Methods for detecting and treating ovarian cancer

Abstract

A novel group of 7 marker metabolites (7MetP) is described, which comprises or consists of a diacetyl spermine, a diacetyl spermidine, an N-(3-acetamidopropyl) pyrrolidin-2-one, an N-acetylneuraminic acid ester, an N-acetyl-mannosamine, an N-acetyl-lactosamine, and a hydroxyisobutyric acid.

Description

[0001] This application claims the benefit of priority to U.S. Provisional Application No. 63 / 369,027, filed on July 21, 2022, the disclosure of which is hereby incorporated by reference in its entirety as if written herein.

[0002] This invention was made with government support under Grant Nos. CA200462 and CA217685 awarded by the National Institutes of Health. The government has certain rights in this invention.

[0003] Disclosed herein are methods and related kits for detecting ovarian cancer. Also provided are methods for treating patients susceptible to or suspected of being susceptible to ovarian cancer.

[0004] Approximately 17% of women who undergo transvaginal sonography (TVS) are found to have ovarian cysts and pelvic masses. However, the majority of such masses are benign, and only a small percentage of women are diagnosed with ovarian cancer. Currently, TVS and cancer antigen 125 (CA125), alone or in combination, do not produce sufficient sensitivity and specificity to distinguish benign from malignant ovarian cysts. The high false-positive rate leads to increased patient anxiety and unnecessary surgical procedures, which carry significant morbidity.

[0005] Two risk assessment algorithms, the Ovarian Malignancy Risk Algorithm (ROMA) and the Ovarian Cancer Risk Algorithm (OVERA), were developed to estimate the probability of malignancy in women with a pelvic mass and to determine whether the patient should be referred to a general gynecologist if the mass is likely to be benign or to a gynecologic oncologist if the mass is likely to be malignant. Gynecologic oncologists are specially trained to perform lymph node dissection, omentectomy, and, if disease is found to be extensive, to remove as much cancerous tissue from the bowel surface as possible. Although the OVERA and ROMA algorithms have high sensitivity, they are limited by suboptimal specificity, which may contribute to the high false-positive rates mentioned above. Tests with high sensitivity and specificity in identifying individuals at high risk for malignant ovarian cysts have the potential to better inform clinical decisions and improve patient outcomes.

[0006] Therefore, there is a need for a method or test to help detect ovarian cancer. Disturbances in cellular metabolism are hallmarks of cancer. Several lines of evidence suggest that cellular and systemic metabolic adaptations occur at the earliest stages of cancer development, suggesting that metabolites can be used as cancer biomarkers. Through a deep learning approach to serum metabolic profiles, a novel 7-marker metabolite panel was discovered, containing or consisting of the following: diacetylspermidine, diacetylspermidine, N-(3-acetamidopropyl)pyrrolidin-2-one, N-acetylneuraminic acid ester, N-acetyl-mannosamine, N-acetyl-lactosamine, and hydroxyisobutyric acid, which can distinguish early ovarian cancer from benign disease. This model was combined with the ROMA algorithm utilizing the biomarkers human epididymis protein 4 (HE4) and mucin 16 (CA125), and showed superior ovarian cancer risk prediction than ROMA alone in women with ovarian cysts. Summary of the invention

[0007] Provided herein is a method of treating ovarian cancer in a patient having elevated levels of diacetylspermine (DAS), diacetylspermidine (DiAcSpd), N-(3-acetamidopropyl)pyrrolidin-2-one (N3AP), N-acetylneuraminic acid (NANA), N-acetyl-mannosamine (NAcMan), N-acetyl-lactosamine (NAcLac), and hydroxyisobutyric acid (HBA), and optionally elevated levels of human epididymis protein 4 (HE4) and mucin 16 (CA125), wherein these elevated levels classify the patient as having ovarian cancer, the method comprising administering to the patient a therapeutically effective amount of an ovarian cancer treatment.

[0008] The present invention also provides a method for treating ovarian cancer, comprising: a) identifying a patient having elevated levels of diacetylspermine, diacetylspermidine, N-(3-acetamidopropyl)pyrrolidin-2-one, N-acetylneuraminate, N-acetyl-mannosamine, N-acetyl-lactosamine and hydroxyisobutyric acid and optionally elevated levels of human epididymis protein 4 (HE4) and mucin 16 (CA125), wherein these elevated levels classify the patient as having ovarian cancer; and b) administering to the patient a therapeutically effective amount of an ovarian cancer treatment.

[0009] Also provided herein is a method of distinguishing between ovarian cancer and a benign pelvic mass (BPM) in a subject, the method comprising, in a biological sample obtained from the subject: a) measuring the levels of diacetylspermine, diacetylspermidine, N-(3-acetamidopropyl)pyrrolidin-2-one, N-acetylneuraminate, N-acetyl-mannosamine, N-acetyl-lactosamine and hydroxyisobutyric acid, and optionally the levels of human epididymis protein 4 (HE4) and mucin 16 (CA125) in the biological sample; and b) classifying the subject as having ovarian cancer or BPM based on the measured levels.

[0010] Also provided herein is a method of determining a subject's risk of developing ovarian cancer, the method comprising, in a biological sample obtained from the subject: a) measuring the levels of diacetylspermine, diacetylspermidine, N-(3-acetamidopropyl)pyrrolidin-2-one, N-acetylneuraminate, N-acetyl-mannosamine, N-acetyl-lactosamine and hydroxyisobutyric acid, and optionally the levels of human epididymis protein 4 (HE4) and mucin 16 (CA125) in the biological sample; and b) classifying the subject as being at risk for ovarian cancer or not at risk for ovarian cancer based on the measured level.

[0011] Also provided herein is a method of generating a risk profile for ovarian cancer in a subject, the method comprising: in a biological sample obtained from the subject: a) measuring the levels of diacetylspermine, diacetylspermidine, N-(3-acetamidopropyl)pyrrolidin-2-one, N-acetylneuraminate, N-acetyl-mannosamine, N-acetyl-lactosamine and hydroxyisobutyric acid, and optionally the levels of human epididymis protein 4 (HE4) and mucin 16 (CA125) in the biological sample; and b) classifying the subject as being at risk for ovarian cancer or not at risk for ovarian cancer based on the measured level.

