WO2020099683A1

WO2020099683A1 - Metabolites for use in ovarian cancer detection and diagnosis

Info

Publication number: WO2020099683A1
Application number: PCT/EP2019/081689
Authority: WO
Inventors: Therese Koal; Barbara BURWINKEL; Yuan BAO-WEN; Udo Müller; Simon SCHAFFERER
Original assignee: Biocrates Life Sciences Ag
Priority date: 2018-11-16
Filing date: 2019-11-18
Publication date: 2020-05-22

Abstract

The present invention relates to new metabolite panels for assessing ovarian cancer. Moreover, the present invention relates to a method for assessing ovarian cancer from a subject to be examined, and to a kit for carrying out the method.

Description

Metabolites for use in Ovarian Cancer Detection and Diagnosis

The present invention relates to the use of a combination of metabolites contained in a blood sample of a mammalian subject in the in vitro or ex vivo detection, diagnosis, prognosis and/or therapy monitoring of ovarian cancer. Said combination of metabolites comprises at least one amino acid, at least one biogenic amine, at least one acylcamitine, at least one lysophosphatidylcholine, at least one phosphatidylcholine, at least one sphingolipid and optionally further component(s).

The present invention further relates to an in vitro or ex vivo method for the detection, diagnosis, prognosis and/or therapy monitoring of ovarian cancer, comprising providing a blood sample of a mammalian subject; determining the amount of a combination of metabolites; and comparing said amount with a control sample. The present invention further relates to a kit for performing said method.

BACKGROUND OF THE INVENTION

Ovarian cancer is the seventh most frequently detected cancer among women worldwide and the second most mortality cancer among gynecological cancers (Torre et al., 2015). Epithelial ovarian cancer is the most common pathologic subtype of ovarian cancer (Reid et al., 2017). Due to the lack of well-defined clinical symptoms as well as highly sensitive and specific biomarkers, the disease typically presents at advanced-stage and the 5-year relative survival rate is only 29% (Noone AM et al.). This is a huge contrast with the localized tumor when the 5-year survival rate is 92% among the US population (Noone AM et al.). Therefore, there is an urgent need to achieve ovarian cancer early detection and diagnosis.

Currently, the mostly used tests to screen for ovarian cancer are transvaginal ultrasound (TVUS), blood based biomarkers cancer antigen 125 (CA125) and human epididymis protein 4 (HE4) or their combinations (Molina et al., 201 1; Rauh-Hain et al., 2011). However, all of these detection methods have limitations. Though, transvaginal ultrasound has higher diagnostic accuracy compared with CA125 and HE4, however, due to its invasive and expensive nature and high false positive rate, it is not the perfect candidate (Rauh-Hain et al., 2011). Cancer biomarkers, CA125 and HE4, have been clinically used for the diagnosis of ovarian cancer, but they have been known for poor sensitivity and specificity since established, especially in early diagnosis (Fong et al., 2011). Metabolomics is the newest member of omics family, which is based on the high-throughput identification and quantification of small molecule metabolites and their interactions within biological networks (German et al., 2005). Due to the rapid development of metabolomics technology, our understanding of systems biology is advancing all the time (Wishart, 2007). Metabolites are the closest substances to physiological phenotype, with vast chemical and physical diversity, and highly dynamic (Dunn et al., 2011). Currently, no single analytical method can measure the concentrations of all metabolites, the findings of different studies differ as regard to ovarian cancer research, which may be related to population differences, the diversity of specimen types, metabolomic analytical methods and/or sample sizes (Turkoglu et al., 2016). Thus, it is urgent to identify a panel of biomarkers that are easy to be detected, could indicate the risk of cancer at early stage, or could identify individuals who are at high risk of developing cancer.

Therefore, it is an object of the invention to provide panels of small molecule biomarkers indicative of cancer, and methods for using the biomarkers for the diagnosis of subjects that have cancer, or that have an increased risk for developing cancer. It is still another obj ect of the invention to provide methods for detecting changes in plasma metabolites that are predictive of ovarian cancer.

US20120004854A1 describes a spectrum of metabolic biomarkers for ovarian cancer in a mammalian subject, the metabolic biomarker panels contain different compositions and numbers of metabolites, which leading to different levels of accuracy, specificity, and/or sensitivity.

All in all, development of new plasma/serum biomarkers to improve ovarian cancer and metastatic ovarian cancer early detection and further for prognosis track, which would have higher diagnostic and life-long surveillance accuracy, is of considerable clinical importance.

