFA, CST3, CCNA1, CCND1, FNI, MY06, KLK3 and PSCA for prediction of prostate cancer biochemical recurrence after surgery using prostate tissue specimens from 36 patients with biochemical recurrence after surgery and 104 patients without biochemical recurrence after surgery during follow-up in a prostate tissue cohort. P value is shown.

[0101] Table 27 Univariate and multivariate cox regression analysis of a 23-Gene Panel comprising of PTEN, PIP5K1A, CDK1, TMPRSS2, ANXA3, HIF1A, FGFR1, BIRC5, AMACR, CRISPS, PMP22, GOLPH2, EZH2, GSTPI, PCA3, VEGFA, CST3, CCNA1, CCND1, FNI, MY06, KLK3 and PSCA, Gleason score and cancer stage for prediction of cancer recurrence-free survival after surgery by using urine samples collected from 42 patients with biochemical recurrence after surgery and 372 patients without biochemical recurrence after surgery during follow-up in a retrospective urine cohort. Hazard ratio (HR) is shown.

[0102] Table 28 Sensitivity, specificity, positive predictive value and negative predictive value of a 23- Gene Panel comprising of PTEN, PIP5K1A, CDK1, TMPRSS2, ANXA3, HIF1A, FGFR1, BIRC5, AMACR, CRISP3, PMP22, GOLPH2, EZH2, GSTPI, PCA3, VEGFA, CST3, CCNA1, CCND1, FNI, MY06, KLK3 and PSCA for prediction of prostate cancer recurrence after surgery using urine samples collected from 46 patients with prostate cancer recurrence and 474 patients without prostate cancer recurrence after surgery during follow-up in a retrospective urine cohort. P value is shown.

[0103] Table 29 Sensitivity, specificity, positive predictive value and negative predictive value of an 18- Gene Panel comprising of PTEN, CDK1, TMPRSS2, HIF1A, FGFR1, BIRC5, CRISPS, FNI, HPN, PSCA, PMP22, IMTK2, EZH2, GSTPI, PCA3, VEGFA, ANXA3 and KLK3, for prediction of prostate cancer recurrence after surgery using urine samples collected from 46 patients with prostate cancer recurrence and 474 patients without prostate cancer recurrence after surgery during follow-up in a retrospective urine cohort. P value is shown.

[0104] Table 30 Sensitivity, specificity, positive predictive value and negative predictive value of a 24- Gene Panel comprising of PIP5K1A, CCND1, GSTPI, CST3, CCNA1, IMTK2, MY06, HPN, CDK1, PSCA, PTEN, G0IM1, PMP22, EZH2, FGFR1, FNI, VEGFA, TMPRSS2, ANXA3, CRISPS, BIRC5, AMACR, HIF1A, KLK3, Gleason score, cancer stage, and their combination for predicting cancer remission after treatment by using prostate tissue specimens collected from 160 patients with remission and 65 patients with no remission in a prostate tissue specimen TCGA Cohort. P value is shown. [0105] Table 31 Univariate and multivariate cox regression analysis of a 24-Gene Panel comprising of PIP5K1A, CCND1, GSTP1, CST3, CCNA1, IMTK2, MY06, HPN, CDK1, PSCA, PTEN, G0IM1, PMP22, EZH2, FGFR1, FN1, VEGFA, TMPRSS2, ANXA3, CRISP 3, BIRC5, AMACR, HIF1A, KLK3, Gleason score, and cancer stage for prediction of patient survival by using prostate tissue specimens collected from 482 patients with survival and 9 patients without survival in a prostate tissue specimen TCGA Cohort. Hazard ratio (HR) is shown.

[0106] Table 32 Sensitivity, specificity, positive predictive value and negative predictive value of a 24- Gene Panel comprising of PIP5K1A, CCND1, GSTP1, CST3, CCNA1, IMTK2, MY06, HPN, CDK1, PSCA, PTEN, G0IM1, PMP22, EZH2, FGFR1, FN1, VEGFA, TMPRSS2, ANXA3, CRISP3, BIRC5, AMACR, HIF1A, KLK3, Gleason score, and cancer stage for prediction of patient survival by using prostate tissue specimens collected from 482 patients with survival and 9 patients without survival in a prostate tissue specimen TCGA Cohort. P value is shown.

[0107] Table 33 Sensitivity, specificity, positive predictive value and negative predictive value of a 25- Gene Panel comprising of PIP5K1A, CCND1, GSTP1, CST3, CCNA1, IMTK2, MY06, HPN, CDK1, PSCA, PTEN, G0IM1, PMP22, EZH2, FGFR1, FN1, VEGFA, TMPRSS2, ANXA3, CRISP3, BIRC5, AMACR, HIF1A, KLK3 and PCA3 for predicting if a prostate cancer patient will have >5 year survival time or <5 year survival time using prostate tissue specimens collected from 59 patients with >5 year survival time and 81 patients with <5 year survival time in a prostate tissue specimen cohort. P value is shown.

[0108] Figure 1 ROC (Receiver Operating Characteristic) curve of a 5-Gene Panel comprising of GSTP1, LMTK2, HPN, G0LM1 and PMP22 for diagnosis of prostate cancer using prostate tissue specimens collected from 88 patients diagnosed with prostate cancer by pathological analysis of prostate tissue and 56 patients diagnosed with benign prostate by pathological analysis of prostate tissue. Value of AUC (Area Under the Curve) is shown.

[0109] Figure 2 ROC (Receiver Operating Characteristic) curve of an 8-Gene Panel comprising of MY06, IMTK2, PCA3, GSTP1, HPN, CCND1, FN1 and PMP22 for cancer diagnosis using prostate tissue specimens collected from 72 patients diagnosed with high risk, aggressive prostate cancer by pathological analysis and 15 patients diagnosed with low risk, indolent prostate cancer by pathological analysis. Value of AUC (Area Under the Curve) is shown.

[0110] Figure 3 ROC (Receiver Operating Characteristic) curve of a 25-Gene Panel comprising of PIP5K1A, CCND1, GSTP1, CST3, CCNA1, IMTK2, MY06, HPN, CDK1, PSCA, PTEN, G0IM1, PMP22, EZH2, FGFR1, FN1, VEGFA, TMPRSS2, ANXA3, CRISP 3, BIRC5, AMACR, HIF1A, KLK3 PCA3 for prostate cancer diagnosis using urine samples collected from 520 patients diagnosed with prostate cancer by pathological analysis and 94 patients diagnosed with benign prostate by pathological analysis in a retrospective urine cohort. Value of AUC (Area Under the Curve) is shown.

[0111] Figure 4 ROC (Receiver Operating Characteristic) curve of a 25-Gene Panel comprising of PIP5K1A, CCND1, GSTP1, CST3, CCNA1, IMTK2, MY06, HPN, CDK1, PSCA, PTEN, GOIM1, PMP22, EZH2, FGFR1, FN1, VEGFA, TMPRSS2, ANXA3, CRISP 3, BIRC5, AMACR, HIF1A, KLK3 PCA3 for prostate cancer diagnosis using urine samples collected from 207 patients diagnosed with prostate cancer by pathological analysis and 189 patients diagnosed with benign prostate by pathological analysis in a prospective urine cohort. Value of AUC (Area Under the Curve) is shown.

[0112] Figure 5 ROC (Receiver Operating Characteristic) curve of a 14-Gene-Panel comprising oiPMP22, G0IM1, IMTK2, EZH2, GSTP1, PCA3, VEGFA, CST3, PTEN, PIP5K1A, CDK1, TMPRSS2, ANXA3 and CCND1 for prostate cancer diagnosis using urine samples from 202 patients diagnosed with prostate cancer by pathological analysis and 191 patients diagnosed with benign prostate by pathological analysis in a combined urine cohort. Value of AUC (Area Under the Curve) is shown.

[0113] Figure 6 ROC (Receiver Operating Characteristic) curve of a 25-Gene Panel comprising of PIP5K1A, CCND1, GSTPI, CST3, CCNA1, EMTK2, MY06, HPN, CDK1, PSCA, PTEN, GO1M1, PMP22, EZH2, FGFR1, FN1, VEGFA, TMPRSS2, ANXA3, CRISP 3, BIRC5, AMACR, HIF1A, KLK3 a PCA3 for determining if a subject has prostate cancer or benign prostate using prostate tissue specimens collected from 55 patients diagnosed with prostate cancer by pathological analysis and 99 patients diagnosed with benign prostate by pathological analysis in a prostate tissue cohort. Value of AUC (Area Under the Curve) is shown.

[0114] Figure 7 ROC (Receiver Operating Characteristic) curves of a 25-Gene Panel comprising of PIP5K1A, CCND1, GSTPI, CST3, CCNA1, EMTK2, MY06, HPN, CDK1, PSCA, PTEN, G0IM1, PMP22, EZH2, FGFR1, FN1, VEGFA, TMPRSS2, ANXA3, CRISP3, BIRC5, AMACR HIF1A, KLK3 and PCA3, PSA and their combination for determining if a subject has prostate cancer or benign prostate using urine samples collected from 193 patients diagnosed with prostate cancer by pathological analysis and 222 patients diagnosed with benign prostate diagnosed by pathological analysis. A. ROC curve of the 25-Gene Panel. B. ROC curve of PSA. C. ROC curve of combining the 25-Gene Panel and PSA. Value of AUC (Area Under the Curve) is shown.

[0115] Figure 8 ROC (Receiver Operating Characteristic) curve of a 14-Gene Panel comprising o PMP22, G0IM1, IMTK2, EZH2, GSTPI, PCA3, VEGFA, CST3, PTEN, PIP5K1A, CDK1, TMPRSS2, ANXA3 and CCND1 for determining if a subject has higher risk or lower risk prostate cancer using urine samples collected from 149 patients diagnosed with higher risk prostate cancer by pathological analysis and 53 patients diagnosed with lower risk prostate cancer by pathological analysis in a urine cohort. Value of AUC (Area Under the Curve) is shown.

[0116] Figure 9 ROC (Receiver Operating Characteristic) curve of a 25-Gene Panel comprising of PIP5K1A, CCND1, GSTPI, CST3, CCNA1, LMTK2, MY06, HPN, CDK1, PSCA, PTEN, G0IM1, PMP22, EZH2, FGFR1, FN1, VEGFA, TMPRSS2, ANXA3, CRISP 3, BIRC5, AMACR, HIF1A, KLK3 an PCA3 for determining if a subject has clinically significant prostate cancer or clinically insignificant prostate cancer using urine samples collected from 162 patients diagnosed with clinically significant cancer by pathological analysis and 45 patients diagnosed with clinically insignificant cancer by pathological analysis in a prospective urine cohort. Value of AUC (Area Under the Curve) is shown.

[0117] Figure 10 ROC (Receiver Operating Characteristic) curve of a 24-Gene Panel comprising of CCND1, HIF1A, FGFR1, BIRC5, AMACR, CRISP 3, FN1, HPN, MY06, PSCA, PMP22, G0EM1, EMTK2, EZH2, GSTPI, PCA3, VEGFA, CST3, PTEN, PIP5K1A, CDKI, TMPRSS2, ANXA3, and CCNA1 for determining if a subject has clinically significant prostate cancer or clinically insignificant prostate cancer using urine samples collected from 164 patients diagnosed with clinically significant cancer by pathological analysis and 43 patients diagnosed with clinically insignificant cancer by pathological analysis in a prospective urine cohort. Value of AUC (Area Under the Curve) is shown.

[0118] Figure 11 ROC (Receiver Operating Characteristic) curve of a 24-Gene Panel comprising of CCND1, HIF1A, FGFR1, BIRC5, AMACR, CRISP 3, FN1, HPN, MY06, PSCA, PMP22, G0EM1, EMTK2, EZH2, GSTPI, PCA3, VEGFA, CST3, PTEN, PIP5K1A, CDKI, TMPRSS2, ANXA3, and CCNA1 for determining if a subject has clinically significant prostate cancer or clinically insignificant prostate cancer using urine samples collected from 272 patients diagnosed with clinically significant cancer by pathological analysis and 248 patients diagnosed with clinically insignificant cancer by pathological analysis in a retrospective urine cohort. Value of AUC (Area Under the Curve) is shown. [0119] Figure 12 ROC (Receiver Operating Characteristic) curve of a 24-Gene Panel comprising of CCND1, HIF1A, FGFR1, BIRC5, AMACR, CRISPS, FNI, HPN, MY06, PSCA, PMP22, G0IM1, IMTK2, EZH2, GSTPI, PCA3, VEGFA, CST3, PTEN, PIP5KIA, CDK1, TMPRSS2, ANXA3, and CCNA1 for determining if a subject has clinically significant prostate cancer or clinically insignificant prostate cancer using urine samples collected from 434 patients in a combined prospective and retrospective urine cohort. Value of AUC (Area Under the Curve) is shown.

[0120] Figure 13 ROC (Receiver Operating Characteristic) curve of a 24-Gene Panel comprising of CCND1, HIF1A, FGFR1, BIRC5, AMACR, CRISPS, FNI, HPN, MY06, PSCA, PMP22, GOIM1, EMTK2, EZH2, GSTPI, PCA3, VEGFA, CST3, PTEN, PIP5KIA, CDK1, TMPRSS2, ANXA3, and CCNA1 for determining if a subject has clinically significant prostate cancer or clinically insignificant prostate cancer using prostate tissue specimens collected from 45 patients diagnosed with clinically significant cancer by pathological analysis and 104 patients diagnosed with clinically insignificant cancer by pathological analysis in a prostate tissue specimen cohort. Value of AUC (Area Under the Curve) is shown.

[0121] Figure 14 ROC (Receiver Operating Characteristic) curve of an 18-Gene Panel comprising PTEN, CDK1, TMPRSS2, HIF1A, FGFR1, BIRC5, CRISPS, FNI, HPN, PSCA, PMP22, IMTK2, EZH2, GSTP1, PCA3, VEGFA, ANXA3, and KLK3 for prediction of prostate cancer metastasis using prostate tissue specimens from 19 patients diagnosed with metastatic prostate cancer by bone scan and 131 patients diagnosed without metastatic prostate cancer by bone scan during follow-up. Value of AUC (Area Under the Curve) is shown.

[0122] Figure 15 Kaplan-Meier plots of an 18-Gene Panel comprising of PTEN, CDK1, TMPRSS2, HIF1A, FGFR1, BIRC5, CRISP 3, FNI, HPN, PSCA, PMP22, IMTK2, EZH2, GSTPI, PCA3, VEGFA, ANXA3, and KLK3 for predicting metastasis-free survival by using urine samples collected from 59 patients diagnosed with metastatic prostate cancer by bone scan and 148 patients diagnosed without metastatic prostate cancer by bone scan during follow-up in a prospective urine cohort. Uog rank P value is shown.

[0123] Figure 16 ROC (Receiver Operating Characteristic) curves of an 18-Gene Panel comprising PTEN, CDK1, TMPRSS2, HIF1A, FGFR1, BIRC5, CRISPS, FNI, HPN, PSCA, PMP22, IMTK2, EZH2, GSTPI, PCA3, VEGFA, ANXA3, and KLK3 for prediction of prostate cancer metastasis using urine samples collected from 59 patients diagnosed with metastatic prostate cancer by bone scan and 148 patients diagnosed without metastatic prostate cancer by bone scan during follow-up in a prospective urine cohort. Value of AUC (Area Under the Curve) is shown.

[0124] Figure 17 ROC (Receiver Operating Characteristic) curves of an 18-Gene Panel comprising PTEN, CDK1, TMPRSS2, HIF1A, FGFR1, BIRC5, CRISPS, FNI, HPN, PSCA, PMP22, IMTK2, EZH2, GSTPI, PCA3, VEGFA, ANXA3, and KLK3 for prediction of prostate cancer metastasis using urine samples collected from 8 patients diagnosed with metastatic prostate cancer by bone scan and 512 patients diagnosed without metastatic prostate cancer by bone scan during follow-up in a retrospective urine cohort. Value of AUC (Area Under the Curve) is shown.

[0125] Figure 18 ROC (Receiver Operating Characteristic) curves of an 18-Gene Panel comprising of PTEN, CDK1, TMPRSS2, HIF1A, FGFR1, BIRC5, CRISPS, FNI, HPN, PSCA, PMP22, IMTK2, EZH2, GSTPI, PCA3, VEGFA, ANXA3 and KLK3, PSA, Gleason score and their combination for prediction of prostate cancer metastasis using urine samples collected from 59 patients diagnosed with metastatic prostate cancer by bone scan and 148 patients diagnosed without metastatic prostate cancer by bone scan during follow-up in a prospective urine cohort. A. ROC curve of the 18-Gene Panel. B. ROC curve of PSA. C. ROC curve of Gleason score. D. ROC curve of combining the 18-Gene Panel with PSA and Gleason score. Value of AUC (Area Under the Curve) is shown.

[0126] Figure 19 ROC (Receiver Operating Characteristic) curve of a 23-Gene Panel comprising of PTEN, PIP5K1A, CDK1, TMPRSS2, ANXA3, HIF1A, FGFR1, BIRC5, AMACR, CRISPS, PMP22, GO1M1, EZH2, GSTPI, PCA3, VEGFA, CST3, CCNA1, CCND1, FN1, MY06, KLK3 and PSCA for prediction of prostate cancer metastasis using urine samples collected from 67 patients diagnosed with metastatic prostate cancer by bone scan and 660 patients diagnosed without metastatic prostate cancer by bone scan during follow-up in a combined prospective and retrospective urine cohort. Value of AUC (Area Under the Curve) is shown.

[0127] Figure 20 ROC (Receiver Operating Characteristic) curve of a 24-Gene Panel comprising of CCND1, HIF1A, FGFR1, BIRC5, AMACR, CRISPS, FNI, HPN, MY06, PSCA, PMP22, G0IM1, IMTK2, EZH2, GSTPI, PCA3, VEGFA, CST3, PTEN, PIP5K1A, CDKI, TMPRSS2, ANXA3, and CCNA1 for prediction of prostate cancer metastasis using urine samples collected from 67 patients diagnosed with metastatic prostate cancer by bone scan and 660 patients diagnosed without metastatic prostate cancer by bone scan during follow-up in a combined prospective and retrospective urine cohort. Value of AUC (Area Under the Curve) is shown.

[0128] Figure 21 Kaplan-Meier plot of an 18-Gene Panel comprising of PTEN, CDK1, TMPRSS2, HIF1A, FGFR1, BIRC5, CRISP 3, FNI, HPN, PSCA, PMP22, IMTK2, EZH2, GSTPI, PCA3, VEGFA, ANXA3 and KLK3 for prediction of metastatic castration-resistant prostate cancer-free survival using urine samples collected from 14 patients diagnosed with metastatic castration-resistant prostate cancer and 59 patients diagnosed without metastatic castration-resistant prostate cancer during follow-up in a prospective mCRPC PCa Cohort. Uog rank P value is shown.

[0129] Figure 22 ROC (Receiver Operating Characteristic) curves of an 18-Gene Panel comprising of PTEN, CDKI, TMPRSS2, HIF1A, FGFR1, BIRC5, CRISPS, FNI, HPN, PSCA, PMP22, IMTK2, EZH2, GSTPI, PCA3, VEGFA, ANNAS and KLK3, Gleason score, PSA, and their combination for prediction of metastatic castration-resistant prostate cancer using urine samples collected from 14 patients diagnosed with metastatic castration-resistant prostate cancer and 59 patients diagnosed without metastatic castrationresistant prostate cancer during follow-up in a prospective mCRPC PCa Cohort. A. ROC curve of the 18- Gene Panel. B. ROC curve of Gleason score. C. ROC curve of PSA. D. ROC curve of combining the 18- Gene Panel with Gleason score and PSA. Value of AUC (Area Under the Curve) is shown.

[0130] Figure 23 Kaplan-Meier plot of an 18-Gene Panel comprising of PTEN, CDK1, TMPRSS2, HIF1A, FGFR1, BIRC5, CRISPS, FNI, HPN, PSCA, PMP22, IMTK2, EZH2, GSTPI, PCA3, VEGFA, ANXA3 and KLK3 for prediction of metastatic castration-resistant prostate cancer using urine samples collected from 14 patients diagnosed with metastatic castration-resistant prostate cancer and 25 patients diagnosed without metastatic castration-resistant prostate cancer during follow-up in a prospective mCRPC MET Cohort. Uog rank P value is shown.