[0012] Also provided herein is a method of calculating a biomarker score or a risk score for ovarian cancer in a patient, the method comprising: a) measuring the levels of diacetylspermine, diacetylspermidine, N-(3-acetamidopropyl)pyrrolidin-2-one, N-acetylneuraminate, N-acetyl-mannosamine, N-acetyl-lactosamine and hydroxyisobutyric acid, and optionally the levels of human epididymis protein 4 (HE4) and mucin 16 (CA125) in the biological sample; and b) using the numerical values ​​of the measured levels in a deep learning model (DLM) to calculate the biomarker score or risk score.

[0013] Also provided herein is a method of risk stratifying a patient at risk for developing ovarian cancer, the method comprising, in a biological sample obtained from a subject: a) measuring the levels of diacetylspermine, diacetylspermidine, N-(3-acetamidopropyl)pyrrolidin-2-one, N-acetylneuraminate, N-acetyl-mannosamine, N-acetyl-lactosamine and hydroxyisobutyric acid, and optionally the levels of human epididymis protein 4 (HE4) and mucin 16 (CA125) in the biological sample; and b) determining, by a processor circuit, a risk score for the patient, wherein the risk score is determined by a scoring function derived from metabolite profiles of biological samples collected from a plurality of individuals undergoing ovarian cancer monitoring. BRIEF DESCRIPTION OF THE DRAWINGS

[0014] Figure 1 The predictive performance of 7MetP in distinguishing early-stage ovarian cancer from benign pelvic masses in an independent test set is depicted.

[0015] Figure 2 A schematic workflow for the analysis of different datasets is depicted.

[0016] Figure 3 Depicted is a Spearman correlation heatmap of the metabolites of 7MetP in the training set. DETAILED DESCRIPTION

[0017] Provided herein is a method of treating ovarian cancer in a patient having elevated levels of diacetylspermine, diacetylspermidine, N-(3-acetamidopropyl)pyrrolidin-2-one, N-acetylneuraminate, N-acetyl-mannosamine, N-acetyl-lactosamine, and hydroxyisobutyric acid, and optionally elevated levels of human epididymis protein 4 (HE4) and mucin 16 (CA125), wherein the elevated levels classify the patient as having ovarian cancer, the method comprising administering to the patient a therapeutically effective amount of an ovarian cancer treatment.

[0018] The present invention also provides a method for treating ovarian cancer, comprising: a) identifying a patient having elevated levels of diacetylspermine, diacetylspermidine, N-(3-acetamidopropyl)pyrrolidin-2-one, N-acetylneuraminate, N-acetyl-mannosamine, N-acetyl-lactosamine and hydroxyisobutyric acid and optionally elevated levels of human epididymis protein 4 (HE4) and mucin 16 (CA125), wherein these elevated levels classify the patient as having ovarian cancer; and b) administering to the patient a therapeutically effective amount of an ovarian cancer treatment.

[0019] Also provided herein is a method of distinguishing between ovarian cancer and a benign pelvic mass (BPM) in a subject, the method comprising, in a biological sample obtained from the subject: a) measuring the levels of diacetylspermine, diacetylspermidine, N-(3-acetamidopropyl)pyrrolidin-2-one, N-acetylneuraminate, N-acetyl-mannosamine, N-acetyl-lactosamine and hydroxyisobutyric acid, and optionally the levels of human epididymis protein 4 (HE4) and mucin 16 (CA125) in the biological sample; and b) classifying the subject as having ovarian cancer or BPM based on the measured levels.

[0020] Also provided herein is a method of determining a subject's risk of developing ovarian cancer, the method comprising, in a biological sample obtained from the subject: a) measuring the levels of diacetylspermine, diacetylspermidine, N-(3-acetamidopropyl)pyrrolidin-2-one, N-acetylneuraminate, N-acetyl-mannosamine, N-acetyl-lactosamine and hydroxyisobutyric acid, and optionally the levels of human epididymis protein 4 (HE4) and mucin 16 (CA125) in the biological sample; and b) classifying the subject as being at risk for ovarian cancer or not at risk for ovarian cancer based on the measured level.

[0021] Also provided herein is a method of generating a risk profile for ovarian cancer in a subject, the method comprising: in a biological sample obtained from the subject: a) measuring the levels of diacetylspermine, diacetylspermidine, N-(3-acetamidopropyl)pyrrolidin-2-one, N-acetylneuraminate, N-acetyl-mannosamine, N-acetyl-lactosamine and hydroxyisobutyric acid, and optionally the levels of human epididymis protein 4 (HE4) and mucin 16 (CA125) in the biological sample; and b) classifying the subject as being at risk for ovarian cancer or not at risk for ovarian cancer based on the measured level.

[0022] Also provided herein is a method of calculating a biomarker score or a risk score for ovarian cancer in a patient, the method comprising: a) measuring the levels of diacetylspermine, diacetylspermidine, N-(3-acetamidopropyl)pyrrolidin-2-one, N-acetylneuraminate, N-acetyl-mannosamine, N-acetyl-lactosamine and hydroxyisobutyric acid, and optionally the levels of human epididymis protein 4 (HE4) and mucin 16 (CA125) in the biological sample; and b) using the numerical values ​​of the measured levels in a deep learning model (DLM) to calculate the biomarker score or risk score.

[0023] Also provided herein is a method of risk stratifying a patient at risk for developing ovarian cancer, the method comprising, in a biological sample obtained from a subject: a) measuring the levels of diacetylspermine, diacetylspermidine, N-(3-acetamidopropyl)pyrrolidin-2-one, N-acetylneuraminate, N-acetyl-mannosamine, N-acetyl-lactosamine and hydroxyisobutyric acid, and optionally the levels of human epididymis protein 4 (HE4) and mucin 16 (CA125) in the biological sample; and b) determining, by a processor circuit, a risk score for the patient, wherein the risk score is determined by a scoring function derived from metabolite profiles of biological samples collected from a plurality of individuals undergoing ovarian cancer monitoring.

[0024] In some embodiments, the DLM comprises an artificial neural network having one to three hidden layers and one to three nodes per layer.

[0025] In some embodiments, the DLM comprises an artificial neural network having three hidden layers and three nodes per layer.

[0026] In some embodiments, the method further comprises measuring the levels of human epididymis protein 4 (HE4) and mucin 16 (CA125) or identifying a patient having elevated levels of human epididymis protein 4 (HE4) and mucin 16 (CA125).

[0027] In some embodiments, the levels of HE4 and CA125 are determined by immunoassay.

[0028] In some embodiments, the levels of HE4 and CA125 are used to calculate a predictive index (PI) for premenopausal women using the formula: PI=-12.0+2.38*ln[HE4]+0.0626*ln[CA125].