Thus, there is a need for novel as well as additional diagnostic, predictive and prognostic markers that aid to distinguish early stage group, to distinguish high risk group, to monitor therapy and to predict outcome. Such markers would help to choose better screening and treatment strategies.

SUMMARY OF THE INVENTION

According to the present invention this object is solved by the use of a combination of metabolites contained in a sample of a mammalian subject, the combination of metabolites comprising or consisting of 1) at least one amino acid selected from the group of alanine (Ala), asparagine (Asn), citrulline (Cit), glutamate (Glu), tryptophan (Trp), and

2) at least one biogenic amine selected from the group of asymmetric dimethylarginine (ADMA), c/s -4 - h y d ro x y p ro 1 i n c (c4-OH-Pro), symmetric dimethylarginine (SDMA),

3) at least one acylcamitine selected from the group of acylcarnitine C4:0, acylcamitine C16: l ,

4) lysophosphatidylcholine Cl 8:2,

5) at least one phosphatidylcholine selected from the group of PC(34:5), PC(35: 1), PC(42:4), PC-O(30:0), PC-0(32:3), and

6) at least one sphingolipid selected from SM(32:2), SM(42: 1)

in the in vitro or ex vivo detection, diagnosis, prognosis and/or therapy monitoring of ovarian cancer.

Hence, the present invention refers to a combination of metabolites as defined above for use in the assessment of ovarian cancer.

According to the present invention this object is solved by an in vitro or ex vivo method for the detection, diagnosis, prognosis and/or therapy monitoring of ovarian cancer, comprising the following steps:

(a) providing a sample of a mammalian subject,

(b) determining the amount of a combination of metabolites as defined above (1-6), and

(c) comparing said amount determined in (b) with a control sample.

Hence, the present invention refers to a combination of metabolites as defined above for use in the assessment of ovarian cancer, wherein the assessment comprises detection, diagnosis, prognosis and/or therapy monitoring of ovarian cancer.

According to the present invention this object is solved by a kit for performing the in vitro or ex vivo method for the detection, diagnosis, prognosis and/or therapy monitoring of ovarian cancer according to the present invention,

comprising a device having one or more wells and one or more inserts impregnated with at least one internal standard. DESCRIPTION OF THE PREFERRED EMBODIMENTS OF THE INVENTION

Before the present invention is described in more detail below, it is to be understood that this invention is not limited to the particular methodology, protocols and reagents described herein as these may vary. It is also to be understood that the terminology used herein is for the purpose of describing particular embodiments only, and is not intended to limit the scope of the present invention which will be limited only by the appended claims. Unless defined otherwise, all technical and scientific terms used herein have the same meanings as commonly understood by one of ordinary skill in the art. For the purpose of the present invention, all references cited herein are incorporated by reference in their entireties.

Concentrations, amounts, and other numerical data may be expressed or presented herein in a range format. It is to be understood that such a range format is used merely for convenience and brevity and thus should be interpreted flexibly to include not only the numerical values explicitly recited as the limits of the range, but also to include all the individual numerical values or sub-ranges encompassed within that range as if each numerical value and sub-range is explicitly recited. This same principle applies to ranges reciting only one numerical value. Furthermore, such an interpretation should apply regardless of the breadth of the range or the characteristics being described. Also, it is to be understood that ranges may differ depending on the institute/facility where the measurements are being performed, methodology of measurement, type of tissue, and technique of tissue collection.

Metabolic biomarkers for ovarian cancer

As discussed above, the present invention provides the use of a combination of metabolites contained in a blood sample of a mammalian subject in the in vitro or ex vivo detection, diagnosis, prognosis and/or therapy monitoring of ovarian cancer.

According to the invention, the combination of metabolites comprises or consists of

at least one amino acid selected from the group of alanine (Ala), asparagine (Asn), citrulline (Cit), glutamate (Glu), tryptophan (Trp), and at least one biogenic amine selected from the group of asymmetric dimethylarginine (ADMA), cis-4-hydroxyproline (c4-OH-Pro), symmetric dimethylarginine (SDMA),

at least one acylcamitine selected from the group of acylcamitine C4:0, acylcamitine

C16: l,

lysophosphatidylcholine Cl 8:2

at least one phosphatidylcholine selected from the group of PC(34:5), PC(35: 1), PC(42:4), PC-O(30:0), PC-0(32:3), and

at least one sphingolipid selected from SM(32:2), SM(42: 1)

Preferably, the at least one amino acid is selected from

alanine (Ala), asparagine (Asn), citrulline (Cit), glutamate (Glu) and tryptophan (Trp) more preferably from alanine (Ala), citrulline (Cit), glutamate (Glu) and tryptophan (Trp).