[0131] Figure 24 ROC (Receiver Operating Characteristic) curves of an 18-Gene Panel comprising of PTEN, CDKI, TMPRSS2, HIF1A, FGFR1, BIRC5, CRISPS, FNI, HPN, PSCA, PMP22, IMTK2, EZH2, GSTPI, PCA3, VEGFA, ANXA3 and KLK3, Gleason score, PSA, and their combination for prediction of metastatic castration-resistant prostate cancer using urine samples collected from 14 patients diagnosed with metastatic castration-resistant prostate cancer and 25 patients diagnosed without metastatic castrationresistant prostate cancer during follow-up in a prospective mCRPC MET Cohort. A. ROC curve of the 18- Gene Panel. B. ROC curve of Gleason score. C. ROC curve of PSA. D. ROC curve of combining the 18- Gene Panel with Gleason score and PSA. Value of AUC (Area Under the Curve) is shown. [0132] Figure 25 ROC (Receiver Operating Characteristic) curve of a 23-Gene Panel comprising of PTEN, PIP5K1A, CDK1, TMPRSS2, ANXA3, HIF1A, FGFR1, BIRC5, AMACR, CRISPS, PMP22, GOLPH2, EZH2, GSTPI, PCA3, VEGFA, CST3, CCNA1, CCND1, FN1, MY06, KLK3 and PSCA for prediction of prostate cancer biochemical recurrence after surgery using prostate tissue specimens from 36 patients with biochemical recurrence after surgery and 104 patients without biochemical recurrence after surgery during follow-up in a prostate tissue cohort. Value of AUC (Area Under the Curve) is shown.

[0133] Figure 26 Kaplan-Meier plot of a 23-Gene Panel comprising of PTEN, PIP5K1A, CDK1, TMPRSS2, ANXA3, HIF1A, FGFR1, BIRC5, AMACR, CRISPS, PMP22, GOLPH2, EZH2, GSTPI, PCA3, VEGFA, CST3, CCNA1, CCND1, FNI, MY06, KLK3 and PSCA, Gleason score and cancer stage for prediction of cancer recurrence-free survival after surgery by using urine samples collected from 42 patients with biochemical recurrence after surgery and 372 patients without biochemical recurrence after surgery during follow-up in a retrospective urine cohort, a. Kaplan-Meier plot of the 23-Gene Panel, b. Kaplan-Meier plot of Gleason score, c. Kaplan-Meier plot of cancer stage. Log rank P value is shown.

[0134] Figure 27 ROC (Receiver Operating Characteristic) curve of a 23-Gene Panel comprising of PTEN, PIP5K1A, CDK1, TMPRSS2, ANXA3, HIF1A, FGFR1, BIRC5, AMACR, CRISPS, PMP22, GOLPH2, EZH2, GSTPI, PCA3, VEGFA, CST3, CCNA1, CCND1, FNI, MY06, KLK3 and PSCA for prediction of prostate cancer recurrence after surgery using urine samples collected from 46 patients with prostate cancer recurrence and 474 patients without prostate cancer recurrence after surgery during follow-up in a retrospective urine cohort. Value of AUC (Area Under the Curve) is shown.

[0135] Figure 28 ROC (Receiver Operating Characteristic) curve of an 18-Gene Panel comprising of PTEN, CDK1, TMPRSS2, HIF1A, FGFR1, BIRC5, CRISPS, FNI, HPN, PSCA, PMP22, IMTK2, EZH2, GSTPI, PCA3, VEGFA, ANXA3 and KLK3 for prediction of prostate cancer recurrence after surgery using urine samples collected from 46 patients with prostate cancer recurrence and 474 patients without prostate cancer recurrence after surgery during follow-up in a retrospective urine cohort. Value of AUC (Area Under the Curve) is shown.

[0136] Figure 29 ROC (Receiver Operating Characteristic) curve of a 24-Gene Panel comprising of PIP5K1A, CCND1, GSTPI, CST3, CCNA1, IMTK2, MY06, HPN, CDK1, PSCA, PTEN, G0IM1, PMP22, EZH2, FGFR1, FNI, VEGFA, TMPRSS2, ANXA3, CRISP 3, BIRC5, AMACR, HIF1A, KLK3, Gleason score, cancer stage, and their combination for predicting cancer remission after treatment by using prostate tissue specimens collected from 160 patients with remission and 65 patients with no remission in a prostate tissue specimen TCGA Cohort. A. ROC curve of the 24-Gene Panel. B. ROC curve of Gleason score. C. ROC curve of cancer stage. D. ROC curve of combining the 24-Gene Panel with Gleason score and cancer stage. Value of AUC (Area Under the Curve) is shown.

[0137] Figure 30 Kaplan-Meier plot of a 24-Gene Panel comprising of PIP5K1A, CCND1, GSTPI, CST3, CCNA1, LMTK2, MY06, HPN, CDKI, PSCA, PTEN, G0IM1, PMP22, EZH2, FGFR1, FNI, VEGFA, TMPRSS2, ANXA3, CRISP3, BIRC5, AMACR, HIF1A, KLK3 for prediction of patient survival by using prostate tissue specimens collected from 482 patients with survival and 9 patients without survival in a prostate tissue specimen TCGA Cohort. Log rank p value is shown.

[0138] Figure 31 ROC (Receiver Operating Characteristic) curve of a 24-Gene Panel comprising of PIP5K1A, CCND1, GSTPI, CST3, CCNA1, LMTK2, MY06, HPN, CDK1, PSCA, PTEN, G0LM1, PMP22, EZH2, FGFR1, FNI, VEGFA, TMPRSS2, ANXA3, CRISP 3, BIRC5, AMACR, HIF1A, KLK3, Gleason score, and cancer stage for prediction of patient survival by using prostate tissue specimens collected from 482 patients with survival and 9 patients without survival in a prostate tissue specimen TCGA Cohort. A. ROC curve of the 24-Gene Panel. B. ROC curve of Gleason score. C. ROC curve of cancer stage. D. ROC curve of combining the 24-Gene Panel with Gleason score and cancer stage. Value of AUC (Area Under the Curve) is shown.

[0139] Figure 32 ROC (Receiver Operating Characteristic) curve of a 25-Gene Panel comprising of PIP5K1A, CCND1, GSTP1, CST3, CCNA1, EMTK2, MY06, HPN, CDK1, PSCA, PTEN, GO1M1, PMP22, EZH2, FGFR1, FN1, VEGFA, TMPRSS2, ANXA3, CRISP 3, BIRC5, AMACR, HIF1A, KLK3 a PCA3 for predicting if a prostate cancer patient will have >5 year survival time or <5 year survival time using prostate tissue specimens collected from 59 patients with >5 year survival time and 81 patients with <5 year survival time in a prostate tissue cohort. Value of AUC (Area Under the Curve) is shown.

DETAILED DESCRIPTION OF THE INVENTION

[0140] Unless defined otherwise, all technical and scientific terms used herein generally have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs. Generally, the nomenclature used herein and the laboratory procedures are those well-known and commonly employed in the art. The techniques and procedures are generally performed according to conventional methods in the art and various general references, which are provided throughout this document.

[0141] Accurate cancer diagnostic tests have tremendous clinical benefit. A test with high sensitivity can be used to find and treat cancer before the cancer becomes aggressive and lethal to avoid “false-negative” diagnosis and “under-treatment”. A test with high specificity can eliminate “false-positive” diagnosis and “over-treatment” so non-cancer patients do not get mis-diagnosed or treated.

[0142] Many gene mutations and alterations contribute to cancer tumorigenesis, progression and metastasis. Therefore no single gene or clinical parameter can provide accurate cancer diagnosis or prognosis. This invention is in part based on discoveries of novel combinations of genes that can be used for accurate cancer diagnosis and prognosis. The present invention provides compositions and methods for cancer diagnosis and prognosis.

[0143] In one aspect, the present invention provides a method for cancer screening, diagnosis, or determining the need for biopsy, comprising the following steps: (a) providing a biological sample from a subject; (b) measuring in the sample expression levels of a panel of genes comprising at least three or more genes selected from the group of genes consists of PIP5K1A, CCND1, GSTP1, CST3, CCNA1, EMTK2, MY06, HPN, CDK1, PSCA, PTEN, GO1M1, PMP22, EZH2, FGFR1, FN1, VEGFA, TMPRSS2, ANXA3, CRISP 3, BIRC5, AMACR, HIF1A, KLK3, PCA3, KLK2, MSMB, FLT1, MMP9, AR, TERT, PGC, SPINK1, STATS, STAT5, and TFF3,- (c) determining a diagnostic score by (i) calculating the relative expression level of each gene as compared to one or more housekeeping genes, (ii) combining the calculated relative expression level of each gene with a predefined algorithm to determine a diagnostic score; and (d) comparing the diagnostic score with a predetermined cancer diagnostic score cutoff value, and if the diagnostic score is higher than the cancer diagnostic score cutoff value, then the subject is diagnosed to have cancer or need for biopsy; or if the diagnostic score is equal to or lower than the cancer diagnostic score cutoff value, then the subject is diagnosed to have no cancer or no need for biopsy.

[0144] For many cancers, early diagnosis is advantageous to treat patient before the tumor becomes aggressive and/or even metastatic. In some embodiments, the present invention provides a method for cancer screening test. Such test can be performed annually or biannually on a subject over certain age (e.g. annual prostate cancer screening test for men over the age of 50, bladder cancer screening test for men or women over the age of 50). [0145] For many patients taking cancer screening test, many false positive patients undergo unnecessary biopsy. In some embodiments, the present invention provides a method to determine if a subject needs to take biopsy after cancer screening test. Such test can reduce unnecessary biopsy and prevent overdiagnosis.

[0146] In some embodiments, the panel of genes consists of PIP5K1A, CCNDI, GSTPI, CST3, CCNA1, LMTK2, MY06, HPN, CDKI, PSCA, PTEN, GOIMI, PMP22, EZH2, FGFRI, FNI, VEGFA, TMPRSS2, ANXA3, CRISP 3, BIRC5, AMACR, HIF1A, KLK3, and PCA3.

[0147] In another embodiment, the group of genes consists of CCND1, HIF1A, FGFR1, BIRC5, AMACR, CRISP3, FNI, HPN, MY06, PSCA, PMP22, GOIMI, EMTK2, EZH2, GSTP1, PCA3, VEGFA, CST3, PTEN, PIP5KIA, CDKI, TMPRSS2, ANXA3 and CCNA1.

[0148] In another embodiment, the group of genes consists of PIP5K1A, CCND1, GSTP1, CST3, CCNA1, EMTK2, MY06, HPN, CDK1, PSCA, PTEN, G0EM1, PMP22, EZH2, FGFR1, FNI, VEGFA, TMPRSS2, ANXA3, CRISP 3, BIRC5, AMACR, HIF1A and KLK3.

[0149] In another embodiment, the group of genes consists of PTEN, PIP5K1A, CDK1, TMPRSS2, ANXA3, HIF1A, FGFRI, BIRC5, AMACR, CRISP3, PMP22, GOIMI, EZH2, GSTPI, PCA3, VEGFA, CST3, CCNA1, CCNDI, FNI, MY06, KLK3 and PSCA.

[0150] In another embodiment, the group of genes consists of PTEN, CDK1, TMPRSS2, HIF1A, FGFRI, BIRC5, CRISP 3, FNI, HPN, PSCA, PMP22, LMTK2, EZH2, GSTP , PCA3, VEGFA, ANXA3 and KLK3.

[0151] Some cancers are slow-growing. Patients with low risk, clinically insignificant cancer do not need to be treated immediately and can be placed on active surveillance, while patients with high risk, clinically significant cancer should be treated immediately. Such stratification is clinically meaningful and important for cancer treatment decision-making and active surveillance management. Therefore it is necessary to identify high risk, clinically significant cancer and low risk, clinically insignificant cancer to avoid overtreatment of low risk, clinically insignificant cancer patients and under-treatment of high risk, clinically significant cancer patients.

[0152] In one aspect, the present invention provides a method for determining if a subject has high risk, clinically significant cancer and needs immediate treatment or low risk, clinically insignificant cancer and needs active surveillance, comprising the following steps: (a) providing a biological sample from a subject who has been diagnosed to have cancer; (b) measuring in the sample expression levels of a panel of genes comprising at least three or more genes selected from the group of genes consists of PIP5K1A, CCNDI, GSTPI, CST3, CCNA1, LMTK2, MY06, HPN, CDK1, PSCA, PTEN, GOIMI, PMP22, EZH2, FGFRI, FNI, VEGFA, TMPRSS2, ANXA3, CRISP3, BIRC5, AMACR, HIF1A, KLK3, PCA3, KLK2, MSMB, FLT1, MMP9, AR, TERT, PGC, SPINKI, STAT 3, STAT 5, and TFF3,' (c) determining a stratification score by (i) calculating the relative expression level of each gene as compared to one or more housekeeping genes, (ii) combining the calculated relative expression level of each gene with a predefined algorithm to determine a stratification score; and (d) comparing the stratification score with a predetermined high risk stratification score cutoff value, and if the stratification score is higher than the high risk stratification score cutoff value, then the subject is diagnosed to have high risk, clinically significant cancer and needs immediate treatment; or if the stratification score is equal to or lower than the high risk stratification score cutoff value, then the subject is diagnosed to have low risk, clinically insignificant cancer and needs active surveillance. [0153] In some embodiments, the present invention provides a method for monitoring cancer progression during active surveillance to determine if a subject has cancer progression and needs immediate treatment or has no cancer progression and will continue active surveillance without treatment.

[0154] In some embodiments, the group of genes consists of PIP5K1A, CCND1, GSTPI, CST3, CCNA1, EMTK2, MY06, HPN, CDK1, PSCA, PTEN, GO1M1, PMP22, EZH2, FGFR1, FN1, VEGFA, TMPRSS2, ANXA3, CRISP 3, BIRC5, AMACR, HIF1A, KLK3 an PCA3.

[0155] In another embodiment, the group of genes consists of CCND1, HIF1A, FGFR1, BIRC5, AMACR, CRISP3, FNI, HPN, MY06, PSCA, PMP22, G0LM1, LMTK2, EZH2, GSTPI, PCA3, VEGFA, CST3, PTEN, PIP5K1A, CDKI, TMPRSS2, ANXA3 and CCNA1.

[0156] In another embodiment, the group of genes consists of PIP5K1A, CCND1, GSTPI, CST3, CCNA1, LMTK2, MY06, HPN, CDKI, PSCA, PTEN, G0LM1, PMP22, EZH2, FGFR1, FNI, VEGFA, TMPRSS2, ANXA3, CRISP 3, BIRC5, AMACR, HIF1A and KLK3.

[0157] In another embodiment, the group of genes consists of PTEN, PIP5K1A, CDK1, TMPRSS2, ANXA3, HIF1A, FGFR1, BIRC5, AMACR, CRISP3, PMP22, G0IM1, EZH2, GSTPI, PCA3, VEGFA, CST3, CCNA1, CCND1, FNI, MY06, KLK3 and PSCA.

[0158] In another embodiment, the group of genes consists of PTEN, CDK1, TMPRSS2, HIF1A, FGFR1, BIRC5, CRISP 3, FNI, HPN, PSCA, PMP22, EMTK2, EZH2, GSTPI, PCA3, VEGFA, ANXA3 and KLK3.

[0159] Metastatic cancer is incurable and currently there is no effective and safe treatment for metastatic cancer. Most patients die within a few years or a few months after metastatic cancer diagnosis. Many metastatic cancer patients need aggressive and effective treatments, thus if a patient is predicted to have metastatic cancer in the future, then the patient can be treated with aggressive and effective treatments before cancer metastasis occurs or can be detected, thus the treatment may be more effective and/or treatment resistance may be prevented. Prediction of cancer metastasis may prevent the development of metastasis and reduce cancer-related death.

[0160] In one aspect, the present invention provides a method for predicting future cancer metastasis in a subject diagnosed to have cancer, comprising the following steps: (a) providing a biological sample from a subject who has been diagnosed to have cancer; (b) measuring in the sample expression levels of a panel of genes comprising at least three or more genes selected from the group of genes consists of PIP5K1A, CCND1, GSTPI, CST3, CCNA1, IMTK2, MY06, HPN, CDK1, PSCA, PTEN, GOIM1, PMP22, EZH2, FGFR1, FNI, VEGFA, TMPRSS2, ANXA3, CRISP3, BIRC5, AMACR, HIF1A, KLK3, PCA3, KLK2, MSMB, FLTI, MMP9, AR, TERT, PGC, SPINK1, STAT3, STAT5, and TFF3; (c) determining a metastasis score by (i) calculating the relative expression level of each gene as compared to one or more housekeeping genes, (ii) combining the calculated relative expression level of each gene with a predefined algorithm to determine a metastasis score; and (d) comparing the metastasis score with a predetermined metastatic cancer score cutoff value, and if the metastasis score is higher than the predetermined metastatic cancer score cutoff value, then the subject is predicted to have metastatic cancer in the future; or if the metastasis score is equal to or lower than the predetermined metastatic cancer score cutoff value, then the subject is predicted to have no metastatic cancer in the future.

[0161] During and after metastatic cancer treatment, it is important to measure metastatic cancer treatment efficacy by determining if metastatic cancer/residual metastatic cancer exists after treatment to decide if further treatment is necessary or if a different treatment is needed. Such information can better guide treatment decision-making and improve treatment outcome. [0162] In some embodiments, the method can be used for measuring metastatic cancer treatment efficacy during or after treatment by comparing a metastasis score obtained from a subject during or after metastatic cancer treatment, and if the metastasis score is higher than a predetermined metastatic cancer score cutoff value, then the subject is determined to still have metastatic cancer; or if the metastasis score is equal to or lower than a predetermined metastatic cancer score cutoff value, then the subject is determined to have no metastatic cancer.

[0163] In some embodiments, the group of genes for predicting cancer metastasis or measuring metastatic cancer treatment efficacy during or after treatment consists of PIP5K1A, CCNDI, GSTPI, CST3, CCNA1, LMTK2, MYO6. HPN, CDKI, PSCA, PTEN, GOIMI, PMP22, EZH2, FGFRI, FNI, VEGFA, TMPRSS2, ANXA3, CRISP 3, BIRC5, AMACR, HIF1A, KLK3 and PCA3.

[0164] In another embodiment, the group of genes consists of CCND1, HIF1A, FGFR1, BIRC5, AMACR, CRISP3, FNI, HPN, MY06, PSCA, PMP22, GOIMI, LMTK2, EZH2, GSTP1, PCA3, VEGFA, CST3, PTEN, PIP5K1A, CDKI, TMPRSS2, ANXA3 and CCNA1.

[0165] In another embodiment, the group of genes consists of PIP5K1A, CCND1, GSTP1, CST3, CCNA1, EMTK2, MY06, HPN, CDK1, PSCA, PTEN, G0EM1, PMP22, EZH2, FGFR1, FNI, VEGFA, TMPRSS2, ANXA3, CRISP 3, BIRC5, AMACR, HIF1A and KLK3.

[0166] In another embodiment, the group of genes consists of PTEN, PIP5K1A, CDK1, TMPRSS2, ANXA3, HIF1A, FGFRI, BIRC5, AMACR, CRISP3, PMP22, GOIMI, EZH2, GSTPI, PCA3, VEGFA, CST3, CCNA1, CCNDI, FNI, MY06, KLK3 and PSCA.

[0167] In another embodiment, the group of genes consists of PTEN, CDK1, TMPRSS2, HIF1A, FGFRI, BIRC5, CRISP 3, FNI, HPN, PSCA, PMP22, LMTK2, EZH2, GSTPI, PCA3, VEGFA, ANXA3 and KLK3.

