METHODS FOR PREDICTING PROSTATE CANCER AND USES THEREOF

CROSS-REFERENCE TO RELATED APPLICATIONS

[0001] This application claims the benefit of U.S. Provisional Application No. 62/978,184, filed February 18, 2020, and U.S. Provisional Application No. 62/819,325, filed March 15, 2019, which are incorporated by reference herein in their entirety.

FIELD OF THE INVENTION

[0002] The present invention relates to compositions and methods for diagnosing, prognosing, monitoring, and treating a patient with prostate cancer. In particular, the invention relates to the use of small non-coding RNAs (sncRNAs) such as miRNA and snoRNA as expression signatures for identifying a clinically significant prostate cancer.

BACKGROUND

[0003] The current method of screening for prostate cancer includes a digital rectal examination followed by a prostate-specific antigen (PSA) test. The former is invasive and the latter requires drawing of a blood sample from the subject.

[0004] Patients with a suspicious DRE and/or an elevated PSA level are subjected to a systematic 12-needle core biopsy or Magnetic Resonance Imaging (MRI)-guided targeted needle biopsy. This standard diagnosis strategy is invasive, imprecise and associated with significant costly morbidities, most notably bacterial infections.

[0005] The PSA test has significant drawbacks. In addition to indicating prostate cancer, elevated PSA levels may also indicate urinary tract infection or prostatitis (an inflammation or the prostate or benign prostatic hyperplasia or BPH). The test overdiagnoses prostate cancer, and many men are unnecessarily subjected to core needle biopsies. The prostate tissue collected during the biopsy is then examined by a pathologist and assigned a Gleason score that assesses the grade of the disease. The Gleason score is the sum of two numbers: (1) a primary grade assigned by the pathologist based on the pathologist’s determination of the grade of the tumor in the most common pathology (2) a secondary grade based on the determination of the grade of the tumor in the next most prominent pathology. For each area, a score of one to five is assigned based on how aggressive the tumor appears and the two numbers are added togther to provide the final Gleason Score. A tumor with cells that appear close to normal is assigned a low Gleason score (six or below, reported as Gleason 3+3) whereas a tumor with cells that appear clearly different from those of a normal prostate is assigned a higher Gleason score (seven or above). Low-grade tumors based on low Gleason scores are less likely to be aggressive; whereas tumors with high Gleason score are more likely to be aggressive and metastasize. There are aspects of the Gleason Scoring System that have been problematic since they were implemented - most noteably the fact that tumors that are Gleason 3+4 and Gleason 4+3 are both reported as Gleason 7, even though the clinical outcomes of these groups are clearly different. A recent refinement of the Gleason Score, referred to as Grade Grouping, has been adopted to eliminate this issue (Grade Group 1 (GG1) encompasses Gleason 3+3; GG2 - Gleason 3+4; GG3 - Gleason 4+3; GG4 - Gleason 4+4 and GG5 - Gleason 5+4 or higher. This change to the scoring system has simplified the reporting of the histopathology of prostate cancer and has eliminated the ambiguity associated with“Gleason 7” tumors, making classifying outcome more straight forward.

[0006] Approximately 50-70% of patients recommended for core needle biopsy on the basis of “elevated” PSA (>3 ng/mL) have negative biopsies, while 14% of men with PSA < 3 ng/mL have prostate cancer but are not routinely biopsied because of their low PSA levels. The combination of PSA screening and core needle biopsy is both invasive and has poor performance characteristics, which leaves physicians and patients with no reliable measures on which to base their choices of treatment options. The result is that many men needlessly opt for clinical intervention, very often prostatectomy. It also hinders the development of new prognostic tools since the Gleason Score "gold standard" is not itself a reliable indicator of prostate tumor progression.

[0007] This problem has been recognized for at least 30 years. It remains a major issue today. The intervening years have seen many attempts to develop prognostic markers for aggressive disease, including ploidy, nuclear morphology and nuclear matrix architecture, microarray-based transcriptome analyses, DNA methylation status, and detection of gene fusions such as the TMPRSS2:ETS family fusions. None of these methodologies have proved to be significantly better than Gleason Scores as indicators of prostate tumor progression. Furthermore, they do not identify the cancer stage or grade adequately. [0008] Tests have been developed that are designed to distinguish cancer states using mRNA expression profiles. However, each demonstrates significant shortcomings. First, with only a few exceptions, these assays used tumor material derived from radical prostatectomy specimens, and therefore are at best predictive of early tumor recurrence. While potentially useful for making post- surgical clinical decisions related to continuing clinical decisions, they do not help distinguish prostate cancer grades prior to surgery. Secondly, a number of these genomic approaches have focused on specific pathways that have been implicated in prostate cancer progression, including the androgen receptor (AR) modulated gene expression, epithelial- stromal interactions, and cell cycle. These assays assume that all prostate tumors progress along a common pathway. Other commercially available biomarker assays use mRNA expression profiles generated by real-time PCR of a small subset of genes.

[0009] To date there has been very few genome wide transcriptome studies of sncRNAs in prostate cancer. One study compared the miRNA and snoRNA signatures in (i) freshly frozen radical prostatectomy samples and (ii) adjacent normal tissue from the same patient using Illumina/Solexa deep sequencing and microarray analysis on the Affymetrix miRNA ^® v.2 microarrays that contains 723 human miRNAs catalogued in Sanger miRBase v.10.1. (Wellcome Sanger Institute). This study provides a valuable data set for comparing the complement of sncRNAs expressed in prostate cancer and peritumoral benign tissue but is not useful for the rational design of a panel of sncRNAs that is prognostic and/or predictive for tumor progression prior to clinical intervention. It is also handicapped as a general screening technology, since the technique requires micro- dissected flash frozen material that is only available after surgery, and therefore cannot be used for diagnosis.

[00010] Accordingly, an improved method of predicting screening and classifying prostate cancer is needed. The present disclosure relates to a non-invasive (by eliminating or reducing the unnecessary core needle biopsy) method for screening the presence or absence of prostate cancer that is highly sensitive and specific.

[00011] The method also provides a platform for disease management that is useful for the diagnosis, classification, prognosis, and monitoring of the progression and treatment of the disease. The disclosed method is based on the interrogation of a large set of at least 200 small non-coding RNAs (sncRNAs) isolated from urinary exosomes in combination with the Sentinel™ PCa, Sentinel™ CS and Sentinel™ HG tests. The Sentinel™ PCa, Sentinel™ CS and Sentinel™ HG tests are based on algorithmic analyses and comparisons of snRNA sequences catalogued from a large target population having no evidence of prostate cancer (NEPC) or having prostate cancer (GG1-GG5) for the Sentinel™ PCa Test; having low grade cancer (GG1) versus intermediate and high grade cancer (GG2-GG5) for the Sentinel™ CS Test; and having low and favorable intermediate grade cancer (GG1+GG2) versus unfavorable intermediate and high grade (GG3- GG5) cancer for the Sentinel™ HG Test . The three Sentinel™ Tests, that can be performed on a single urine sample, are used to sequentially determine whether a patient has prostate cancer or not, and whether patients with prostate cancer have low- or favorable intermediate-grade disease that can be monitored on active surveillance protocols or high-grade disease that needs immediate treatment.

SUMMARY OF THE DISCLOSURE

[00012] In one aspect, the disclosure provides a method of screening a subject for prostate cancer comprising: (i) obtaining a biological sample from the subject, (ii) detecting the aggregate expression profile of a signature collection of small non-coding RNAs (sncRNAs) from the biological sample wherein the collection of sncRNAs comprises SEQ ID NOs: 1-280, (iii) correlating the aggregate expression profile of SEQ ID NOs: 1-280 from the subject by comparing the aggregate expression level of SEQ ID NOs: 1-280 in a training data set from a target population having no evidence of prostate cancer (NEPC) or having prostate cancer; and (iv) classifying the subject as NEPC or has prostate cancer based on the results in (iii). This procedure is embodied in the Sentinel™ PCa test.

[00013] In yet another aspect, the disclosure provides a method of determining whether a patient diagnosed as having cancer has low-grade (GG1) or intermediate or high grade disease (GG2- GG5) or not, comprising: (i) obtaining a biological sample from a subject, (ii) detecting the aggregate expression profile of a signature collection of small non-coding RNAs (sncRNAs) from the biological sample wherein the collection of sncRNAs comprises SEQ ID NOs: 281-560, (iii) correlating the aggregate expression profile of SEQ ID NOs: 281-560 from the subject by comparing the aggregate expression profile of SEQ ID NOs: 281-560 in a training data set from a target population known to have low risk, low grade (GG1) or intermediate and high grade, intermediate and high risk prostate cancer (GG2-GG5) and (iv) classifying the subject as GG1 or GG2-GG5 based on the results obtained from (iii). This procedure is embodied in the Sentinel™ CS test.