Preferably, the at least one biogenic amine is selected from

asymmetric dimethylarginine (ADMA), cis-4-hydroxyproline (c4-OH-Pro) and symmetric dimethylarginine (SDMA)

more preferably from asymmetric dimethylarginine (ADMA) and cis-4-hydroxyproline (c4-OH- Pro).

Preferably, the at least one acylcamitine is selected from

butyrylcamitine (AC(4:0)) and hexadecenoylcamitine (AC(16: 1)).

Most preferably, the lysophosphatidylcholine Cl 8:2 is selected.

Preferably, the at least one phosphatidylcholine is selected from

phosphatidylcholine PC(34:5), PC(35: 1), PC(42:4), PC-O(30:0) and PC-0(32:3) more preferably from phosphatidylcholine PC(34:5), PC(35: 1) and PC(42:4).

More preferably, the at least one sphingolipid is selected from

sphingolipid SM(32:2) and SM(42: 1).

The detection, diagnosis and/or prognosis of ovarian cancer preferably comprises: ovarian cancer early detection,

ovarian cancer early diagnosis,

disease monitoring,

differential diagnostic,

prognosis track,

distinguishing early stage group patients and/or high-risk group patients,

predicting therapy outcome,

risk classification,

or combinations thereof.

The combination of metabolites is particularly suitable for assessing ovarian cancer, wherein the assessment is for the early stage detection and diagnosis, for the prognosis, for differential diagnostic.

The sample is preferably selected from blood, plasma, serum or a mixture thereof.

The mammalian subject is preferably human.

According to the invention, the use preferably comprises the detection of the presence of the metabolites in said sample and comparing it to a control sample.

Said control sample is preferably the sample of a healthy subject.

The metabolites from the disclosed set/combination in ovarian cancer patients are significantly different from the healthy controls, which is regarded as the assessment for the diagnosis and/or risk classification.

Preferably, the metabolites are measured based on a quantitative analytical method,

preferably

(a) chromatography,

(b) spectroscopy,

(c) mass analyzers/ spectrometry,

and combinations thereof.

Wherein: (a) Chromatography preferably comprises GC, CE, LC, HPLC, and UHPLC.

(b) Spectroscopy preferably comprises UV/Vis, IR, NIR and NMR.

(c) Mass analyzers/spectrometry preferably comprises ESI, Quadrupole Mass Analyzers, Ion Trap Mass Analyzers, TOF (Time of Flight) Mass Analyzer, Orbitrap mass analyzer, Magnetic Sector Mass Analyzer, Electrostatic Sector Mass Analyzer, Ion Cyclotron Resonance (ICR) and combinations thereof,

including single quadrupole (Q) and triple quadrupole (QqQ), QqTOF, TOF-TOF, Q-Orbitrap, APCI-QqQ, APCI-QqTOF, MAFDI-QqQ, MAFDI- QqTOF, and MAFDI-TOF-TOF.

Methods for diagnosis and/or prognosis

As discussed above, the present invention provides an in vitro or ex vivo method for the detection, diagnosis, prognosis and/or therapy monitoring of ovarian cancer.

Said method comprises the following steps:

(a) providing a blood sample of a mammalian subject,

(b) determining the amount of a combination of metabolites

and

(c) comparing said amount determined in (b) with a control sample

The combination of metabolites of step (b) is as defined herein above.

Said combination of metabolites comprises or consists of

one amino acid selected from the group of alanine (Ala), asparagine (Asn), citrulline (Cit), glutamate (Glu), tryptophan (Trp), and

one biogenic amine selected from the group of asymmetric dimethylarginine (ADMA), cis-4-hydroxyproline (c4-OH-Pro), symmetric dimethylarginine (SDMA),

one acylcamitine selected from the group of acylcamitine C4:0, acylcarnitine C16: l , lysophosphatidylcholine Cl 8:2,

one phosphatidylcholine selected from the group of PC(34:5), PC(35: 1), PC(42:4), PC- 0(30:0), PC-0(32:3), and

one sphingolipid selected from SM(32:2), SM(42: 1). Preferably, the at least one amino acid is selected from

Preferably, the at least one biogenic amine is selected from

Preferably, the at least one acylcarnitine is selected from

butyrylcamitine (AC(4:0)) and hexadecenoylcamitine (AC(16: 1)).