[0168] Some patients receiving cancer treatments (e.g. surgery, chemotherapy, radiation therapy) develop treatment-resistance within a few years after treatment. Prediction of treatment-resistance is important for treatment-decision making so patients can receive aggressive and effective treatment before the development of treatment-resistance for better outcome, which may prevent or delay treatment-resistance.

[0169] In one aspect, the present invention provides a method to predict the development of treatmentresistant cancer, comprising the following steps: (a) providing a biological sample from a subject who has been diagnosed to have cancer; (b) measuring in the sample expression levels of a panel of genes comprising at least three or more genes selected from the group of genes consists of PIP5K1A, CCNDI, GSTPI, CST3, CCNA1, EMTK2, MY06, HPN, CDK1, PSCA, PTEN, GOIMI, PMP22, EZH2, FGFRI, FNI, VEGFA, TMPRSS2, ANXA3, CRISP3, BIRC5, AMACR, HIF1A, KLK3, PCA3, KLK2, MSMB, FLT1, MMP9, AR, TERT, PGC, SPINKI, STAT3, STAT5, and TFF3,' (c) determining a treatment-resistance score by (i) calculating the relative expression level of each gene as compared to one or more housekeeping genes, (ii) combining the calculated relative expression level of each gene with a predefined algorithm to determine a treatment-resistance score; and (d) comparing the treatment-resistance score with a predetermined treatment-resistance score cutoff value, and if the treatment-resistance score is higher than the predetermined treatment-resistance score cutoff value, then the subject is predicted to have treatmentresistance in the future; or if the treatment-resistance score is equal to or lower than the predetermined treatment-resistance score cutoff value, then the subject is predicted to have no treatment-resistance in the future.

[0170] In some embodiments, the treatment-resistant cancer is castration-resistant prostate cancer. [0171] In another embodiment, the treatment-resistant cancer is metastatic castration-resistant prostate cancer.

[0172] In some embodiments, the group of genes consists of PIP5K1A, CCND1, GSTPI, CST3, CCNA1, EMTK2, MY06, HPN, CDK1, PSCA, PTEN, G0IM1, PMP22, EZH2, FGFR1, FNI, VEGFA, TMPRSS2, ANXA3, CRISP 3, BIRC5, AMACR, HIF1A, KLK3 an PCA3.

[0173] In another embodiment, the group of genes consists of CCND1, HIF1A, FGFR1, BIRC5, AMACR, CRISP3, FNI, HPN, MY06, PSCA, PMP22, GOLMI, LMTK2, EZH2, GSTPI, PCA3, VEGFA, CST3, PTEN, PIP5K1A, CDKI, TMPRSS2, ANXA3 and CCNA1.

[0174] In another embodiment, the group of genes consists of PIP5K1A, CCND1, GSTPI, CST3, CCNA1, LMTK2, MY06, HPN, CDKI, PSCA, PTEN, GOLMI, PMP22, EZH2, FGFR1, FNI, VEGFA, TMPRSS2, ANXA3, CRISP 3, BIRC5, AMACR, HIF1A and KLK3.

[0175] In another embodiment, the group of genes consists of PTEN, PIP5K1A, CDK1, TMPRSS2, ANXA3, HIF1A, FGFR1, BIRC5, AMACR, CRISP3, PMP22, GOLMI, EZH2, GSTPI, PCA3, VEGFA, CST3, CCNA1, CCND1, FNI, MY06, KLK3 and PSCA.

[0176] In another embodiment, the group of genes consists of PTEN, CDK1, TMPRSS2, HIF1A, FGFR1, BIRC5, CRISP 3, FNI, HPN, PSCA, PMP22, EMTK2, EZH2, GSTPI, PCA3, VEGFA, ANXA3 and KLK3.

[0177] Many patients have recurrent cancer within a few years after treatment. Prediction of cancer recurrence after treatment has significant clinical benefit as such information can give treatment guidance so patients predicted to have cancer recurrence after treatment can receive further treatment or different treatment to prevent cancer recurrence.

[0178] In one aspect, the present invention provides a method for predicting if a subject diagnosed to have cancer will have cancer recurrence after treatment, comprising the following steps: (a) providing a biological sample from a subject who has been diagnosed to have cancer; (b) measuring in the sample expression levels of a panel of genes comprising at least three or more genes selected from the group of genes consists of PIP5K1A, CCND1, GSTPI, CST3, CCNA1, IMTK2, MY06, HPN, CDK1, PSCA, PTEN, GOLMI, PMP22, EZH2, FGFR1, FNI, VEGFA, TMPRSS2, ANXA3, CRISP3, BIRC5, AMACR, HIF1A, KLK3, PCA3, KLK2, MSMB, FLTI, MMP9, AR, TERT, PGC, SPINK1, STAT3, STAT5, and TFF3; (c) determining a recurrence score by (i) calculating the relative expression level of each gene as compared to one or more housekeeping genes, (ii) combining the calculated relative expression level of each gene with a predefined algorithm to determine a recurrence score; and (d) comparing the recurrence score with a predetermined recurrent cancer score cutoff value, and if the recurrence score is higher than the predetermined recurrent cancer score cutoff value, then the subject is predicted to have cancer recurrence in the future; or if the recurrence score is equal to or lower than the predetermined recurrent cancer score cutoff value, then the subject is predicted to have no cancer recurrence in the future.

[0179] In some embodiments, the group of genes consists of PIP5K1A, CCND1, GSTPI, CST3, CCNA1, LMTK2, MY 06. HPN, CDK1, PSCA, PTEN, GOLMI, PMP22, EZH2, FGFR1, FNI, VEGFA, TMPRSS2, ANXA3, CRISP 3, BIRC5, AMACR, HIF1A, KLK3 and PCA3.

[0180] In another embodiment, the group of genes consists of CCND1, HIF1A, FGFR1, BIRC5, AMACR, CRISP3, FNI, HPN, MY06, PSCA, PMP22, G0EM1, EMTK2, EZH2, GSTPI, PCA3, VEGFA, CST3, PTEN, PIP5K1A, CDKI, TMPRSS2, ANXA3 and CCNAL [0181] In another embodiment, the group of genes consists of PIP5K1A, CCND1, GSTPI, CST3, CCNA1, LMTK2, MY06, HPN, CDK1, PSCA, PTEN, GOIM1, PMP22, EZH2, FGFR1, FNI, VEGFA, TMPRSS2, ANXA3, CRISP 3, BIRC5, AMACR, HIF1A and KLK3.

[0182] In another embodiment, the group of genes consists of PTEN, PIP5K1A, CDK1, TMPRSS2, ANXA3, HIF1A, FGFR1, BIRC5, AMACR, CRISP3, PMP22, G0IM1, EZH2, GSTPI, PCA3, VEGFA, CST3, CCNA1, CCND1, FNI, MY06, KLK3 and PSCA.

[0183] In another embodiment, the group of genes consists of PTEN, CDK1, TMPRSS2, HIF1A, FGFR1, BIRC5, CRISP 3, FNI, HPN, PSCA, PMP22, LMTK2, EZH2, GSTPI, PCA3, VEGFA, ANXA3 and KLK3.

[0184] During or after cancer treatment, detecting the presence or absence of cancer using a diagnostic test can better monitor the treatment efficacy and outcome. A cancer diagnostic test can be used to determine if there is cancer/residual cancer during or after treatment, and such information can be used to decide if further treatment is necessary or if a different treatment is needed. In addition, timely detection of cancer recurrence after treatment is essential for patients to receive immediate treatment to prevent cancer progression, metastasis and development of treatment resistance. A diagnostic test can be used to detect cancer recurrence after treatment so recurrent cancer can be detected and treated promptly.

[0185] In one aspect, the present invention provides a method to determine treatment efficacy by detecting cancer/residual cancer during or after treatment or detecting cancer recurrence after treatment, comprising the following steps: (a) providing a biological sample from a subject during or after cancer treatment; (b) measuring in the sample expression levels of a panel of genes comprising at least three or more genes selected from the group of genes consists of PIP5K1A, CCND1, GSTPI, CST3, CCNA1, EMTK2, MY06, HPN, CDK1, PSCA, PTEN, G0IM1, PMP22, EZH2, FGFR1, FNI, VEGFA, TMPRSS2, ANXA3, CRISP3, BIRC5, AMACR, HIF1A, KLK3, PCA3, KLK2, MSMB, FLTI, MMP9, AR, TERT, PGC, SPINK1, STAT3, STAT5, and TFF3,- (c) determining a cancer diagnostic score by (i) calculating the relative expression level of each gene as compared to one or more housekeeping genes, (ii) combining the calculated relative expression level of each gene with a predefined algorithm to determine a cancer diagnostic score; and (d) comparing the cancer diagnostic score with a predetermined cancer diagnostic score cutoff value, and if the cancer diagnostic score is higher than the predetermined cancer diagnostic score cutoff value, then the subject is determined to have cancer/residual cancer during or after treatment or cancer recurrence after treatment; or if the cancer diagnostic score is equal to or lower than the predetermined cancer diagnostic score cutoff value, then the subject is determined to have no cancer/residual cancer during or after treatment or no cancer recurrence after treatment.

[0186] In some embodiments, the group of genes consists of PIP5K1A, CCND1, GSTPI, CST3, CCNA1, EMTK2, MY06, HPN, CDK1, PSCA, PTEN, G0EM1, PMP22, EZH2, FGFR1, FNI, VEGFA, TMPRSS2, ANXA3, CRISP 3, BIRC5, AMACR, HIF1A, KLK3 an PCA3.

[0187] In another embodiment, the group of genes consists of CCND1, HIF1A, FGFR1, BIRC5, AMACR, CRISP3, FNI, HPN, MY06, PSCA, PMP22, G0EM1, EMTK2, EZH2, GSTPI, PCA3, VEGFA, CST3, PTEN, PIP5K1A, CDK1, TMPRSS2, ANXA3 and CCNA1.

[0188] In another embodiment, the group of genes consists of PIP5K1A, CCND1, GSTPI, CST3, CCNA1, LMTK2, MY06, HPN, CDK1, PSCA, PTEN, GOIM1, PMP22, EZH2, FGFR1, FNI, VEGFA, TMPRSS2, ANXA3, CRISP 3, BIRC5, AMACR, HIF1A and KLK3. [0189] In another embodiment, the group of genes consists of PTEN, PIP5K1A, CDK1, TMPRSS2, ANXA3, HIF1A, FGFR1, BIRC5, AMACR, CRISPS, PMP22, GO1M1, EZH2, GSTP1, PCA3, VEGFA, CST3, CCNAI, CCND1, FNI, MY06, KLK3 and PSCA.

[0190] In another embodiment, the group of genes consists of PTEN, CDK1, TMPRSS2, HIF1A, FGFR1, BIRC5, CRISPS, FN , HPN, PSCA, PMP22, LMTK2, EZH2, GSTPI, PCA3, VEGFA, ANXA3 and KLK3.

[0191] After the patients have received treatment, some will go into remission while others have treatmentresistant cancer. Thus, prediction of cancer remission after the treatment may assist cancer treatment decision-making to prevent ineffective management of aggressive cancer and treatment failure. Currently, few accurate tests exist to predict cancer remission after treatment.

[0192] In one aspect, the present invention provides a method to predict cancer remission after treatment, comprising the following steps: (a) providing a biological sample from a subject before cancer treatment; (b) measuring in the sample expression levels of a panel of genes comprising at least three or more genes selected from the group of genes consists of PIP5K1A, CCND1, GSTPI, CST3, CCNA1, LMTK2, MY06, HPN, CDKI, PSCA, PTEN, GOLM1, PMP22, EZH2, FGFRI, FNI, VEGFA, TMPRSS2, ANXA3, CRISP3, BIRC5, AMACR, HIF1A, KLK3, PCA3, KLK2, MSMB, FLT1, MMP9, AR, TERT, PGC, SPINK1, STAT3, STAT5, and TFF3,- (c) determining a remission score by (i) calculating the relative expression level of each gene as compared to one or more housekeeping genes, (ii) combining the calculated relative expression level of each gene with a predefined algorithm to determine a remission score; and (d) comparing the remission score with a predetermined remission score cutoff value and partial remission score cutoff value, and if the remission score is higher than a predetermined remission score cutoff value, then the subject is predicted to have cancer remission after treatment in the future; if the remission score is equal to or lower than a predetermined remission score cutoff value but higher than a predetermined partial remission score cutoff value, then the subject is predicted to have partial remission after treatment in the future; or if the remission score is equal to or lower than a predetermined partial remission score cutoff value, then the subject is predicted to not have cancer remission after treatment in the future.

[0193] In some embodiments, the group of genes consists of PIP5K1A, CCND1, GSTPI, CST3, CCNA1, LMTK2, MY06, HPN, CDKI, PSCA, PTEN, GOIM1, PMP22, EZH2, FGFRI, FNI, VEGFA, TMPRSS2, ANXA3, CRISP 3, BIRC5, AMACR, HIF1A, KLK3 and PCA3.

[0194] In another embodiment, the group of genes consists of CCND1, HIF1A, FGFRI, BIRC5, AMACR, CRISP3, FNI, HPN, MY06, PSCA, PMP22, GO1M1, LMTK2, EZH2, GSTPI, PCA3, VEGFA, CST3, PTEN, PIP5K1A, CDKI, TMPRSS2, ANXA3 and CCNA1.

[0195] In another embodiment, the group of genes consists of PIP5K1A, CCND1, GSTPI, CST3, CCNA1, EMTK2, MY06, HPN, CDKI, PSCA, PTEN, G0EM1, PMP22, EZH2, FGFRI, FNI, VEGFA, TMPRSS2, ANXA3, CRISP 3, BIRC5, AMACR, HIF1A and KLK3.

[0196] In another embodiment, the group of genes consists of PTEN, PIP5K1A, CDKI, TMPRSS2, ANXA3, HIF1A, FGFRI, BIRC5, AMACR, CRISPS, PMP22, G0IM1, EZH2, GSTPI, PCA3, VEGFA, CST3, CCNAI, CCNDI, FNI, MY06, KLK3 and PSCA.

[0197] In another embodiment, the group of genes consists of PTEN, CDKI, TMPRSS2, HIF1A, FGFRI, BIRC5, CRISP 3, FNI, HPN, PSCA, PMP22, LMTK2, EZH2, GSTPI, PCA3, VEGFA, ANXA3 and KLK3.

[0198] Patients diagnosed with cancer, especially non -metastatic cancer, need long-term prognostic information to assist cancer treatment and management, especially at cancer diagnosis. Currently in clinical practice, most methods use short-term and surrogate outcomes, such as biochemical recurrence. However, using survival itself* as an endpoint to predict patients’ long-term outcome is more robust and clinically meaningful. Presently, no accurate prognostic method for cancer survival has been established.

[0199] In one aspect, the present invention provides a method to predict survival of a subject diagnosed as having cancer, comprising the following steps: (a) providing a biological sample from a subject diagnosed as having cancer; (b) measuring in the sample expression levels of a panel of genes comprising at least three or more genes selected from the group of genes consists of PIP5K1A, CCND1, GSTPI, CST3, CCNA1, EMTK2, MY06, HPN, CDK1, PSCA, PTEN, GO1M1, PMP22, EZH2, FGFR1, FN1, VEGFA, TMPRSS2, ANXA3, CRISP3, BIRC5, AMACR, HIF1A, KLK3, PCA3, KLK2, MSMB, FLT1, MMP9, AR, TERT, PGC, SPINK1, STAT3, STAT5, and TFF3,' (c) determining a survival score by (i) calculating the relative expression level of each gene as compared to one or more housekeeping genes, (ii) combining the calculated relative expression level of each gene with a predefined algorithm to determine a survival score; and (d) comparing the survival score with a predetermined survival score cutoff value, and if the survival score is higher than the predetermined survival score cutoff value, then the subject is predicted to survive cancer in the future; or if the survival score is equal to or lower than the predetermined survival score cutoff value, then the subject is predicted to not survive or die of cancer in the future.

[0200] In some embodiments, the group of genes consists of PIP5K1A, CCND1, GSTPI, CST3, CCNA1, LMTK2, MY06, HPN, CDK1, PSCA, PTEN, G0LM1, PMP22, EZH2, FGFR1, FN1, VEGFA, TMPRSS2, ANXA3, CRISP 3, BIRC5, AMACR, HIF1A, KLK3 an PCA3.

[0201] In another embodiment, the group of genes consists of CCND1, HIF1A, FGFR1, BIRC5, AMACR, CRISP3, FN1, HPN, MY06, PSCA, PMP22, G0LM1, LMTK2, EZH2, GSTPI, PCA3, VEGFA, CST3, PTEN, PIP5K1A, CDKI, TMPRSS2, ANXA3 and CCNA1.

[0202] In another embodiment, the group of genes consists of PIP5K1A, CCND1, GSTPI, CST3, CCNA1, LMTK2, MY06, HPN, CDKI, PSCA, PTEN, G0LM1, PMP22, EZH2, FGFR1, FN1, VEGFA, TMPRSS2, ANXA3, CRISP 3, BIRC5, AMACR, HIF1A and KLK3.

[0203] In another embodiment, the group of genes consists of PTEN, PIP5K1A, CDK1, TMPRSS2, ANXA3, HIF1A, FGFR1, BIRC5, AMACR, CRISP3, PMP22, G0IM1, EZH2, GSTPI, PCA3, VEGFA, CST3, CCNA1, CCND1, FN1, MY06, KLK3 and PSCA.

[0204] In another embodiment, the group of genes consists of PTEN, CDK1, TMPRSS2, HIF1A, FGFR1, BIRC5, CRISP 3, FN1, HPN, PSCA, PMP22, EMTK2, EZH2, GSTPI, PCA3, VEGFA, ANXA3 and KLK3.

[0205] In one aspect, the present invention provides a method for predicting survival time of a subject diagnosed as having cancer, comprising the following steps: (a) providing a biological sample from a subject diagnosed as having cancer; (b) measuring in the sample expression levels of a panel of genes comprising at least three or more genes selected from the group of genes consists of PIP5K1A, CCND1, GSTPI, CST3, CCNA1, EMTK2, MY06, HPN, CDK1, PSCA, PTEN, G0EM1, PMP22, EZH2, FGFR1, FN1, VEGFA, TMPRSS2, ANXA3, CRISP3, BIRC5, AMACR, HIF1A, KLK3, PCA3, KLK2, MSMB, FLT1, MMP9, AR, TERT, PGC, SPINK1, STAT3, STAT5, and TFF3,' (c) determining a 5-year survival score by

(i) calculating the relative expression level of each gene as compared to one or more housekeeping genes,

(ii) combining the calculated relative expression level of each gene with a predefined algorithm to determine a 5-year survival score; (d) determining a 10-year survival score by (i) calculating the relative expression level of each gene as compared to one or more housekeeping genes, (ii) combining the calculated relative expression level of each gene with a predefined algorithm to determine a 10-year survival score; (e) determining a 20-year survival score by (i) calculating the relative expression level of each gene as compared to one or more housekeeping genes, (ii) combining the calculated relative expression level of each gene with a predefined algorithm to determine a 20-year survival score; and (f) predicting the subject as either (i) will have less than 5-year survival time if the 5-year survival score is equal to or lower than the predetermined 5-year survival score cutoff value, or (ii) will have 5-10 year survival time if the 5-year survival score is higher than the predetermined 5-year survival score cutoff value but the 10-year survival score is equal to or lower than the predetermined 10-year survival score cutoff value, or (iii) will have 10- 20 year survival time if the 10-year survival score is higher than the predetermined 10-year survival score cutoff value but the 20-year survival score is equal to or lower than the predetermined 20-year survival score cutoff value, or (iv) will have more than 20-year survival time if the 20-year survival score is higher than the predetermined 20-year survival score cutoff value.

[0206] In some embodiments, the group of genes consists of PIP5K1A, CCND1, GSTPI, CST3, CCNA1, LMTK2, MY06, HPN, CDK1, PSCA, PTEN, G0LM1, PMP22, EZH2, FGFR1, FN1, VEGFA, TMPRSS2, ANXA3, CRISP 3, BIRC5, AMACR, HIF1A, KLK3 an PCA3.

