WO2019118443A1

WO2019118443A1 - Predictive biomarkers for adverse effects of radiation therapy

Info

Publication number: WO2019118443A1
Application number: PCT/US2018/064924
Authority: WO
Inventors: Anatoly Dritschilo; Amrita K. CHEEMA; Scott Grindrod; Xiaogang ZHONG
Original assignee: Georgetown University
Current assignee: Georgetown University
Priority date: 2017-12-11
Filing date: 2018-12-11
Publication date: 2019-06-20
Anticipated expiration: 2020-06-11

Abstract

The present invention relates to methods of determining if a subject has an increased risk of having an adverse reaction to receiving radiation therapy for prostate cancer. The methods comprise analyzing at least one sample from the subject to determine a value of the subject's metabolite profile, and comparing the value of the subject's metabolite profile with the value obtained from subjects determined to define a normal metabolite profile, to determine if the subject's metabolite profile is altered compared to a normal metabolite profile. A difference in the value of the subject's metabolite profile compared to those defined as having a normal metabolite profile is indicative that the subject has an increased risk of having an adverse reaction to receiving radiation therapy for prostate cancer.

Description

PREDICTIVE BIOMARKERS FOR ADVERSE EFFECTS OF RADIATION

THERAPY

Statement Regarding Federally Sponsored Research or Development

[0001] This invention was made with Government support under grant no. HHSN261201600027C awarded by National Institutes of Health. The government has certain rights in the invention.

Background of the Invention

Field of the Invention

[0002] The present invention relates to methods of determining if a subject has an increased risk of having an adverse reaction to receiving radiation therapy for cancer, for example prostate cancer. The methods comprise analyzing at least one sample from the subject to determine a value of the subject's metabolite profile, and comparing the value of the subject's metabolite profile with the value obtained from subjects determined to define a normal metabolite profile, to determine if the subject's metabolite profile is altered compared to a normal metabolite profile. A difference in the value of the subject's metabolite profile compared to those defined as having a normal metabolite profile is indicative that the subject has an increased risk of having an adverse reaction to receiving radiation therapy for cancer, for example prostate cancer

Background of the Invention

[0003] Radiation therapy (RT) is an effective modality as a primary treatment of cancers, or as an adjuvant to surgery or chemotherapy. Risks, benefits and late effects of radiation therapy are observed in the heterogeneous clinical responses of patients receiving curative radiation therapy. In principle, all cancers can be controlled if sufficient radiation doses can be delivered to tumors; however, in practice, the achievable radiation doses are frequently limited by toxicities that may result following exposure of normal tissues to high radiation doses. Organ-specific tissue injuries following prostate irradiation may include acute toxicities (such as cyctitis and enteritis), late toxicities (such as rectal of bladder bleeding) and broad toxicities such as bone marrow depletion or soft tissue necrosis. Strategies to improve the therapeutic index of RT have focused on conformal technologies to target tumors more exactly, limit the volume of exposed normal tissues and limit the doses delivered to normal tissues. The technology includes computer assisted shaping of the radiation doses; examples include intensity modulated radiation therapy (IMRT), stereotactic radiosurgery (SRS) and high-dose rate (HDR) brachytherapy, as well as the use of particle therapy (proton beam therapy (PBT) or carbon ions). Despite the sophistication of current technologies, the need to deliver high doses of radiation to tumors results in normal tissue toxicities in subsets of patients.

[0004] Variations in patients' normal tissue sensitivities to radiation have been attributed to genetic factors, including mutations in genes associated with DNA repair processes, immunological diseases, and connective tissue diseases. Extreme examples are provided by the genetic syndromes of ataxia- telangiectasia, Nijmegen breakage syndrome and the clinical syndromes of scleroderma and systemic Lupus erythematosus.

[0005] The past decade has seen major developments in the treatment of cancer, including technical improvements in radiotherapy. A fraction of patients treated for cancer, however, experience radiation treatment related acute and late effects that adversely affect quality of life and also lead to tumor recurrence episodes. The manifestation of these symptoms takes months to develop and raises an urgent need for developing smarter strategies for symptom anticipation and management. Application of personalized medicine for patient treatment has further highlighted the need for clinical biomarkers to predict response and to direct therapy.

[0006] Prostate cancer patients that are susceptible to radiation induced adverse effects carry a biochemical fingerprint that could be characterized using blood based metabolomics. Furthermore, these molecular changes may provide insight into specific pathway perturbations that could be used to instruct clinical therapeutics. Based on a retrospective outcome study, a biomarker panel was delineated for prediction of radiation response in patients treated for prostate cancer. Such biomarkers may aid in early detection of tissue toxicity in cancer patients, informing clinical decisions for treatment and follow-up management in patients at risk.

Summary of the Invention

[0007] The present invention relates to methods of determining if a subject has an increased risk of having an adverse reaction to receiving radiation therapy for cancer, for example prostate cancer. The methods comprise analyzing at least one sample from the subject to determine a value of the subject's metabolite profile, and comparing the value of the subject's metabolite profile with the value obtained from subjects determined to define a normal metabolite profile, to determine if the subject's metabolite profile is altered compared to a normal metabolite profile. A difference in the value of the subject's metabolite profile compared to those defined as having a normal metabolite profile is indicative that the subject has an increased risk of having an adverse reaction to receiving radiation therapy for cancer, for example prostate cancer

[0008] In certain embodiments, the adverse reaction in tumor recurrence. In other embodiments, the adverse reaction is rectal toxicity. In other embodiments, the adverse reaction is urinary toxicity.

Brief Description of the Drawings

[0009] FIGURE 1 depicts predictive biomarkers of Recurrence Episodes in Prostate Cancer. (Panel A). ROC curve for an eight-metabolite panel for classification of patients with recurrence episodes. (Panel B). Biomarker panel (8 metabolites) predictive of recurrence.

[0010] FIGURE 2 depicts box and whisker plots of plasma 8 metabolite panel results for normal and recurrence groups in the prostate cancer cohort.