Most preferably, lysophosphatidylcholine Cl 8:2 is selected.

Preferably, the at least one phosphatidylcholine is selected from

phosphatidylcholine PC(34:5), PC(35: 1), PC(42:4), PC-0(30:0) and PC-0(32:3) more preferably from phosphatidylcholine PC(34:5), PC(35: 1) and PC(42:4).

More preferably, the at least one sphingolipid is selected from

sphingolipid SM(32:2) and SM(42: 1).

The sample is preferably selected from blood, plasma, serum.

The mammalian subject is preferably human.

As discussed above, the detection, diagnosis and/or prognosis of ovarian cancer preferably comprises:

ovarian cancer early detection,

ovarian cancer early diagnosis,

disease monitoring,

differential diagnostic, prognosis track,

distinguishing early stage group patients and/or high-risk group patients,

predicting therapy outcome,

risk classification,

or combinations thereof.

As discussed above, the metabolites are preferably measured based on a quantitative analytical method,

preferably

(a) chromatography,

(b) spectroscopy,

(c) mass analyzers/ spectrometry,

and combinations thereof.

Wherein:

(a) Chromatography preferably comprises GC, CE, LC, HPLC, and UHPLC.

(b) Spectroscopy preferably comprises UV/Vis, IR, NIR and NMR.

including single quadrupole (Q) and triple quadrupole (QqQ), QqTOF, TOF-TOF, Q-Orbitrap, APCI-QqQ, APCI-QqTOF, MAFDI-QqQ, MAFDI- QqTOF, and MAFDI-T OF-TOF .

Kit for performing the method of the invention

As discussed above, the present invention provides a kit (kit for use) or use of a kit for performing the in vitro or ex vivo method for the detection, diagnosis, prognosis and/or therapy monitoring of ovarian cancer according to the present invention,

comprising a device having one or more wells and one or more inserts impregnated with at least one internal standard.

In one embodiment, the kit further comprises

antibodies that target the combination of metabolites of the present invention.

Hence, the invention relates to a kit adapted for carrying out the method, wherein the kit comprises a device which contains one or more wells and one or more inserts impregnated with at least one internal standard. Such a device is in detail described in WO 2007/003344 and WO 2007/003343.

Specific method of differential diagnosis between ovarian cancer and breast cancer

A combination of metabolites for use in the differential diagnosis between ovarian cancer and breast cancer, wherein the combination of metabolites comprises or consists of:

1) at least one amino acid selected from the group of arginine (Arg), glutamate (Glu), and

2) at least one biogenic amine selected from the group of asymmetric dimethylarginine (ADMA), cis -4 - h y d ro x y p ro 1 i n c (c4-OH-Pro), symmetric dimethylarginine (SDMA),

3) at least one acylcamitine selected from the group of acylcamitine C4:0-OH,

acylcamitine C5:0DC, acylcamitine Cl 2:0,

4) lysophosphatidylcholine Cl 8:2,

5) at least one phosphatidylcholine selected from the group of PC(34:5), PC-0(34:3). Further description of preferred embodiments

Metabolites related to cellular respiration, carbohydrate, lipid, protein and nucleotide metabolism were significantly altered in ovarian cancer (Ke et al., 2015; Ke et al., 2016; Turkoglu et al., 2016; Zhang et al., 2015). However, ovarian cancer metabolomics research is still in its infancy, and one of the main tasks was to explore reproducible and better performance metabolic biomarkers.

The inventors have now identified 134 metabolites that significantly differed between ovarian cancer and healthy controls in training cohort, among which they identified 74 metabolites that also significantly differ in the validation cohort. ROC analysis was built by penalized LASSO logistic regression model (with penalty parameter tuning conducted by 10-fold cross-validation), was used to compute the least redundant and most informative panel of metabolites that can discriminate ovarian cancer and healthy control groups. A good discriminatory accuracy using 18 metabolites (Ala. Asn, Cit Glu. Trp. ADMA. c4-OH-Pro. SDMA, AC(4:0), AC(16: 1), LPC(18:2). PC(34:5h PC(35: lh PC(42:4L PC-O(30:0), PC-0(32:3), SM(32:2h SM(42: lfi with AUC = 0.99 (95% Cl: 0.97 ~ 1.00) for LASSO logistic regression model was built (Figure 1). In addition, individual metabolites also have ability in distinguishing ovarian cancer from healthy controls with high AUCs (Figure 2).

A further biostatistical analysis comprised univariate and multivariate methods for describing the statistical properties of individual metabolites and combination of metabolites.