[0207] In another embodiment, the group of genes consists of CCND1, HIF1A, FGFR1, BIRC5, AMACR, CRISP3, FNI, HPN, MY06, PSCA, PMP22, G0LM1, LMTK2, EZH2, GSTPI, PCA3, VEGFA, CST3, PTEN, PIP5K1A, CDKI, TMPRSS2, ANXA3 and CCNA1.

[0208] In another embodiment, the group of genes consists of PIP5K1A, CCND1, GSTPI, CST3, CCNA1, LMTK2, MY06, HPN, CDKI, PSCA, PTEN, G0LM1, PMP22, EZH2, FGFR1, FNI, VEGFA, TMPRSS2, ANXA3, CRISP 3, BIRC5, AMACR, HIF1A and KLK3.

[0209] In another embodiment, the group of genes consists of PTEN, PIP5K1A, CDK1, TMPRSS2, ANXA3, HIF1A, FGFR1, BIRC5, AMACR, CRISP3, PMP22, G0IM1, EZH2, GSTPI, PCA3, VEGFA, CST3, CCNA1, CCND1, FNI, MY06, KLK3 and PSCA.

[0210] In another embodiment, the group of genes consists of PTEN, CDK1, TMPRSS2, HIF1A, FGFR1, BIRC5, CRISP 3, FNI, HPN, PSCA, PMP22, EMTK2, EZH2, GSTPI, PCA3, VEGFA, ANXA3 and KLK3.

[0211] In one aspect, the expression levels of a panel of genes are expression levels of mRNA, DNA methylation, protein, peptide, or their combination.

[0212] In some embodiments, the RNA can be isolated from a sample, reversely transcribed and quantified by real time qRT-PCR.

[0213] In another embodiment, the reversely transcribed cDNA from RNA in a sample can be preamplified before quantified by real time qRT-PCR.

[0214] In one aspect, the sample obtained from a subject can be blood, urine, ascites, other body fluids, tissue and cell.

[0215] In another aspect, the sample obtained from a subject is urine.

[0216] The invention provides a method for using a kit to make measurement of expression levels of genes in a panel.

[0217] The invention further provides a process for how the kit is prepared. In some embodiments, the kit comprises of reagents for mRNA isolation, cDNA reverse transcription, preamplification of cDNA, and PCR detection. [0218] In one aspect, the present invention provides an algorithm to use the calculated relative expression level of each gene in a panel to calculate a diagnostic or prognostic score to make a diagnosis or prognosis.

[0219] In some embodiments, an algorithm for a gene panel includes:

Diagnostic or Prognostic Score=Cp-CNon

Whereas Ap is positive diagnosis or prognosis constant, BNOII is negative diagnosis or prognosis constant, n is number of genes in the panel, CtSi through QSn are relative Ct values of gene 1 through gene n, Xi through Xn are positive diagnosis or prognosis regression coefficients of gene 1 through gene n, Xi*i through Xn*n are gene 1 and gene 1 cross positive diagnosis or prognosis regression coefficients through gene n and gene n cross positive diagnosis or prognosis regression coefficients, Y i through Yn are negative diagnosis or prognosis regression coefficients of gene 1 through gene n, and Y 1*1 through Yn*n are gene 1 and gene 1 cross negative diagnosis or prognosis regression coefficients through gene n and gene n cross negative diagnosis or prognosis regression coefficients. The sample is diagnosed or prognosed to be positive when Diagnostic or Prognostic Score is >0, whereas the sample is diagnosed or prognosed to be negative when Diagnostic or Prognostic Score is <0.

[0220] In some embodiments, the present invention provides a computer program to perform data analysis and diagnosis or prognosis, comprising the following steps: (a) receiving gene expression data on test genes in a panel; (b) determining a diagnostic or prognostic score by (i) calculating the relative expression level of each gene in said panel as compared to one or more housekeeping genes, (ii) combining the calculated relative expression level of each gene using a predefined algorithm to determine a diagnostic or prognostic score; and (c) comparing the value of the diagnostic or prognostic score to a predetermined diagnostic or prognostic score cutoff value to make diagnosis or prognosis, and displaying the result.

[0221] In one aspect, the present invention provides a method to use an algorithm to make a diagnosis or prognosis by combining expression levels of a panel of genes and one or more of other cancer diagnostic or prognostic test results.

[0222] In some embodiments, the other cancer diagnostic or prognostic test includes, but not limited to, prostate-specific antigen (PSA), total PSA, free PSA, percent-free PSA, PSA density, PSA velocity, total Gleason score, primary Gleason score, secondary Gleason score, tertiary Gleason pattern 5 (TGP5), clinical tumor stage, biopsy core with cancer, ERG fusion status, fraction of genome altered, copy number alteration, copy number variation, copy number cluster, lymph node involvement, number of lymph nodes removed for examination, number of lymph nodes with tumor, seminal vesicle invasion, extra-capsular extension, surgical margin status, and nomograms.

[0223] In one aspect, the said cancer is prostate cancer.

[0224] While preferred embodiments of the present invention have been shown and described herein, it will be obvious to those skilled in the art that such embodiments are provided by way of example only. Numerous variations, changes, and substitutions will now occur to those skilled in the art without departing from the invention. It should be understood that various alternatives to the embodiments of the invention described herein may be employed in practicing the invention. It is intended that the claims define the scope of the invention and that methods and structures within the scope of these claims and their equivalents be covered thereby. EXAMPLES

[0225] The following examples are provided in order to demonstrate and further illustrate certain preferred embodiments and aspects of the present invention and are not to be construed as limiting the scope thereof.

EXAMPLE 1

[0226] The diagnostic performance of a 5 -Gene Panel comprising of GSTP1, LMTK2, HPN, G0LM1 and PMP22 was tested for prostate cancer diagnosis using prostate tissue specimens.

Patients and Methods

[0227] 144 prostate tissue specimens used in the study were obtained from TissueScan Prostate Tissue qPCR Array (OriGene Technologies, Rockville, MD, USA). The prostate tissues were surgically removed by prostatectomy and flash frozen within 30 minutes of ischemia. Comprehensive pathology reports were provided which included diagnosis, tumor grade, TNM classification, minimum stage grouping, detailed sample cellularity (tumor cells, normal epithelial and luminal cells, stroma and necrosis in percentage) and Gleason score. The pathological diagnosis of prostate cancer or benign prostate (including benign prostatic hyperplasia (BPH), prostatitis and normal) was based on pathological analysis of the specimen.

[0228] Prostate cancer samples were selected based on tumor content to include minimally 50% of tumor as determined by microscopic pathology analysis. Most of the normal samples were taken from patients without pathological diseases, while some were taken from adjacent normal regions of diseased patients. Most of the BPH samples were taken from BPH patients, while some were taken from adjacent BPH regions of prostate cancer patients.

[0229] The tissue samples were processed for RNA purification and the quality of RNA was examined to show minimal to no degradation by Agilent Bioanalyzer analysis. cDNA was then generated by reverse transcription and normalized with a housekeeping gene beta-actin to form cDNA arrays. All specimens were collected under IRB approved protocols and all human subjects were fully informed and explicitly asked for their consent to future research use of their samples.

[0230] The gene expression levels of a 5-Gene Panel were measured. The primers and probes of all genes in the panel were predesigned assays purchased from Integrated DNA Technologies (San Diego, California, USA). Duplex PCR assays of the genes in each panel were validated with seven -point, 10-fold serial- diluted standard curves with a range from 1000 ng to 1 pg RNA for each singleplex and duplex assay. Each 10 pl reaction consisted of cDNA equivalent to 20 ng of total RNA for the 1000 ng standard curve point down to 50 fg for the 1 pg standard curve point, and 500 nM each of forward and reverse amplification primers and 250 nM probe.

[0231] To assess expression levels of the genes in the panel, real time qRT-PCR was performed in each reaction well containing 3-4 ng of cDNA from TissueScan Prostate Tissue qPCR Array. Real time PCR amplification of cDNA was performed on ABI 7900HT Fast Real Time PCR System (Applied Biosystems, Foster City, CA, USA). The PCR reaction was performed in 30 pl volume consists of 3-4 ng of cDNA, 15 pl 2 x TaqMan® universal PCR master mixes (Life Technologies, Foster City, CA, USA), 1500 nM each of forward and reverse amplification primers and 750 nM probe. The cycling conditions were set as the following: 95°C for 10 minutes for polymerase activation, followed by 40 cycles of 95°C for 15 seconds and 60°C for 1 minute.

[0232] The level of a housekeeping gene beta-actin mRNA was measured in each specimen for gene expression normalization to control variations of cDNA quantity in the patient specimens. The cycle threshold (Ct) value of each gene in the panels was divided by the Ct value of the beta-actin mRNA as the normalized mRNA expression value of the gene (C $=Ct (sample)ZCt (actin)). For each gene, duplicate PCRs were performed to average the Ct values.

[0233] A diagnostic score was calculated by combining Cfcvalues of all genes in the panel with a predefined algorithm to discriminate prostate cancer and benign prostate. Then compare the diagnostic score with a predetermined cancer diagnostic score cutoff value to make a diagnosis. The diagnosis of all specimens with the gene panel was then compared to the pathological diagnosis of the specimens and the receiver operator curve (ROC) analysis was performed using a statistical software program (XLSTAT). The diagnostic performance measures including sensitivity, specificity, positive predictive value (PPV) and negative predictive value (NPV) were calculated. The P value was obtained from statistical comparative test Mann-Whitney Test using XLSTAT.

Results

[0234] The result showed that the 5-Gene Panel consisting of GSTP1, LMTK2, HPN, G0LM1 and PMP22 was able to distinguish prostate cancer from benign prostate. As shown in Table 1, the 5-Gene Panel was able to distinguish prostate cancer from benign prostate with very high sensitivity of 96.6% and specificity of 94.6% (p<0.0001). The positive predictive value (PPV) reached 96.6% and the negative predictive value (NPV) reached 94.6%. The ROC analysis was performed to measure the classification power of the 5- Gene Panel for cancer diagnosis and the result showed AUC of the ROC curve to be 0.996 (Figure 1), an extremely high value for prostate cancer diagnosis.

Table 1

Positive Negative Total

Cancer 85 3 88

Non-Cancer 3 53 56

Sensitivity 96.6%

Specificity 94.6%

PPV 96.6%

NPV 94.6%

PO.OOOl

Figure 1

EXAMPLE 2

[0235] The diagnostic performance of an 8-Gene Panel comprising of comprising of MY06, LMTK2, PCA3, GSTP1, HPN, CCND1, FN1 and PMP22 was tested for diagnosing high risk, aggressive and low risk, indolent prostate cancer using prostate tissue specimens.

Patients and Methods

[0236] 87 prostate cancer tissue specimens used in the study were obtained from TissueScan Prostate Tissue qPCR Array (OriGene Technologies, Rockville, MD, USA) and processed to measure the expression levels of the eight genes in the panel as described in the previous example. The pathological diagnosis of aggressive and indolent prostate cancer was based on Gleason score. The patients with Gleason score > 7 were diagnosed as having high risk, aggressive prostate cancer, while the patients with Gleason score < 7 were diagnosed as having low risk, indolent prostate cancer.

[0237] A stratification score was calculated by combining Cts values of the eight genes in the panel with a predefined algorithm to discriminate high risk, aggressive prostate cancer and low risk, indolent prostate cancer. Then compare the stratification score with a predetermined high risk stratification score cutoff value to make a diagnosis. The diagnosis of all specimens with the gene panel was then compared to the pathological diagnosis of the risk of the specimens and the diagnostic performance of the panel was assessed by discriminant analysis using a statistical software program (XLSTAT).

Results

[0238] The result showed that the 8-Gene Panel consisting of MY06, LMTK2, PCA3, GSTP1, HPN, CCND1, FN1 and PMP22 was able to distinguish high risk, aggressive prostate cancer from low risk, indolent prostate cancer. As shown in Table 2, the 8-Gene Panel was able to distinguish high risk, aggressive prostate cancer from low risk, indolent prostate cancer with sensitivity of 90.3% and specificity of 93.3% (p<0.0001). The positive predictive value (PPV) reached 98.5% and the negative predictive value (NPV) reached 66.7%. The ROC analysis was performed to measure the classification power of the 8- Gene Panel and the result showed AUC of the ROC curve to be 0.950 (Figure 2), a high value for prostate cancer stratification.

Table 2 Positive Negative Total

Aggressive Cancer 65 7 72

Indolent Cancer 1 14 15

Sensitivity 90.3%

Specificity 93.3%

PPV 98.5%

NPV 66.7%

PO.OOOl

Figure 2

EXAMPLE 3

[0239] The diagnostic performance of a 25-Gene Panel comprising of HIF1A, FGFR1, BIRC5, AMACR, CR1SP3, FN1, HPN, MY06, PSCA, PMP22, GO1M1, LMTK2, EZH2, GSTP1, PCA3, VEGFA, CST3, PTEN, PIP5K1A, CDK1, TMPRSS2, ANXA3, CCNA1, CCND1 and KLK3 was tested for prostate cancer diagnosis using patient urine samples collected without digital rectal examination from a retrospective urine study.

Patients and Methods

[0240] The 614 patient urine study was approved by an Institutional Review Board. With informed consent of the patients, urine samples were collected before needle biopsy, radical prostatectomy or electroprostatectomy. ~15 ml urine samples were centrifuged at 1000 xg and the cell pellets were flash frozen and stored at -80°C. The pathological diagnosis of prostate cancer or benign prostate (including men with BPH and/or prostatitis) was based on pathological analysis of biopsy or surgical specimen.

[0241] The frozen urine pellet was thawed at 37°C and resuspended in cold PBS followed by centrifugation at 1000 xg for 10 min. Quick-RNA MicroPrep Kit was used to purify total RNA from the cell pellet following the manufacturer’s procedure (Zymo Research, Irvine, CA, USA). 100 ng purified RNA was then used for reverse transcription of cDNA using either High Capacity cDNA Reverse Transcription Kit (Life Technologies, Foster City, CA, USA) or iScript Reverse Transcription Supermix for real time qRT- PCR (Bio-Rad, Hercules, CA, USA) following the manufacturers’ protocols. The cDNA from reverse transcription was preamplified using TaqMan® PreAmp Master Mix (Thermo Fisher Scientific, Waltham, MA, USA) or Prostate cancer PreAmplification Mix (Hao Rui Jia Biotech Ltd., Beijing, China) according to the manufacturers’ directions. Real-time qRT-PCR was performed to assess mRNA expression levels using predesigned primers and probe assays from Integrated DNA Technologies (San Diego, CA, USA) on ABI Quantstudio 6, ABI 7500 or ABI 7900 Real-Time PCR System (Thermo Fisher Scientific, Waltham, MA, USA). The PCR reaction was set in 10 pl volume, which contains preamplified cDNA transcribed from 0.2 ng of purified RNA, 5 pl of 2x TaqMan® Universal PCR Master Mix (Thermo Fisher Scientific, Waltham, MA, USA) or PrimeTime® Gene Expression Master Mix (Integrated DNA Technologies, San Diego, CA, USA), 500 nM each of forward and reverse amplification primers, and 250 nM of probe. The real-time qRT-PCR was performed using the following cycling condition: 10 minutes at 95°C for polymerase activation, and 40 cycles of 15 seconds at 95°C and 1 minute at 60°C. For each gene, triplicate PCR were performed. All of the gene expression measurement and calculation were performed blindly without prior knowledge of patient information.

[0242] The mRNA expression level of the housekeeping gene beta-actin was measured in each urine sample and used for gene expression normalization to control variation of cDNA quantity in the patient samples. The cycle threshold (Ct) value of each gene in the 25 -Gene Panel was divided by the Ct value of the betaactin and then multiplied by 1000 as the normalized gene expression value (CtS=Ct(sample)/Ct(actin)* 1000). For each gene, average Ct value from triplicate PCR was used.

[0243] A diagnostic score was calculated by combining Cfcvalues of all genes in the panel with a predefined algorithm to discriminate prostate cancer and benign prostate. Then compare the diagnostic score with a predetermined cancer diagnostic score cutoff value to make a diagnosis. The diagnosis of all samples with the gene panel was then compared to the pathological diagnosis of the samples and the diagnostic performance was tested by discriminant analysis using a statistical software program (XESTAT).

Results

[0244] The result showed that the 25-Gene Panel comprising of HIF1A, FGFR1, BIRC5, AMACR, CRISPS, FN1, HPN, MY06, PSCA, PMP22, GOLM1, LMTK2, EZH2, GSTP1, PCA3, VEGFA, CST3, PTEN, PIP5K1A, CDK1, TMPRSS2, ANXA3, CCNA1, CCND1 and KLK3 was able to distinguish prostate cancer from benign prostate. As shown in Table 3, the 25-Gene Panel was able to distinguish prostate cancer from benign prostate with sensitivity of 92.5% and specificity of 91.5% (p<0.0001). The positive predictive value (PPV) reached 98.4% and the negative predictive value (NPV) reached 68.8%. The ROC analysis was performed to measure the classification power of the 25-Gene Panel for cancer diagnosis and the result showed AUC of the ROC curve to be 0.946 (Figure 3), a high value for prostate cancer diagnosis.

Table 3

Positive Negative Total

Cancer 481 39 520

Non-Cancer 8 86 94

Sensitivity 92.5%

Specificity 91.5%

PPV 98.4%

NPV 68.8%

PO.OOOl

Figure 3

EXAMPLE 4

[0245] The diagnostic performance of a 25-Gene Panel comprising of HIF1A, FGFR1, BIRC5, AMACR, CRISP3, FN1, HPN, MY06, PSCA, PMP22, GOLM1, LMTK2, EZH2, GSTP1, PCA3, VEGFA, CST3, PTEN, PIP5K1A, CDK1, TMPRSS2, ANXA3, CCNA1, CCND1 and KLK3 was tested for prostate cancer diagnosis using patient urine samples collected without digital rectal examination from a prospective urine study.

Patients and Methods

[0246] The 396 patient urine study was approved by an Institutional Review Board. With informed consent of the patients, urine samples were collected before needle biopsy, radical prostatectomy or electroprostatectomy. 10-45 ml urine samples were voided into 50 ml urine collection tubes containing DNA/RNA preservative AssayAssure (Thermo Fisher Scientific, Waltham, MA, USA) or U-Preserve (Hao Rui Jia Biotech Ltd., Beijing, China) and stored at 4°C until processing within seven days. The urine sample was centrifugation at 1000xg for 10 min and the pellet was washed with phosphate-buffered saline (PBS) followed by a second centrifugation at 1000xg for 10 min. The cell pellet was processed for RNA purification or immediately frozen on dry ice and stored at -80°C until future purification. The pathological diagnosis of prostate cancer or benign prostate (including men with BPH and/or prostatitis) was based on pathological analysis of biopsy or surgical specimen.

[0247] The gene expression quantification of the urine samples was performed as described in the retrospective urine study. A diagnostic score was calculated by combining Cts values of all genes in the panel with the same algorithm as used in the retrospective urine study to discriminate prostate cancer and benign prostate. Then compare the diagnostic score with the same cancer diagnostic score cutoff value to make a diagnosis. The diagnostic performance was measured by discriminant analysis using XLSTAT.

Results

[0248] The result showed that the 25-Gene Panel comprising of HIF1A, FGFR1, BIRC5, AMACR, CRISPS, FN1, HPN, MYO6. PSCA, PMP22, G0IM1, LMTK2, EZH2, GSTP1, PCA3, VEGFA, CST3, PTEN, PIP5K1A, CDK1, TMPRSS2, ANXA3, CCNA1, CCND1 and KLK3 was able to distinguish prostate cancer from benign prostate. As shown in Table 4, the 25-Gene Panel was able to distinguish prostate cancer from benign prostate with sensitivity of 85.0%, specificity of 94.7% (p<0.0001), positive predictive value (PPV) of 94.6% and negative predictive value (NPV) of 85.2%. The ROC analysis was performed to measure the classification power of the 25-Gene Panel for cancer diagnosis and the result showed AUC of the ROC curve to be 0.901 (Figure 4).