[0029] In some embodiments, the levels of HE4 and CA125 are used to calculate the prediction index (PI) for postmenopausal women using the formula: PI=-8.09+1.04*ln[HE4]+0.732*ln[CA125].

[0030] In some embodiments, the prediction index (PI) is used to calculate the Risk of Ovarian Malignancy Algorithm (ROMA) score, and the formula is:

[0031] In some embodiments, a combined model score is calculated using logistic regression of the ROMA score and the biomarker score.

[0032] In some embodiments, the ovarian cancer is early stage (eg, stage I or stage II).

[0033] In some embodiments, the ovarian cancer is advanced (eg, stage III or stage IV).

[0034] In some embodiments, the levels of diacetylspermine, diacetylspermidine, N-(3-acetamidopropyl)pyrrolidin-2-one, N-acetylneuraminate, N-acetyl-mannosamine, N-acetyl-lactosamine, and hydroxyisobutyric acid, and optionally HE4 and CA125, are elevated relative to a reference patient or group not suffering from ovarian cancer.

[0035] In some embodiments, the levels of diacetylspermine, diacetylspermidine, N-(3-acetamidopropyl)pyrrolidin-2-one, N-acetylneuraminate, N-acetyl-mannosamine, N-acetyl-lactosamine, and hydroxyisobutyric acid, and optionally HE4 and CA125, are elevated relative to a reference patient or group with benign pelvic mass (BPM).

[0036] In some embodiments, the subject presents with a pelvic mass.

[0037] In some embodiments, each of diacetylspermine, diacetylspermidine, N-(3-acetamidopropyl)pyrrolidin-2-one, N-acetylneuraminate, N-acetyl-mannosamine, N-acetyl-lactosamine, and hydroxyisobutyric acid, and optionally HE4 and CA125, produces a detectable signal.

[0038] In some embodiments, the detectable signal is detectable by spectroscopy.

[0039] In some embodiments, the spectroscopy is selected from ultraviolet-visible spectroscopy, mass spectrometry, nuclear magnetic resonance (NMR) spectroscopy, proton NMR spectroscopy, nuclear magnetic resonance (NMR) spectroscopy, gas chromatography-mass spectrometry (GC-MS), liquid chromatography-mass spectrometry (LC-MS), correlation spectroscopy (COSY), nuclear Overhauser effect spectroscopy (NOESY), rotating coordinate nuclear Overhauser effect spectroscopy (ROESY), time-of-flight LC-MS (LC-TOF-MS), liquid chromatography-tandem mass spectrometry (LC-MS / MS) and capillary electrophoresis-mass spectrometry.

[0040] In some embodiments, the spectroscopy is mass spectrometry.

[0041] In some embodiments, the mass spectrometry is LC-TOF-MS.

[0042] In some embodiments, the treatment is selected from surgery, chemotherapy, immunotherapy, radiation therapy, targeted therapy, or a combination thereof.

[0043] In some embodiments, the measured levels are used to calculate a biomarker score or risk profile based on sensitivity and specificity values ​​corresponding to the subject's risk of developing ovarian cancer.

[0044] In some embodiments, the sensitivity and specificity values ​​are Figure 1 There is no significant difference in the curves.

[0045] In some embodiments, the difference between the sensitivity and specificity values ​​is less than 10%.

[0046] In some embodiments, the difference between the sensitivity and specificity values ​​is less than 5%.

[0047] In some embodiments, the difference between the sensitivity and specificity values ​​is less than 1%.

[0048] In some embodiments, the cutoff value comprises an AUC (95% CI) of at least 0.76.

[0049] In some embodiments, the method further comprises assigning the patient to an appropriate risk group based on the calculated risk score.

[0050] In some embodiments, there are at least two risk groups.

[0051] In some embodiments, the AUC of a method is greater than the AUC of a different single biomarker, multiple biomarkers, panels, assays, or algorithms incorporating combinations thereof.

[0052] In some embodiments, the AUC is greater than 0.76.

[0053] In some embodiments, the AUC is between 0.76 and 0.95.

[0054] In some embodiments, the AUC is about 0.88.

[0055] In some embodiments, the AUC is about 0.86.

[0056] In some embodiments, the AUC is between 0.82 and 0.93.

[0057] In some embodiments, the AUC is about 0.87.

[0058] In some embodiments, the positive predictive value (PPV) of a method is greater than the AUC of a different single biomarker, multiple biomarkers, panel, assay, or algorithm incorporating a combination thereof.

[0059] In some embodiments, the PPV is greater than 0.67.

[0060] In some embodiments, the PPV is between 0.67 and 0.87.

[0061] In some embodiments, the PPV is about 0.79.

[0062] In some embodiments, the algorithm is the Risk of Ovarian Malignancy Algorithm (ROMA).

[0063] In some embodiments, the biomarkers are only HE4 and CA125.

[0064] In some embodiments, the cutoff point for each method is used for classification.

[0065] In some embodiments, different biomarkers, panels, assays or algorithms are analyzed by the same statistical method.

[0066] In some embodiments, the levels of diacetylspermine, diacetylspermidine, N-(3-acetamidopropyl)pyrrolidin-2-one, N-acetylneuraminate, N-acetyl-mannosamine, N-acetyl-lactosamine, and hydroxyisobutyric acid are measured according to a given threshold value or thresholds.

[0067] In some embodiments, the value exceeds one or more thresholds, then the patient is classified as being at risk for ovarian cancer.

[0068] In some embodiments, values ​​below one or more thresholds classify the patient as not at risk for ovarian cancer.

[0069] In some embodiments, values ​​below one or more thresholds classify the patient as having BPM.

[0070] In some embodiments, the patient is subsequently referred for further ovarian cancer screening or treatment.

[0071] In some embodiments, the screening is selected from endoscopic ultrasound, magnetic resonance imaging (MRI), and computed tomography (CT).

[0072] In some embodiments, screening is performed annually.

[0073] In some embodiments, screening is performed semi-annually.

[0074] In some embodiments, the methods disclosed above and herein may optionally further comprise a combination of metabolites selected from those disclosed in Table 4 below with a p-value < 0.05 (minimum) when comparing all cases to controls or comparing early cases to controls. definition

[0075] As used herein, the following terms have the indicated meanings.