[0011] FIGURE 3 depicts plasma metabolite index (PMI) plot representation demonstrating group stratification. The PMI results are based on the logistic regression model are illustrated as a boxplot that distinguishes between prostate cancer patients who developed recurrence and those that remained cancer free post SBRT. Solid black horizontal lines represent the mean value, while the whiskers denote the spread within a group. Orange and light blue dots represent PC patients who received hormone therapy or not, respectively. The higher index values (left vertical axis) are associated with an increased risk of recurrence in the prostate cancer cohort. The confidence interval (right vertical axis) of predicting risk of recurrence, transitions from 90 to 100% at a relative index value of 2.

[0012] FIGURE 4 depicts predictive biomarkers of rectal toxicity episodes, post SBRT in the prostate cancer cohort. Panel A. ROC curve for a six metabolite panel for classification of patients with radiation proctitis. Panel B. Biomarker panel (6 metabolites) predictive of rectal toxicity.

[0013] FIGURE 5 depicts box and whisker plots of plasma six-metabolite panel results for normal and radiation proctitis groups in the prostate cancer cohort.

[0014] FIGURE 6 depicts plasma metabolite index (PMI) plot representation demonstrating group stratification. The PMI results are based on the logistic regression model that are illustrated as a boxplot that distinguishes between prostate cancer patients who developed rectal toxicity and those that remained normal, post SBRT. Solid black horizontal lines represent the mean value, while the whiskers denote the spread within a group. Orange and light blue dots represent PC patients who received hormone therapy or not, respectively. The higher index values (left vertical axis) are associated with an increased risk of radiation proctitis (RP) in the prostate cancer cohort. The confidence interval (right vertical axis) of predicting risk of RP, transitions from 90 to 100% at a relative index value of 5.

[0015] FIGURE 7 depicts predictive biomarkers of urinary toxicity episodes, post SBRT in the prostate cancer cohort. Panel A. ROC curve for a nine metabolite panel for classification of patients with urinary toxicity episodes. Panel B. Biomarker panel (9 metabolites) predictive of urinary toxicity.

[0016] FIGURE 8 depicts box and whisker plots of nine-plasma metabolite panel results for normal and group experiencing urinary toxicity in the prostate cancer cohort.

[0017] FIGURE 9 depicts plasma metabolite index (PMI) plot representation demonstrating group stratification. The PMI results are based on the logistic regression model are illustrated as a boxplot that distinguishes between prostate cancer patients who developed urinary toxicity and those that remained normal post SBRT. Solid black horizontal lines represent the mean value, while the whiskers denote the spread within a group. Orange and light blue dots represent PC patients who received hormone therapy or not, respectively. The higher index values (left vertical axis) are associated with an increased risk of urinary toxicity in the prostate cancer cohort. The confidence interval (right vertical axis) of predicting risk of recurrence, transitions from 90 to 100% at a relative index value of 7.5.

Detailed Description of the Invention

[0018] The present invention relates to methods of determining if a subject has an increased risk of having an adverse reaction to receiving radiation therapy for cancer, for example prostate cancer. The methods comprise analyzing at least one sample from the subject to determine a value of the subject's metabolite profile, and comparing the value of the subject's metabolite profile with the value obtained from subjects determined to define a normal metabolite profile, to determine if the subject's metabolite profile is altered compared to a normal metabolite profile. A difference in the value of the subject's metabolite profile compared to those defined as having a normal metabolite profile is indicative that the subject has an increased risk of having an adverse reaction to receiving radiation therapy for cancer, for example prostate cancer. [0019] In certain embodiments, the adverse reaction in tumor recurrence. In other embodiments, the adverse reaction is rectal toxicity. In other embodiments, the adverse reaction is urinary toxicity.

[0020] As used herein, the term subject or "test subject" indicates a mammal, in particular a human or non-human primate. As used herein, the phrase cancer is understood in the art and is used to mean such things as an abnormal growth, such as hyperplasia, neoplasia or a tumor, to name a few. According to the methods of the present invention, the subject having the cancer would be a candidate for radiation therapy. Thus, the cancer as used herein does not include those cancers for which radiation therapy is not an option. In one embodiment, the cancer is prostate cancer. As used herein, the phrase "prostate cancer" is well-understood in the art and means a cancer in the prostate gland. As used herein, the "prostate cancer" can be a pre-cancerous lesion with a low Gleason score, including lesions having a Grade 1, Grade 2, Grade 3, Grade 4 or Grade 5 Gleason score. In addition, "prostate cancer" includes Stage 1, Stage 2, Stage 3, Stage 4 or even recurrent prostate cancer.

[0021] As used herein, the term means "increased risk" is used to mean that the test subject has an increased chance of developing or acquiring an adverse reaction to radiation therapy for cancer, for example prostate cancer, compared to a normal individual. The increased risk may be relative or absolute and may be expressed qualitatively or quantitatively. For example, an increased risk may be expressed as simply determining the subject's metabolite profile and placing the patient in an

"increased risk" category, based upon previous population or individual studies. Alternatively, a numerical expression of the subject's increased risk may be determined based upon the metabolite profile. As used herein, examples of expressions of an increased risk include but are not limited to, odds, probability, odds ratio, p-values, attributable risk, metabolite index score, relative frequency, positive predictive value, negative predictive value, and relative risk.

[0022] For example, the correlation between a subject's metabolite profile and the likelihood of suffering from an adverse reaction to radiation therapy for cancer may be measured by an odds ratio (OR) and by the relative risk (RR). If P(R⁺) is the probability of developing an adverse reaction to radiation therapy for cancer for individuals with the risk profile (R) and P(R ) is the probability of developing an adverse reaction to radiation therapy for cancer for individuals without the risk profile, then the relative risk is the ratio of the two probabilities: RR=P(R⁺)/P(R ).

[0023] In case-control studies, however, direct measures of the relative risk often cannot be obtained because of sampling design. The odds ratio allows for an approximation of the relative risk for low- incidence diseases and can be calculated: 0R=(F7(l-F⁺))/(F /(l-F )), where F⁺ is the frequency of a risk profile in cases studies and F is the frequency of risk profile in controls. F⁺ and F can be calculated using the metabolite profile frequencies of the study.