Table 4

Positive Negative Total

Cancer 176 31 207

Non-Cancer 10 179 189

Sensitivity 85.0%

Specificity 94.7%

PPV 94.6%

NPV 85.2%

PO.OOOl

Figure 4

EXAMPLE 5

[0249] The diagnostic performance of a 14-Gene Panel comprising of PMP22, G0LM1, LMTK2, EZH2, GSTP1, PCA3, VEGFA, CST3, PTEN, PIP5K1A, CDK1, TMPRSS2, ANXA3 and CCND1 was tested for prostate cancer diagnosis using patient urine samples collected without digital rectal examination from a prospective urine study.

Patients and Methods

[0250] The 393 patients prospective urine cohort was used to assess the ability of the 14-Gene Panel for prostate cancer diagnosis. A diagnostic score was calculated by combining Cts values of all genes in the panel with an algorithm to discriminate prostate cancer and benign prostate. Then compare the diagnostic score with a predetermined cancer diagnostic score cutoff value to make a diagnosis. The diagnostic performance was measured by discriminant analysis using XLSTAT.

Results [0251] The result showed that the 14-Gene Panel comprising o PMP22, G0EM1, EMTK2, EZH2, GSTP1, PCA3, VEGFA, CST3, PTEN, PIP5K1A, CDK1, TMPRSS2, ANXA3 and CCND1 was able to distinguish prostate cancer from benign prostate. As shown in Table 5, the 14-Gene Panel was able to distinguish prostate cancer from benign prostate with sensitivity of 80.7%, specificity of 74.9% (p<0.0001), positive predictive value (PPV) of 77.3%, and negative predictive value (NPV) of 78.6%. The ROC analysis was performed to measure the classification power of the 14-Gene Panel for cancer diagnosis and the result showed AUC of the ROC curve to be 0.854 (Figure 5).

Table 5

Positive Negative Total

Cancer 163 39 202

Non-Cancer 48 143 191

Sensitivity 80.7%

Specificity 74.9%

PPV 77.3%

NPV 78.6%

PO.OOOl

Figure 5

EXAMPLE 6

[0252] The diagnostic performance of a 25-Gene Panel comprising of HIF1A, FGFR1, BIRC5, AMACR, CRISP3, FN1, HPN, MY06, PSCA, PMP22, GOLM1, LMTK2, EZH2, GSTP1, PCA3, VEGFA, CST3, PTEN, PIP5K1A, CDK1, TMPRSS2, ANXA3, CCNA1, CCND1 and KLK3 was tested for prostate cancer diagnosis using prostate tissue specimens.

Patients and Methods

[0253] The GSE17951 prostate tissue specimen cohort includes quantitative mRNA expression data of prostate cancer and benign prostate specimens obtained from Affymetrix U133Plus2 array. The Prostate cancer tissues (n=56) in the cohort were collected from patient biopsy specimens and the benign prostate tissues (n=98) were obtained from prostate autopsy specimens of patients with benign disease. The gene expression levels of the 25 genes in the panel were obtained from the database and normalized with betaactin expression level.

[0254] A diagnostic score was calculated by combining gene expression values of all genes in the panel with a predefined algorithm to discriminate prostate cancer and benign prostate. Then compare the diagnostic score with a predetermined cancer diagnostic score cutoff value to make a diagnosis. The diagnosis of all samples with the gene panel was then compared to the pathological diagnosis of the samples and the diagnostic performance was measured by discriminant analysis using XLSTAT.

Results

[0255] The result showed that the 25-Gene Panel comprising of HIF1A, FGFR1, BIRC5, AMACR, CRISP 3, FN1, HPN, MYO6, PSCA, PMP22, GO1M1, LMTK2, EZH2, GSTP1, PCA3, VEGFA, CST3, PTEN, PIP5K1A, CDK1, TMPRSS2, ANXA3, CCNA1, CCND1 and KLK3 was able to distinguish prostate cancer from benign prostate using prostate tissue specimens. As shown in Table 6, the 25-Gene Panel was able to distinguish prostate cancer from benign prostate with sensitivity of 100% and specificity of 96.0% (p<0.0001). The positive predictive value (PPV) reached 93.2% and the negative predictive value (NPV) reached 100%. The ROC analysis was performed to measure the classification power of the 25-Gene Panel and the result showed AUC of the ROC curve to be 0.998 (Figure 6), a high value for cancer diagnosis.

Table 6

Positive Negative Total

Cancer 55 0 55

Non-cancer 4 95 99

Total 59 95 154

Sensitivity (95% CI) 100% (100-100%)

Specificity (95% CI) 96.0% (99.8-92.1%)

Figure 6

EXAMPLE 7 [0256] The diagnostic performance of a 25-Gene Panel comprising of HIF1A, FGFR1, BIRC5, AMACR, CR1SP3, FN1, HPN, MY06, PSCA, PMP22, GO1M1, EMTK2, EZH2, GSTP1, PCA3, VEGFA, CST3, PTEN, PIP5K1A, CDK1, TMPRSS2, ANXA3, CCNA1, CCND1 and KLK3, PSA, and their combination was tested for prostate cancer diagnosis using patient urine samples collected without digital rectal examination from the combined retrospective and prospective studies.

Patients and Methods

[0257] 415 patients from the retrospective and prospective study cohorts with PSA data was used as PSA cohort to assess the ability of the 25-Gene Panel, PSA and their combination for prostate cancer diagnosis. A diagnostic score was calculated by combining Cts values of all genes in the panel with the predefined algorithm to discriminate prostate cancer and benign prostate. The diagnostic performance of PSA, the 25- Gene Panel, and their combination was assessed by discriminant analysis using XLSTAT.

Results

[0258] The result showed that the 25-Gene Panel comprising of HIF1A, FGFR1, BIRC5, AMACR, CRISP3, FN1, HPN, MY06, PSCA, PMP22, G0IM1, LMTK2, EZH2, GSTP1, PCA3, VEGFA, CST3, PTEN, PIP5K1A, CDK1, TMPRSS2, ANXA3, CCNA1, CCND1 and KLK3 was able to distinguish prostate cancer from benign prostate. As shown in Table 7, the 25-Gene Panel was able to distinguish prostate cancer from benign prostate with sensitivity of 88.6%, specificity of 93.2% (p<0.0001), odds ratio (OR) of 107.3 and AUC of the ROC curve of 0.939 (Figure 7A). In contrast, PSA had low sensitivity of 36.3%, odds ratio of 6.7 and AUC of 0.710 (Figure 7B). When they were combined, the diagnostic performance was improved with higher sensitivity of 94.8% and AUC of 0.961 (Figure 7C).

Table 7

Sensitivity Specificity PPV NPV OR AUC

EXAMPLE 8

[0259] The ability of a 25-Gene Panel comprising otHIFlA, FGFR1, BIRC5, AMACR, CRISP 3, FN1, HPN, MYO6, PSCA, PMP22, G0IM1, IMTK2, EZH2, GSTP1, PCA3, VEGFA, CST3, PTEN, PIP5K1A, CDK1, TMPRSS2, ANXA3, CCNA1, CCND1 and KLK3 to show the absence of prostate cancer after the tumors had been surgically removed by radical prostatectomy was tested in a prospective urine study.

Patients and Methods

[0260] Ten patients undergoing prostatectomy were recruited to collect urine samples several days before and after surgery. The diagnosis of the urine samples collected before and after the surgery was performed using the 25-Gene Panel and compared with the pathological diagnosis of prostate cancer.

Results

[0261] The result showed that nine out of ten urine samples (90%) were diagnosed to be non-cancer after radical prostatectomy (Table 8), which was consistent with successful surgery in most patients. The one patient diagnosed to be prostate cancer may still have residual cancer lesion after the surgery and need additional treatment. The result suggests that the 25-Gene Panel comprising of HIF1A, FGFR1, BIRC5, AMACR, CRISP3, FN1, HPN, MY06, PSCA, PMP22, G0IM1, IMTK2, EZH2, GSTP1, PCA3, VEGFA, CST3, PTEN, PIP5K1A, CDK1, TMPRSS2, ANXA3, CCNA1, CCND1 and KLK3 has potential to be used as an accurate and simple test to measure efficacy of radical prostatectomy treatment.

Table 8

Pre-Surgery Urine Post-Surgery Urine

Patient A Prostate cancer Prostate cancer

Patient B Prostate cancer Non-Prostate cancer

Patient C Prostate cancer Non-Prostate cancer

Patient D Prostate cancer Non-Prostate cancer

Patient E Prostate cancer Non-Prostate cancer

Patient F Prostate cancer Non-Prostate cancer

Patient G Prostate cancer Non-Prostate cancer

Patient H Prostate cancer Non-Prostate cancer Patient I Prostate cancer Non-Prostate cancer

Patient J Prostate cancer Non-Prostate cancer

% Non-Prostate cancer 0 90.0%

EXAMPLE 9

[0262] The diagnostic performance of a 14-Gene Panel comprising of PMP22, G0LM1, LMTK2, EZH2, GSTP1, PCA3, VEGFA, CST3, PTEN, PIP5K1A, CDK1, TMPRSS2, ANXA3 and CCND1 was tested for diagnosing higher risk and lower risk prostate cancer using patient urine samples collected without digital rectal examination from a prospective urine study.

Patients and Methods

[0263] The urine sample cohort comprising of 202 prostate cancer patients was obtained from a prospective study approved by the Institutional Review Board. The urine samples were collected from seven hospitals collaborated in the study before needle biopsy, radical prostatectomy or electro-prostatectomy from patients with informed consent.

[0264] The pathological diagnosis of prostate cancer risk was defined based on National Comprehensive Cancer Network (NCCN) guidelines. NCCN recommends patients with very high, high and unfavorable intermediate risk to receive immediate treatment, while most patients with very low, low and favorable intermediate risk are suggested to be placed on active surveillance. Therefore in this study, we classified patients with very high, high and unfavorable intermediate risk as higher risk prostate cancer who needs treatment and patients with very low, low and favorable intermediate risk to be lower risk patients who need active surveillance. The higher risk prostate cancer patients in our study were classified as meeting any of the following criteria: Gleason score >7, Gleason score 4+3=7, cancer stage >T3, PSA >20 ng/mL at diagnosis, and more than half of biopsy core with cancer. The rest of the patients were classified as lower risk prostate cancer.

[0265] The urine sample processing and gene expression quantification were performed as in the previous examples. For prostate cancer risk stratification, the CtS values of the 14 genes in the panel were used to generate a classification score (Stratification D Score) for each urine sample using a stratification algorithm. The sample was diagnosed to be higher risk prostate cancer when Stratification D Score was >0, whereas the sample was diagnosed to be lower risk prostate cancer when Stratification D Score was <0. The diagnosis of each sample by the 14-Gene Panel was compared to their pathological diagnosis of higher and lower risk to assess the diagnostic performance by discriminant analysis using XLSTAT.

Results

[0266] The result showed that the 14-Gene Panel comprising o PMP22, G0EM1, LMTK2, EZH2, GSTP1, PCA3, VEGFA, CST3, PTEN, PIP5K1A, CDK1, TMPRSS2, ANXA3 and CCND1 was able to distinguish higher risk and lower risk prostate cancer. As shown in Table 9, the 14-Gene Panel was able to distinguish higher risk and lower risk prostate cancer with sensitivity of 83.2% and specificity of 79.3% (p<0.0001), positive predictive value (PPV) of 91.9% and negative predictive value (NPV) of 62.7%. The ROC analysis was performed to measure the classification power of the 14-Gene Panel and the result showed AUC of the ROC curve to be 0.897 (Figure 8).

Table 9 Positive Negative Total

Higher Risk Cancer 124 25 149

Lower Risk Cancer H 42 53

Sensitivity 83.2%

Specificity 79.3%

PPV 91.9%

NPV 62.7%

PO.OOOl

Figure 8

EXAMPLE 10

[0267] The diagnostic performance of a 25-Gene Panel comprising of HIF1A, FGFR1, BIRC5, AMACR, CR1SP3, FN1, HPN, MYO6, PSCA, PMP22, GO1M1, LMTK2, EZH2, GSTP1, PCA3, VEGFA, CST3, PTEN, PIP5K1A, CDK1, TMPRSS2, ANXA3, CCNA1, CCND1 and KLK3 was tested for diagnosing clinically significant prostate cancer and clinically insignificant prostate cancer using urine samples collected without digital rectal examination from a prospective urine study.

Patients and Methods

[0268] The 207 prostate cancer patient prospective urine study was used to assess the 25-Gene Panel for diagnosing clinically significant and clinically insignificant prostate cancer. The pathological diagnosis of clinically significant cancer and clinically insignificant cancer group was defined based on the National Comprehensive Cancer Network (NCCN) guidelines. NCCN recommends patients with very high, high and unfavorable intermediate risk to receive immediate treatment, while most patients with very low, low and favorable intermediate risk are suggested to be placed on active surveillance. Therefore in this study, we classified patients with very high, high and unfavorable intermediate risk as clinically significant cancer patients who need treatment and patients with very low, low and favorable intermediate risk to be clinically insignificant cancer patients who need active surveillance. The diagnosis of clinically significant prostate cancer was classified as meeting any of the criteria: Gleason score >7, Gleason score 4+3=7, cancer staging >T3, PSA >20 ng/mL at diagnosis, and more than half of biopsy cores with cancer. The rest of the patients were classified as having clinically insignificant prostate cancer. [0269] The gene expression quantification was performed as described in the previous examples. A clinically significant cancer score was calculated by combining Cts values of all genes in the panel with a predefined algorithm to discriminate clinically significant and clinically insignificant prostate cancer. The diagnosis of all samples with the gene panel was then compared to the pathological diagnosis of the samples and the diagnostic performance was measured by discriminant analysis using XLSTAT.

Results

[0270] The result showed that the 25 Gene-Panel comprising of HIF1A, FGFR1, BIRC5, AMACR, CRISP 3, FN1, HPN, MY06, PSCA, PMP22, GO1M1, LMTK2, EZH2, GSTP1, PCA3, VEGFA, CST3, PTEN, PIP5K1A, CDK1, TMPRSS2, ANXA3, CCNA1, CCND1 and KLK3 was able to distinguish clinically significant prostate cancer and clinically insignificant prostate cancer with high accuracy. As shown in Table 10, the 25 -Gene Panel was able to distinguish clinically significant cancer and clinically insignificant cancer with sensitivity of 87.0%, specificity of 97.8% (p<0.0001), positive predictive value (PPV) of 99.3%, and negative predictive value (NPV) of 67.7%. The ROC analysis was performed to measure the classification power of the 25-Gene Panel and the result showed AUC of the ROC curve to be 0.958 (Figure 9), a high value for cancer stratification and subtyping.

Table 10

Positive

Total

Clinically Significant Cancer 141

162

Clinically Insignificant Cancer 1

45

Sensitivity 87.0%

Specificity 97.8%

PPV 99.3%

NPV _ 67.7%

PO.OOOl

Figure 9

EXAMPLE 11 [0271] The diagnostic performance of a 24-Gene Panel comprising of CCND1, HIF1A, FGFR1, BIRC5, AMACR, CRISPS, FN1, HPN, MYO6, PSCA, PMP22, G0IM1, IMTK2, EZH2, GSTP1, PCA3, VEGFA, CST3, PTEN, PIP5K1A, CDK1, TMPRSS2, ANXA3 and CCNA1 was tested for diagnosing clinically significant prostate cancer and clinically insignificant prostate cancer using urine samples collected without digital rectal examination from a prospective urine study.

Patients and Methods

[0272] In the 207 patient prospective urine cohort, the diagnostic performance of a 24-Gene Panel for distinguishing clinically significant and clinically insignificant prostate cancer was tested.

Results

[0273] The result showed that the 24-Gene Panel comprising of CCND1, HIF1A, FGFR1, BIRC5, AMACR, CRISP3, FNI, HPN, MY06, PSCA, PMP22, GOIM1, IMTK2, EZH2, GSTP1, PCA3, VEGFA, CST3, PTEN, PIP5K1A, CDK1, TMPRSS2, ANXA3 and CCNA1 was able to distinguish clinically significant and clinically insignificant prostate cancer with high accuracy. As shown in Table 11, the 24-Gene Panel was able to distinguish clinically significant and insignificant cancer with sensitivity of 86.0%, specificity of 97.7% (p<0.0001), positive predictive value (PPV) of 99.3%, negative predictive value (NPV) of 64.6%. The ROC analysis was performed to measure the classification power of the 24-Gene Panel and the result showed AUC of the ROC curve to be 0.959 (Figure 10), a high value for cancer stratification and subtyping.

Table 11

Positive

Total

Clinically Significant Cancer 141

164

Clinically Insignificant Cancer 1

43

Sensitivity 86.0%

Specificity 97.7%

PPV 99.3%

NPV _ 64.6%

PO.OOOl

Figure 10

EXAMPLE 12

[0274] The diagnostic performance of a 24-Gene Panel comprising of CCND1, HIF1A, FGFR1, BIRC5, AMACR, CRISP3, FN1, HPN, MY06, PSCA, PMP22, G0IM1, LMTK2, EZH2, GSTP1, PCA3, VEGFA, CST3, PTEN, PIP5K1A, CDK1, TMPRSS2, ANXA3 and CCNA1 was tested for diagnosing clinically significant prostate cancer and clinically insignificant prostate cancer using urine samples collected without digital rectal examination from a retrospective urine study.

Patients and Methods

[0275] In the retrospective urine cohort comprising of 520 prostate cancer patients, the diagnostic performance of a 24-Gene Panel was tested for diagnosing clinically significant and insignificant cancer.

Results

[0276] The result showed that the 24-Gene Panel comprising of CCND1, HIF1A, FGFR1, BIRC5, AMACR, CRISP3, FN1, HPN, MY06, PSCA, PMP22, G0IM1, LMTK2, EZH2, GSTP1, PCA3, VEGFA, CST3, PTEN, PIP5K1A, CDK1, TMPRSS2, ANXA3 and CCNA1 was able to distinguish clinically significant and insignificant prostate cancer with high accuracy. As shown in Table 12, the 24-Gene Panel had sensitivity of 83.8%, specificity of 94.4% (p<0.0001), positive predictive value (PPV) of 94.3%, negative predictive value (NPV) of 84.2%, and AUC of the ROC curve of 0.916 (Figure 11).

Table 12

Positive Negative Total

Clinically Significant Cancer 228 44 272

Clinically Insignificant Cancer 14 234 248

Sensitivity 83.8%

Specificity 94.4%

PPV 94.3%

NPV 84.2%

PO.OOOl

Figure 11

EXAMPUE 13

[0277] The diagnostic performance of a 24-Gene Panel comprising of CCND1, HIF1A, FGFR1, BIRC5, AMACR, CRISP3, FN1, HPN, MY06, PSCA, PMP22, G0IM1, LMTK2, EZH2, GSTP1, PCA3, VEGFA, CST3, PTEN, PIP5K1A, CDK1, TMPRSS2, ANXA3 and CCNA1, cancer stage, Gleason score and their combination was tested for diagnosing clinically significant and clinically insignificant prostate cancer using urine samples collected without digital rectal examination from the combined cohorts.

Patients and Methods

[0278] The diagnostic performance of a 24-Gene Panel, cancer stage, Gleason score and their combination was tested in the combined retrospective and prospective cohort with cancer stage and Gleason score information (n = 423).

Results

[0279] The result showed that the 24-Gene Panel comprising of CCND1, HIF1A, FGFR1, BIRC5, AMACR, CRTSP3, FN1, HPN, MY06, PSCA, PMP22, GO1M1, EMTK2, EZH2, GSTP1, PCA3, VEGFA, CST3, PTEN, PIP5K1A, CDK1, TMPRSS2, ANXA3 and CCNA1 was able to distinguish clinically significant and clinically insignificant prostate cancer with high accuracy. As shown in Table 13, the 24-Gene Panel had sensitivity of 85.0%, specificity of 94.9%, positive predictive value (PPV) of 95.1%, negative predictive value (NPV) of 84.6%, and AUC of the ROC curve of 0.892. In contrast, cancer stage had low sensitivity of 72.3% and AUC of 0.874, and Gleason score had low specificity of 23.5% and AUC of 0.578. However, when they were combined, the diagnostic performance was improved with sensitivity of 95.7%, specificity of 96.9%, positive predictive value (PPV) of 97.3%, negative predictive value (NPV) of 94.1%, and AUC of 0.966 (Figure 12A-D).