[00014] In yet another aspect, the disclosure provides a method of determining whether a patient diagnosed as having cancer has high-grade (GG3-GG5) or not (low or intermediate grade disease (GG1 + GG2), comprising: (i) obtaining a biological sample from a subject, (ii) detecting the aggregate expression profile of a signature collection of small non-coding RNAs (sncRNAs) from the biological sample wherein the collection of sncRNAs comprises SEQ ID NOs: 561-840, (iii) correlating the aggregate expression profile of SEQ ID NOs: 561-840 from the subject by comparing the aggregate expression profile of SEQ ID NOs: 561-840 in a training data set from a target population known to have high grade, high risk prostate cancer (GG3-GG5) or low- or intermediate-risk cancer (GG1+GG2) and (iv) classifying the subject as GG3-GG5 or GG1+GG2 based on the results obtained from (iii). This procedure is embodied in the Sentinel™ HG test.

[00015] In yet another aspect, the disclosure provides a method for treating a prostate cancer comprising: (i) obtaining a biological sample from a subject, (ii) detecting the aggregate expression profile of a signature collection of small non-coding RNAs (sncRNAs) from the biological sample wherein the collection of sncRNAs comprises SEQ ID NOs: 281-840, (iii) correlating the aggregate expression profile of SEQ ID NOs: 1-840 from the subject by comparing the aggregate expression profile of SEQ ID NOs: 281-840 in a training data set from a target populations having NEPC, GG1, GG2, GG3, GG4 or GG5 prostate cancer, (iv) classifying the subject as having low- intermediate-grade, prostate cancer (GG1-GG2) or high-grade prostate cancer (GG3-GG5) based on the results obtained from (iii), and (v) treating the subject classified as having high-risk prostate cancer by administering one or more chemotherapeutic agents, hormones, immuno therapeutic, radiation, cryotherapy, surgery or a combination thereof.

[00016] In a further aspect, the disclosure provides a method for determining the likelihood of survival, disease recurrence or response to treatment for a subject with prostate cancer comprising: (i) obtaining a biological sample from a patient, (ii) detecting the aggregate expression profile of a signature collection of small non-coding RNAs (sncRNAs) from the biological sample wherein the collection of sncRNAs comprises SEQ ID NOs: 1-840, (iii) comparing the aggregate expression profile of SEQ ID NOs: 1-840 after treatment with that prior to treatment, (iv) correlating the aggregate expression profile of SEQ ID Nos: 1-840 from the subject by comparing the aggregate expression profile of SEQ ID Nos: 1-840 in a training data set from a target populations having no evidence of prostate cancer (NEPC) or having prostate cancer and the aggregate expression profile of SEQ ID Nos: 1-840 in a training data set from a target populations having Grade Group 1, 2, 3 or 4-5; and (v) determining the likelihood of survival, disease recurrence or response to treatment in a subject treated for prostate cancer.

[00017] In one aspect, the disclosure provides a method for predicting future prostate cancer in a subject comprising: (i) obtaining a biological sample from a patient, (ii) detecting the aggregate expression profile of a signature collection of small non-coding RNAs (sncRNAs) from the biological sample wherein the collection of sncRNAs comprises SEQ ID NOs: 1-280, (iii) correlating the aggregate expression profile of SEQ ID Nos: 1-280 from the subject by comparing the aggregate expression profile of SEQ ID Nos: 1-280 in a training data set from a target population from a target populations having Grade Group 1, 2, 3 or 4-5; (iv) determining the likelihood of a subject at risk of having Grade Group 2 -5 prostate cancer based on the results obtained from (iii), and (iv) treating the subject predicted with a high risk of developing aggressive prostate cancer by administering one or more chemotherapeutic agents, hormones, immuno therapeutic, radiation, cryotherapy, surgery or a combination thereof.

[00018] In another aspect, the disclosure provides a system for determining whether a patient has no cancer or has cancer and classifying the subject with cancer as (i) indolent (low grade, GG1), (ii) intermediate or high grade (GG2-GG5), (iii) low/intermediate risk (GG1-GG2) or (iv) aggressive (high grade, GG3-GG5) prostate cancer comprising at least three processors configured to (a) interrogate sncRNA sequences for informative sequences, (b) determine and compare a Sentinel Score to determine if the subject has prostate cancer or no prostate cancer and to classify the prostate cancer stage group. BRIEF DESCRIPTION OF THE DRAWINGS

[00019] The patent or application file contains at least one drawing executed in color. Copies of this patent or patent application publication with color drawing(s) will be provided by the Office upon request and payment of the necessary fees.

[00020] Figure 1: Exosomes are small extracellular, membrane bound vesicles that are formed in the early endosomes and released from cells. Exosomes contain proteins, mRNAs and an array of sncRNAs [miRNAs and C/D box and H/ACA Box small nucleolar RNA (snoRNAs)] that reflect the biology of the cell.

[00021] Figure 2: Using a commercial kit, cell-free urine is collected from the patient. The exosomes in the urine are captured, and the RNA is extracted from the exosomes. The RNA yield is measured using the Qubit assay (ThermoFisher) and the sample quality is assessed using the Agilent 2100 bioanalyzer. The resultant sncRNA levels are interrogated using custom made OpenArray™ (Thermo Fisher) plates that are specifically designed for the miR Scientific Sentinel™ PCa, Sentinel™ CS or Sentinel™ HG Tests. Interrogate is a common term of art for the simultaneous analysis of a large number of sequences in a biological sample. The resultant readout for amplification curves for snoRNA and microRNA (collectively called sncRNAs), are then analyzed and used to diagnose the patient, and when cancer is present to classify the disease, and monitor treatment accordingly.

[00022] Figure 3 illustrates the complexity of using more than a single entity to establish an unbiased statistical approach to identify important interactions to identify individual and combinations of sncRNAs and to correlate the grade grouping or prostate cancer phenotype. (See, [00048] -[00051]).

[00023] Figure 4 shows a schematic for screening and diagnosis of a patient suspected of having prostate cancer.

[00024] The patient first provides a urine sample for a three-layered Sentinel™ Analysis. Total urinary exosomal RNA is extracted and interrogated for expression of sncRNAs (SEQ ID NOs: 1- 280) specific for the Sentinel™ PCa Test; sncRNAs (SEQ ID NOs: 281-560) specific for the Sentinel™ CS Test; and sncRNAs (SEQ ID NOs: 561-840) specific for the Sentinel™ HG Test. The expression signature will be used to classify patients into those with prostate cancer, and those without prostate cancer (Sentinel™ PCa Test, 1st layer). Patients with negative score will return every 12 months for monitoring.

[00025] In the second layer, patients with positive Sentinel™ PCa Score (those with prostate cancer) will be subjected to a secondary analysis that classifies them into having clinically insignificant (GG1) tumors or clinically significant tumors (GG2-GG5) using the Sentinel™ CS Test. Patients with clinically insignificant (GG1) tumors will be recommended for active surveillance (AS) and monitored continuously with quarterly Sentinel™ CS tests, to establish that the tumor has not progressed to GG2 or higher. Patients with clinically significant tumors (GG2- GG5) will be referred for immediate therapy.

[00026] For some patients, a third classification layer, the Sentinel™ HG Test, will further classify patients as having GG1-GG2 tumors or GG3-GG5 tumors. This test is designed to identify patients with GG3-GG5 cancers that need immediate intervention. Patients with GG1 or GG2 can be monitored by quarterly Sentinel™ HG Test to identify patients that progress to GG3 and therefore need therapeutic intervention. The availability of the Sentinel™ CS and Sentinel™ HG provides both patients and health care providers with the individualized information for treatment decision making.

[00027] Figure 5 shows the output of the Discovery PCa Experiments. The PCa Discovery Studies use very carefully defined patient cohort [NEPC =89; Cancer (GG1-GG5) =146] with well characterized histopathology to identify the most informative sequences among the 6,599 sncRNAs interrogated on the miR4.0 microarrays.

[00028] Figure 5 (left panel): Scatter plot of no cancer (NEPC) and cancer (GG1-GG5) status in the training data set. Positive Discovery PCa Score is indicative of having prostate cancer and a negative Discovery PCa Score is indicative of no cancer. The cancer status as determined by histopathology of core biopsies is shown in blue (no cancer) and red (cancer) circles.

[00029] Figure 5 (right panel): Identification of informative sncRNAs for the Sentinel™ PCa Test Identification of the most informative sncRNA entities (top 35 are shown, each circle represents a single entity) for discriminating between no cancer and cancer status using the proprietary Selection Algorithm. The resultant Sentinel™ PCa Test interrogates 280 sncRNA including the 145 most informative sncRNA sequences, which comprises of 60 snoRNA and 85 miRNA entities as shown in the bar graph. Green: miRNA entities; Yellow: snoRNA entities.

[00030] Figure 6 (left panel) shows the output of the Discovery CS Experiments. The CS Discovery Studies use very carefully defined patient cohorts [GG1=90; GG2-GG5=56] with well characterized histopathology to identify the most informative sequences among the 6,599 sncRNAs interrogated on the miR4.0 microarrays. Positive Discovery CS Score is indicative of having GG2-GG5 cancer (yellow circles) and a negative Discovery CS Score is indicative of having GG1 cancer (green circles)

[00031] Figure 6 (right panel): Identification of informative sncRNAs for the Sentinel™ CS Test Identification of the most informative sncRNA entities (top 35 are shown, each circle represents a single entity) for discriminating between GG1 and GG2-GG5 cancer status using the proprietary Selection Algorithm. The resultant Sentinel™ CS Test interrogates 280 sncRNA including the 145 most informative sncRNA sequences, which comprises of 66 snoRNA and 130 miRNA entities as shown in the bar graph. Green: miRNA entities; Yellow: snoRNA entities.