[0024] The attributable risk (AR) can also be used to express an increased risk. The AR describes the proportion of individuals in a population exhibiting an adverse reaction to radiation therapy for cancer due to a specific member of the metabolite risk profile. AR may also be important in quantifying the role of individual components (specific member) in reaction etiology and in terms of the public health impact of the individual marker. The public health relevance of the AR measurement lies in estimating the proportion of cases of an adverse reaction to radiation therapy for cancer in the population that could be prevented if the profile or individual component were absent. AR may be determined as follows: AR=P_E(RR-1)/(P_E(RR-1)+1), where AR is the risk attributable to a profile or individual component of the profile, and P_E is the frequency of exposure to a profile or individual component of the profile within the population at large. RR is the relative risk, which can be approximated with the odds ratio when the profile or individual component of the profile under study has a relatively low incidence in the general population.

[0025] In one embodiment, the increased risk of a patient can be determined from p-values that are derived from association studies. Specifically, associations with specific profiles can be performed using regression analysis by regressing the metabolite profile with an adverse reaction to radiation therapy for cancer. In addition, the regression may or may not be corrected or adjusted for one or more factors.

The factors for which the analyses may be adjusted include, but are not limited to age, sex, weight, ethnicity, geographic location, fasting state, general health of the subject, alcohol or drug consumption, caffeine or nicotine intake and circadian rhythms, and the subject's Prostate Specific Antigen (PSA) status to name a few.

[0026] Increased risk can also be determined from p-values that are derived using logistic regression. Binomial (or binary) logistic regression is a form of regression which is used when the dependent is a dichotomy and the independents are of any type. Logistic regression can be used to predict a dependent variable on the basis of continuous and/or categorical independents and to determine the percent of variance in the dependent variable explained by the independents; to rank the relative importance of independents; to assess interaction effects; and to understand the impact of covariate control variables. Logistic regression applies maximum likelihood estimation after transforming the dependent into a "logit" variable (the natural log of the odds of the dependent occurring or not). In this way, logistic regression estimates the probability of a certain event occurring. These analyses are conducted with the program SAS.

[0027] SAS ("statistical analysis software") is a general purpose package (similar to Stata and SPSS) created by Jim Goodnight and N.C. State University colleagues. Ready-to-use procedures handle a wide range of statistical analyses, including but not limited to, analysis of variance, regression, categorical data analysis, multivariate analysis, survival analysis, psychometric analysis, cluster analysis, and nonparametric analysis.

[0028] Accordingly, select embodiments of the present invention comprise the use of a computer comprising a processor and the computer is configured or programmed to generate one or more metabolite profiles and/or to determine statistical risk. The methods may also comprise displaying the one or more profiles and/or risk profiles on a screen that is communicatively connected to the computer. In another embodiment, two different computers can be used: one computer configured or programmed to generate one or more metabolite profiles and a second computer configured or programmed to determine statistical risk. Each of these separate computers can be communicatively linked to its own display or to the same display.

[0029] As used herein, the phrase "metabolite profile" means the combination of a subject's metabolites found in the peripheral blood or portions thereof, such as but not limited to plasma or serum. The metabolite profile is a collection of measurements, such as but not limited to a quantity or concentration, for individual metabolites taken from a test sample of the subject. Examples of test samples or sources of components for the metabolite profile include, but are not limited to, biological fluids, which can be tested by the methods of the present invention described herein, and include but are not limited to whole blood, such as but not limited to peripheral blood, serum, plasma,

cerebrospinal fluid, urine, semen, lymph fluids, and various external secretions of the respiratory, intestinal and genitourinary tracts, tears, saliva, milk, white blood cells, myelomas and the like. Test samples to be assayed also include but are not limited to tissue specimens including normal and abnormal tissue.

[0030] Techniques to assay levels of individual components of the metabolite profile from test samples are well known to the skilled technician, and the invention is not limited by the means by which the components are assessed. In one embodiment, levels of the individual lipid components of the metabolite profile are assessed using mass spectrometry in conjunction with ultra-performance liquid chromatography (UPLC), high-performance liquid chromatography (HPLC), gas chromatography (GC), gas chromatography/mass spectroscopy (GC/MS), and UPLC to name a few. Other methods of assessing levels of the individual components include biological methods, such as but not limited to ELISA assays.

[0031] The assessment of the levels of the individual components of the metabolite profile can be expressed as absolute or relative values and may or may not be expressed in relation to another component, a standard an internal standard or another molecule of compound known to be in the sample. If the levels are assessed as relative to a standard or internal standard, the standard may be added to the test sample prior to, during or after sample processing.

[0032] To assess levels of the individual components of the metabolite profile, a sample is taken from the subject. The sample may or may not processed prior assaying levels of the components of the metabolite profile. For example, whole blood may be taken from an individual and the blood sample may be processed, e.g., centrifuged, to isolate plasma or serum from the blood. The sample may or may not be stored, e.g., frozen, prior to processing or analysis.

[0033] As used herein, the term "adverse reaction" as it relates to radiation therapy for cancer is cancer recurrence after receiving the therapy, rectal toxicity or urinary toxicity. As used herein, the recurrence of cancer, for example prostate cancer, can occur any time after the subject receives the therapy and would be considered in remission or even "cured." As used herein, a subject is "cured" of cancer, for example prostate cancer, at least initially, if the subject is in remission and remains cancer free for at least 5 years from the time of the prescribed therapy to remove or destroy the diseased tissue.

[0034] As used herein, the rectal toxicity should be attributable to the subject receiving radiation therapy for prostate cancer. As is well-understood, rectal toxicity can occur as a result of other factors besides receiving radiation. Thus, the methods of the present invention include determining if a subject has or has a risk of developing rectal toxicity from a source other than receiving radiation therapy for cancer, for example prostate cancer. If the subject is free of rectal toxicity prior to receiving radiation therapy for cancer, the methods of the present invention can be performed on the subject that is a candidate for receiving radiation therapy for cancer.

[0035] As used herein, the urinary toxicity should be attributable to the subject receiving radiation therapy for cancer. As is well-understood, urinary toxicity can occur as a result of other factors besides receiving radiation. Thus, the methods of the present invention include determining if a subject has or has a risk of developing urinary toxicity from a source other than receiving radiation therapy for cancer. If the subject does not have urinary toxicity prior to receiving radiation therapy for cancer, the methods of the present invention can be performed on the subject that is a candidate for receiving radiation therapy for cancer, for example prostate cancer.