Moreover, the inventive combination is justified by a univariate analysis, wherein the following ranking is identified:

Table 0:

In a further approach the inventors have analysed the specific role of sphingomyelins and it was found that sphingomyelins decisively contribute to the specificity and sensitivity of the inventive combination (cf. Fig. 5a and 5b).

The inventors have also identified 72 metabolites that significantly differed between ovarian cancer and primary breast cancer in training cohort, among which they identified 49 metabolites that also significantly differ in the validation cohort. ROC analysis was built by penalized LASSO logistic regression model (with penalty parameter tuning conducted by 10- fold cross-validation), was used to compute the least redundant and most informative panel of metabolites that can discriminate ovarian cancer and primary breast cancer. A good discriminatory accuracy using 11 metabolites (Are. Glu. ADMA c4-OH-Pro. SDMA. AC(4:0-OH), AC(5:0-DCh AC(12:0h LPC(18:2h PC(34:5L PC-0(34:3)) with AUC = 0.97 (95% Cl: 0.95 ~ 1.00) for LASSO logistic regression model was built (Figure 3). In addition, individual metabolites also have ability in distinguishing ovarian cancer from primary breast cancer with high AUCs (Figure 4).

In the case of any lipids it should be noted that, the detected signal is a sum of several isobaric lipids with the same molecular weight (±0.5 Da range) within the same class. This is because of the limitation of the mass resolution in the preferably employed MS/MS measurements. For example, the signal of PC aa C36:6 can arise from different lipid species that have different fatty acid composition (e.g. PC 16: 1/20:5 versus PC 18:4/18:2), various positioning of fatty acids sn- l/sn-2 (e.g. PC 18:4/18:2 versus PC18:2/18:4) and different double bond positions and stereochemistry in those fatty acid chains (e.g. PC(18:4(6Z,9Z,12Z,15Z)/18:2(9Z,12Z)) versus PC( 18 :4(9Z, 11 Z, 13Z, 15Z)/18 :2(9Z, 12Z)).

The term“ovarian cancer” refers to a type of gynecologic tumors having no or few symptoms in the early stage of a patient, in particular female patient. Ovarian cancer is a malignant disease of the ovary in the genital tract in women (cf. Pschyrembel, de Gruyter, 263rd edition (2012), Berlin).

Within the scope of this invention, the term“patient, in particular female patient” is understood to mean any test subject (human or mammal), with the provision that the test subject is tested for ovarian cancer. The term“female patient” is understood to mean any female test subject. In one embodiment of the method according to the invention body fluid, preferably blood, is drawn from the patient to be examined, optionally full blood or serum, or available plasma, and the assessment, in particular diagnosis is carried out in vitro/ex vivo, e.g. outside of the human or animal body.

The following examples and drawings illustrate the present invention without, however, limiting the same thereto.

BRIEF DESCRIPTION OF THE FIGURES

Figure 1: Receiver operating characteristics (ROC) curve of the classifier for the differentiation between healthy controls and ovarian cancer patients.

Performance on the training cohort is shown to the left (A) whereas the performance on the validation cohort is shown to the right (B). The continuous line represents the discrimination power of the classifier. Corresponding AUC and 95% confidence interval (Cl) are shown. ROC curves are colorized according to the cut-off, the scale of which is represented at the right side of the plot. The x-axis and y-axis show the false positive rate (1 -specificity) and the true positive rate (sensitivity), respectively. The dash-dotted grey diagonal line indicates no discrimination power, i.e. random classification.

The multiparametric panel based on panalized LASSO logistic regression model with 18 variables has best performance: Ala, Asn, Cit, Glu, Trp, ADMA, c4-OH-Pro, SDMA, AC(4:0), AC(16: 1), LPC(18:2), PC(34:5), PC(35: 1), PC(42:4), PC-O(30:0), PC-0(32:3), SM(32:2), SM(42: 1)

Figure 2: Receiver operating characteristics (ROC) curve of the classifier for the differentiation between healthy controls and ovarian cancer patients with single features. Performances of single features to differentiate ovarian cancer patients from healthy controls. The corresponding features are indicated in the title of plot. The continuous red lines represent the discrimination power of the classifier for training cohort, while the dotted blue lines represent the results for validation cohort. Corresponding AUC are shown. The x-axis and y-axis show the false positive rate (1 -specificity) and the true positive rate (sensitivity), respectively. The diagonal grey lines indicate no discrimination power, i.e. random classification.