[00032] Figure 7 (left panel) shows the output of the Discovery HG Experiments. The HG Discovery Studies use very carefully defined patient cohorts [GG1+GG2=181; GG3-GG5=55] with well characterized histopathology to identify the most informative sequences among the 6,599 sncRNAs interrogated on the miR4.0 microarrays. Positive Discovery HG Score is indicative of having GG3-GG5 cancer (purple circles) and a negative Discovery HG Score is indicative of having GG1+GG2 cancer (brown circles).

[00033] Figure 7 (right panel): Identification of informative sncRNAs for the Sentinel™ HG Test. Identification of the most informative sncRNA entities (top 35 are shown, each circle represents a single entity) for discriminating between GG1 and GG2-GG5 cancer status using the proprietary Selection Algorithm. The resultant Sentinel™ CS Test interrogates 280 sncRNA including the 196 most informative sncRNA sequences, which comprises of 66 snoRNA and 130 miRNA entities as shown in the bar graph. Green: miRNA entities; Yellow: snoRNA entities. [00034] Figures 8A-8C show the clinical validation of high throughput OpenArray™ interrogation of urinary exosomal sncRNA using the Sentinel™ PCa Test. The data from a case-control study of 1436 men (836 subjects in the training group used to cross-validate the interrogation of sncRNAs identified in the Discovery PCa phase and 600 independent subjects used in the validation study) are shown.

[00035] Figure 8A: Scatter plot of cancer status in the validation group set that examines 600 patients (300 no cancer; 300 cancer). Classification of no cancer (black circle) and cancer (green circle) patients where a positive Sentinel™ PCa Score is indicative of having prostate cancer and a negative Sentinel™ PCa Score is indicative of no cancer.

[00036] Figure 8B: Sorted plot of cancer status in the validation group set that examines 600 patients. Classification of no cancer (black circle) and cancer (green circle) patients where a positive Sentinel™ PCa Score is indicative of having prostate cancer and a negative Sentinel™ PCa Score is indicative of no cancer.

[00037] Figure 8C: Receiver Operator Curve (ROC) for Sentinel™ PCa Test. The ROC curve for the analysis of 600 patients in testing group shown in Figures 8A and 8B was calculated by successively calculating (1 -specificity) for different user-defined false negative rates. Performance characteristics reported in Table 6 (see, [000112]) were from the user defined false negative rate of 0.05 (shown in red).

[00038] Figures 9A-9C show the clinical validation of high throughput OpenArray™ interrogation of urinary exosomal sncRNA using the Sentinel™ CS Test. The data from a case-control study of 1436 men (836 subjects in the training group used to cross-validate the interrogation of sncRNAs identified in the Discovery CS phase and 600 independent subjects used in the validation study) are shown.

[00039] Figure 9A: Scatter plot of cancer status in the validation group set that examines 300 prostate cancer patients (146 GGl-low grade and 154 GG2-GG5 intermediate and high grade). Classification of low grade (teal circle) and intermediate and high-grade cancer (orange circle) patients where a positive Sentinel™ CS Score is indicative of having high-grade prostate cancer and a negative Sentinel™ CS Score is indicative of low-grade cancer. [00040] Figure 9B: Sorted plot of cancer status in the validation group set as shown in Figure 9A that examines 300 prostate cancer patients. Classification of low grade (teal circle) and high-grade cancer (orange circle) patients where a positive Sentinel™ CS Score is indicative of having high- grade prostate cancer and a negative Sentinel™ CS Score is indicative of low-grade cancer.

[00041] Figure 9C: Receiver Operator Curve (ROC) for Sentinel™ CS Test. The ROC curve for the analysis of 300 prostate cancer patients, as shown in Figures 9A and 9B was calculated by successively calculating (1 -specificity) for different user-defined false negative rates. Performance characteristics reported in Table 6 (see, [000112]) were from the user defined false negative rate of 0.05 (shown in red).

[00042] Figures 10 A- IOC show the clinical validation of high throughput OpenArray™ interrogation of urinary exosomal sncRNA using the Sentinel™ HG Test. The data from a case- control study of 1436 men (836 subjects in the training group used to cross-validate the interrogation of the same sncRNAs identified in the Discovery HG phase and 600 independent subjects used in the validation study) are shown.

[00043] Figure 10A: Scatter plot of cancer status in the validation group set that examines 300 prostate cancer patients (200 GG1+GG2 low grade and 100 GG3-GG5 intermediate and high grade). Classification of low grade (teal circle) and intermediate and high-grade cancer (orange circle) patients where a positive Sentinel™ CS Score is indicative of having high-grade prostate cancer and a negative Sentinel™ CS Score is indicative of low-grade cancer.

[00044] Figure 10B: Sorted plot of cancer status in the validation group set as shown in Figure 10A that examines 300 prostate cancer patients. Classification of low grade (blue circle) and high- grade cancer (red circle) patients where a positive Sentinel™ HG Score is indicative of having high grade prostate cancer and a negative Sentinel™ HG Score is indicative of low-grade cancer.

[00045] Figure IOC: Receiver Operator Curve (ROC) for Sentinel™ HG Test. The ROC curve for the analysis of 300 prostate cancer patients, as shown in Figures 10A and 10B was calculated by successively calculating (1 -specificity) for different user-defined false negative rates. Performance characteristics reported in Table 6 (see, [000112]) were from the user defined false negative rate of 0.05 (shown in red). DETAILED DESCRIPTION

[00046] The present subject matter may be understood more readily by reference to the following detailed description that forms a part of this disclosure. It is to be understood that this invention is not limited to the specific products, methods, conditions or parameters described and/or shown herein, and that the terminology used is for the purpose of describing particular aspect and embodiments by way of example only and is not intended to be limiting of the claimed invention.

[00047] The disclosure relates to a method for screening, diagnosing and treating prostate cancer in a subject. The method provides robust tests for (1) classifying a male patient with unknown prostate cancer status and (2) accurately distinguishing between prostate cancer grades in biological samples from patients. The method is based on the detection and correlation of the aggregate expression profiles of a collection of sncRNAs from the patient’s biological sample to determine if a patient has prostate cancer or not using Sentinel™ PCa Test. For the patient identified as having prostate cancer, the exosomal sncRNA is further interrogated using the Sentinel™ Clinical Significant (CS) Test to distinguish patients with clinically significant or aggressive (GG2-GG5) from those with clinically insignificant or indolent (GG1) prostate cancer, and the Sentinel™ High Grade (HG) Test to identify patients with high grade, high risk (GG3- GG5) prostate cancer.

[00048] The disclosed method is based on an unbiased statistical approach developed to identify important interactions of individual sequences and combinations of sequences that correlate best to a phenotype of interest. The approach is based on (i) the modulating effects of miRNA on mRNA and (ii) the influence of snoRNAs on mRNA translatability through post-transcriptional modification of ribosomal RNAs, tRNAs and other nuclear RNAs, which lead to new protein products that alters protein function and phenotype.

[00049] The disclosed computational/statistical approach analyses urinary exosomal sncRNAs to provide a very granular analysis of the critical associations between sncRNAs that leads to the identification of Sentinel sequences that accurately predict the prostate cancer phenotype. This is illustrated in the accompanying Figure 3. For example, in the single entity analysis, the expression level of individual sncRNAs is correlated to the Grade Group of the prostate cancer (the phenotype). For each sncRNA entity there are two informative outcomes: either an increase in the expression level of the entity relative to the control pathology ( e.g ., no cancer) or a decreased expression level. No change of expression between the two phenotypes indicates that there is no association (1) with either phenotypes, and (2) the entity is not useful as a marker of either phenotype. Thus, when single entity is used in the analysis, there are only two

informative outcomes, leaving all of the possible sncRNA interactions unexplored.

[00050] In the two entities analysis, examining the association of the expression changes for all possible interactions between two entities results in 8 different informative outcomes and 1 non- informative outcome (when neither entity is differentially expressed in their phenotype) (see, Figure 3“Interrogation of 2 Entities). Thus, in the context of the Sentinel™ Tests, when the association of all possible combinations of two sncRNA entities with a specific grade group are compared, there are 8 different ways that will lead to a meaningful association between pairs of sncRNAs and grade groups. This provides a more detailed analysis that uncovers hidden associations between expression levels of sncRNAs and Grade Grouping.

[00051] Using the same approach of three or four or more sncRNA entities provides a very granular analysis of the association between sncRNA expression and phenotype (grade grouping), making it possible to assess a patient of unknown disease status and predict the individual disease status using the expression levels of urinary exosomal sncRNA selected by the algorithm.