Table 13

Sensitivity Specificity PPV NPV

Cancer Stage 72.3%, 99.5% 99.4% 75.6%

Gleason Score 85.0% 23.5% 56.1% 57.5 A

24-Gene 85.0% 94.9% 95.1% 84.6%

Combination 94.7% 96.9% 97.3% 94.1%

Figure 12

EXAMPLE 14

[0280] The diagnostic performance of a 24-Gene Panel comprising of CCND1, HIF1A, FGFR1, BIRC5, AMACR, CRISP3, FN1, HPN, MY06, PSCA, PMP22, GOLM1, LMTK2, EZH2, GSTP1, PCA3, VEGFA, CST3, PTEN, PIP5K1A, CDK1, TMPRSS2, ANXA3 and CCNA1 was tested for diagnosing clinically significant prostate cancer and clinically insignificant prostate cancer using prostate tissue specimens.

Patients and Methods

[0281] For prostate tissue specimen cohort, mRNA expression Z-Scores of the genes in the panels were obtained from the MSKCC dataset at www.cbioportal.com. The mRNA expression Z-Scores of the 24 genes in the panel were used to generate a clinically significant cancer score by combining Z-Scores of all genes in the panel with a predefined algorithm to discriminate clinically significant prostate cancer and clinically insignificant prostate cancer. Then compare the clinically significant cancer score with a predetermined clinically significant cancer score cutoff value to make a diagnosis.

[0282] The diagnosis of all samples with the gene panel was then compared to the pathological diagnosis of the samples and the diagnostic performance was measured by discriminant analyzing using XLSTAT.

Results [0283] The result showed that the 24-Gene Panel comprising of CCND1, HIF1A, FGFR1, BIRC5, AMACR, CR1SP3, FN1, HPN, MY06, PSCA, PMP22, GO1M1, EMTK2, EZH2, GSTP1, PCA3, VEGFA, CST3, PTEN, PIP5K1A, CDK1, TMPRSS2, ANXA3 and CCNA1 was able to distinguish clinically significant and insignificant prostate cancer with high accuracy in the prostate tissue specimen cohort. As shown in Table 14, the 24-Gene Panel was able to distinguish clinically significant and insignificant cancer with sensitivity of 71.1%, specificity of 98.1% (p<0.0001), positive predictive value (PPV) of 94.1%, negative predictive value (NPV) of 88.7%, and AUC of the ROC curve of 0.976 (Figure 13).

Table 14

Positive

Total

Clinically Significant Cancer 32

45

Clinically Insignificant Cancer 2 102 104

Sensitivity 71.1%

Specificity 98.1%

PPV 94.1%

NPV _ 88.7%

PO.OOOl

Figure 13

EXAMPLE 15

[0284] The diagnostic performance of an 18-Gene Panel comprising of PTEN, CDK1, TMPRSS2, HIF1A, FGFR1, BIRC5, CRISP3, FN1, HPN, PSCA, PMP22, LMTK2, EZH2, GSTP1, PCA3, VEGFA, ANXA3 and KLK3 was tested for prediction of prostate cancer metastasis using prostate tissue specimens.

Patients and Methods

[0285] A dataset of prostate tissue cohort MSKCC Prostate Oncogenome Project was obtained from cBioPortal (www.cbioportal.com) database and used in the study. The cohort contains transcriptome profiles of 218 prostate cancer tissue specimens (181 primaries and 37 metastases). The specimens were obtained from 218 patients treated by radical prostatectomy (RP) with at least 70% tumor content. The transcriptome measurements including mRNA were conducted without amplification. The quantitative mRNA expression Z-Scores of the genes in the panel were obtained from the dataset along with clinicopathological information including cancer metastasis and Gleason scores. Patients without Z-Score of the genes in the panel or without metastasis information were excluded from the cohort, resulting in a cohort of 150 patients including 19 metastases.

[0286] A metastasis score was calculated by combining Z-Score of all genes in the panel with a predefined algorithm to discriminate metastatic and non-metastatic cancer. Then compare the metastasis score with a predetermined metastatic cancer score cutoff value to make a prediction. The prediction of all specimens with the gene panel was then compared to the imaging diagnosis of the specimens during follow-up and the prognostic performance was measured by discriminant analysis using XLSTAT.

Results

[0287] The result showed that the 18-Gene Panel comprising of PTEN, CDK1, TMPRSS2, HIF1A, FGFR1, BIRC5, CRISP 3, FN1, HPN, PSCA, PMP22, LMTK2, EZH2, GSTP1, PCA3, VEGFA, ANXA3 and KLK3 was able to predict metastasis. As shown in Table 15, the 18-Gene Panel was able to distinguish metastatic prostate cancer from non-metastatic prostate cancer with high sensitivity of 100%, specificity of 100% (p<0.0001), positive predictive value (PPV) of 100%, negative predictive value (NPV) of 100%, and AUC of the ROC curve of 1 (Figure 14), showing extremely high predictive power.

Table 15

Positive Negative Total

Metastatic Cancer 19 0 19

Non-Metastatic Cancer 0 131 131

Sensitivity 100%

Specificity 100%

PPV 100%

NPV 100%

PO.OOOl

Figure 14

EXAMPLE 16

[0288] The prognostic performance of an 18-Gene Panel comprising of PTEN, CDK1, TMPRSS2, HIF1A, FGFR1, BIRC5, CRISP3, FN1, HPN, PSCA, PMP22, LMTK2, EZH2, GSTP1, PCA3, VEGFA, ANXA3 and KLK3 was tested for prediction of prostate cancer metastasis in newly diagnosed cancer patients using urine samples collected without digital rectal examination from a prospective urine study.

Patients and Methods

[0289] In the 207 patient prospective urine study cohort, the patients were tested for cancer metastasis by imaging with computed tomography (CT), magnetic resonance (MRI), X-ray and bone scan at diagnosis. During the average 5 year follow-up, the patients were assessed every 3 months for serum PSA, bone scan, and CT, MRI or positron emission tomography (PET) to monitor metastasis.

[0290] The urine sample processing and gene expression quantification were performed as described in the previous examples. A metastasis score was calculated by combining Cts values of all genes in the panel with a predefined algorithm to discriminate metastatic prostate cancer and non-metastatic prostate cancer. Univariate and multivariate Cox proportional hazards regression analyses and Kaplan-Meier survival plot of metastatic cancer-free survival for the 18-Gene Panel were conducted using SPSS (IBM, Armonk, New York).

Results

[0291] In Cox regression analysis, the 18-Gene Panel had high predictive power with hazard ratio (HR) of 68.64 (95% CI 16.70-282.16) (p<0.001) in univariate analysis and 53.45 (95% CI 12.90-221.43) (p<0.001) in multivariate analysis after adjusted for Gleason score and PSA (Table 16). In contrast, Gleason score and PSA had much lower HR in univariate and multivariate analyses.

[0292] Kaplan-Meier plot of metastasis-free survival was performed to show a large and statistically significant difference in survival between the 18-Gene Panel Positive Group with -50% survival at 12 months and the Negative Group with -97% survival at 80 months (log rank p<0.001) (Figure 15). The result showed that the 18-Gene Panel comprising of PTEN, CDK1, TMPRSS2, HIF1A, FGFR1, BIRC5, CRISP 3, FN1, HPN, PSCA, PMP22, EMTK2, EZH2, GSTP1, PCA3, VEGFA, ANXA3 and KLK3 was able to predict metastasis-free survival with high accuracy.

Table 16

Univariate Multivariate

Variable HR (95% CI) P Value HR (95% CI) P Value

18-Gene Panel 68.64 (16.70-282.16) <0.001 53.45 (12.90-221.43) <0.001

Gleason Score 2.17 (1.68-2.78) <0.001 1.75 (1.33-2.29) <0.001

PSA 1.00 (1.00-1.00) <0.001 1.00 (1.00-1.00) 0.418

Figure 15

EXAMPLE 17

[0293] The prognostic performance of an 18-Gene Panel comprising of PTEN, CDK1, TMPRSS2, HIF1A, FGFR1, BIRC5, CRISP 3, FNI, HPN, PSCA, PMP22, LMTK2, EZH2, GSTPI, PCA3, VEGFA, ANXA3 and KLK3 was tested for prediction of prostate cancer metastasis using urine samples collected without digital rectal examination from a prospective urine study.

Patients and Methods

[0294] In the 207 patient prospective urine study cohort, a metastasis score of the 18-Gene Panel was calculated by combining Cts values of all genes in the panel with a predefined algorithm to discriminate metastatic prostate cancer and non-metastatic prostate cancer. Then compare the metastasis score with a predetermined metastatic cancer score cutoff value to make a prediction. The prediction of all samples with the gene panel was then compared to the imaging diagnosis of the samples during follow-up and the predictive performance was measured by discriminant analysis using XLSTAT.

Results

[0295] The result showed that the 18-Gene Panel comprising of PTEN, CDK1, TMPRSS2, HIF1A, FGFR1, BIRC5, CRISP 3, FNI, HPN, PSCA, PMP22, LMTK2, EZH2, GSTPI, PCA3, VEGFA, ANXA3 and KLK3 was able to predict metastatic and non-metastatic prostate cancer. As shown in Table 17, the 18-Gene Panel was able to predict metastatic prostate cancer with sensitivity of 96.6%, specificity of 84.5% (p<0.0001), positive predictive value (PPV) of 71.3%, negative predictive value (NPV) of 98.4%, and AUC of the ROC curve of 0.957 (Figure 16).

Table 17

Positive

Total

Metastatic Cancer 57

59

Non-Metastatic Cancer 23 125 148

Sensitivity 96.6% Specificity 84.5%

PPV 71.3%

NPV 98.4%

PO.OOOl

Figure 16

EXAMPLE 18

[0296] The prognostic performance of an 18-Gene Panel comprising of PTEN, CDK1, TMPRSS2, HIF1A, FGFR1, BIRC5, CRISP 3, FN1, HPN, PSCA, PMP22, LMTK2, EZH2, GSTP1, PCA3, VEGFA, ANXA3 and KLK3 was tested for prediction of prostate cancer metastasis using urine samples collected without digital rectal examination from a retrospective urine study.

Patients and Methods

[0297] In the 520 patient retrospective urine study cohort, the predictive performance of the 18-Gene Panel was measured by discriminant analysis using XLSTAT.

Results

[0298] The result showed that the 18-Gene Panel comprising of PTEN, CDK1, TMPRSS2, HIF1A, FGFR1, BIRC5, CRISP 3, FN1, HPN, PSCA, PMP22, LMTK2, EZH2, GSTP1, PCA3, VEGFA, ANXA3 and KLK3 was able to predict metastatic and non-metastatic prostate cancer in the retrospective cohort. As shown in Table 18, the 18-Gene Panel had sensitivity of 87.5%, specificity of 97.3% (p<0.0001), positive predictive value (PPV) of 33.3%, negative predictive value (NPV) of 99.8%, and AUC of the ROC curve of 0.991 (Figure 17), which was a high value for predicting prostate cancer metastasis.

Table 18

Positive Negative Total

Metastatic Cancer 7 1 8

Non-Metastatic Cancer 14 498 512

Sensitivity 87.5% Specificity 97.3%

PPV 33.3%

NPV 99.8%

PO.OOOl

Figure 17

EXAMPUE 19

[0299] The prognostic performance of an 18-Gene Panel comprising of PTEN, CDK1, TMPRSS2, HIF1A, FGFR1, BIRC5, CRISP3, FN1, HPN, PSCA, PMP22, LMTK2, EZH2, GSTP1, PCA3, VEGFA, ANXA3 and KLK3, PSA, Gleason score and their combination was tested for prediction of prostate cancer metastasis in newly diagnosed patients using urine samples collected without digital rectal examination from a prospective urine study.

Patients and Methods

[0300] The 207 patient prospective urine study with average 5 year follow-up was used to assess the predictive performance of the 18-Gene Panel, PSA, Gleason score and their combination for prediction of prostate cancer metastasis. Univariate and multivariate logistic regression analyses were performed to test the predictive performance using XLSTAT.

Results

[0301] The result showed that the 18-Gene Panel comprising of PTEN, CDK1, TMPRSS2, HIF1A, FGFR1, BIRC5, CRISP 3, FN1, HPN, PSCA, PMP22, EMTK2, EZH2, GSTP1, PCA3, VEGFA, ANXA3 and KLK3 was able to predict metastatic and non-metastatic prostate cancer with high accuracy. As shown in Table 19, the 18-Gene Panel had high sensitivity of 96.6%, specificity of 84.5% (p<0.0001), positive predictive value (PPV) of 71.3%, negative predictive value (NPV) of 98.4% and AUC of the ROC curve of 0.915. PSA had very low sensitivity of 15.5% and AUC of 0.761, while Gleason score had low sensitivity of 67.8% and AUC of 0.786. When they were combined, the predictive performance was found with low sensitivity of 77.6% and high AUC of 0.976 (Figure 18A-D).

Table 19 Sensitivity Specificity _ PPV _ NPV

PSA 15.5% 98.0% 75.0% 74.5%

Gleason Score 67.8% 82.4% 60.6% 86.5% 18-Gene Panel 96.6% 84.5% 71.3% 98.4% Combination 77.6% 97.3% 91.8% 91.6%

Figure 18

EXAMPLE 20

[0302] The prognostic performance of a 23-Gene Panel comprising of PTEN, PIP5K1A, CDK1, TMPRSS2, ANXA3, HIF1A, FGFR1, BIRC5, AMACR, CRISP3, PMP22, GOLPH2, EZH2, GSTP1, PCA3, VEGFA, CST3, CCNA1, CCND1, FN1, MY06, KLK3 and PSCA was tested for prediction of prostate cancer metastasis using urine samples collected without digital rectal examination from the combined prospective and retrospective urine studies.

Patients and Methods

[0303] The retrospective urine cohort comprising of 520 patients and the prospective urine cohort comprising of 207 patients was combined to form a 727 patients cohort. The predictive performance of a 23-Gene Panel was measured by discriminant analysis in the combination cohort.

Results [0304] The result showed that the 23-Gene Panel comprising of PTEN, PIP5K1A, CDK1, TMPRSS2, ANXA3, HIF1A, FGFR1, BIRC5, AMACR, CRISP3, PMP22, GOLPH2, EZH2, GSTP1, PCA3, VEGFA, CST3, CCNA1, CCND1, FN1, MYO6. KLK3 and PSCA was able to predict metastatic and non-metastatic prostate cancer in the combination cohort. As shown in Table 20, the 23-Gene Panel had sensitivity of 86.6%, specificity of 94.2% (p<0.0001), positive predictive value (PPV) of 60.4%, negative predictive value (NPV) of 98.6%, and AUC of the ROC curve of 0.970 (Figure 19).

Table 20

Positive Negative Total

Metastatic Cancer 58 9 67

Non-Metastatic Cancer 38 622 660

Sensitivity 86.6%

Specificity 94.2%

PPV 60.4%

NPV 98.6%

PO.OOOl

Figure 19

EXAMPLE 21

[0305] The prognostic performance of a 24-Gene Panel comprising of CCND1, HIF1A, FGFR1, BIRC5, AMACR, CRISP3, FN1, HPN, MY06, PSCA, PMP22, G0IM1, LMTK2, EZH2, GSTP1, PCA3, VEGFA, CST3, PTEN, PIP5K1A, CDK1, TMPRSS2, ANXA3 and CCNA1 was tested for prediction of prostate cancer metastasis using urine samples collected without digital rectal examination from the combined prospective and retrospective urine studies.

Patients and Methods

[0306] In the 727 patients combination cohort, the predictive performance of a 24-Gene Panel was assessed by discriminant analyzing using XLSTAT. Results

[0307] The result showed that the 24-Gene Panel comprising of CCND1, HIF1A, FGFR1, BIRC5, AMACR, CRISP3, FN1, HPN, MYO6, PSCA, PMP22, GO1M1, LMTK2, EZH2, GSTPI, PCA3, VEGFA, CST3, PTEN, PIP 5 KI A, CDK1, TMPRSS2, ANNA 3 and CCNA1 was able to predict metastatic and non-metastatic prostate cancer in the combination cohort. As shown in Table 21, the 24-Gene Panel was able to predict metastatic prostate cancer with sensitivity of 89.6%, specificity of 95.5% (p<0.0001), positive predictive value (PPV) of 66.7%, negative predictive value (NPV) of 98.9%, and AUC of the ROC curve of 0.974 (Figure 20).

Table 21

Positive Negative Total

Metastatic Cancer 60 7 67

Non-Metastatic Cancer 30 630 660

Sensitivity 89.6%

Specificity 95.5%

EXAMPLE 22

[0308] The prognostic performance of an 18-Gene Panel comprising of PTEN, CDK1, TMPRSS2, HIF1A, FGFR1, BIRC5, CRISP 3, FNI, HPN, PSCA, PMP22, LMTK2, EZH2, GSTPI, PCA3, VEGFA, ANXA3 and KLK3 for prediction of metastatic castration-resistant prostate cancer was assessed using urine samples collected without digital rectal examination from a prospective mCRPC PCa Cohort.

Patients and Methods

[0309] In the 207 patients prospective cohort, the patients were assessed periodically for castration-resistant prostate cancer (CRPC) following the EAU-EANM-ESTRO-ESUR-ISUP-SIOG Guidelines during the follow-up. Castration-resistant prostate cancer was defined as having castrate serum testosterone levels <50 ng/dL with three consecutive rises in PSA levels of two 50% increase obtained at least one week apart and PSA >2 ng/mL, and/or radiological progression of two or more new bone lesions detected by bone scan or soft tissue lesions using Response Evaluation Criteria in Solid Tumors (RECIST). The treatment response was assessed based on PCWG2. The appearance of two or more new bone lesions or soft tissue lesions was assessed using RECIST 1.1. Specifically, the new bone or soft tissue lesions were measurable lesions in at least one dimension with a minimum size of 10 mm by CT scan, 10 mm caliper measurement by clinical exam, or 20 mm by chest X-ray. At the completion of the average 6 year follow-up, 73 patients with CRPC information formed a mCRPC PCa Cohort while 134 patients without CRPC information were excluded.

[0310] The gene expression levels of the 18-Gene Panel in the urine samples were measured as described in the previous examples. A CRPC score was calculated by combining C& values of all genes in the classifier with mCRPC Prediction Algorithm to discriminate CRPC and non-CRPC. Then compare the CRPC score with a predetermined CRPC score cutoff value to make a prediction.

[0311] Univariate and multivariate Cox proportional hazards regression analyses and Kaplan-Meier survival plot of metastatic castration-resistance-free survival for the 18-Gene Panel, Gleason score and PSA were conducted using SPSS (IBM, Armonk, New York).

Results

[0312] After cancer diagnosis, the patients received different treatments such as surgery, radiation and hormone therapies. 14 metastatic cancer patients developed castration-resistant prostate cancer (mCRPC) after the treatment within the average 6 year follow-up. The 18-Gene Panel comprising of PTEN, CDK1, TMPRSS2, HIF1A, FGFR1, BIRC5, CRISPS, FN1, HPN, PSCA, PMP22, LMTK2, EZH2, GSTP1, PCA3, VEGFA, ANXA3 and KLK3 was evaluated for predicting CRPC. Cox regression analysis was performed to show high mCRPC prognostic power with HR of 36.07 (95% CI 4.71-275.90) in univariate (p<0.001) and 36.89 (95% CI 4.80-283.41) in multivariate analysis (p<0.001) (Table 22). Neither Gleason score nor PSA showed any prognostic ability for mCRPC.