[0036] When assessing if a subject is at risk of developing cancer recurrence after receiving radiation therapy for cancer, for example prostate cancer, individual components of the metabolite profile include but are not limited to phosphatidylcholine acyl-alkyl C40:l (PC ae C40:l), phosphatidylcholine acyl-alkyl C40:6 (PC ae C4Q:6), phosphatidylcholine acyl-alkyl C42:l (PC ae C42:l), Arginine, Adenosine, phosphatidylcholine diacyl C26:0 (PC aa C26:0), phosphatidylcholine acyl-alkyl C36:2 (PC ae C36:2) and lysophosphatidylcholine acyl C26:l (LysoPC a C26:l). Metabolite C species, e.g., C3, denote acyicarnitines (ACs). Phosphocholine (PC) metabolites display combined numbers of carbon atoms for their two acyl groups (snl and sn2 positions), e.g., C38, whereas the combined number of double bonds (unsaturation) is displayed after the colon, e.g., C38:6. Acyl group linkages to choline backbone for PCs feature ester (a) or ether (e) linkage, e.g., PC ae C42:l.

[0037] In one embodiment, when treating a subject having cancer or determining the risk of cancer recurrence, for example prostate cancer recurrence after receiving radiation therapy for cancer, the individual levels of each of the metabolites are lower than those compared to normal levels. In another embodiment, one, two, three, four, five, six, seven or eight of the levels of each of the metabolites are lower than normal levels while others, if any, are higher than normal levels. In another embodiment, the individual levels of each of the metabolites are higher than those compared to normal levels. In another embodiment, one, two, three, four, five, six, seven or eight of the levels of each of the metabolites are higher than normal levels while others, if any, are higher than normal levels.

[0038] The levels of depletion or augmentation of the metabolites compared to normal levels can vary when treating a subject having cancer or determining the risk of cancer recurrence after receiving radiation therapy for cancer, for example prostate cancer. In one embodiment, the levels of any one or more of the metabolites is at least 1.05, 1.1, 1.2, 1.3, 1.4, 1.5, 1.6, 1.7, 1.8, 1.9, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20 lower than normal levels. For the purposes of the present invention, the number of "times" the levels of a metabolite is lower or higher over normal can be a relative or absolute number of times. In the alternative, the levels of the metabolites may be normalized to a standard and these normalized levels can then be compared to one another to determine if a metabolite is lower or higher.

[0039] For the purposes of the present invention, when treating a subject having cancer or determining the risk of cancer recurrence after receiving radiation therapy for cancer, the metabolite profile comprises at least two, three, four, five, six, seven or all eight of the metabolites listed above. If two metabolites are used in generating the metabolite profile, any combination of the two listed above can be used. If three metabolites are used in generating the metabolite profile, any combination of three of the metabolites listed above can be used. If four metabolites are used in generating the metabolite profile, any combination of four of the metabolites listed above can be used. If five metabolites are used in generating the metabolite profile, any combination of five of the metabolites listed above can be used. If six metabolites are used in generating the metabolite profile, any combination of six of the metabolites listed above can be used. If seven metabolites are used in generating the metabolite profile, any combination of seven of the metabolites listed above can be used. Of course, all eight metabolites can be used in generating the metabolite profile to treat a subject having cancer or to determine risk of cancer recurrence after receiving radiation therapy for cancer.

[0040] When assessing if a subject is at risk of developing rectal toxicity after receiving radiation therapy for cancer, individual components of the metabolite profile include but are not limited to phosphatidylcholine acyl-alkyl C36:l (PC ae C36:l), phosphatidylcholine acyl-alkyl C42:G (PC ae C42:0), sphingomyelin C2Q:2 (SM C20:2), O-Acetyl-L-Carnitine, 2-Aminoadipic acid and the ratio of

chenodeoxycholic acid to deoxycholic acid (CDCA/DCA).

[0041] In one embodiment, when treating a subject having cancer or determining the risk of developing rectal toxicity after receiving radiation therapy for cancer, the individual levels of each of the metabolites are lower than those compared to normal levels. In another embodiment, one, two, three, four, five or six of the levels of each of the metabolites are lower than normal levels while others, if any, are higher than normal levels. In another embodiment, the individual levels of each of the metabolites are higher than those compared to normal levels. In another embodiment, one, two, three, four, five or six of the levels of each of the metabolites are higher than normal levels while others, if any, are higher than normal levels.

[0042] The levels of depletion or augmentation of the metabolites compared to normal levels can vary when treating a subject having cancer or determining the risk of developing rectal toxicity after receiving radiation therapy for cancer. In one embodiment, the levels of any one or more of the metabolites is at least 1.05, 1.1, 1.2, 1.3, 1.4, 1.5, 1.6, 1.7, 1.8, 1.9, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18,

19, 20 lower than normal levels. For the purposes of the present invention, the number of "times" the levels of a metabolite is lower or higher over normal can be a relative or absolute number of times. In the alternative, the levels of the metabolites may be normalized to a standard and these normalized levels can then be compared to one another to determine if a metabolite is lower or higher.

[0043] For the purposes of the present invention, when treating a subject having cancer or determining the risk of developing rectal toxicity after receiving radiation therapy for cancer, the metabolite profile comprises at least two, three, four, five or all six of the metabolites listed above. If two metabolites are used in generating the metabolite profile, any combination of the two listed above can be used. If three metabolites are used in generating the metabolite profile, any combination of three of the metabolites listed above can be used. If four metabolites are used in generating the metabolite profile, any combination of four of the metabolites listed above can be used. If five metabolites are used in generating the metabolite profile, any combination of five of the metabolites listed above can be used. Of course, all six metabolites can be used in generating the metabolite profile to treat a subject having cancer or to determine risk of developing rectal toxicity after receiving radiation therapy for cancer.