[0076] When a range of values ​​is disclosed and the notation "from n1 ... to n2" or "between n1 ... and n2" (where n1 and n2 are numbers) is used, then unless otherwise stated, this notation is intended to include these numbers themselves and the range between them. This range can be integer or continuous between these end values ​​and include these end values. For example, the range "from 2 to 6 carbons" is intended to include two, three, four, five, and six carbons, because carbon appears in integer units. By way of example, the range "from 1 to 3 μM (micromolar)" (which is intended to include 1 μM, 3 μM, and all numbers between the two) is compared to any number of significant figures (e.g., 1.255 μM, 2.1 μM, 2.9999 μM, etc.).

[0077] As used herein, the term "about" is intended to qualify the numerical value it modifies, indicating that such value may vary within a range. When no specific range is recited (such as a margin of error or standard deviation of a mean value given in a graph or data table), the term "about" should be understood to mean a range that encompasses the recited values, and a range that is included by rounding to that number taking into account significant figures, and a range that encompasses the larger of ±20% of the recited values.

[0078] As used herein, "ovarian cancer" refers to a malignant growth of cells formed in the ovaries. The most common origin of ovarian cancer is epithelial cells (accounting for about 90% of ovarian cancers), which can be divided into several types, including serous ovarian cancer and non-serous ovarian cancer. Serous ovarian cancer is the most common type of epithelial cell ovarian cancer, accounting for about 40% of all ovarian cancers, while non-serous ovarian cancer may include, but is not limited to, endometrioid carcinoma, mucinous carcinoma, and clear cell carcinoma. In some embodiments, the severity of ovarian cancer can vary, represented by stages I to IV. In some embodiments, ovarian cancer can be in an early stage (e.g., stage I or II), or it can be advanced (e.g., stage III or IV).

[0079] When a group is defined as "null", it means that the group does not exist.

[0080] As used herein, the term "subject" or "patient" refers to a mammal, preferably a human, for whom classification as ovarian cancer positive or ovarian cancer negative is desired and for whom further treatment may be provided.

[0081] As used herein, a "reference patient," "reference subject," or "reference group" refers to a group of patients or subjects to which a test sample from a patient or subject suspected of having or at risk of having ovarian cancer can be compared. In some embodiments, such a comparison can be used to determine whether the test subject has ovarian cancer. A reference patient or group can serve as a control for testing or diagnostic purposes. As described herein, a reference patient or group can be a sample obtained from a single patient, or can represent a group of samples, such as a pooled group of samples.

[0082] As used herein, "healthy" refers to an individual in which no evidence of ovarian cancer is found, i.e., the individual does not have ovarian cancer. Such individuals can be classified as "ovarian cancer negative," or as having healthy ovaries or normal, unimpaired ovarian function. Healthy patients or subjects have no symptoms of ovarian cancer or other ovarian diseases, but may have benign pelvic masses, i.e., a combination of adenomas and cysts. In some embodiments, healthy patients or subjects can be used as reference patients for comparison with diseased or suspected diseased samples to determine ovarian cancer in a patient or patient group.

[0083] As used herein, the term "treatment" or "treating" refers to the administration of a drug or the performance of a medical procedure on a subject for the purpose of prophylaxis / prevention or cure, or, in the case of an affliction, to reduce the extent or likelihood of the occurrence or recurrence of an infirmity or disorder or condition or event. In connection with the present disclosure, the term may also mean the administration of a pharmacological substance or formulation, or the performance of a non-pharmacological approach, including, but not limited to, radiation therapy and surgery. Pharmacological substances as used herein may include, but are not limited to, anti-cancer drugs, including chemotherapeutic drugs, polyamine inhibitors, hormone therapy, and targeted therapy. Examples of chemotherapeutic drugs for ovarian cancer include paclitaxel (e.g., albumin-bound paclitaxel or nab-paclitaxel, trade name ), Hexamethylmelamine Capecitabine Cyclophosphamide Etoposide (VP-16), gemcitabine Ifosfamide Irinotecan (CPT-11, ), liposomal irinotecan Liposomal doxorubicin Melphalan, pemetrexed Topotecan and vinorelbine and combination chemotherapy regimens, including cisplatin + paclitaxel, TIP (paclitaxel / taxol, ifosfamide and cisplatin / cisplatin), VeIP (vinblastine, ifosfamide and cisplatin / cisplatin), VIP (etoposide / VP-16, ifosfamide and cisplatin / cisplatin), VAC (vincristine, dactinomycin and cyclophosphamide) and PEB (cisplatin / cisplatin, etoposide and bleomycin). Examples of polyamine inhibitors that have been used for anticancer treatment or are being explored for anticancer treatment in clinical trials include, but are not limited to, eflornithine Examples of hormone therapy for ovarian cancer include luteinizing hormone-releasing hormone (LHRH) agonists (e.g., goserelin Leuprorelin ), tamoxifen, and aromatase inhibitors (such as letrozole Anastrozole and exemestane ). Examples of targeted therapies for ovarian cancer include angiogenesis inhibitors such as bevacizumab (Avastin), and (poly (ADP)-ribose polymerase) (PARP) inhibitors such as olaparib (Lynparza), rucaparib (Rubraca), and niraparib (Zejula). The terms "pharmacological substances" and "anticancer therapy" may also include substances used in immunotherapy, such as checkpoint inhibitors. Treatment may include a variety of pharmacological substances or a variety of treatment methods, including but not limited to surgery and chemotherapy.

[0084] As used herein, "amount" or "level" refers to a generally quantifiable measurement of a biomarker described herein, wherein the measurement enables comparison of markers between samples and / or comparison with a control sample. In some embodiments, the amount or level is quantifiable and refers to the level of a specific marker in a biological sample (e.g., blood, serum, urine, etc.), as determined by a laboratory method or test such as an immunoassay (e.g., an antibody), mass spectrometry, or liquid chromatography. In some embodiments, the marker can be present in a sample in an increased amount or a decreased amount. Marker comparisons can be based on direct measurements of biomarker levels described herein (e.g., by protein quantification or gene expression analysis) or can be based on, for example, measurements of reporter molecules, biomarker-receptor complexes, biomarker-relay-receptor complexes, etc.

[0085] As used herein, the term "ELISA" refers to an enzyme-linked immunosorbent assay. The assay typically involves contacting a fluorescently labeled protein sample with antibodies that have specific affinity for those proteins. Detection of these proteins can be accomplished using a variety of methods, including but not limited to laser fluorescence assays.