Development of Sentinel™ PCa, and Sentinel™ HG Test Sentinel™ CS Platforms

[00052] The Sentinel™ PCa Test is a classification platform or algorithm based on the analysis of a collection of signature sncRNA ( i.e ., miRNAs and snoRNAs sequences) levels. The predictive value of each sequence is defined via a data-driven Selection Algorithm that is independent of the a priori determined biological role of the sequences in prostate biology. The Selection Algorithm is trained on a dataset consisting of: (1) control subjects who presented in urology for conditions unrelated to prostate cancer; (2) subjects with suspicion of prostate cancer known to not have prostate cancer based on the biopsy results; and (3) patients diagnosed with prostate cancer and whose core needle biopsy histopathology was reported as Grade Groups 1 through 5 (GG1-GG5). [00053] To establish robust datasets for the Sentinel™ Tests, exosomal sncRNAs obtained from the urinary exosomes of these training set of patients were interrogated using the Affymetrix miR 4.0 microarrays to define expression signatures. These studies using selected subjects with well characterized histopathology are referred to as the Discovery PCa, Discovery CS, and Discovery HG Tests. The patients included in the“no cancer” group were carefully selected from age-matched men who were seen at urology clinics for issues unrelated to urological oncology, and from men who had one or more 12-needle diagnostic core needle biopsy that showed no evidence of prostate cancer (NEPC). For patients in the“cancer” cohort, the pathological grade group classification of the core needle biopsies each tumor was thoroughly assessed. These carefully select groups of patients (no cancer and cancer group in different stages of cancer) form the training set in the development of the Discovery PCa, CS and HG Tests. The demographics of the 235 patients used for the Discovery experiments are shown in Table 4 (see, [000103] -[000104])

The Selection Algorithm

[00054] The most informative sncRNA sequences that discriminate between cancer and no cancer were identified using Selection algorithm which determines which sncRNA sequences are differentially represented between patients that do not have prostate cancer (NEPC) and those with prostate cancer (GG1-GG5). This is exemplified below for the Discovery PCa Test. The Selection algorithm tests how well the levels of the urinary exosomal sncRNAs correlate with pathological stage of the disease (cancer/no cancer) in a large population of participants [89 subjects with NEPC and 146 patients with cancer (GG1-GG5)] with carefully defined pathology. The Selection algorithm individually assesses how well each of the 6,599 sncRNAs interrogated on the miR4.0 arrays correlates the known pathology of the tumor. Since many sncRNA are coordinately modulated, the algorithm then assesses all combinations of 2 sncRNAs, 3 sncRNAs or 4 sncRNAs, followed by examination of each individual sncRNA using a leave-one-out strategy to assess the importance of each individual sncRNA in the pathology of the disease. As would be expected leaving out most of the 6,599 sncRNA sequences from the Selection algorithm has no impact on the distinction between having prostate cancer and no prostate cancer because they are not differentially associated with either pathology. The impact of the sncRNAs assessment can be visualized using the importance plot shown in Figure 5 (right panel). The importance plots show the following: (1) some exosomal sncRNAs are present in different levels in different pathologies, (2) the sncRNAs are snoRNAs and miRNAs indicating that one type of sncRNA is insufficient for the disclosed analysis and (3) the diagnosis using the algorithm does not change with more than 280 sncRNA sequences in the classification assessment.

[00055] The informative sequences for the Discovery CS Test (which differentiates between low- risk (GG1) and intermediate- and high-risk prostate cancer (GG2-GG5), were identified using the appropriate Grade Groups and the same Selection strategy (Figures 6). The patient population for this analysis included 89 subjects with NEPC and 146 patients with cancer (GG1-GG5)]

[00056] The informative sequences for the Discovery HG Test (which differentiates between low- and intermediate-risk (GG1+GG2) and high-grade, high-risk (GG3-GG5) prostate cancers were similarly determined (Figure 7). The patient population for this analysis included 181 patients with GG1+GG2 cancer and 55 patients with GG3-GG5 cancer. (Figure 7).

[00057] It is important to note that while some of the sncRNAs are common between the tests, their relative importance in the Classification of disease status varies from test to test ( i.e ., Discovery PCa Test, CS Test, and HG Test).

[00058] For each Test the most informative 280 sncRNAs SEQ. ID NOs: 1-840) were used to design customized OpenArray™ platforms. The OpenArray™ platform for each Sentinel™ Test was further validated in a large case-control study of 1436 patients. The demographics of the subjects used to train and validate the Sentinel™ PCa, Sentinel™ CS and Sentinel™ HG Tests are shown in Table 5 (see [000109] -[000110]). A stratified random sample of 600 subjects was selected to identify the validation dataset; the remaining 836 subjects served as the training dataset. The validation sample of 600 patients was stratified so that an equal number of subjects were biopsy negative versus biopsy positive (300 each), and of the biopsy positive cases, 200 were GG1 + GG2 (146 GG1 and 54 GG2) and 100 GG3 - GG5.

The Sentinel™ PCa Test to identify prostate cancer

[00059] The Sentinel™ PCa, Sentinel™ CS and Sentinel™ HG Tests are based on a Classification Algorithm that takes as input the sncRNA expression signature for each patient with unknown disease status and produces a Sentinel™ Score; the participant is classified by comparing this Score to the pre-determined cutoff value (obtained from cross-validation in the training dataset) that controls sensitivity for classifying a future patient with unknown disease status (but known expression signature), at a user-defined level (typically 95% or greater).

[00060] The Sentinel PCa Score is compared to a calculated cutoff that controls sensitivity for a future patient at a desired level of, for example, 95% to distinguish between having prostate cancer and no prostate cancer for the PCa Test (Figure 7). The Sentinel™ PCa Test utilizes 280 sncRNA (identified by the Discovery PCa Test), of which 145 unique sncRNAs: 60 miRNAs and 85 snoRNAs are highly informative. This defines the classification boundary used to dichotomize patients into prostate cancer or not. The cutoff is determined by the algorithm such that the Sentinel™ PCa score, which dichotomizes the patients into cancer/no cancer, correctly classify the patient as having cancer 19 times out of 20 ( i.e ., with 95% sensitivity).

Table 1: SEQ ID NOs: 1-280 Used in the PCa Test Analysis

The Sentinel™ CS Test to identify low grade (indolent) prostate cancer

[00061] The Sentinel™ Clinically Significant (CS) Test uses a similar classification algorithm to produce a Sentinel™ CS Score that is compared to a calculated cutoff. The cutoff controls sensitivity for a future patient at a desired level (95%), to distinguish between Clinically Significant cancer (GG2-GG5) (if the Sentinel™ CS Score is greater than or equal to the cutoff) and Clinically Insignificant cancer (GG1) (if the Sentinel™ CS Score is less than the cutoff). The algorithm is trained using only the subset of patients known to have prostate cancer in the dataset used to train the Sentinel™ PCa Test. Similarly, using the Classification Algorithm, 280 sncRNAs were used as a basis to define an expression signature for the Sentinel™ CS Test. The Sentinel™ CS Test utilizes 280 sncRNA (identified by the Discovery CS Test, of which 135 unique sncRNAs: 130 miRNA and 66 snoRNAs are highly informative. Table 2: SEQ ID NOs: 281-560 Used in the CS Test Analysis

The Sentinel™ HG Test to identify patients with High Grade prostate cancer

[00062] For individuals classified as having prostate cancer, a similar approach was used to train and validate the Sentinel™ High Grade (HG) which discriminates between GG1 + GG2 (low- and favorable-intermediate risk cancer) and unfavorable-intermediate and high-risk prostate cancer (GG3-GG5). These informative sequences identified form the basis of the Sentinel™ HG Test.

Additional analyses demonstrated that the same cohort of sncRNAs can be used to dichotomize the patients with cancer into GG1+GG2 (low and intermediate risk cancer) from GG3-5 (high risk cancers). This biostatistical analysis forms the basis for the Sentinel™ HG Test, which utilizes 280 sncRNAs, identified by the Discovery HG Test, of which 280 unique sncRNAs: 191 miRNA and 89 snoRNAs are highly informative

Table 3: SEQ ID NOs: 561-840 Used in the HG Test Analysis

[00063] The selection of the sncRNAs in the Sentinel™ PCa, CS and HG Tests is independent of PSA, Gleason Score, or biological pathway analysis, and as such is entirely unbiased. Because the algorithm was validated using sncRNA levels obtained from an independent training set made up of a cohort of participants whose core needle biopsy is positive or negative (PCa Test), or patients labeled as either having advanced disease (GG3-5) or not (No evidence of PCa or GG1- 1) for the CS Test (see Table 4, [000103] -[000104] and Table 5 [000109]-[000110]), this statistical methodology minimizes both Type 1 error (false negative) and Type 2 error (false positive), which ensure that the tests rigorously distinguish between none and low-grade cancer, low and intermediate grade cancer and between intermediate and high-grade disease. Based on the algorithm used in the analysis, the described invention has no false negatives and a very low (<5%) false positive rate.

[00064] Based on the three tests described above, the OpenArray™ platform sequentially interrogates the informative RNA entities present in a single sample of sncRNA extracted from urinary exosomes without compromising sensitivity and specificity of the three tests.