[0044] When assessing if a subject is at risk of developing urinary toxicity after receiving radiation therapy for cancer, individual components of the metabolite profile include but are not limited to lysophosphatidylcholine acyl C20:4 (LysoPC a C20:4), phosphatidylcholine diacyi C34:2 (PC aa C34:2), phosphatidylcholine acyl-alkyl C4G:5 (PC ae C4Q:5), phosphatidylcholine diacyl C36:l (PC aa C36:l), phosphatidylcholine diacyl C40:5 (PC aa C4G:5), phosphatidylcholine acyl-alkyl C40:3 (PC ae C40:3), lysophosphatidylcholine acyl C18:2 (LysoPC a C18:2), lysophosphatidylcholine acyl C20:3 (LysoPC a C20:3) and lysophosphatidylcholine acyl C14:0 (LysoPC a C14:0).

[0045] In one embodiment, when treating a subject having cancer or determining the risk of developing urinary toxicity after receiving radiation therapy for cancer, the individual levels of each of the metabolites are lower than those compared to normal levels. In another embodiment, one, two, three, four, five, six, seven, eight or nine of the levels of each of the metabolites are lower than normal levels while others, if any, are higher than normal levels. In another embodiment, the individual levels of each of the metabolites are higher than those compared to normal levels. In another embodiment, one, two, three, four, five, six, seven, eight or nine of the levels of each of the metabolites are higher than normal levels while others, if any, are higher than normal levels.

[0046] The levels of depletion or augmentation of the metabolites compared to normal levels can vary when treating a subject having cancer or determining the risk of developing urinary toxicity after receiving radiation therapy for cancer. In one embodiment, the levels of any one or more of the metabolites is at least 1.05, 1.1, 1.2, 1.3, 1.4, 1.5, 1.6, 1.7, 1.8, 1.9, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20 lower than normal levels. For the purposes of the present invention, the number of "times" the levels of a metabolite is lower or higher over normal can be a relative or absolute number of times. In the alternative, the levels of the metabolites may be normalized to a standard and these normalized levels can then be compared to one another to determine if a metabolite is lower or higher.

[0047] For the purposes of the present invention, when treating a subject having cancer or determining the risk of developing urinary toxicity after receiving radiation therapy for cancer, the metabolite profile comprises at least two, three, four, five, six, seven, eight or all nine of the metabolites listed above. If two metabolites are used in generating the metabolite profile, any combination of the two listed above can be used. If three metabolites are used in generating the metabolite profile, any combination of three of the metabolites listed above can be used. If four metabolites are used in generating the metabolite profile, any combination of four of the metabolites listed above can be used. If five metabolites are used in generating the metabolite profile, any combination of five of the metabolites listed above can be used. If six metabolites are used in generating the metabolite profile, any combination of six of the metabolites listed above can be used. If seven metabolites are used in generating the metabolite profile, any combination of seven of the metabolites listed above can be used. If eight metabolites are used in generating the metabolite profile, any combination of eight of the metabolites listed above can be used. Of course, all nine metabolites can be used in generating the metabolite profile to treat a subject having cancer or to determine risk of developing urinary toxicity after receiving radiation therapy for cancer.

[0048] Tables 1 through 3 below list exemplary analysis of the metabolites used for each specific adverse reaction to receiving radiation therapy for cancer, for example prostate cancer. Flerein, a "mean fold change" of one (1) indicates no change while values less than one indicate a negative change in the diagnostic group as compared to the normal control (NC). Flerein, values greater than one indicate a positive change in the diagnostic group compared to NC. [0049] Techniques to assay levels of individual components of any non-lipid component of the metabolite profile from test samples are well known to the skilled technician, and the invention is not limited by the means by which the components are assessed. In one embodiment, levels of the individual components of the non-lipid portion of the profile are assessed using quantitative arrays, PCR, Northern Blot analysis, Western Blot analysis, mass spectroscopy, high-performance liquid

chromatography (HPLC, high performance gas chromatography (HPGC) and the like. Other methods of assessing levels of the individual components include biological methods, such as but not limited to ELISA assays. To determine levels of metabolites, it is not necessary that an entire metabolite, e.g., a full length protein or an entire RNA transcript, be present or fully sequenced. In other words, determining levels of, for example, a fragment of protein being analyzed may be sufficient to conclude or assess that an individual component of the metabolite profile, including the lipid and non-lipid portions of the metabolite profile, being analyzed is increased or decreased. Similarly, if, for example, arrays or blots are used to determine metabolite levels, the presence/absence/strength of a detectable signal may be sufficient to assess levels of metabolites.

[0050] The subject's metabolite profile is compared to the profile that is deemed to be a normal metabolite profile. To establish the metabolite profile of a normal individual, an individual or group of individuals may be first assessed for the lack of any observable or noticeable adverse reactions to radiation therapy for cancer, for example prostate cancer. Once established, the metabolite profile of the individual or group of individuals can then be determined to establish a "normal metabolite profile." In one embodiment, a normal metabolite profile can be ascertained from the same subject when the subject is deemed to not have cancer, for example prostate cancer, and is displaying no signs (clinical or otherwise) of cancer. In one embodiment, a "normal" metabolite profile is assessed in the same subject from whom the sample is taken prior to the onset of measureable, perceivable or diagnosed sign of cancer. That is, the term "normal" with respect to a metabolite profile can be used to mean the subject's baseline metabolite profile prior to the onset of cancer or receiving radiation therapy for cancer. The metabolite profile can then be reassessed periodically and compared to the subject's baseline metabolite profile.

[0051] In another embodiment, a normal metabolite profile is assessed in a sample from a different subject or patient (from the subject being analyzed) and this different subject does not have or is not suspected of having cancer or showed no observable adverse reaction to receiving radiation therapy for cancer. In still another embodiment, the normal metabolite profile is assessed in a population of healthy individuals, the constituents of which do not have cancer or showed no observable adverse reaction to receiving radiation therapy for cancer. Thus, the subject's metabolite profile can be compared to a normal metabolite profile generated from a single normal sample or a metabolite profile generated from more than one normal sample.

[0052] Of course, measurements of the individual components, e.g., concentration, ratio, log ratios etc., of the normal metabolite profile can fall within a range of values, and values that do not fall within this "normal range" are said to be outside the normal range. These measurements may or may not be converted to a value, number, factor or score as compared to measurements in the "normal range." For example, a measurement for a specific metabolite that is below the normal range, may be assigned a value or -1, -2, -3, etc., depending on the scoring system devised.