[0086] As used herein, the term "regression" refers to a statistical method that can assign a predicted value to a potential feature of a sample based on an observable trait (or set of observable traits) of the sample. In some embodiments, the feature is not directly observable. For example, the regression method used herein can relate the qualitative or quantitative results of a particular biomarker test or set of biomarker tests for a subject to the probability that the subject is positive for ovarian cancer.

[0087] As used herein, the term "logistic regression" refers to a regression method in which the assignment of a prediction from a model can have one of several allowed discrete values. For example, a logistic regression model used herein can assign a prediction of ovarian cancer positive or ovarian cancer negative to a subject.

[0088] As used herein, the term "biomarker score" refers to a numerical score for a particular subject calculated by inputting the level of a particular biomarker for that subject into a statistical method.

[0089] As used herein, the term "comprehensive score" refers to the sum of normalized values ​​of predetermined markers measured in a sample from a subject. In one embodiment, the normalized values ​​are reported as biomarker scores, and then those biomarker score values ​​are added to provide a comprehensive score for each test subject. When used in the context of a risk categorization table and associated with a stratified grouping based on a range of comprehensive scores in a risk categorization table, the "comprehensive score" is used to determine a "risk score" for each test subject, where a multiplier indicating an increased likelihood of having cancer for the stratified grouping becomes the "risk score."

[0090] As used herein, the term "risk score" refers to a single numerical value that indicates the risk of ovarian cancer in an asymptomatic human subject compared to the known prevalence of ovarian cancer in a disease cohort. In certain embodiments, a composite score is calculated for a human subject and associated with a multiplier that indicates the risk of ovarian cancer, wherein the composite score is associated based on the range of composite scores for each stratified grouping in the risk category table. In this way, the composite score is converted to a risk score based on a multiplier that indicates an increased likelihood that the grouping that best matches the composite score has cancer.

[0091] As used herein, the term "cutoff" or "cutoff point" refers to a mathematical value associated with a particular statistical method that can be used to assign a classification of ovarian cancer positive or ovarian cancer negative to a subject based on the subject's biomarker score.

[0092] As used herein, when a value above or below a cutoff value is "characteristic of ovarian cancer," it means that the subject whose analysis of the sample yielded that value has or is at risk for ovarian cancer.

[0093] As used herein, "use" of a marker for diagnosing ovarian cancer refers to quantifying the level or amount of one or more markers described herein in a biological sample. Quantification can be performed using any known method or technique in the art or described herein. In some embodiments, markers can be used or combined together as a group for statistical comparison with other samples.

[0094] In some embodiments, the amount or level of DAS, NANA, NAcMan, NAcLac, DiAcSpmd, N3AP, and HBA, or the amount or level of DAS, NANA, NAcMan, NAcLac, DiAcSpmd, N3AP, HBA, HE4, and CA125 is about 0.48 to about 0.88, such as about 0.48, about 0.49, about 0.50, about 0.51, about 0.52, about 0.52, about 0.53, about 0.54, about 0.55, about 0.56, about 0.57, about 0.58, about 0.60, about 0.61, about 0.62, about 0.63, about 0.64, about 0.65, about 0.66, about 0.67, about 0.68, about 0.69, about 0.70, about 0.71, about 0.72, about 0.73, about 0.74, about 0.75, about 0.76, about 0.77, about 0.78, about 0.79, about 0.80, The cutoff values ​​of AUC (95% CI) of about 0.58, about 0.59, about 0.60, about 0.61, about 0.62, about 0.63, about 0.64, about 0.65, about 0.66, about 0.67, about 0.68, about 0.69, about 0.70, about 0.71, about 0.72, about 0.73, about 0.74, about 0.75, about 0.76, about 0.77, about 0.78, about 0.79, about 0.80, about 0.81, about 0.82, about 0.83, about 0.84, about 0.85, about 0.86, about 0.87, about 0.88, etc. were compared. In some embodiments, the markers DAS, NANA, NAcMan, NAcLac, DiAcSpmd, N3AP, and HBA are used together as a group, or the markers DAS, NANA, NAcMan, NAcLac, DiAcSpmd, N3AP, HBA, HE4, and CA125 are used together as a group, and the AUC (95% CI) may be 0.66 or greater, including about 0.66, about 0.67, about 0.68, about 0.69, about 0.70, about 0.71, about 0.72, about 0.73, about 0.74, about 0.75, about 0.76, about 0.77, about 0.78, about 0.79, about 0.80, about 0.81, about 0.82, about 0.83, about 0.84, about 0.85, about 0.86, about 0.87, about 0.88, about 0.89, about 0.90, about 0.91, about 0.92, about 0.93, about 0.94, about 0.95, about 0.96, about 0.97, about 0.98, about 0.99, etc.

[0095] In some embodiments, analyzing the markers DAS, NANA, NAcMan, NAcLac, DiAcSpmd, N3AP, and HBA together as a group using a fixed coefficient to diagnose ovarian cancer, or analyzing the markers DAS, NANA, NAcMan, NAcLac, DiAcSpmd, N3AP, HBA, HE4, and CA125 together as a group using a fixed coefficient to diagnose ovarian cancer, can produce an AUC (95% CI) of about 0.82 to about 0.93 for distinguishing ovarian cancer cases from individuals with benign diseases, such as about 0.82, about 0.83, about 0.84, about 0.85, about 0.86, about 0.87, about 0.88, about 0.89, about 0.90, about 0.91, about 0.92, about 0.93, etc. In some embodiments, analyzing these marker groups using a fixed coefficient can produce an AUC (95% CI) of 0.88 for distinguishing ovarian cancer cases from individuals with benign diseases. In some embodiments, analysis of any one of the marker panels described herein for diagnosing ovarian cancer using a fixed coefficient can produce an AUC (95% CI) for distinguishing early ovarian cancer of about 0.76 to about 0.95, such as about 0.76, about 0.77, about 0.78, about 0.79, about 0.80, about 0.81, about 0.82, about 0.83, about 0.84, about 0.85, about 0.86, about 0.87, about 0.88, about 0.89, about 0.90, about 0.91, about 0.92, about 0.93, about 0.94, about 0.95, etc. In some embodiments, analysis of these marker panels using a fixed coefficient can produce an AUC (95% CI) of 0.86 for distinguishing early ovarian cancer.

[0096] In some embodiments, the cutoff value of DAS comprises an AUC (95% CI) of at least 0.76, e.g., about 0.76, about 0.77, about 0.78, about 0.79, about 0.80, about 0.81, about 0.82, about 0.83, about 0.84, about 0.85, about 0.86, about 0.87, about 0.88, about 0.89, about 0.90, about 0.91, about 0.92, about 0.93, about 0.94, about 0.95, about 0.96, about 0.97, about 0.98, about 0.99, etc.