[00065] In one aspect, the disclosure provides a method for diagnosing prostate cancer comprising a platform that allows one to distinguish between clinically significant tumors and indolent tumors and in putting the data based on a subset sncRNAs interrogated into an algorithm that has been validated based on an independent training data set.

[00066] In one aspect, the method for diagnosing prostate cancer in a male patient comprising (1) obtaining a biological sample from the patient, (2) detecting the aggregate expression profile of a collection of signature small non-coding RNAs (sncRNAs) that bind to a plurality of nucleic acids or hybridizing probes selected from the group consisting of SEQ ID NOs: 1 - 280; and (3) correlating the aggregate expression profile of the collection of signature sncRNAs using the PCa test to determine whether the patient is at risk for prostate cancer, i.e., having no evidence of prostate cancer or having prostate cancer.

[00067] In another aspect, the disclosure provides a method for screening prostate cancer using the same. In yet another aspect, the disclosure provides a method for predicting the probability of prostate in a subject.

[00068] For patients identified to be at risk for prostate cancer (i.e., determined to have prostate cancer), the samples are re-analyzed using the Sentinel™ Clinical Significant (CS) Test to distinguish patients with clinically significant or aggressive prostate cancer (GG2-GG5) from patients with clinically insignificant or indolent (GG1) prostate cancer. In one embodiment, the patient is identified as having aggressive prostate cancer when the aggregate or combined expression profile of a plurality or a collection of signature sncRNAs is higher than or equal to the aggregate expression profile in a prostate cancer biological sample, or identifying the subject as having low-grade prostate cancer when the aggregate expression profile of a collection of signature sncRNAs is less than or equal to the aggregate expression profile in a low-grade prostate cancer biological sample.

[00069] In one embodiment the biological sample includes, but is not limited to, prostate tissue, blood, plasma, serum, urine, urine supernatant, urine cell pellet, cerebrospinal fluid, semen, prostatic secretions, and prostate cells. In some embodiments, the biological sample is a urine sample. In yet another embodiment, the samples are exosomes isolated from the urine sample. In a preferred embodiment, the sample is sncRNAs isolated from exosomes derived from the urinary sample.

[00070] Exosomes are small extracellular vesicles (EV) that originate in the endosomal compartment of eukaryotic cells. They are found in biological fluids including blood, urine, semen and cerebrospinal fluid. The biogenesis of exosomes is not well-understood; however, it is generally accepted that they arise at the point at which the early endosomal pathway bifurcates to form late endosomes and multi-vesicle endosomes, the first stages of the exosomal pathway. Exosomes contain sncRNAs including miRNAs and small nucleolar RNAs (snoRNAs) which are derived from the cytoplasm and nucleolar region of the cell respectively. The presence of exosomes and EVs in the tumor microenvironment have been associated with malignancy in a number of tumor types including prostate cancer and other cancers.

[00071] In certain embodiments, sncRNAs isolated from exosomes are derived from semen, blood, prostatic secretions and cerebral spinal fluid. In further embodiments, the exosomes are isolated from the cancer cells including prostate cancer cells, lymphocytes, and cells from prostate tissues.

[00072] The method for isolating exosomes is well known in the art and can be carried out using kits such as the Exosome RNA Isolation Kits (Norgen Biotek Corp., Ontario, CA). sncRNA yields can be quantified by fluorimetry (Qubit, Thermo Fisher Scientific), and the quality of the sncRNAs isolated is assessed using an Agilent 2100 bioanalyzer.

[00073] The RNAs extracted from the isolated exosomes are a mixture of small RNAs referred collectively as small non-coding RNAs (sncRNAs), which include miRNA, snoRNA, scaRNA, siRNA, snRNA, and exRNA. Due to their size (< 200 nucleotides), the sncRNAs are readily extracted from biological samples, including, for example, Formalin-Fixed Paraffin-Embedded (FFPE) tissue or urine. The sncRNAs are not degraded during fixation or extraction, obviating the problems intrinsic in extraction of mRNA from FFPE tissues. Yield of sncRNAs of about 10 ng from the biological samples is sufficient for multiple analyses using the Exosome RNA Isolation Kits of Norgen disclosed above.

[00074] The extracted sncRNAs are reversed transcribed into cDNAs, which are more stable than RNA, thus allowing for longer storage. The resulting cDNAs are hybridized against a selected set or collection of signature sncRNAs probes or genome array or micro-array chips, such as the miR 4.0 arrays (ThermoFisher Scientific) for further analysis. The selection of the informative set of sncRNA probes is independent of PSA, Gleason Score or biological pathways. The selected set or collection of signature sncRNAs comprises SEQ ID NOs: 1-280, 281-560 and 561-840. The number of sncRNA sequences or probes in the set range from 145 - 196, preferably not more than 280 sncRNA sequences for in putting in the Sentinel™ PCa, Sentinel™ CS and Sentinel™ HG Tests. The tests are more precise with up to 280 sncRNA sequences added to the algorithm. Addition of more than 280 sncRNA sequences in the analysis does not increase the precision of the assay.

[00075] Real time-PCR or RT-PCR, often known as qPCR or RT-qPCR, is conducted to quantify the absolute amount of a target sequence or to compare relative amounts of a target sequence between samples. The RT-qPCR monitors amplification of the target in real-time via a target- specific (probes) fluorescent signal emitted during amplification. Because background fluorescence occurs during most RT-qPCR reactions despite the use of sequence specific probes against the targets, the issue of background fluorescence signal can be addressed by considering two values in real-time PCR: (1) the threshold line (C _t ) and (2) the cycle quantification (C _q ). value. The threshold line (C _t ) is the level of detection when a reaction reaches a fluorescent intensity that is above background levels, that is a point where the reaction curve begins the exponential phase (inflexion point). The C _q or cycle quantification value is the PCR cycle number at which the sample’s reaction curve intersects the threshold line. The C _q, value therefore, indicates how many cycles it took to detect a real signal from the samples, i.e., time to event, where the event is the saturation of the fluorescence, indicating the maximum level of detection. Because RT-qPCR runs provide a reaction curve for each sample, there will be many C _q values. The software in the PCR cycler will calculate and chart the C _q value for each of the samples. Values are inverse to the amount of target nucleic acid in the sample, and correlate to the number of target copies in the sample. Lower C _q values (typically below 29 cycles) indicate high amounts of target sequence. Conversely, higher C _q values (above 38 cycles) indicate lower amounts of the target nucleic acid in the sample. However, the time to event value can be obtained when the slope of the reaction curve becomes zero, rather than using the C _q when the slope is at a maximum. In one embodiment, the sncRNAs are interrogated using RT-qPCR. In another embodiment, the sncRNAs are interrogated using qPCR. In a further embodiment, the sncRNAs isolated from the urinary exosomes are interrogated using the Affymetrix GeneChip™ miRNA 4.0 Array following manufacturer’s instructions.

[00076] These informative sequences obtained from the training set (usually 280 sequences) are then transferred to the OpenArray platform. Patient samples of unknown status are then interrogated on the OpenArray platform and the Sentinel Score is determined using the Classification Algorithm. The status of the patient is determined from the Sentinel Score.

[00077] In one embodiment, the sncRNAs levels from a patient of unknown prostate cancer status are interrogated on the OpenArray platform and Sentinel score from a patient of unknown disease status can then be compared to that of the training set to determine the status of the patient

[00078] In one embodiment, the data obtained from the analysis of sncRNAs levels in the test sample from a patient of unknown prostate disease status (test sample) using, for example, by RT- qPCR on the Affymetrix GeneChip™ miRNA 4.0 Array can be compared to those from healthy patients (no evidence of cancer) or healthy cells obtained from the subject with prostate cancer. The data obtained from the analysis of sncRNA levels from the test sample can also be compared to clinical baselines established by analyzing healthy (no cancer) and non-healthy (with genitourinary cancer) patients, and the non-healthy patients be further categorized into different specific cancer types, which can be further categorized into different stages or severity of the specific cancer type (for example, prostate cancer, etc.) and different stages of the specific disease. In one embodiment, the data obtained from the analysis of sncRNAs expression in the test sample using RT-qPCR on the Affymetrix GeneChip™ miRNA 4.0 Array are compared to those from healthy patients (no evidence of cancer) or the data can also be compared to clinical baselines established by analyzing healthy (no cancer) and non-healthy (with having prostate cancer) patients, and the non-healthy patients be further categorized into different stages of prostate cancer (GG1 and GG2-GG5 or GG1+2 and GG3-GG5).