[0053] In one embodiment, the "metabolite profile value" can be a single value, number, factor or score given as an overall collective value to the individual molecular components of the profile, or to the categorical components, i.e., a phosphatidylcholine portion, a biogenic amine portion and/or an amino acid portion. For example, if each component is assigned a value, such as above, the metabolite value may simply be the overall score of each individual or categorical value. For example, if five of the components of the 8PMI metabolite profile are phosphatidylcholine, and three of those components are assigned values of "-2" and two are assigned values of "-1," the phosphatidylcholine portion of the metabolite profile in this example would be -8, with a normal value being, for example, "0." In this manner, the metabolite profile value could be a useful single number or score, the actual value or magnitude of which could be an indication of the actual risk of developing an adverse reaction to radiation therapy for cancer, e.g., the "more negative" the value, the greater the risk of developing an adverse reaction to radiation therapy for cancer.

[0054] In another embodiment the "metabolite profile value" can be a series of values, numbers, factors or scores given to the individual components of the overall profile. In another embodiment, the "metabolite profile value" may be a combination of values, numbers, factors or scores given to individual components of the profile as well as values, numbers, factors or scores collectively given to a group of components, such as a phosphatidylcholine portion, an acylcarnitine portion, a biogenic amine portion and/or an amino acid portion. In another example, the metabolite profile value may comprise or consist of individual values, number, factors or scores for specific component as well as values, numbers, factors or scores for a group on components. [0055] In another embodiment individual values from the metabolites can be used to develop a single score, such as a "combined metabolite index," which may utilize weighted scores from the individual component values reduced to a diagnostic number value. The combined metabolite index may also be generated using non-weighted scores from the individual component values. When the "combined metabolite index" exceeds a specific threshold level, determined by a range of values developed similarly from control subjects, the individual has a high risk, or higher than normal risk, of developing an adverse reaction to radiation therapy for cancer, whereas maintaining a normal range value of the "combined metabolite index" would indicate a low or minimal risk of developing an adverse reaction to radiation therapy for cancer. In this embodiment, the threshold value would be or could be set by the combined metabolite index from one or more normal subjects.

[0056] In another embodiment, the value of the metabolite profile can be the collection of data from the individual measurements and need not be converted to a scoring system, such that the "metabolite profile value" is a collection of the individual measurements of the individual components of the profile.

[0057] In specific embodiments, a subject is diagnosed of having an increased risk of an adverse reaction to radiation therapy for cancer if the subject's nine, eight, seven, six, five, four, three, two or even one of the metabolites herein are at abnormal levels.

[0058] If it is determined that a subject has an increased risk of an adverse reaction to radiation therapy for cancer, the attending health care provider may subsequently prescribe or institute a treatment program or prescribe a different treatment for cancer. In this manner, the present invention also provides for methods of screening individuals as candidates for treatment of an adverse reaction to radiation therapy for cancer. The attending healthcare worker may begin treatment, based on the subject's metabolite profile, before there are perceivable, noticeable or measurable signs of an adverse reaction to radiation therapy for cancer in the individual.

[0059] Similarly, the invention provides methods of treating a subject having cancer, for example prostate cancer. The treatment methods include obtaining a subject's metabolite or composite profile as defined herein and prescribing a treatment regimen to the subject if the metabolite and/or composite profile indicate that the subject is at risk of suffering from an adverse reaction to radiation therapy for cancer, for example prostate cancer. [0060] Suitable radiation therapies for prostate cancer are well-known, and the methods disclosed and described herein are not dependent on the specific type of radiation therapy for cancer.

[0061] The objective of this study was to employ a high through put metabolomics approach for delineating a biomarker panel predictive of radiation induced adverse effects in patients treated for prostate cancer. Such biomarkers aid in early detection of tissue toxicity in cancer patients, so that intervention can be initiated early in patients at risk. Metabolite signatures were developed for prediction of adverse responses to radiation therapy in a cohort of patients undergoing stereotactic body radiation therapy (SBRT) for prostate cancer. Subsets of these patients developed urinary toxicity (UT) (N=8) and rectal toxicity (RT) (N=6).

[0062] Individuals sensitive to radiation toxicities carry a biochemical fingerprint that can be identified as a distinct plasma profile. Analysis of banked plasma samples was correlated with clinical outcomes and symptom assessment to identify markers of genitourinary and gastrointestinal late effects for future validation in a larger clinical population. Using stable isotope labeling multiple reaction monitoring based molecular phenotyping approach, high accuracy predictive algorithms were developed for recurrence, urinary toxicity and rectal toxicity episodes in this cohort. This analysis was performed using the pre-radiation samples from this set of patients. The panels can be useful to predict late effects of radiation therapy and lay the foundation for the development of strategies by which toxicity may be detected at an early stage and mitigated with intervention therapies.

[0063] All patents and publications mentioned in this specification are indicative of the level of those skilled in the art to which the invention pertains. All patents and publications cited herein are incorporated by reference to the same extent as if each individual publication was specifically and individually indicated as having been incorporated by reference in its entirety.

Examples

[0064] Patients were enrolled at MedStar-Georgetown University Hospital into IRB protocol 2012- 1175; an approved quality of life clinical trial. The protocol permits longitudinal collection of clinical samples, symptom monitoring and quality of life data which have contributed to interim published reports of clinical outcomes including GU and Gl acute and late effects. This study population was a part of ongoing recruitment of PC patients coming in through the referral network to MedStar-Georgetown University Hospital (MGUH). The study participants included men of varying races aged 35-70 years, residing in Washington DC and surrounding areas, who were diagnosed with localized prostate cancer by biopsy. Patients were recruited from the Departments of Radiation Medicine and Urology at the MGUH.

[0065] Technical aspects of stereotactic body radiation therapy (SBRT) treatment planning and radiation delivery have been previously described. Briefly, ultrasound guided placement of gold fiducial markers is performed 2 or more weeks prior to thin cut CT and high resolution MRI imaging. The clinical target volume (CTV) includes the prostate and proximal seminal vesicles, to the bifurcation. The prescribed doses of 35-36.25 Gy are delivered in five fractions of 7-7.25 Gy over 2 weeks. Symptom management medications were prescribed based on the treating physician's clinical judgment and urinary symptoms were managed with alpha-adrenergic antagonists and bothersome bowel symptoms were managed with anti-diarrheal medication (loperamide).