[0097] In some embodiments, the cutoff value of NANA comprises at least 0.58, such as about 0.58, about 0.59, about 0.60, about 0.61, about 0.62, about 0.63, about 0.64, about 0.65, about 0.66, about 0.67, about 0.68, about 0.69, about 0.70, about 0.71, about 0.72, about 0.73, about 0.74, about 0.75, about 0.76, about 0.77, about 0.78, about 0.79, about 0.80, about 0.81, about 0.82, about 0.83, about 0.84, about 0.85, about 0.86, about 0.87, about 0.88, about 0.89, about 0.90, about 0.91, about 0.92, about 0.93, about 0.94, about 0.95, about 0.96, about 0.97, about 0.98 The AUC (95% CI) was about 0.77, about 0.78, about 0.79, about 0.80, about 0.81, about 0.82, about 0.83, about 0.84, about 0.85, about 0.86, about 0.87, about 0.88, about 0.89, about 0.90, about 0.91, about 0.92, about 0.93, about 0.94, about 0.95, about 0.96, about 0.97, about 0.98, about 0.99, etc.

[0098] In some embodiments, the cutoff value of NAcMan comprises at least 0.50, such as about 0.50, about 0.51, about 0.52, about 0.53, about 0.54, about 0.55, about 0.56, about 0.57, about 0.58, about 0.59, about 0.60, about 0.61, about 0.62, about 0.63, about 0.64, about 0.65, about 0.66, about 0.67, about 0.68, about 0.69, about 0.70, about 0.71, about 0.72, about The AUC (95% CI) of about 0.73, about 0.74, about 0.75, about 0.76, about 0.77, about 0.78, about 0.79, about 0.80, about 0.81, about 0.82, about 0.83, about 0.84, about 0.85, about 0.86, about 0.87, about 0.88, about 0.89, about 0.90, about 0.91, about 0.92, about 0.93, about 0.94, about 0.95, about 0.96, about 0.97, about 0.98, about 0.99, etc.

[0099] In some embodiments, the cutoff value of NAcLac comprises at least 0.48, such as about 0.48, about 0.49, about 0.50, about 0.51, about 0.52, about 0.53, about 0.54, about 0.55, about 0.56, about 0.57, about 0.58, about 0.59, about 0.60, about 0.61, about 0.62, about 0.63, about 0.64, about 0.65, about 0.66, about 0.67, about 0.68, about 0.69, about 0.70, about 0.71, about The AUC (95% CI) of the present invention is about 0.72, about 0.73, about 0.74, about 0.75, about 0.76, about 0.77, about 0.78, about 0.79, about 0.80, about 0.81, about 0.82, about 0.83, about 0.84, about 0.85, about 0.86, about 0.87, about 0.88, about 0.89, about 0.90, about 0.91, about 0.92, about 0.93, about 0.94, about 0.95, about 0.96, about 0.97, about 0.98, about 0.99, etc.

[0100] In some embodiments, the cutoff value of DiAcSpmd comprises at least 0.60, such as about 0.60, about 0.61, about 0.62, about 0.63, about 0.64, about 0.65, about 0.66, about 0.67, about 0.68, about 0.69, about 0.70, about 0.71, about 0.72, about 0.73, about 0.74, about 0.75, about 0.76, about 0.77, AUC (95% CI) of about 0.78, about 0.79, about 0.80, about 0.81, about 0.82, about 0.83, about 0.84, about 0.85, about 0.86, about 0.87, about 0.88, about 0.89, about 0.90, about 0.91, about 0.92, about 0.93, about 0.94, about 0.95, about 0.96, about 0.97, about 0.98, about 0.99, etc.

[0101] In some embodiments, the cutoff value of N3AP comprises at least 0.49, such as about 0.49, about 0.50, about 0.51, about 0.52, about 0.53, about 0.54, about 0.55, about 0.56, about 0.57, about 0.58, about 0.59, about 0.60, about 0.61, about 0.62, about 0.63, about 0.64, about 0.65, about 0.66, about 0.67, about 0.68, about 0.69, about 0.70, about 0.71, about 0.72. , about 0.73, about 0.74, about 0.75, about 0.76, about 0.77, about 0.78, about 0.79, about 0.80, about 0.81, about 0.82, about 0.83, about 0.84, about 0.85, about 0.86, about 0.87, about 0.88, about 0.89, about 0.90, about 0.91, about 0.92, about 0.93, about 0.94, about 0.95, about 0.96, about 0.97, about 0.98, about 0.99, etc. AUC (95% CI) is about 0.

[0102] In some embodiments, the cutoff value of HBA includes an AUC (95% CI) of at least 0.64, such as about 0.64, about 0.65, about 0.66, about 0.67, about 0.68, about 0.69, about 0.70, about 0.71, about 0.72, about 0.73, about 0.74, about 0.75, about 0.76, about 0.77, about 0.78, about 0.79, about 0.80, about 0.81, about 0.82, about 0.83, about 0.84, about 0.85, about 0.86, about 0.87, about 0.88, about 0.89, about 0.90, about 0.91, about 0.92, about 0.93, about 0.94, about 0.95, about 0.96, about 0.97, about 0.98, about 0.99, etc.

[0103] As used herein, a subject at "risk of ovarian cancer" is a subject who may not yet demonstrate overt symptoms of ovarian cancer, but is producing biomarker levels that indicate the subject has ovarian cancer or may develop ovarian cancer in the near future. A subject who has ovarian cancer or is suspected of having ovarian cancer may be treated for the cancer or suspected cancer.

[0104] As used herein, the term "classification" refers to assigning a subject as being at risk or not at risk for ovarian cancer based on the results of a biomarker score obtained for the subject.

[0105] As used herein, the term "Wilcoxon rank sum test", also known as the Mann-Whitney U test, the Mann-Whitney-Wilcoxon test, or the Wilcoxon-Mann-Whitney test, refers to a specific statistical method for comparing two populations. For example, the test can be used herein to associate an observable trait (particularly a biomarker level) with the absence or risk of ovarian cancer in a population of subjects.

[0106] As used herein, the term "positive predictive value" refers to the proportion of true positives among the positive results obtained by a certain method.

[0107] As used herein, the term "negative predictive value" refers to the proportion of true negatives among negative results obtained by a certain method.