[00079] In some embodiments, the method uses an OpenArray™ technology (ThermoFisher Scientific) to interrogate a panel of sncRNAs ( e.g ., miRNAs, snoRNAs). The OpenArray™ technology uses a microscope slide-sized plate having 48 subarrays. Each subarray has 64 through-holes, and each hole is 300 pm in diameter and 300 pm deep. The holes are treated with hydrophilic and hydrophobic coatings in order to retain reagents in the through-holes via surface tension. The OpenArray™ technology, with its 3,072 through-hole (48x64), provides a system for streamlining real-time PCR studies that use large number of samples, assays or both. The system thus allows processing of samples for gene expression in large number in short time periods using microquantities of samples and reagents. The method employs an algorithm that relies on the expression level of each of the sncRNAs and the grading of the biopsies (at least 12 core needle biopsies). In the case of prostate cancer, the methodology is independent of serum Prostate Specific Antigen (PSA) levels, Gleason Score (neither of which are meaningful markers of tumor progression), or patient age. The methodology is also independent of any analyses of biological pathways. The present method uses sncRNAs isolated from the subject (e.g., urinary sample, urinary exosome, or prostate tissue samples) to stratify men into those that have prostate cancer (both indolent (clinically insignificant) or aggressive (clinically significant)) and those that do not. This methodology can replace serum PSA as the major screening assay for prostate cancer.

[00080] In one aspect, the disclosure provides a method that distinguishes clinically significant prostate cancer based on the aggregate expression profile of a collection of signature sncRNAs interrogated that is subjected to a classification algorithm that is independent of pathology (Gleason Score), tumor volume or PSA. The RNA extracted from biological samples of patients with known cancer outcomes are reverse-transcribed and hybridized against a full-genome array (e.g., Affymetrix GeneChip miR 4.0) containing sncRNAs. Small non-coding RNAs that are differentially regulated in clinically significant prostate tumors are identified. In one embodiment, the absolute value of the signal from the Open Array identifying the sncRNAs that hybridize to the probes for SEQ ID NOs: 281-561, is compared to the aggregate expression profile found in clinically significant (GG2-5) prostate cancer tumors. In another embodiment, the aggregate value of the signal from the Open Array identifying the sncRNAs that hybridize to the probes for SEQ ID NOs: 561-840, is compared to the absolute expression profile found in clinically low and favorable intermediate grade (GG1 + GG2) versus unfavorable intermediate and high grade (GG3- GG5) prostate cancer tumors.

[00081] The disclosed method provides a robust and accurate determination of prostate cancer prognosis in 72-96 hours from the time when the urine sample is received to obtaining a Sentinel Score. In another embodiment, the aggregate expression profile of the identified sncRNAs that bind to SEQ ID NOs: 281-560 and 561-840 is compared to the aggregate expression profile of sncRNAs in clinically significant prostate cancer. In another embodiment, the aggregate expression profile of the identified sncRNAs interrogated that bind to SEQ ID NOs: 281-560 and 561-840 is compared to the aggregate expression profile of sncRNA in clinically significant prostate cancer. In a further embodiment, the aggregate expression profile of the identified sncRNAs that bind to SEQ ID NOs: 281-560 and 561-840 is compared to the aggregate expression profile of sncRNA in clinically significant prostate cancer sample. As a result, appropriate treatment options (or the lack thereof) can be initiated.

[00082] The terms relative aggregate expression profile are used interchangeably. The aggregate expression profile of at least a plurality of sncRNAs are combined and compared to the same aggregate expression profile in clinically significant prostate cancer tissue. In some embodiments at least 40 sncRNAs are combined and compared to the same aggregate expression profile in clinically significant prostate cancer tissue. In some embodiments at least 90 sncRNAs are combined and compared to the same aggregate expression profile in clinically significant prostate cancer tissue. In some embodiments at least 150 sncRNAs are combined and compared to the same aggregate expression profile in clinically significant prostate cancer tissue. In some embodiments at least 200 sncRNAs are combined and compared to the same aggregate expression profile in clinically significant prostate cancer tissue. In a preferred embodiment, at least 224 sncRNAs and not more than 280 sncRNAs are combined and compared to the same aggregate expression profile in clinically significantly prostate cancer tissue. In certain embodiments, a higher aggregate expression profile as compared to the aggregate expression profile in a low-grade prostate cancer tissue indicates the patient has aggressive prostate cancer and treatment is required. In other embodiments, an aggregate expression profile equal to or lower than the aggregate expression profile in a low-grade prostate cancer tissue indicates the patient does not have aggressive prostate cancer and monitoring but not treatment may be required.

[00083] In some embodiments, the aggregate expression profile of selected sncRNAs is an aggregation of various types of modulated expression of the sncRNAs. The modulated expression can be decreased or increased expression profile relative to the same sncRNA in other tissue/tumor types, such as healthy prostate tissue, low-grade prostate cancer tissue, or high-grade prostate cancer tissue.

[00084] In other embodiments, the aggregate expression profile of selected sncRNAs can be an aggregation of the decreased aggregate expression profile of certain sncRNAs as well as an aggregation of the increased aggregate expression profile of other sncRNAs in the same tissue sample. For example, a progression score, or aggregate expression profile of a collection of signature sncRNAs may include one or more sncRNAs with decreased aggregate expression profiles relative to another tissue type or other sncRNAs in the same tissue sample, while one or more of the remaining sncRNAs exhibit increased aggregate expression levels relative to another tissue type or other sncRNAs in the same tissue sample. The aggregate expression profile of the collection of differently modulated sncRNAs provides a sophisticated, unbiased, indication of whether a prostate tumor is clinically significant. Unlike other methods that merely evaluate the presence or absence, or simple increase or decrease of individual target molecules, as compared to normal tissue, the methods disclosed provide a truly unbiased, independent, and multi-variable analysis of a prostate tissue sample thereby allowing for a surprisingly accurate diagnosis of whether a prostate cancer tumor is clinically significant.

[00085] In some aspect, the method provides for the use of the aggregate expression profile of the collection of signature sncRNAs for monitoring metastasis and cancer staging.

[00086] In another aspect, the disclosure provides a method for detecting a urological malignancy based on the aggregate expression profile of the collection of signature sncRNAs interrogated and subject to analysis using the classification algorithm disclosed. In some embodiments the malignancy is cancer of the prostate. [00087] The present disclosure provides an algorithm-based molecular diagnostic assay for predicting a clinical outcome for a patient with prostate cancer. The expression level of one or more sncRNAs may be used alone or arranged into functional gene subsets to calculate a quantitative score that can be used to predict the likelihood of a clinical outcome.

[00088] A“quantitative score” is an arithmetically or mathematically calculated numerical value for aiding in simplifying or disclosing or informing the analysis of more complex quantitative information, such as the correlation of certain expression profile of the disclosed sncRNAs or sncRNAs subsets to a likelihood of a clinical outcome of a prostate cancer patient. A quantitative score may be determined by the application of a specific algorithm. The algorithm used to calculate the quantitative score in the methods disclosed may group the expression profile values of the sncRNAs. The grouping of sncRNAs may be performed at least in part based on knowledge of the relative contribution of the sncRNAs according to physiologic functions or component cellular characteristics, such as in the groups discussed herein. A quantitative score may be determined for a sncRNA group (“sncRNA group score” or the Sentinel™ Score). The formation of groups, in addition, can facilitate the mathematical weighting of the contribution of various aggregate expression profile of genes or gene subsets to the quantitative score. The weighting of a sncRNA or sncRNAs group representing a physiological process or component cellular characteristic can reflect the contribution of that process or characteristic to the pathology of the cancer and clinical outcome, such as recurrence or upgrading/upstaging of the cancer. The present invention provides a number of algorithms for calculating the quantitative scores. For example, the Classification algorithm in the present disclosure works the same way for developing the Sentinel Scores for distinguishing different disease states. The Classification algorithm selects different sncRNA sequences from the training data set for clinically significant and insignificant disease states.

[00089] On the other hand, the Selection algorithm test how well the aggregate expression profile of the urinary exosomal sncRNAs correlate with pathological stage of the disease (cancer/no cancer) in a large population of participants with known pathology. The Selection algorithm individually assessed how well each of the 6,599 sncRNAs interrogated on the miR4.0 arrays correlates the known pathology of the participant. It then iteratively assesses all combinations of 2 sncRNAs, 3 sncRNAs or 4 sncRNAs of the 6,599 sncRNAs interrogated by the miR 4.0 Arrays, followed by examination of each individual sncRNA using a leave-one-out strategy to assess the importance of each individual sncRNA in the pathology of the disease. The Sentinel Score for a patient with unknown disease status is then determined by interrogating selected sncRNAs using the Open Array and the clinical status is determined by comparing the score to that from the training data sets. In an embodiment of the invention, an increase in the quantitative score indicates an increased likelihood of a negative clinical outcome.

[00090] Based on the quantitative score and cumulative or absolute or aggregate expression profile, methods of treatment can also be decided. The methods of treating prostate cancer include surgery for complete surgical removal of prostate tissue, administering an effective dose of radiation, and administering a therapeutically effective amount of a medication for the treatment of prostate cancer, or a combination of the above.