[0066] Stable isotope labeled multiple reaction monitoring mass spectrometry (SID-MRM-MS) was used for quantitation of 350 metabolites. Metabolite extraction was performed using 25 pL of plasma sample was mixed with 175 pL of 40% acetonitrile in 25% methanol and 35% water containing internal standards [stable isotope labeled). The samples were incubated on ice for 10 minutes and centrifuged at 14,000 rpm at 4°C for 20 minutes. The supernatant was transferred to a fresh tube and used for UPLC-QQQ -MS analysis. Each plasma sample (2 pL) was injected onto a reverse-phase CSH C18 1.7pM 2.1 x 100 mm column using an Acquity UPLC online with a triple quadrupole MS (Xevo TQ-S, Waters Corporation, USA) G2-QTOF system operating in the MRM mode.

[0067] Following data pre-processing and ion annotation, the m/z values of the measured metabolites were normalized with log transformation that stabilizes variance, followed by quantile normalization to achieve uniform empirical distribution of intensities (measure of metabolite abundance) across samples.

[0068] After the data pre-processing and ion annotation, the m/z values of the measured metabolites from tissue samples were normalized with log transformation that stabilized the variance, followed by quantile normalization to make the empirical distribution of intensities the same across samples.

Differential expression between various patient groups was assessed using analysis of variance

(ANOVA). Multiple comparisons were adjusted using the Bonferroni correction. The heat maps were generated for the significant metabolites using the log 2 transformed values of fold changes and hierarchically clustered by Pearson correlation. [0069] Among these differentially expressed metabolites identified, feature selection was performed using a regularized learning technique, which uses the least absolute shrinkage and selection operator (LASSO) penalty. Differential expression between various patient groups was assessed using t-test constrained by p-value <0.05. Among these differentially expressed metabolites, each m/z value was scored for annotation against the HMDB, Metlin, MMCD and Lipid Maps databases within a 5 ppm mass tolerance.

[0070] After the data pre-processing and ion annotation, the m/z values of the measured metabolites from plasma samples were normalized with log transformation that stabilized the variance, followed by quantile normalization to make the empirical distribution of intensities the same across samples.

Differential expression between various patient groups was assessed using analysis of variance (ANOVA). Multiple comparisons were adjusted using the Bonferroni correction.

[0071] Among these differentially expressed metabolites identified, feature selection was performed using a regularized learning technique, which uses the least absolute shrinkage and selection operator (LASSO) penalty. The classification performance of the biomarker panel was assessed using the area under the ROC (receiver operating characteristic) curve (AUC). A logistic model with the selected biomarker panel was, adjusting for age, race, and PSA, applied, and the ROC curve was understood as a plot of the probability of classifying correctly positive samples against the rate of incorrectly classifying true negative samples.

[0072] Feature selection was performed using a ROC regularized learning technique, which uses the least absolute shrinkage and selection operator (LASSO) penalty. The regularization path was obtained over a grid of values for the tuning parameter lambda through 10-fold cross-validation. Then the optimal value of the tuning parameter lambda, obtained by the cross-validation procedure was used to fit the model. Finally, all the features with non-zero coefficients were retained as the candidate biomarker panel. This technique is known to reduce overfitting and variance in classification.

[0073] The classification performance of the biomarker panel was assessed using the area under the ROC (receiver operating characteristic) curve (AUC). The ROC curve can be understood as a plot of the probability of classifying correctly positive samples against the rate of incorrectly classifying true negative samples. Therefore, the AUC measure of an ROC plot is a measure of predictive accuracy. Due to the perfect separation for the classification, the panel was also evaluated using the hidden logistic regression model with the maximum estimated likelihood (MEL) estimator, and the AUC scores were similar. The individual markers were also analyzed, and the AUC score was estimated for the regression with each marker, all of them showing high discriminative value for distinguishing "high" vs "low" and "recur" vs "low" patient group, to rule out correlation with the patients' hormone therapy status.

[0074] Leveraging the longitudinally collected clinical outcome data, a retrospective outcome analysis was performed on this cohort to determine if there was a differentiation between the low risk, high risk and recurrence categories of patients by comparing their metabolic profiles at baseline. The abundance measurements for metabolites were expressed as intensity units that were initially normalized using log transformation and quantile normalization.

[0075] The cohort of SBRT treated patients is shown in Figure 1. Of 105 patients, 10 developed biochemical recurrences with an average time of 18 months. To develop a predictive panel of recurrence, the pre-radiation plasma metabolite profiles were compared to a sub-set of patients who remained cancer free during this time. The panel was adjusted for age, PSA levels and Gleason's grade.

A sub-set analyses was performed in patients not receiving hormone therapy to rule out the influence of hormone therapy on the marker panel, in high risk and recurrence patient categories.

[0076] Example 2

[0077] An eight metabolite panel with an AUC > 98% (Figure 1, Panels A and B) was able to accurately discriminate between those who reported recurrence episodes (N=10) as compared to those who remained cancer free (N=20) during the follow up period. Differential expression of these biomarkers in the two comparative groups was visualized as box plots (Figure 2). Finally, the resulting combined classifier allowed the development of a plasma metabolite index (PMI), which was obtained by mapping the 8 metabolites to the hyperplane that maximizes the margin between different groups (Figure 3). Based on the model with the biomarker panel, the eight-member PMI (8PMI) helps identify the degree to which individuals are 'at-risk' of an outcome (recurrence, rectal or bladder late effects). The natural log of odds in the model was transformed to 0-100 index value using a linear mapping.

[0078] The 8PMI, listed in Table 1 below, was able to accurately discriminate between those who reported recurrence episodes (N=10) as compared to those who remained cancer free (N=20) during the follow up period.