[0108] As used herein, the term "sensitivity" refers to the ability of a test to correctly identify those with the disease (i.e., the true positive rate) in the context of various biochemical assays. By comparison, as used herein, the term "specificity" refers to the ability of a test to correctly identify those without the disease (i.e., the true negative rate) in the context of various biochemical assays. Sensitivity and specificity are statistical measures of the performance of a binary classification test (i.e., a classification function). Sensitivity quantifies the avoidance of false negatives, while specificity does the same for false positives.

[0109] As used herein, "fixed coefficients" or "fixed model coefficients" refer to a statistical method of standardizing coefficients to allow comparison of the relative importance of each coefficient in a regression model. In some embodiments, the fixed coefficients involve using the same beta coefficients from a logistic regression model to generate a composite score for the developed combined rule, which is ultimately used to make a clinical decision based on one or more decision thresholds.

[0110] As used herein, "sample" refers to a test substance to be tested for the presence and level or concentration of a biomarker as described herein. A sample can be any suitable substance according to the present disclosure, including but not limited to blood, serum, plasma, or any portion thereof.

[0111] As used herein, "metabolite" refers to a small molecule that is an intermediate and / or product of cellular metabolism. Metabolites can perform a variety of functions in cells, for example, structural effects, signal transduction effects, stimulation and / or inhibition of enzymes. In some embodiments, metabolites can be non-protein, plasma-derived metabolite markers, for example, including but not limited to DAS, NANA, NAcMan, NAcLac, DiAcSpmd, N3AP and HBA. In some embodiments, useful metabolites as described herein can be "polyamines", i.e., organic compounds with more than two amino groups. In some embodiments, polyamines as described herein are plasma polyamines. In some embodiments, polyamines that can be used in this group and method include but are not limited to DAS, NANA, NAcMan, NAcLac, DiAcSpmd, N3AP and HBA. These polyamines can be combined with other markers (e.g., CA125 or HE4) to enhance ovarian cancer detection as described herein.

[0112] As used herein, the term "7-marker metabolite panel" or "7MetP" refers to a group of seven biomarkers, including DAS, NANA, NAcMan, NAcLac, DiAcSpmd, N3AP and HBA, that can be used to detect ovarian cancer in patients suspected of having ovarian cancer. In some embodiments, the 7-marker metabolite panel can be evaluated in combination with additional markers, such as plasma polyamines, to enhance the detection of ovarian cancer in biological samples from patients suspected of having ovarian cancer. Useful plasma polyamines include, but are not limited to, N3AP, AcSpmd, DiAcSpmd and / or DAS. Additional polyamines are known in the art and may be included as deemed appropriate by the clinician.

[0113] As used herein, the term "ROC" refers to a receiver operating characteristic, which is a graphical diagram used herein to measure the performance of a diagnostic method at various cutoff points. A ROC diagram can be constructed from the fractions of true positives and false positives at different cutoff points.

[0114] As used herein, the term "AUC" refers to the area under the ROC curve. AUC can be used to estimate the predictive power of a diagnostic test. Generally, the larger the AUC, the stronger the predictive power and the lower the frequency of prediction errors. The possible values ​​of AUC range from 0.5 to 1.0, where a value of 1.0 represents an error-free prediction method.

[0115] As used herein, the term "p-value" or "p" refers to the probability that the biomarker score distributions for ovarian cancer-positive and ovarian cancer-negative subjects are identical in the context of a Wilcoxon rank sum test. Typically, a p-value close to zero indicates that a particular statistical method has a high predictive ability in classifying a subject.

[0116] As used herein, the term "CI" refers to a confidence interval, ie, an interval in which a certain value can be predicted depending on a certain confidence level. As used herein, the term "95% CI" refers to an interval in which a certain value can be predicted depending on a 95% confidence level.

[0117] As used herein, the term "disease progression" or "early disease progression" is defined as an upgrade in Gleason score and / or an increase in tumor volume on surveillance biopsy within 18 months of initiation of active surveillance.

[0118] The phrase "therapeutically effective" is intended to qualify the amount of active ingredient used to treat a disease or disorder or to affect a clinical endpoint. List of abbreviations

[0119] AUC = area under the curve; DAS = N1,N12-diacetylspermine; DiAcSpmd = N1,N8-diacetylspermidine; HBA = hydroxyisobutyric acid; HILIC = hydrophilic interaction liquid chromatography; HPLC = high performance liquid chromatography; N3AP = N-(3-acetamidopropyl)pyrrolidin-2-one; NANA = N-acetylneuraminic acid ester; NAcMan = N-acetyl-mannosamine; NAcLac = N-acetyl-lactosamine; OvCa = ovarian cancer; ROC = receiver operating characteristic; SEM = standard error of the mean; TCGA = The Cancer Genome Atlas; UPLC = ultra-performance liquid chromatography; UPLC / MS = ultra-performance liquid chromatography / mass spectrometry. Examples

[0120] The following examples are included to illustrate embodiments of the present disclosure. The following examples are presented only by way of illustration and to assist those of ordinary skill in the use of the present disclosure. These examples are not intended to limit the scope of the present disclosure in any way. It will be appreciated by those skilled in the art based on the present disclosure that many changes may be made to the specific embodiments disclosed and still obtain the same or similar results without departing from the spirit and scope of the present invention. Example 1: Sample set

[0121] Blood samples were collected preoperatively from patients admitted for surgery based on ultrasound findings of a mass, elevated CA125, or positive biopsy at the University of Texas MD Anderson Cancer Center (MDACC) and Fred Hutchinson Cancer Research Center (FHCRC, IRB 4563) under an IRB / ethics committee-approved protocol (LAB04-0687) with informed consent. All patients were fasting at the time of blood collection. Samples were processed the same day, usually within 4 hours of blood draw, aliquoted according to standardized operating procedures to minimize freeze-thaw cycle effects, and stored at −80°C until use. The sample set included plasma from 59 patients with stage I-II invasive epithelial ovarian cancer, 160 patients with stage III-IV invasive epithelial ovarian cancer, and 190 patients with benign pelvic masses. Biopsy samples were reviewed by a qualified pathologist for the diagnosis of cancer or benign pelvic conditions. Detailed patient and tumor characteristics are provided in Table 1 . Information on histological ovarian cancer subtypes and benign etiologies is provided in Table 2 . All participants consented to the use of samples in ethically approved secondary research. Table 1. Patient and tumor characteristics Individuals with benign pelvic masses (BPM) For continuous variables, the Wilcoxon rank sum test was used to determine statistical significance, while for categorical variables, Fisher's exact test or 2 Statistical significance was determined using the trend test. Two-sided p values ​​are reported. Table 2. Characteristics of ovarian cancer and benign pelvic masses Example 2: Metabolomics analysis Primary metabolites and biogenic amines