[00091] The algorithm-based assay and associated information provided by the practice of the methods of the present invention facilitate optimal treatment decision-making in prostate cancer. For example, such a clinical tool would enable physicians to identify patients who have a low likelihood of having an aggressive cancer and therefore would require no further medical intervention except for a routine follow-up or active surveillance every 3 months, 6 months or 12 months. Patients with no cancer do not require medical intervention return for follow-up once every year. Patients at risk for developing aggressive cancer require medical intervention, which includes but is not limited to treatment with one or more chemotherapeutic agents (e.g., taxotere, cabazitaxel, docetaxel, mitoxantrone, epirubicin, paclitaxel and estramustine, etc.), hormone therapy (e.g., lutenizing hormone releasing hormone agonists to prevent production of testosterone such as leuprorelin, goserelin and triptorelin or anti-androgen drugs that prevent testosterone from reaching the cancer cells, e.g., bicalutaminde and nilutamide), immunotherapeutics, radiation, cryotherapy, surgery or a combination thereof.

[00092] Patients who undergo treatment are monitored using the disclosed method to determine the patients’ response to treatment. In one aspect, the disclosure provide a method for determining the patient’s response to treatment comprising: (i) obtaining a biological sample from a patient, (ii) detecting the aggregate expression profile of a signature collection of small non-coding RNAs (sncRNAs) from the biological sample wherein the collection of sncRNAs comprises SEQ ID NOs: 1-280, 281-560 and 561-840 (iii) correlating the aggregate expression profile of sncRNAs of SEQ ID NOs: 1-280, 281-560 and 561-840 from the subject after treatment by comparing the aggregate expression profile of SEQ ID NOs: 1-280, 281-560 and 561-840 to that prior to treatment, (iv) determining if the patient is responsive to the treatment, and if there is a need for modification of the treatment. In one embodiment, the method further compares the resulting aggregate expression profiles of a signature collection of sncRNA from (iii) above is then compared to the aggregate expression profile of a signature collection of sncRNA for the large training data set from a target population having prostate cancer with known Grade groups to determine if the (a) patient prostate cancer is stable (no apparent change compared to the Grade group), (b) the patient is responsive to the treatment, i.e., patient gets better (the results show tumors that resembles tumors with lower grade group) or (c) the patient is non-responsive (patient gets worse when the results show tumors that resembles tumor of higher Grade group) based on the aggregate expression profile of a collection of signature sncRNAs and Sentinel Score for that Grade group, and if there is a need for modification of the treatment. Treatment modification includes but not limited to adjusting the concentration or amounts of chemotherapeutic agents, radiation, immunotherapeutic or hormone administered, adding or removing one or more of agents used.

[00093] In another aspect, the disclosure provides a method for determining the disease recurrence, disease progression or likelihood of survival based on the aggregate expression profile of a collection of signature sncRNAs comprising SEQ. ID. NOs: 1-280, 281-560 and 561-840 by comparing the aggregate expression profile of SEQ. ID. NOs: 1-280, 281-560 and 561-840 in a training dataset and the patient’s earlier profile.

[00094] In another aspect, the disclosure provides a system for determining whether a patient has no cancer or has cancer and classifying the subject with cancer as (i) indolent (low grade, GG1), (ii) intermediate or high grade (GG2-GG5), (iii) low/intermediate risk (GG1-GG2) or (iv) aggressive (high grade, GG3-GG5) prostate cancer comprising at least three processors configured to (a) interrogate sncRNA sequences for informative sequences, (b) determine and compare a Sentinel Score to determine if the subject has prostate cancer or no prostate cancer and to classify subject determined to have cancer to the various Grade groups, e.g., low grade, intermediate/high grade, low/intermediate risk or aggressive grade cancer. Subjects determined to have no evidence of cancer do not require medical intervention and would return for follow-up once every year. Subjects determined to have low grade or low/intermediate grade prostate cancer would require no medical intervention except for a routine follow-up or active surveillance every 3, 6, or 12 months, and subjects determined to have intermediate/high grade or aggressive prostate cancer require medical intervention.

EXAMPLES

[00095] This and other aspects of the present invention are further illustrated by the following non limiting examples.

EXAMPLE 1

Study Populations

[00096] Two independent patient cohorts were used for the development and validation of the Sentinel™ PCa and Sentinel™ CS Tests. The clinical and demographic characteristics of the 233 participants used to develop the Sentinel™ PCa to classify patients as having cancer or no cancer was based on the statistical analysis of a collection of signatures snRNAs. For patients classified as having cancer, patients with GG1 (indolent, low risk cancer) are distinguished from GG2-5 (respectively as intermediate, high-risk and aggressive cancers) using the Sentinel™ CS Tests, which is also based on the statistical analysis of another collection of signatures the sncRNAs using a second Classification algorithm to classify tumors into GG1 versus GG2-5. The sncRNAs in both tests are interrogated by the Affymetrix miR 4.0 array.

Urine Collection and Processing

[00097] Urine samples for the development of the Sentinel™ PCa and CS Tests and the US-based cohort of the retrospective study were collected on the day of visit for clinical workup at two clinical sites: Albany Medical Center (Albany, NY, USA) and SUNY Downstate Medical Center (Brooklyn, NY, USA). Remaining samples for the retrospective study were retrieved from the GUBioBank, University Health Network, Toronto CA and shipped frozen at -20°C in bulk to the miR Scientific laboratories. Patient information was collected and anonymized as approved by Institutional Review Board at each participating site. Prostate cancer diagnosis was obtained by histopathological grading of core-needle biopsies; the percentage of tumor per core and number of positive cores were used to assess the grade group (GG).

[00098] Urine samples were centrifuged to remove free cells and debris. RNA was extracted using Exosome RNA Isolation Kits (Norgen Biotek, ON) according to the manufacturer's instructions. sncRNA yields were quantified by fluorimetry (Qubit, Thermo Fisher Scientific) and RNA samples were stored at -80°C until analysis.

Microarray analysis of total exosomal sncRNAs

[00099] sncRNAs were interrogated using the Affymetrix GeneChip™ miR 4.0 Array following the manufacturer's instructions. MAIME-compliant raw data files for the 235 patients analyzed on these arrays have been deposited in NCBI’s Gene Expression Omnibus. {Edgar R et al. Gene Expression Omnibus: NCBI gene expression and hybridization array data repository. Nucleic Acid Res 2002 30:207). Six thousand five hundred and ninety-nine sncRNAs in the training set were interrogated on the Affymetrix GeneChip™ miRNA 4.0 Array.

[000100] The small non-coding RNA entities interrogated for each participant were analyzed using proprietary Selection and Classification Algorithms. The most informative sequences for distinguishing between cancer and non-cancer subjects (SEQ ID NOs: 1-280) and between Grade Group 1 and Grade Group 2-5 patients were identified. (SEQ ID NOs: 281-842)

QuantStudio OpenArray™-based interrogation of exosomal sncRNAs

[000101] cDNA synthesis, pre-amplification of selected miRNAs: For analysis of exosomal miRNA, total sncRNA was reverse transcribed in separate reactions with three specific miRNA stem- loop primer pools with the TaqMan™ Micro RNA Reverse Transcription Kit (Thermo Fisher Scientific) as recommended by the manufacturer. The miRNA cDNA pools were enriched individually with Pre-Amp primer pools for 16 cycles (95°C for 10 min, 55°C for 2 min, 72°C for 2 min, 95°C for 15 sec and 60°C for 4 min repeated 16 cycles, 99.9°C for 10 min), and interrogated on the QuantStudio OpenArray™ on three 56-entity sub-arrays following the manufacturer’s recommendations . [000102] cDNA synthesis, pre-amplification and interrogation of selected snoRNAs: Total sncRNA was reverse transcribed with High-Capacity cDNA Reverse Transcription Kit with a single Pre-Amp primer pool (Thermo Fisher Scientific) as recommended by the manufacturer. snoRNA cDNA products were enriched by preamplification (95°C for 10 min, 95°C for 15 sec and 60°C for 4 min repeated for 14 and 18 cycles respectively, and 99°C for 10 min) and interrogated on two 56-entity sub-arrays.

Statistical Analysis:

[000103] The Sentinel™ PCa Test is based on Classification Algorithm that has been trained on a cohort of participants whose core-needle biopsy is positive or negative. The Classification Algorithm takes as input the sncRNA expression signature for a participant with unknown disease status and produces a Sentinel™ Score; the participant is classified by comparing the Sentinel™ PCa Score to the pre-determined cutoff value that maintains the sensitivity for classifying a future patient with unknown disease status (but known expression signature), at a user-defined level (95% or greater). A second classification algorithm, the Sentinel™ CS Test operates analogously to the Sentinel™ PCa Test. However, the classification algorithm for the Sentinel™ CS Test is trained on a cohort of patients labeled as low grade (GG1) and a second cohort of patients labeled as favorable intermediate to high grade prostate cancer (GG2-GG5). The third classification algorithm, the Sentinel™ HG Test is trained on a cohort of patients determined to be low- and favorable intermediate-risk (GG1+GG2) prostate cancer and a second cohort characterized as unfavorable intermediate risk and high-risk (GG3-GG5)

[000104] The Sentinel™ Testing paradigm operates in two or three layers. First, it uses the sncRNA signature from the participant’s urine to input to the classification rule of the Sentinel™ PCa Test to determine if cancer is present; second, for those patients diagnosed with cancer, the Sentinel™ CS Test determines whether the cancer is low risk (GG1) or not; thirdly the Sentinel™ HG Test determine whether the tumor is unfavorable intermediate or high risk (GG3-GG5) or not. (See Figure 4) Table 4: Demographics and Clinical Characteristics of Cohort Used to Develop Classification Algorithm.