Figure 3: Receiver operating characteristics (ROC) curve of the classifier for the differentiation between ovarian cancer and primary breast cancer patients.

The multiparametric panel based on panalized LASSO logistic regression model with 11 variables has best performance: Arg, Glu, ADMA, c4-OH-Pro, SDMA, AC(4:0-OH), AC(5:0-DC), AC(12:0), LPC(18:2), PC(34:5), PC-0(34:3) Figure 4: Receiver operating characteristics (ROC) curve of the classifier for the differentiation between ovarian cancer patients and primary breast cancer patients with single features.

Performances of single features to differentiate ovarian cancer patients from primary breast cancer patients. The corresponding features are indicated in the title of plot. The continuous red lines represent the discrimination power of the classifier for training cohort, while the dotted blue lines represent the results for validation cohort. Corresponding AUC are shown. The x-axis and y-axis show the false positive rate (1 -specificity) and the true positive rate (sensitivity), respectively. The diagonal grey lines indicate no discrimination power, i.e. random classification.

Figure 5a: Area under the Curve (AUC) of classes of metabolites

In addition to amino acids, acylcamitines and glycerophospholipids, the class of sphingomyelins also shows a very high potential for the separation of Ctrl vs cancer. With the sphingomyelins alone one already achieves an AUC of 0.9119 for the separation of control vs ovarian cancer.

Figure 5b: Box plot of the top ranked SM(32:2) showing fold change (healthy controls (left) vs. ovarian cancer (right) resp. 2 samples HB2015 and HB2016)

EXAMPFES

1. Materials and Methods

1.1 Clinic blood sample collection and metabolites measurement

In this research project, samples from three types of study cohorts were investigated; all studies were approved by the Ethical Committee. Clinical data, detailed questionnaire data and follow up information were available for all patients. All subjects were female and of Caucasian origin thus matched for gender and ethnicity. Written consent was obtained from all participants prior to phlebotomy.

Ovarian cancer patients : The ovarian cancer cohort included patients of sporadic and initial diagnosis of ovarian cancer. Blood was collected from them before they underwent any therapy or surgery. This cohort included 34 patients for training cohort, which consisting of 2 FIGO stage I, 1 FIGO stage II, 29 FIGO stage III and 2 FIGO stage IV samples. While in the validation cohort, 35 samples were included, with 5 FIGO stage I, 5 FIGO stage II, 15 FIGO stage III and 6 FIGO stage IV samples. In addition, 4 samples didn’t have certain FIGO staging information.

Primary breast cancer patients : The primary breast cancer cohort consisted of patients with sporadic and initial diagnosis of breast cancer. Blood was collected from them before they underwent any therapy or surgery. Histopathological features were determined from tumour tissue obtained from both the initial biopsy and surgical resection. In case of discrepancy between the two, the results of the latter were considered valid. This cohort included 80 patients for training cohort and 109 patients for validation cohort.

Healthy control individuals : Healthy individuals with no history of malignant diseases. With no auto-immune disease and no current inflammatory condition (based on self-report). In addition to the blood samples, the volunteers were asked to fill in a questionnaire regarding their lifestyles. This cohort included 100 patients for training cohort and 50 patients for validation cohort.

Metabolite quantification by targeted metabolomics was done with the AbsolutelDQ® p400 HR Kit (herein referred to as“p400 kit”) according to the manufacturer’s instructions (Biocrates Life Sciences AG. Austria).

In total, 12 classes of metabolites were measured for each sample in the p400 kit, namely:

acylcamitines,

amino acids,

biogenic amines,

monosaccharides,

sphingomyelins,

diglycerides,

triglycerides,

ly sopho sphatidylcho lines ,

phosphatidylcholines,

ceramides, and

cholesteryl esters.

For details, see Table 1 below. 1.2 Quality control

Coefficients of variability for 5 triplicated samples were calculated to check the robustness of the results. Then sample quality control and metabolite quality control were performed according to the concentration of specific metabolites and limit of detection (LOD), respectively. Concentration values below LOD were replaced by imputed values afterwards.

1.3 Biomarkers candidate identification

Potential co-founders, such as age, stage, hormone status and other characteristics were identified and adjusted. Then multiple pairwise comparisons or regressions between sample groups were performed to select significantly variable metabolites as biomarker candidates. Multiple testing adjustments were used in all testing with Bonferroni correction to control the false discovery rate (FDR) at the level of 0.05.