Table 22

Univariate Multivariate

Variable HR (95% CI) P Value HR (95% CI) P Value

18-Gene Panel 36.07 (4.71-275.90) <0.001 36.89 (4.80-283.41) <0.001

Gleason Score 0.66 (0.22-1.98) 0.457 0.55 (0.17-1.77) 0.318

PSA 0.51 (0.11-2.27) 0.373 0.73 (0.16-3.39) 0.687

[0313] Kaplan-Meier plotting of mCRPC-free survival revealed large and statistically significant difference between the 18-Gene Panel risk groups (log rank p=0.000) (Figure 21).

Figure 21

[0314] The mCRPC predictive performance of the 18-Gene Panel by logistic regression analysis showed high accuracy with sensitivity of 92.86% (95% CI 79.37-106.35%), specificity of 100%, and AUC of 0.969 (95% CI 0.904-1.034) (p=0.000) (Table 23, Figure 22A). Gleason score and PSA could not detect mCRPC with 0% sensitivity and low AUC (Table 23, Figure 22B-C). In addition, combining the 18-Gene Panel with Gleason score and PSA increased the predictive accuracy slightly with increased AUC of 0.987 (Table 23, Figure 22D). The result showed that the 18-Gene Panel had very high accuracy for prediction of metastatic castration-resistant prostate cancer.

Table 23

Univariate Multivariate

P Sensitivity Specificity AUC P Sensitivity Specificity AUC

Value (95% CI) (95% CI) (95% CI) Value (95% CI) (95% CI) (95% CI)

18-Gene 0.000 92.86% 100% (100- 0.969 0.000

Panel (79.37- 100%) (0.904-

106.35%) 1.034)

Gleason Score 0.785 0% (0-0%) 100% (100- 0.689 0.710

100%) (0.523-

0.855)

PSA 0.262 % (0-0%) 100% (100- 0.605 0.443

100%) (0.433-

0.777)

Combination - - - - <0.000 92.86% 100% (100- 0.987

1 (79.37- 100%) (0.945-

106.35%) 1.029)

Figure 22

EXAMPLE 24

[0315] The prognostic performance of an 18-Gene Panel comprising of PTEN, CDK1, TMPRSS2, HIF1A, FGFR1, BIRC5, CRISP 3, FNI, HPN, PSCA, PMP22, LMTK2, EZH2, GSTPI, PCA3, VEGFA, ANXA3 and KLK3 for prediction of metastatic castration-resistant prostate cancer was assessed using urine samples collected without digital rectal examination in a prospective mCRPC MET Cohort.

Patients and Methods

[0316] In the prospective study as described in the previous example, among the metastatic patients, 39 had CRPC information and 14 metastatic cancer patients developed CRPC (mCRPC) after the treatment within the average 6 year follow-up. These patients formed a mCRPC MET Cohort. The procedures to assess the performance of the 18-Gene Panel for prediction of metastatic castration -resistant prostate cancer were as described in the previous example.

Results

[0317] The 18-Gene Panel comprising of PTEN, CDK1, TMPRSS2, HIF1A, FGFR1, BIRC5, CRISP 3, FNI, HPN, PSCA, PMP22, LMTK2, EZH2, GSTPI, PCA3, VEGFA, ANXA3 and KLK3 was evaluated for predicting mCRPC in the metastatic patients in the mCRPC MET Cohort (Table 24). The same 18 gene expression quantities and mCRPC Prediction Algorithm were used to classify mCRPC in the cohort. HR in Cox regression analysis was 25.51 (95% CI 3.33-195.36) (p=0.002) and 26.12 (95% CI 3.40-200.57) (p=0.002) in univariate and multivariate analysis respectively (Table 24).

Table 24

Univariate Multivariate

Variable HR (95% CI) P Value HR (95% CI) P Value

18-Gene Panel 25.51 (3.33-195.36) 0.002 26.12 (3.40-200.57) 0.002

Gleason Score 0.63 (0.21-1.89) 0.406 0.58 (0.19-1.82) 0.351

PSA 1.03 (0.23-4.65) 0.967 1.41 (0.30-6.75) 0.666

[0318] Kaplan-Meier plot of mCRPC-free survival showed large and statistically significant difference between the two 18-Gene Panel risk groups (log rank p=0.000) (Figure 23).

Figure 23

[0319] High mCRPC prediction accuracy of the 18-Gene Panel was shown by logistic regression analysis with sensitivity of 92.86% (95% CI 79.37-106.35%), specificity of 92.00% (95% CI 81.37-102.63%), and AUC of 0.986 (95% CI 0.941-1.031) (p=0.001) (Table 25, Figure 24A). Gleason score and PSA had no predictive power for mCRPC (Table 25, Figure 24B-C). When they were combined, the predictive power was decreased (Table 25, Figure 24D). The result showed that the the 18-Gene Panel had very high accuracy for prediction of metastatic castration-resistant prostate cancer in metastatic cancer patients.

Table 25

Univariate Multivariate

P Value Sensitivity Specificity AUC P Value Sensitivity Specificity AUC

(95% CI) (95% CI) (95% (95% CI) (95% CI) (95%

CI) CI)

18-Gene 0.001 92.86% 92.00% 0.986 0.006

Panel (79.37- (81.37- (0.941-

106.35%) 102.63%) 1.031) Gleason 0.980 % (0-0%) 100% 0.683 0.801

Score (100- (0.501-

100%) 0.865)

PSA 0.701 % (0-0%) 100% 0.554 0.781

(100- (0.361-

100%) 0.747)

Combination - - - - <0.0001 92.86% 100% 0.970

EXAMPLE 25

[0320] The prognostic performance of a 23-Gene Panel comprising of PTEN, PIP5K1A, CDK1, TMPRSS2, ANXA3, HIF1A, FGFR1, BIRC5, AMACR, CRISP3, PMP22, GOLPH2, EZH2, GSTP1, PCA3, VEGFA, CST3, CCNA1, CCND1, FN1, MY06, KLK3 and PSCA was tested for prediction of prostate cancer biochemical recurrence after surgery using prostate tissue specimens.

Patients and Methods [0321] A dataset of prostate tissue cohort MSKCC Prostate Oncogenome Project was obtained from cBioPortal (www.cbioportal.com) database and used in the study. The dataset contains transcriptome profiles of 218 prostate cancer tissue specimens ( 181 primaries and 37 metastases). The quantitative mRNA expression Z-Scores of genes in the 23-Gene Panel were obtained from the dataset along with clinicopathological information including biochemical recurrence (BCR) after radical prostatectomy (defined as consecutive PSA rise above 0.2 ng/mL twice according to NCCN guidelines), and Gleason scores. Patients without Z-Score of the genes in the panel or without recurrence information were excluded from the cohort, resulting in a cohort of 140 patients including 36 patients with recurrence.

[0322] A recurrence score was calculated by combining Z-Score values of all genes in the panel with a predefined algorithm to discriminate recurrent and non-recurrent cancer. Then compare the recurrence score with a predetermined recurrent cancer score cutoff value to make a prediction. The prediction of all specimens with the gene panel was then compared to the recurrence information collected during followup and the prognostic performance was measured by discriminant analysis using XLSTAT.

Results

[0323] The result showed that the 23-Gene Panel comprising of PTEN, PIP5K1A, CDK1, TMPRSS2, ANXA3, HIF1A, FGFR1, BIRC5, AMACR, CRISPS, PMP22, GOLPH2, EZH2, GSTP1, PCA3, VEGFA, CST3, CCNA1, CCND1, FN1, MYO6, KLK3 and PSCA was able to distinguish biochemical recurrent prostate cancer from non-recurrent prostate cancer. As shown in Table 26, the 23-Gene Panel was able to distinguish biochemical recurrent from non-recurrent prostate cancer with high sensitivity of 86.1%, specificity of 100% (p<0.0001), positive predictive value (PPV) of 100%, negative predictive value (NPV) of 95.4%, and AUC of the ROC curve of 0.903 (Figure 25).

Table 26

Positive Negative Total

Recurrent Cancer 31 5 36

Non-Recurrent Cancer 0 104 104

Sensitivity 86.1%

Specificity 100%

PPV 100%

NPV 95.4%

PO.OOOl

Figure 25

EXAMPLE 26

[0324] The ability of a 23-Gene Panel comprising of PTEN, PIP5K1A, CDK1, TMPRSS2, ANXA3, HIF1A, FGFR1, BIRC5, AMACR, CRISP3, PMP22, GOLPH2, EZH2, GSTP1, PCA3, VEGFA, CST3, CCNA1, CCND1, FN1, MY06, KLK3 and PSCA, cancer stage and Gleason score to predict cancer recurrence-free survival was assessed by Cox regression analysis and Kaplan-Meier survival plot in patients using urine samples collected without digital rectal examination from a retrospective urine study.

Patients and Methods

[0325] A multi-center retrospective study was conducted with Institutional Review Board approval to test archived urine samples collected without prior digital rectal examination. In the 520 patients retrospective cohort, all prostate cancer patients who had radical prostatectomy or other treatments were assessed periodically for biochemical recurrence (BCR, defined as consecutive PSA rise above 0.2 ng/mL twice according to NCCN guidelines) during the follow-up period.

[0326] A recurrence score was calculated by combining CtS values of all genes in the panel with a predefined algorithm to discriminate recurrent cancer and non-recurrent cancer. Then compare the recurrence score with a predetermined recurrent cancer score cutoff value to make a prediction. Univariate and multivariate Cox regression analyses of BCR-free survival for the 23-Gene Panel as well as cancer stage and Gleason score were conducted using SPSS (IBM, Armonk, New York). Kaplan-Meier survival plot of BCR-free survival for the 23-Gene Panel as well as cancer stage and Gleason score were conducted using SPSS.

Results

[0327] The result showed in the univariate analysis that the 23-Gene Panel comprising of PTEN, PIP5K1A, CDK1, TMPRSS2, ANXA3, HIF1A, FGFR1, BIRC5, AMACR, CR1SP3, PMP22, GOLPH2, EZH2, GSTP1, PCA3, VEGFA, CST3, CCNA1, CCND1, FN1, MY06, KLK3 and PSCA had HR of 1730.9 (Table 27) with statistical significance in prediction of cancer recurrence (p=0.014). In the multivariate analysis, the 23- Gene Panel had HR of 1795.0 (95% CI 4.30-7.49E+5) (p=0.015). This suggests that the 23-Gene Panel had high predictive power for cancer recurrence. In contrast, cancer stage and Gleason score had much lower HR with no statistical significance in univariate and multivariate analyses.

[0328] Kaplan-Meier survival analysis for the two groups of no BCR (23-Gene Negative) and BCR (23- Gene Panel Positive) patients showed much longer BCR-free survival in the 23-Gene Negative patients (with 100% BCR-free survival at 120 months) than in the 23 -Gene Positive patients (with -60% BCR-free survival at 48 months) (log rank p=0.000). In contrast, the difference of BCR-free survival between patients with Gleason score <7 and patients with Gleason score >7 was small (log rank p=0. 137). The difference between cancer stage I/II and cancer stage III/IV groups was statistically significant (log rank p=0.013), yet much smaller than that of the 23-Gene Panel groups (Figure 26a-c). This showed that the 23-Gene Panel was more accurate at predicting BCR-free survival than cancer stage and Gleason score.

Table 27

Univariate Multivariate

Variable HR (95% CI) P-value HR (95% CI) P-value

Cancer Stage 22.19 (0.09-5.34E+3) 0.268 10.11 (0.05-2.21E+3) 0.400

Gleason Score 20.76 (0.00-2.17E+5) 0.521 106.03 (0.00-2.74E+7) 0.463

23-Gene Panel 1730.9 (4.52-6.63E+5) 0.014 1795.01 (4.30-7.49E+5) 0.015

EXAMPLE 27

[0329] The prognostic performance of a 23-Gene Panel comprising of PTEN, PIP5K1A, CDK1, TMPRSS2, ANXA3, HIF1A, FGFR1, BIRC5, AMACR, CRISP3, PMP22, GOLPH2, EZH2, GSTP1, PCA3, VEGFA, CST3, CCNA1, CCND1, FN1, MY06, KLK3 and PSCA for prediction of prostate cancer biochemical recurrence after surgery was tested by discriminant analysis using patient urine samples collected without digital rectal examination from a retrospective urine study.

Patients and Methods

[0330] In the 520 patient urine study cohort, the predictive performance of the 23-Gene Panel was assessed by discriminant analysis using XLSTAT.

Results

[0331] The result showed that the 23-Gene Panel comprising of PTEN, PIP5K1A, CDK1, TMPRSS2, ANXA3, HIF1A, FGFR1, BIRC5, AMACR, CRISP3, PMP22, GOLPH2, EZH2, GSTP1, PCA3, VEGFA, CST3, CCNA1, CCND1, FN1, MY06, KLK3 and PSCA was able to accurately predict cancer recurrence after surgery. As shown in Table 28, the 23-Gene Panel was able to distinguish biochemical recurrent from non-recurrent prostate cancer with high sensitivity of 100%, specificity of 86.3% (p<0.0001), positive predictive value (PPV) of 45.2%, negative predictive value (NPV) of 100%, and AUC of the ROC curve of 0.929 (Figure 27).

Table 28

Positive

Total

Recurrent Cancer 46 0 46

Non-Recurrent Cancer 63 411 474

Sensitivity 100% Specificity 86.3%

PPV 45.2%

NPV 100%

PO.OOOl

Figure 27

EXAMPLE 28

[0332] The prognostic performance of a 18-Gene Panel comprising of PTEN, CDK1, TMPRSS2, HIF1A, FGFR1, BIRC5, CRISP 3, FNI, HPN, PSCA, PMP22, LMTK2, EZH2, GSTPI, PCA3, VEGFA, ANXA3 and KLK3 for prediction of prostate cancer biochemical recurrence after surgery was tested using urine samples collected without digital rectal examination from a retrospective urine study.

Patients and Methods

[0333] In the 520 patient urine study cohort, the predictive performance of the 18-Gene Panel was assessed by discriminant analysis using XLSTAT.

Results

[0334] The result showed that the 18-Gene Panel comprising of PTEN, CDK1, TMPRSS2, HIF1A, FGFR1, BIRC5, CRISP 3, FNI, HPN, PSCA, PMP22, LMTK2, EZH2, GSTPI, PCA3, VEGFA, ANXA3 and KLK3 was able to predict cancer recurrence after surgery. As shown in Table 29, the 18-Gene Panel was able to distinguish biochemical recurrent from non-recurrent prostate cancer with sensitivity of 89. 1%, specificity of 82.5% (p<0.0001), positive predictive value (PPV) of 33.1%, negative predictive value (NPV) of 98.7%, and AUC of the ROC curve of 0.925 (Figure 28).

Table 29

Positive

Total

Recurrent Cancer 41 5 46

Non-Recurrent Cancer S3 391 474

Sensitivity 89.1% Specificity 82.5%

PPV 33.1%

NPV 98.7%

PO.OOOl

Figure 28

EXAMPLE 28

[0335] The prognostic performance of a 24-Gene Panel comprising of PIP5K1A, CCND1, GSTP1, CST3, CCNA1, LMTK2, MY06, HPN, CDK1, PSCA, PTEN, G0LM1, PMP22, EZH2, FGFR1, FN1, VEGFA, TMPRSS2, ANXA3, CRISP 3, BIRC5, AMACR, HIF1A and KLK3 for prediction of cancer remission after treatment was assessed in prostate tissue specimens from a TCGA cohort.

Patients and Methods

[0336] For a 228 patient prostate tissue specimen cohort, mRNA expression Z-Scores of the genes in the 24-Gene Panel were obtained from the TCGA dataset at www.cbioportal.com. Prostate cancer remission was defined as complete response according to the RECIST 1.1 guidelines. The mRNA expression Z- Scores of the 24 genes in the panel were used to generate a remission score by combining Z-Scores of all genes in the panel with a predefined algorithm to discriminate cancer remission and non-remission. Then compare the remission score with a predetermined remission score cutoff value to make a prediction. The prediction of all samples with the panel was then compared to the remission data of the samples obtained from the database to measure predictive performance using univariate and multivariate logistic regression analyses. The P value was obtained from statistical comparative test Mann-Whitney Test using a statistical software program (XLSTAT).

Results

[0337] The 24-Gene Panel comprising of PIP5K1A, CCND1, GSTP1, CST3, CCNA1, IMTK2, MY06, HPN, CDK1, PSCA, PTEN, G0IM1, PMP22, EZH2, FGFR1, FN1, VEGFA, TMPRSS2, ANXA3, CRISP3, BIRC5, AMACR, HIF1A and KLK3 was evaluated for predicting cancer remission after treatment in prostate tissue specimens. As shown in Table 30, the 24-Gene Panel was able to predict cancer remission with sensitivity of 93.21%, specificity of 80.30%, and AUC of 0.987 (p<0.0001) (Figure 29A). Gleason score and cancer stage had lower predictive accuracy than the 24-Gene Panel. When they were combined, the predictive accuracy was lowered (Table 30, Figure 29B-D). The result showed that the the 24-Gene Panel had high accuracy for prediction of cancer remission after treatment.

Table 30

Univariate Multivariate

P Value Sensitivity Specificity AUC P Value Sensitivity Specificity AUC

EXAMPLE 29

[0338] The prognostic performance of a 24-Gene Panel comprising of PIP5K1A, CCND1, GSTP1, CST3, CCNA1, LMTK2, MY06, HPN, CDK1, PSCA, PTEN, G0LM1, PMP22, EZH2, FGFR1, FN1, VEGFA, TMPRSS2, ANXA3, CRISP 3, BIRC5, AMACR, HIF1A and KLK3 for prediction of patient survival was assessed in prostate tissue specimens in a TCGA Cohort.

Patients and Methods

[0339] For the 498 prostate tissue specimen cohort, mRNA expression Z-Scores of the genes in the 24- Gene Panel were obtained from the TCGA Cohort at www.cbioportal.com. The mRNA expression Z- Scores of the 24 genes in the panel were used to generate a survival score by combining Z-Scores of all genes in the panel with a predefined algorithm to discriminate patient survival and death. Then compare the survival score with a predetermined survival score cutoff value to make a prediction.

[0340] Univariate and multivariate Cox proportional hazards regression analyses and Kaplan-Meier survival plot of survival for the 24-Gene Panel, Gleason score and pre-operative PSA were conducted using SPSS (IBM, Armonk, New York). The prediction of all samples with the panel was then compared to the survival data of the samples obtained from the database to measure the predictive performance by univariate and multivariate logistic regression analyses using a statistical software program (XLSTAT).

Results

[0342] The 24-Gene Panel comprising of PIP 5K1 A, CCND1, GSTPI, CST3, CCNA1, IMTK2, MY06, HPN, CDKI, PSCA, PTEN, G0IM1, PMP22, EZH2, FGFR1, FNI, VEGFA, TMPRSS2, ANXA3, CRISP3, BIRC5, AMACR, HIF1A and KLK3 was evaluated for predicting patient survival in prostate tissue specimens in the TCGA Cohort. Cox regression analysis was performed to show high survival prognostic power with HR of 18.93 (95% CI 5.06-70.91) in univariate (p=0.000) and 11.64 (95% CI 2.97-45.55) in multivariate analysis (p=0.000) (Table 31). In contrast, Gleason score and cancer stage showed low prognostic ability for survival with very low HR (Table 31).

Table 31

Univariate Multivariate

Variable HR (95% CI) P-value HR (95% CI) P-value

[0343] Kaplan-Meier plot of survival for the 24-Gene Panel showed large and statistically significant difference between the two 24-Gene Panel survival groups (log rank p=0.000) (Figure 30).

Figure 30 24-Gene Classifier Overall Survival

Time (Months)

Numbers at Risk

[0344] The predictive accuracy of the 24-Gene Panel for patient survival was tested and the result showed high sensitivity of 95.70%, specificity of 100%, and AUC of 0.894 (p<0.0001) (Table 32, Figure 31A). Gleason score and cancer stage had lower predictive accuracy than the 24-Gene Panel (Table 32, Figure 31B-C). The result showed that the the 24-Gene Panel had high accuracy for prediction of patient survival using prostate tissue specimens in the TCGA Cohort.