Table 1. Biomarker panel predictive of tumor recurrence in PC patients

[0079] Example 3

[0080] A predictive biomarker panels of adverse outcomes of radiation therapy was also developed. A retrospective outcome analysis was performed on a sub-set of patients who reported rectal toxicity (N=6) by comparing their pre-radiation plasma metabolite profiles with a randomly selected sub-set of prostate cancer patients who remained normal during the post-radiation monitoring period. A six metabolite biomarker panel yielded an AUC of 98.3% (Figure 4, Panels A and B). The relative abundance of biomarkers in the two comparative groups is illustrated in Figure 5 while the six-member PMI (6PMI) (Figure 6), helped stratify patients who later developed rectal toxicity from those who did not develop radiation induced adverse symptoms, underscoring the power of this technology.

[0081] A retrospective outcome analysis was performed on a sub-set of patients who reported rectal toxicity after radiation therapy (N=6). A six-member metabolite biomarker panel detailed in Table 2 yielded an AUC of 98.3% (Figure 4).

Table 2. Six metabolite panel predictive of rectal toxicity

[0082] Example 4

[0083] Using LASSO, a nine metabolite panel with AUC >98% (Figure 7, Panels A and B) was established for patients' pre-radiation profiles and those patients who reported urinary toxicity. The pattern of expression of individual markers was visualized as a box plot (Figure 8), while a nine-member PMI (9PMI) that was developed using logistic regression helped stratify the two comparative groups unambiguously (Figure 9).

[0084] The classification algorithm was developed that would be predictive of urinary flare (N=8). It is noteworthy that some of the patients who developed urinary flare also developed tumor recurrence. Table 3 lists the nine-member metabolite panel that discriminates the sub-set of patients from those who did not. The ROC curve showed an AUC of 98.1% (Figure 7).

Table 3: Biomarker panel predictive of urinary flare

Claims

What is Claimed is:

1. A method of determining if a subject has an increased risk of having an adverse reaction to receiving radiation therapy for cancer, the method comprising a) analyzing at least one sample from the subject to determine a value of the subject's

metabolite profile, and b) comparing the value of the subject's metabolite profile with the value obtained from

subjects determined to define a normal metabolite profile, to determine if the subject's metabolite profile is altered compared to a normal metabolite profile, wherein a difference in the value of the subject's metabolite profile compared to those defined as having a normal metabolite profile is indicative that the subject has an increased risk of having an adverse reaction to receiving radiation therapy for cancer.

2. The method of claim 1, wherein the metabolite profile comprises one or more components selected from the group consisting of PC ae C4Q:1, PC ae C40:6, PC ae C42:l, Arginine, Adenosine, PC aa C26:0, PC ae C36:2 and LysoPC a C26:l.

3. The method of claim 2, wherein the adverse reaction is tumor recurrence after receiving the radiation therapy.

4. The method of claim 1, wherein the metabolite profile comprises one or more components selected from the group consisting of PC ae C36:l, PC ae C42:0, SM C2Q:2, O-Acety!-L-Carnitine, 2-Aminoadipic acid and CDCA/DCA.

5. The method of claim 4, wherein the adverse reaction to the administered radiation therapy is proctitis.

6. The method of claim 1, wherein the metabolite profile comprises one or more components selected from the group consisting of LysoPC aC2Q:4, PC aa C34:2, PC ae C40:5, PC aa C36:l, PC aa C40:5, PC ae C40:3, LysoPC a C18:2, LysoPC a C20:3 and LysoPC a C14:Q.

7. The method of claim 6, wherein the adverse reaction to the administered radiation therapy is urinary toxicity.

8. The method of any of claims 1 to 7, wherein the normal metabolite profile comprises the subject's metabolite profile prior to receiving radiation treatment for cancer.

9. The method of any of claim 1 to 7, wherein the normal metabolite profile comprises a

metabolite profile generated from a population of individuals that did not have tumor recurrence after receiving radiation treatment for cancer.

10. A method of treating a subject having cancer, the method comprising a) analyzing at least one sample from the subject to determine a value of the subject's

subjects determined to define a normal metabolite profile, to determine if the subject's metabolite profile is altered compared to a normal metabolite profile, c) administering radiation therapy to the subject, if the value of the subject's metabolite profile compared to those defined as having a normal metabolite profile is indicative that the subject does not have an increased risk of having an adverse reaction to the administered radiation therapy.

11. The method of claim 10, wherein the metabolite profile comprises one or more components selected from the group consisting of PC ae C40:l, PC ae C4Q:6, PC ae C42:l, Arginine, Adenosine, PC aa C26:Q, PC ae C36:2 and LysoPC a C26:l.

12. The method of claim 11, wherein the adverse reaction to the administered radiation therapy is tumor recurrence after having received the administered radiation therapy.

13. The method of claim 10, wherein the metabolite profile comprises one or more components selected from the group consisting of PC ae C36:l, PC ae C42:0, SM C20:2, G-Acetyl-L-Carnitine, 2-Aminoadipic acid and CDCA/DCA.

14. The method of claim 13, wherein the adverse reaction to the administered radiation therapy is rectal toxicity.

15. The method of claim 10, wherein the metabolite profile comprises one or more components selected from the group consisting of LysoPC aC20:4, PC aa C34:2, PC ae C4G:5, PC aa C36:l, PC aa C40:5, PC ae C40:3, LysoPC a C18:2, LysoPC a C20:3 and LysoPC a C14:0.

16. The method of claim 15, wherein the adverse reaction to the administered radiation therapy is urinary toxicity.

17. A method of measuring levels of metabolites in a subject, the method comprising obtaining a sample from a subject at a first time point and determining individual levels of PC ae C4Q:1, PC ae C4Q:6, PC ae C42:l, Arginine, Adenosine, PC aa C26:0, PC ae C36:2 and LysoPC a C26:l at said first time point.

18. A method of measuring levels of metabolites in a subject, the method comprising obtaining a sample from a subject at a first time point and determining individual levels of PC ae C36:l, PC ae C42:0, SM C20:2, O-Acetyi-L-Carnitine, 2-Aminoadipic acid and CDCA/DCA at said first time point.

19. A method of measuring levels of metabolites in a subject, the method comprising obtaining a sample from a subject at a first time point and determining individual levels LysoPC aC20:4, PC aa C34:2, PC ae C40:5, PC aa C36:l, PC aa C40:5, PC ae C4Q:3, LysoPC a C18:2, LysoPC a C2G:3 and LysoPC a C14:Q at said first time point.