1.4 Feature selection

Based on the selected significant metabolites, LASSO regression model was built to reduce variables. The model was built with the training cohort to select the best lambda, which was realized by 10-fold cross-validation performed 100 times. The classification ability of resulting variables was further evaluated by different models.

1.5 ROC analysis: diagnostic value of candidate metabolites

Multiple machine learning methods including Support Vector Machine (SVM), Generalized Boosted Regression Models (GBM), Boosted Logistic Regression (Logit), Classification And Regression Trees (CART) and Random Forest (RF) were employed to construct prediction model using previously selected metabolites. Their performance was evaluated using receiver operating characteristic curve (ROC). All these were done with R software (Team, 2015).

For measuring or determining the metabolite amounts targeted metabolomics is used to quantify the metabolites in the sample including the analyte classes as shown in T able 1. The quantification is carried out using in the presence of isotopically labelled internal standards and determined by the methods as described above. A list of analytes including their abbreviations (BC codes) being suitable as metabolites to be measured according to the invention is indicated in the following Table 1. Table 1. List of metabolites measured in Biocrates p400 kit:

Table la. Amino Acids

Table lb. Biogenic Amines

Table lc. Monosaccharides

Table Id. Acylcamitines

Table le. Diglycerides

Table If. Triglycerides

Table lg. Lysophosphatidylcholines

Table lh. Phosphatidylcholines

Table li. Sphingomyelins

Table lj. Ceramides

Table lk. Cholesteryl Esters

The features disclosed in the foregoing description, in the claims and/or in the accompanying figures may both separately and in any combination thereof be material for realizing the invention in diverse forms thereof. REFERENCES

Dunn, W.B., Broadhurst, D.I., Atherton, H.J., Goodacre, R., and Griffin, J.L. (2011). Systems level studies of mammalian metabolomes: the roles of mass spectrometry and nuclear magnetic resonance spectroscopy. Chem Soc Rev 40, 387-426.

Fong, M.Y., McDunn, J., and Kakar, S.S. (2011). Identification of Metabolites in the Normal Ovary and Their Transformation in Primary and Metastatic Ovarian Cancer. PLoS One 6.

German, J.B., Hammock, B.D., and Watkins, S.M. (2005). Metabolomics: building on a century of biochemistry to guide human health. Metabolomics 1, 3-9.

Ke, C., Hou, Y., Zhang, H., Fan, L, Ge, T., Guo, B., Zhang, F., Yang, K., Wang, J., Lou, G., et al. (2015). Large- scale profiling of metabolic dysregulation in ovarian cancer. Int J Cancer 136, 516-526.

Ke, C., Li, A., Hou, Y., Sun, M., Yang, K., Cheng, J., Wang, J., Ge, T., Zhang, F., Li, Q., et al. (2016). Metabolic phenotyping for monitoring ovarian cancer patients. Sci Rep 6, 23334.

Molina, R., Escudero, J.M., Auge, J.M., Filella, X., Foj, L., Torne, A., Lejarcegui, J., and Pahisa, J. (2011).

HE4 a novel tumour marker for ovarian cancer: comparison with CA 125 and ROMA algorithm in patients with gynaecological diseases. Tumour Biol 32, 1087-1095.

Noone AM, Howlader N, Krapcho M, Miller D, Brest A, Yu M, Ruhl J, Tatalovich Z, Mariotto A, Lewis DR, et al. SEER Cancer Statistics Review, 1975-2015, National Cancer Institute.

Rauh-Hain, J.A., Krivak, T.C., Del Carmen, M.G., and Olawaiye, A.B. (2011). Ovarian cancer screening and early detection in the general population. Rev Obstet Gynecol 4, 15-21.

Reid, B.M., Permuth, J.B., and Sellers, T.A. (2017). Epidemiology of ovarian cancer: a review. Cancer Biol Med 14, 9-32.

Team, R.C. (2015). R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria https://wwwR-proiectorg/.

Torre, L.A., Bray, F., Siegel, R.L., Ferlay, J., Lortet-Tieulent, J., and Jemal, A. (2015). Global cancer statistics, 2012. CA Cancer J Clin 65, 87-108.

Turkoglu, O., Zeb, A., Graham, S., Szyperski, T., Szender, J.B., Odunsi, K., and Bahado-Singh, R. (2016). Metabolomics of biomarker discovery in ovarian cancer: a systematic review of the current literature. Metabolomics 12.

Wishart, D.S. (2007). Current progress in computational metabolomics. Brief Bioinform 8, 279-293.