Table 32

Univariate Multivariate

P Value Sensitivity Specificity AUC P Value Sensitivity Specificity AUC

Figure 31

EXAMPLE 30

[0345] The prognostic performance of a 25-Gene Panel comprising of HIF1A, FGFR1, BIRC5, AMACR, CRISP3, FN1, HPN, MY06, PSCA, PMP22, G0LM1, LMTK2, EZH2, GSTP1, PCA3, VEGFA, CST3, PTEN, PIP5K1A, CDK1, TMPRSS2, ANXA3, CCNA1, CCND1 and KLK3 for prediction of prostate cancer patient survival time was tested using prostate tissue specimens.

Patients and Methods

[0346] For prostate tissue specimen cohort (n=140), mRNA expression Z-Scores of the genes in the panels were downloaded from the MSKCC dataset at www.cbioportal.com. The mRNA expression Z-Scores of the 25 genes in the panel were used to generate a 5 -year survival score by combining Z-Scores of all genes in the panel with a predefined algorithm to discriminate patients with survival time >5 years and patients with survival time <5 years. Then compare the 5 -year survival score with a predetermined 5 -year survival score cutoff value to make a prediction of >5 years survival or <5 year survival. The survival prediction of all samples with the gene panel was then compared to the survival time of the patients and the predictive performance was tested by discriminant analysis using XLSTAT.

Results

[0347] The result showed that the 25-Gene Panel comprising of Hl Fl A, FGFR1, B1RC5, AMACR, CRISP 3, FN1, HPN, MY06, PSCA, PMP22, G0LM1, LMTK2, EZH2, GSTP1, PCA3, VEGFA, CST3, PTEN, PIP5K1A, CDK1, TMPRSS2, ANXA3, CCNA1, CCND1 and KLK3 was able to accurately predict cancer patient survival time. As shown in Table 33, the 25 -Gene Panel was able to distinguish survival time >5 years from survival time <5 years with sensitivity of 96.3%, specificity of 91.5% (p<0.0001), positive predictive value (PPV) of 94.0%, negative predictive value (NPV) of 94.7%, and AUC of the ROC curve of 0.991 (Figure 32).

Table 33

Positive Negative Total

>5 years Survival 54 5 59

<5 years Survival 3 78 81

Sensitivity 96.3%

Specificity 91.5%

PPV 94.0%

NPV 94.7%

PO.OOOl

Figure 32

REFERENCES

• Boyd LK, Mao X, Lu Y. The complexity of prostate cancer: genomic alterations and heterogeneity. Nature Reviews Urology 2012;9:652-64.

• Guo J, Yang J, Zhang X, et al. A Panel of Biomarkers for Diagnosis of Prostate Cancer Using Urine Samples. Anticancer Res. 2018;38(3): 1471-77.

• Guo J, Liu D, Zhang X, Johnson H, Feng X, Zhang H, et al. Establishing a Urine-Based Biomarker Assay for Prostate Cancer Risk Stratification. Front Cell Dev Biol. 2020;8:597961.

• Guo J, Zhang X, Johnson H, Feng X, Zhang H, Simoulis A, et al. A 23-Gene Classifier urine test for prostate cancer prognosis. Clin Transl Med. 2021; 1 l(3):e340.

• Guo J, Zhang X, Xia T, Johnson H, Feng X, Simoulis A, et al. Non-invasive Urine Test for Molecular Classification of Clinical Significance in Newly Diagnosed Prostate Cancer Patients. Front Med. 2021;721554. • Guo J, Johnson H, Zeng Q, Zhang X, Zhang H, Xia T, Feng X, et al. A Novel 18-Gene Classifier Urine Test for Prediction of Metastasis and Castration-resistance in Newly Diagnosed Prostate Cancer Patients. Manuscript in Preparation.

• Johnson H, Guo J, Zhang X, Zhang H, Simoulis A, Wu AHB, et al. Development and validation of a 25-Gene Panel urine test for for prostate cancer diagnosis and potential treatment follow-up. BMC Med. 2020;18(l):376.

• Kirby M, Hirst C, Crawford ED: Characterising the castration-resistant prostate cancer population: a systematic review. Int J Clin Pract 2011;65(11): 1180-1192.

• Osses DF, Roobol MJ, Schoots IG. Prediction Medicine: Biomarkers, Risk Calculators and Magnetic Resonance Imaging as Risk Stratification Tools in Prostate Cancer Diagnosis. Int J Mol Sci. 2019;20(7).

• Rycaj K, Tang DG: Molecular determinants of prostate cancer metastasis. Oncotarget. 2017;8(50):88211-88231.

• Shigeta K, Kosaka T, Hongo H et al: Castration-resistant prostate cancer patients who had poor response on first androgen deprivation therapy would obtain certain clinical benefit from early docetaxel administration. Int J Clin Oncol 2019;24(5):546-553.

• Smith RP, Malkowicz SB, Whittington R, VanArsdalen K, Tochner Z, Wein AJ. Identification of Clinically Significant Prostate Cancer by Prostate-Specific Antigen Screening. Arch Intern Med. 2004;164: 1227-1230.

• Thomsen FB, Jakobsen H, Langkilde NC, Borre M, Jakobsen EB, Frey A, et al. Active Surveillance for Localized Prostate Cancer: Nationwide Observational Study. J Urol. 2019;201(3):520-527.

• Truong M, Yang B, Jarrard DF. Towards the Detection of Prostate Cancer in Urine: A Critical Analysis. J Urol. 2013;189(2):422-29.

• Vickers A, Carlsson SV. Toward Responsible, Informed Decision Making for Prostate Cancer Treatment Decisions. J Clin Oncol. 2019:JC01900989.

• Xiao K, Guo J, Zhang X, et al. Use of two gene panels for prostate cancer diagnosis and patient risk stratification. Tumour Biol. 2016;37(8): 10115-22.

• Xiao K, Guo J, Zhang X, Feng X, Zhang H, Cheng Z, et al. Use of two gene panels for prostate cancer diagnosis and patient risk stratification. Tumour Biol. 2016;37(8): 10115-22.

Claims

CLAIMS What is claimed is:

1. A method for cancer screening, diagnosis, or determining the need for biopsy in a subject, comprising:

(a) providing a biological sample from a subject;

(b) measuring, in said biological sample, expression levels of a panel of genes comprising at least three or more genes selected from the group of genes consisting of PIP5K1A, CCND1, GSTP1, CST3, CCNA1, LMTK2, MYO6. HPN, CDK1, PSCA, PTEN, GOLM1, PMP22, EZH2, FGFR1, FN1, VEGFA, TMPRSS2, ANXA3, CRTSP3, BIRC5, AMACR, HIF1A, KLK3, PCA3, KLK2, MSMB, FLT1, MMP9, AR, TERT, PGC, SPINK1, STAT3, STAT5, and TFF3;

(c) determining a diagnostic score by (i) calculating the relative expression level of each gene as compared to one or more housekeeping genes, (ii) combining the calculated relative expression level of each gene with a predefined algorithm to determine a diagnostic score; and

(d) determining said subject as either (i) having cancer or needing biopsy if said diagnostic score is higher than a predetermined cancer diagnostic score cutoff value, or (ii) not having cancer or not needing biopsy if said diagnostic score is equal to or lower than a predetermined cancer diagnostic score cutoff value.

2. A method for determining if a subject has high risk, clinically significant cancer and needs immediate treatment or low risk, clinically insignificant cancer and needs active surveillance, comprising:

(a) providing a biological sample from a subject diagnosed as having cancer;

(b) measuring, in said biological sample, expression levels of a panel of genes comprising at least three or more genes selected from the group of genes consisting of PIP5K1A, CCND1, GSTP1, CST3, CCNA1, EMTK2, MY06, HPN, CDK1, PSCA, PTEN, GO1M1, PMP22, EZH2, FGFR1, FN1, VEGFA, TMPRSS2, ANXA3, CRISP3, BIRC5, AMACR, HIF1A, KLK3, PCA3, KLK2, MSMB, FLT1, MMP9, AR, TERT, PGC, SPINK1, STAT3, STAT5, and TFF3;

(c) determining a stratification score by (i) calculating the relative expression level of each gene as compared to one or more housekeeping genes, (ii) combining the calculated relative expression level of each gene with a predefined algorithm to determine a stratification score; and

(d) determining said subject as either (i) having high risk, clinically significant cancer and needs immediate treatment if said stratification score is higher than a predetermined high risk stratification score cutoff value, or (ii) having low risk, clinically insignificant cancer and needs active surveillance if said stratification score is equal to or lower than a predetermined high risk stratification score cutoff value.

3. The method of claim 2, wherein the method can be used for monitoring cancer progression during active surveillance to determine if said subject has cancer progression and needs immediate treatment or has no cancer progression and will continue active surveillance without treatment.

4. A method for predicting if a subject diagnosed as having cancer will have metastatic cancer in the future, comprising:

(a) providing a biological sample from a subject diagnosed as having cancer;

(b) measuring, in said biological sample, expression levels of a panel of genes comprising at least three or more genes selected from the group of genes consisting of PIP5K1A, CCND1, GSTP1, CST3, CCNA1, EMTK2, MY06, HPN, CDK1, PSCA, PTEN, G0IM1, PMP22, EZH2, FGFR1, FN1, VEGFA, TMPRSS2, ANXA3, CRISP3, BIRC5, AMACR, HIF1A, KLK3, PCA3, KLK2, MSMB, FLT1, MMP9, AR, TERT, PGC, SPINK1, STAT3, STAT5, and TFF3;

74 (c) determining a metastasis score by (i) calculating the relative expression level of each gene as compared to one or more housekeeping genes, (ii) combining the calculated relative expression level of each gene with a predefined algorithm to determine a metastasis score; and

(d) predicting said subject either (i) will have metastatic cancer in the future if said metastasis score is higher than a predetermined metastatic cancer score cutoff value, or (ii) will not have metastatic cancer in the future if said metastasis score is equal to or lower than a predetermined metastatic cancer score cutoff value.

5. The method of claim 4, wherein the method can be used for measuring metastatic cancer treatment efficacy during or after treatment by comparing a metastasis score obtained from a subject undergoing or after metastatic cancer treatment and determining said subject either (i) still has metastatic cancer if said metastasis score is higher than a predetermined metastatic cancer score cutoff value, or (ii) has no metastatic cancer if said metastasis score is equal to or lower than a predetermined metastatic cancer score cutoff value.

6. A method for predicting the development of treatment-resistant cancer in a subject, comprising:

(a) providing a biological sample from a subject before cancer treatment;

(b) measuring, in said biological sample, expression levels of a panel of genes comprising at least three or more genes selected from the group of genes consisting of PIP5K1A, CCND1, GSTP1, CST3, CCNA1, LMTK2, MY06, HPN, CDK1, PSCA, PTEN, GOLM1, PMP22, EZH2, FGFR1, FN1, VEGFA, TMPRSS2, ANXA3, CRISP3, BIRC5, AMACR, HIF1A, KLK3, PCA3, KLK2, MSMB, FLT1, MMP9, AR, TERT, PGC, SPINK1, STAT3, STAT5, and TFF3,-

(c) determining a treatment-resistance score by (i) calculating the relative expression level of each gene as compared to one or more housekeeping genes, (ii) combining the calculated relative expression level of each gene with a predefined algorithm to determine a treatment-resistance score; and

(d) predicting said subject either (i) will have treatment-resistance if said treatment-resistance score is higher than a predetermined treatment-resistance score cutoff value, or (ii) will not have treatmentresistance if said treatment-resistance score is equal to or lower than the predetermined treatment-resistance score cutoff value.

7. A method for predicting if a subject diagnosed as having cancer will have recurrent cancer after treatment in the future, comprising:

(a) providing a biological sample from a subject before cancer treatment;

(c) determining a recurrence score by (i) calculating the relative expression level of each gene as compared to one or more housekeeping genes, (ii) combining the calculated relative expression level of each gene with a predefined algorithm to determine a recurrence score; and

(d) predicting said subject either (i) will have cancer recurrence after treatment in the future if said recurrence score is higher than a predetermined recurrent cancer score cutoff value, or (ii) will not have cancer recurrence after treatment in the future if said recurrence score is equal to or lower than a predetermined recurrent cancer score cutoff value.

8. A method for determining treatment efficacy by detecting cancer/residual cancer during or after treatment or detecting cancer recurrence after treatment, comprising:

(a) providing a biological sample from a subject during or after cancer treatment;

75 (b) measuring, in said biological sample, expression levels of a panel of genes comprising at least three or more genes selected from the group of genes consisting of PIP5K1A, CCND1, GSTP1, CST3, CCNA1, LMTK2, MY06, HPN, CDK1, PSCA, PTEN, GOLM1, PMP22, EZH2, FGFR1, FN1, VEGFA, TMPRSS2, ANXA3, CRISP3, BIRC5, AMACR, HIF1A, KLK3, PCA3, KLK2, MSMB, FLT1, MMP9, AR, TERT, PGC, SPINK1, STAT3, STAT5, and TFF3;

(c) determining a cancer diagnostic score by (i) calculating the relative expression level of each gene as compared to one or more housekeeping genes, (ii) combining the calculated relative expression level of each gene with a predefined algorithm to determine a cancer diagnostic score; and

(d) determining said subject as either (i) having cancer/residual cancer during or after treatment or cancer recurrence after treatment if said cancer diagnostic score is higher than a predetermined cancer diagnostic score cutoff value, or (ii) having no cancer/residual cancer during or after treatment or no cancer recurrence after treatment if said cancer diagnostic score is equal to or lower than a predetermined cancer diagnostic score cutoff value.

9. A method for cancer monitoring to predict cancer remission after treatment in a subject, comprising:

(a) providing a biological sample from a subject before cancer treatment;

(c) determining a remission score by (i) calculating the relative expression level of each gene as compared to one or more housekeeping genes, (ii) combining the calculated relative expression level of each gene with a predefined algorithm to determine a remission score; and

(d) predicting said subject either (i) will have cancer remission after treatment if said remission score is higher than a predetermined remission score cutoff value, (ii) will have partial cancer remission after treatment if said remission score is equal to or lower than the predetermined remission score cutoff value but higher than a predetermined partial remission score cutoff value, or (iii) will not have cancer remission after treatment if said remission score is equal to or lower than the predetermined partial remission score cutoff value.

10. A method for predicting survival of a subject, comprising:

(a) providing a biological sample from a subject diagnosed as having cancer;

(b) measuring, in said biological sample, expression levels of a panel of genes comprising at least three or more genes selected from the group of genes consisting of PIP5K1A, CCND1, GSTP1, CST3, CCNA1, EMTK2, MY06, HPN, CDK1, PSCA, PTEN, G0EM1, PMP22, EZH2, FGFR1, FN1, VEGFA, TMPRSS2, ANXA3, CRISP3, BIRC5, AMACR, HIF1A, KLK3, PCA3, KLK2, MSMB, FLT1, MMP9, AR, TERT, PGC, SPINK1, STAT3, STAT5, and TFF3;

(c) determining a survival score by (i) calculating the relative expression level of each gene as compared to one or more housekeeping genes, (ii) combining the calculated relative expression level of each gene with a predefined algorithm to determine a survival score; and

(d) predicting said subject either (i) will survive cancer if said survival score is higher than a predetermined survival score cutoff value, or (ii) will not survive or die of cancer if said survival score is equal to or lower than the predetermined survival score cutoff value.

11. A method for predicting survival time of a subject, comprising:

(a) providing a biological sample from a subject diagnosed as having cancer;

(b) measuring, in said biological sample, expression levels of a panel of genes comprising at least three or more genes selected from the group of genes consisting of PIP5K1A, CCND1, GSTP1, CST3, CCNA1, LMTK2, MY06, HPN, CDK1, PSCA, PTEN, GOLM1, PMP22, EZH2, FGFR1, FN1, VEGFA, TMPRSS2,

76 ANXA3, CRISP3, BIRC5, AMACR, HIF1A, KLK3, PCA3, KLK2, MSMB, FLT1, MMP9, AR, TERT, PGC, SPINK1, STAT3, STAT5, and TFF3,-

(c) determining a 5 -year survival score by (i) calculating the relative expression level of each gene as compared to one or more housekeeping genes, (ii) combining the calculated relative expression level of each gene with a predefined algorithm to determine a 5-year survival score;

(d) determining a 10-year survival score by (i) calculating the relative expression level of each gene as compared to one or more housekeeping genes, (ii) combining the calculated relative expression level of each gene with a predefined algorithm to determine a 10-year survival score;

(e) determining a 20-year survival score by (i) calculating the relative expression level of each gene as compared to one or more housekeeping genes, (ii) combining the calculated relative expression level of each gene with a predefined algorithm to determine a 20-year survival score; and

(f) predicting said subject either (i) will have less than 5-year survival time if said 5-year survival score is equal to or lower than a predetermined 5 -year survival score cutoff value, or (ii) will have 5 - 10 year survival time if said 5-year survival score is higher than a predetermined 5-year survival score cutoff value but said 10-year survival score is equal to or lower than a predetermined 10-year survival score cutoff value, or (iii) will have 10-20 year survival time if said 10-year survival score is higher than a predetermined 10-year survival score cutoff value but said 20-year survival score is equal to or lower than a predetermined 20- year survival score cutoff value, or (iv) will have more than 20-year survival time if said 20-year survival score is higher than a predetermined 20-year survival score cutoff value.

12. The method of claims 1-11, wherein the said group of genes consists of PIP5K1A, CCND1, GSTPI, CST3, CCNA1, LMTK2, MY06, HPN, CDK1, PSCA, PTEN, G0IM1, PMP22, EZH2, FGFR1, FNI, VEGFA, TMPRSS2, ANXA3, CRISP 3, BIRC5, AMACR, HIF1A, KLK3 and PCA3.

13. The method of claim 1-11, wherein the said group of genes consists of CCND1, HIF1A, FGFR1, BIRC5, AMACR, CRISP3, FNI, HPN, MY06, PSCA, PMP22, G0IM1, IMTK2, EZH2, GSTP1, PCA3, VEGFA, CST3, PTEN, PIP5K1A, CDKI, TMPRSS2, ANXA3 and CCNA1.

14. The method of claims 1-11, wherein the said group of genes consists of PIP5K1A, CCND1, GSTPI, CST3, CCNA1, LMTK2, MY06, HPN, CDKI, PSCA, PTEN, G0IM1, PMP22, EZH2, FGFR1, FNI, VEGFA, TMPRSS2, ANXA3, CRISP 3, BIRC5, AMACR, HIF1A and KLK3.

15. The method of claims 1-11, wherein the said group of genes consists of PTEN, PIP5K1A, CDK1, TMPRSS2, ANXA3, HIF1A, FGFR1, BIRC5, AMACR, CRISP3, PMP22, G0IM1, EZH2, GSTPI, PCA3, VEGFA, CST3, CCNA1, CCND1, FNI, MY06, KLK3 and PSCA.

16. The method of claims 1-11, wherein the said group of genes consists of PTEN, CDK1, TMPRSS2, HIF1A, FGFR1, BIRC5, CRISP3, FNI, HPN, PSCA, PMP22, IMTK2, EZH2, GSTPI, PCA3, VEGFA, ANXA3 and KLK3.

17. The method of claims 1-16, wherein the expression levels of mRNA, DNA methylation, protein, peptide, or their combination of a panel of genes were measured in a said biological sample and an algorithm is used to make a diagnosis or prognosis by using the measured expression levels of the panel of genes.

18. The method of claim 17, wherein the said biological sample from a subject includes, but not limited to, blood, urine, ascites, other body fluids, tissue and cell from the subject.

19. The method of claim 17, wherein a kit is provided to measure mRNA, DNA methylation, protein or peptide levels of a panel of genes in the said sample.

77

20. The method of claim 17, wherein a system is provided to make data analysis and diagnosis or prognosis, comprising a computer program for:

(a) receiving expression data of a panel of genes being tested;

(b) determining an expression test score by (i) calculating the relative expression level of each gene in said panel as compared to one or more housekeeping genes, (ii) combining the calculated relative expression level of each gene with a predefined algorithm to determine a score;

(c) comparing the calculated expression test score to a predetermined diagnostic or prognostic score cutoff value to make a diagnosis or prognosis and display the result of diagnosis or prognosis.

78