^* Under exempt study status, PSA levels were not available for patients with no evidence of cancer.

[000105] Table 4 established the training datasets used to develop the Sentinel™ tests. Of the 235 patients, patients included in the“no cancer” cohort (89 patients) were carefully selected from age- matched men who were seen at urology clinics for issues unrelated to urological oncology (n=58), and from men who had one or more 12-needle diagnostic core needle biopsies that showed no evidence of prostate cancer (n=30).

[000106] Patients in the“cancer” cohort (n=146) were selected based on the histopathology of the core needle biopsies. Of the 146“cancer” cohort, 90 patients were classified as GG1 cancer, 56 patients were classified as GG2-5.

[000107] Of the 6,599 microarray sequences from the training data set interrogated using the proprietary Selection Algorithm to separate that are“informative” for the outcome versus those that are not, only 400-600 are informative. By outcome, it is meant to mean sncRNA sequences that impact the algorithm when each sequence is added to predict whether the subject of unknown disease status has or has no prostate cancer and the stages of the cancer (indolent versus aggressive).

[000108] The statistical analysis used is based on the ability to identify sequences that have hidden associations with outcome that is only observed after conditioning on other sequences. Of the 400- 600 informative sncRNA sequences, 280 sncRNA sequences were used in the Classification algorithm as a basis to define an expression signature for the Discovery PCa Test (Figure 5), and Discovery HG Test (Figure 7) and Discovery CS Test (Figure 6). The subset of informative sncRNAs considered to be of the highest importance were then identified for each Test (Figures5, 7 and 6, respectively (right panels)).

[000109] These 280 sncRNAs were combined to design an OpenArray™ platform that provides the basis for the Sentinel™ PCa and CS Tests. The Sentinel™ PCa Test incorporates the aggregate expression profiles of 84 unique sncRNAs: 60 miRNAs and 24 snoRNAs, for classifying a subject with unknown disease status as having prostate cancer or no prostate cancer. Similarly, the Sentinel™ CS Test utilizes 135 unique sncRNAs: 105 miRNA and 30 snoRNAs for classifying a subject having prostate cancer as having GG1 (indolent) prostate cancer or GG2-GG5 (aggressive) prostate cancer. In addition, 61 sncRNAs (25 miRNAs and 36 snoRNAs) are informative in both Tests. The OpenArray™ platform sequentially interrogates the informative RNA entities present in a single sample of sncRNA extracted from urinary exosomes without compromising sensitivity and specificity of the two tests.

EXAMPLE 2

Validation of the Sentinel™ PCa, Sentinel™ CS and Sentinel™ HG Tests in a Case

Control Patient Cohort

[000110] The performance characteristics of the Sentinel™ PCa and Sentinel™ CS Tests using the OpenArray™ platform was established in a case control study of 1436 patients (Table 5). Table 5: Demographics and Clinical Characteristics of Case-Control Sample Used to Validate the Sentinel™ PCa and Sentinel™ HG Tests.

Under exempt study status, PSA levels were not available for individual patients with no evidence of cancer, all were less than 3.0ng/mL.

[000111] The performance characteristics of the Sentinel™ PCa Test were determined in case control cohort of 600 men whose demographics are shown in Table 5. The scatter plot of the Sentinel™ PCa Scores is shown in Figure 8A, with the corresponding Receiver Operator curve (ROC) curve in Figure 8C. As summarized in Table 6, the Sentinel™ PCa Test correctly classifies 281/300 patients as having cancer and 275/300 patients as having no cancer (Sensitivity 93.7%,

Specificity 91.7%).

[000112] The performance characteristics of the Sentinel™ CS Test were determined in a testing cohort of 600 men. The scatter plot of the Sentinel™ CS Scores is shown in Figure 9 A, with the corresponding Receiver Operator curve (ROC) curve in Figure 9C. As summarized in Table 6, the Sentinel™ CS Test correctly classifies 143/154 patients as high grade (GG3-GG5) and 132/143 as not high grade (Sensitivity 92.9%, Specificity 90.4%). [000113] The performance characteristics of the Sentinel™ HG Test were determined in a testing cohort of 600 men. The scatter plot of the Sentinel™ HG Scores is shown in Figure 10A, with the corresponding Receiver Operator curve (ROC) curve in Figure IOC. As summarized in Table 6, the Sentinel™ CS Test correctly classifies 94/100 patients as high grade (GG3-GG5) and 191/200 as not high grade (GG1+GG2) (Sensitivity 94%, Specificity 95.5%).

Table 6: Empirical Sensitivity, Specificity, PPV and NPV for Sentinel™ PCa, Sentinel™ CS and Sentinel™ HG Tests

^* NPV or Negative Predictive Value is the probability that following a negative test result, that individual will not have that specific disease.

NPV = _ True Negative _

True Negative + False Negative

"Sensitivity of a test is the proportion of people who test positive among all those who actually have the disease.

Sensitivity = _ True Positive _

True Positive + False Negative

^*** The specificity of a test is the proportion of people who test negative among all those who do not actually have that disease.

Specificity = _ True Negative _

True Negative + False Positive

^**** PPV or Positive Predictive Value is the probability that following a positive test result, that individual will truly have the specific disease.

PPV = _ True Positive _

True Positive + False Positive EXAMPLE 3

Safety and Scientific Validity Study to Identify Clinically Insignificant PCa with Sceintific Sentinel™ Platform.

[000114] The purpose the clinical study is to validate the performance characteristics of the Scientific Sentinel™ PCa Test and the Scientific Sentinel™ CS Test to (1) identify patients with prostate cancer in men of age 50-80 years with suspicion of prostate cancer for whom needle biopsy is performed, and (2) to distinguish men of ages 50-80 years with clinically significant prostate cancer (Grade 2 or above) from men with clinically insignificant prostate cancer (Grade Group 1). These classifications will be compared to the results of core needle biopsies, and of radical prostatectomy (where available). The sensitivity, specificity, positive and negative predictive values will be established. This study is a prospective, observational and non- interventional study. The informed participants will provide two or more urine samples over the course of the study and consent to share relevant anonymized clinical data with the study team.

[000115] Participants between the age of 50 and 80 years with suspicion of prostate cancer for whom a core-needle biopsy is performed, and otherwise meeting the inclusion and exclusion criteria, will be enrolled and will provide urine samples for the Sentinel™ PCa/CS Tests. The study will evaluate the properties of the Sentinel™ PCa Test and the Sentinel™ CS Test that is based on the disclosed method using Classification Algorithms to identify future patients with prostate cancer and to classify prostate cancer as clinically significant or clinically insignificant.

[000116] The "gold standard" assessment of cancer will be made from the results of core needle biopsies: participants with no positive cores will be designated "cancer- free"; participants with cancer in one or more cores will be designated as having "Clinically Insignificant" prostate cancer provided all cores with cancer have no greater than Grade Group 1 histopathology; participants will be designated as having "Clinically Significant" prostate cancer if any cores have Grade Groups 2-5.

[000117] Each study participant enrolled will be followed for one year. Participants will provide urine samples during each visit, and all relevant clinical data, including re-biopsies, PSA results and pathology report from radical prostatectomy (if administered as part of clinical care) will be obtained. The follow-up results, if available will be used for outcome analysis. For each urine sample provided, the Sentinel™ PCa and CS tests will be determined and compared with the available 1 year follow up outcome data to inform the sensitivity, specificity, positive and negative predictive values of the tests.

[000118] The Classification Algorithm employed functions by controlling sensitivity at, or above, a pre-specified level, denoted 1-a; for example, the value that has been assumed in this design is a=0.05, so that sensitivity is at least 95% in the population. Note that the value of a represents the false-negative rate of the test, i.e., the test is (incorrectly) negative for a patient who is truly positive.

[000119] To describe how the cutoff the Sentinel™ PCa Score is calculated to control sensitivity, for each participant in the training dataset, the Sentinel™ PCa Score will be calculated using the remaining members of the training dataset and only his small non-coding RNA (sncRNA) sequence; that is, the true disease status of each patient in the training dataset will be blinded, thereby mimicking the setting for classification of a future patient. The cutoff used in the Sentinel™ PCa Test is then calculated so that the empirical sensitivity over patients in the training dataset with prostate cancer corresponds to the value that provides an upper one-sided 95% confidence interval for population sensitivity for a future patient of at least 1-a.

[000120] With this cutoff for the Sentinel™ PCa Score determined from the training dataset a priori , the corresponding values of sensitivity, specificity, positive and negative predictive values will be calculated, along with a corresponding upper 95% confidence interval, on the prospective participants data accrued in this proposed study, with each biopsy result blinded, i.e., using only the participant's sncRNA sequence. Note that these error rates refer to the classification of a future patient with unknown disease status.

[000121] Any patent, patent application publication, or scientific publication, cited in this application, is incorporated by reference in its entirety.