Zhang, H., Ge, T., Cui, X., Hou, Y., Ke, C., Yang, M., Yang, K., Wang, J., Guo, B., Zhang, F., et al. (2015). Prediction of advanced ovarian cancer recurrence by plasma metabolic profiling. Mol Biosyst 11, 516- 521.

Claims

1. Use of a combination of metabolites contained in a sample of a mammalian subject for assessing ovarian cancer, the combination of metabolites comprising

1) at least one amino acid selected from the group of Alanine (Ala), Asparagine (Asn), Citrulline (Cit), glutamate (Glu), tryptophane (Trp), and

2) at least one biogenic amine selected from the group of Asymmetric dimethylarginine (ADMA), c/s-4-Hydroxyproline (c4-OH-Pro), Symmetric dimethylarginine (SDMA),

3) at least one acylcarnitine selected from the group of acylcamitine C4:0, acylcamitine C16: l,

4) lysophosphatidylcholine Cl 8:2

6) at least one sphingolipid selected from SM(32:2), SM(42: 1)

2. The use according to claim 1, wherein the detection, diagnosis and/or prognosis of ovarian cancer comprises:

ovarian cancer detection,

ovarian cancer diagnosis,

disease monitoring,

differential diagnostic,

prognosis track,

distinguishing early stage group patients and/or high risk group patients,

predicting therapy outcome,

risk classification,

or combinations thereof.

3. The use according to any one of claims 1 to 2, wherein the sample is selected from blood, plasma or serum or a mixture thereof.

4. The use according to any one of claims 1 to 3, wherein the mammalian subject is human.

5. The use according to any one of claims 1 to 4, comprising the detection of the presence of the metabolites in said sample and comparing it to a control sample.

6. The use according to any one of claims 1 to 5, wherein the metabolites are measured based on a quantitative analytical method, particularly chromatography, spectroscopy, and mass analyzers/spectrometry .

7. The use according to claim 6, wherein chromatography comprises GC, CE, LC, HPLC, and UHPLC.

8. The use according to claim 6, wherein spectroscopy comprises UV/Vis, IR, NIR and NMR.

9. The use according to claim 6, wherein mass analyzers/spectrometry comprises ESI, Quadrupole Mass Analyzers, Ion Trap Mass Analyzers, TOF (Time of Flight) Mass Analyzer, Orbitrap mass analyzer, Magnetic Sector Mass Analyzer, Electrostatic Sector Mass Analyzer, Ion Cyclotron Resonance (ICR) and combinations thereof, in particular single quadrupole (Q) and triple quadrupole (QqQ), QqTOF, TOF-TOF, Q-Orbitrap, APCI-QqQ, APCI-QqTOF, MAFDI- QqQ, MAFDI- QqTOF, and M AFDI-T OF-TOF .

10. An in vitro or ex vivo method for the detection, diagnosis, prognosis and/or therapy monitoring of ovarian cancer, comprising the following steps:

(a) providing a sample of a mammalian subject,

(b) determining the amount of a combination of metabolites as defined in claim 1, and

(c) comparing said amount determined in (b) with a control sample.

1 1. The method of claim 10, wherein the sample is selected from blood, plasma, serum.

12. The method according to any one of claims 10 to 11, wherein the mammalian subject is human.

13. The method of claim 10, wherein the metabolites are measured based on a quantitative analytical method,

particularly chromatography, spectroscopy, and mass analyzers/ spectrometry,

wherein chromatography particularly comprises GC, CE, LC, HPLC, and UHPLC;

wherein spectroscopy particularly comprises UV/Vis, IR, NIR and NMR; and

wherein mass analyzers/spectrometry particularly comprises ESI, Quadrupole Mass Analyzers, Ion Trap Mass Analyzers, TOF (Time of Flight) Mass Analyzer, Orbitrap mass analyzer, Magnetic Sector Mass Analyzer, Electrostatic Sector Mass Analyzer, Ion Cyclotron Resonance (ICR) and combinations thereof,

in particular single quadrupole (Q) and triple quadrupole (QqQ), QqTOF, TOF-TOF, Q-Orbitrap, APCI-QqQ, APCI-QqTOF, MAFDI-QqQ, MAFDI- QqTOF, and MAFDI-T OF-TOF .

14. A kit for performing the in vitro or ex vivo method for the detection, diagnosis, prognosis and/or therapy monitoring of ovarian cancer according to any one of claims 10 to 13, comprising a device having one or more wells and one or more inserts impregnated with at least one internal standard.

15. The kit of claim 14, further comprising

antibodies that target said metabolites as defined in claim 1.