[0122] Serum metabolites were extracted from pre-aliquoted EDTA plasma (10 μL) with 30 μL LCMS grade methanol (ThermoFisher) in a 96-well microplate (Eppendorf). The plate was heat sealed, vortexed at 750 rpm for 5 min, and centrifuged at 2000 × g for 10 min at room temperature. The supernatant (10 μL) was carefully transferred to a 96-well plate, leaving the precipitated protein. The supernatant was further diluted with 10 μL 100 mM ammonium formate (pH 3). For hydrophilic interaction liquid chromatography (HILIC) analysis, the samples were diluted with 60 μL LCMS grade acetonitrile (ThermoFisher), while for C18 analysis, the samples were diluted with 60 μL water (GenPure ultrapure water system, ThermoFisher). Each sample solution was transferred to a 384-well microplate (Eppendorf) for LCMS analysis. Non-targeted analysis of primary metabolites and biogenic amines

[0123] In a Waters Acquity TM Untargeted metabolomics analysis was performed on a UPLC system with a 2D column regeneration configuration (type I and type H). TM UPLC BEH Amide, 1.7 μm 2.1 × 100 mm, Waters Corporation (Milford, USA) and C18 (Acquity TM UPLC HSS T3, Chromatographic separation was performed at 45 °C on a 1.8 μm, 2.1 × 100 mm, Waters Corporation, Milford, TN, USA) column.

[0124] The quaternary solvent system mobile phase was (A) 0.1% formic acid in water, (B) 0.1% formic acid in acetonitrile, and (D) 100 mM ammonium formate (pH 3). The samples were separated on the HILIC column using the following gradient profile: a starting gradient of 95% B and 5% D increased linearly to 70% A, 25% B, and 5% D in 5 min at a flow rate of 0.4 mL / min, followed by a 1 min isocratic gradient at 100% A at a flow rate of 0.4 mL / min. For the C18 separation, the chromatographic gradient was performed as follows: starting conditions were 100% A, linearly increased to final conditions 5% A, 95% B, followed by a 1 min isocratic gradient at 95% B, 5% D.

[0125] Column regeneration and equilibration were performed using a binary pump. The solvent system mobile phase was (A1) 100 mM ammonium formate (pH 3), (A2) 0.1% formic acid in 2-propanol, and (B1) 0.1% formic acid in acetonitrile. The HILIC column was stripped for 5 min using 90% A2, followed by equilibration for 2 min using 100% B1 at a flow rate of 0.3 mL / min. A reverse phase C18 column was regenerated for 2 min using 95% A1, 5% B1, followed by 5 min column equilibration using 5% A1, 95% B1. Mass spectrometry data acquisition

[0126] Mass spectrometry data were acquired in positive and negative electrospray ionization mode using the “sensitivity” mode in the range of 50–1200 Da for primary metabolites and 100–2000 Da for complex lipids. For electrospray acquisition, the capillary voltage was set to 1.5 kV (positive), 3.0 kV (negative), the sampling cone voltage was 30 V, the source temperature was 120 °C, the cone gas flow was 50 L / h, and the desolvation gas flow rate was 800 L / h, in continuous mode with a scan time of 0.5 s. Leucine enkephalin; 556.2771 Da (positive) and 554.2615 Da (negative) were used for lockspray correction and scanned at 0.5 min. The injection volume for each sample was 3 μL unless otherwise stated. Acquisition was performed with instrument automatic gain control to optimize the sensitivity of the instrument within the sample acquisition time. Data processing

[0127] Data were processed using Progenesis QI (Nonlinear, Waters). Peak extraction and retention time alignment of LC-MS and MSe data were performed using Progenesis QI software (Nonlinear, Waters). Data processing and peak annotation were performed using in-house automated processes. Annotations were determined by matching accurate mass and retention time using a custom library created from authentic standards and by matching experimental tandem mass spectrometry data to NIST MSMS, LipidBlast, or HMDB v3 theoretical fragments; for complex lipids, retention time patterns specific to lipid subclasses were also considered. To correct for injection sequence drift, each feature was normalized using data from duplicate injections of quality control samples collected every 10 injections throughout the run sequence. Measured data were smoothed by locally weighted scatter plot smoothing (LOESS) signal correction (QC-RLSC) as described previously. Values ​​are reported as ratios relative to the median values ​​of historical quality control reference samples for each analytical batch run for a given analyte. Determination of CA125 and HE4

[0128] Serum CA125 and HE4 concentrations were measured using the Architect CA125II assay (Abbott Diagnostics, Abbott Park) and the HE4 EIA assay (Fujirebio Diagnostics, Malvern, PA). To calculate the ROMA score, the predictive index (PI) was calculated using serum HE4 and CA125 II levels and one of the following formulas, depending on the patient's menopausal status: 1. Premenopause: Predictive index (PI) = -12.0 + 2.38*ln[HE4] + 0.0626*ln[CA125] 2. Postmenopausal: Predictive Index (PI) = -8.09 + 1.04*ln[HE4] + 0.732*ln[CA125]

[0129] The following formula uses each patient's prediction index (PI) to calculate the Risk of Ovarian Malignancy Algorithm (ROMA) score: Example 3: Statistical Analysis

[0130] Figure 2An overall schematic workflow of the study is provided. Metabolite selection and model building were performed using metabolic profiles generated from serum samples from FHCRC. Relevant variables to be included in the model were prioritized using the method reported by Gedeon. This approach removes irrelevant or noisy variables by analyzing the relative weight of each variable in the entire data matrix. The importance score is calculated by dividing the absolute value of the weight of an input connected to the output by the total absolute value of all weights from that input. When applied to deep learning models, the method is recursively extended backwards through the layers by considering the influence of neurons on connected nodes and then multiplying the derived weights by the influence of a given node on the target output and summing over all connected nodes.

[0131] Here, P jk It represents the average contribution of node j in one layer to node k in the next layer. w is the weight of the connection and nh is the number of nodes in the next layer.

[0132] The contribution of the input neuron to the output is:

[0133] Using this approach, 20 iterations were performed, with each iteration slightly adjusting the hyperparameters and recalculating the relative variable importance score for each metabolite. Metabolites that consistently produced a relative variable importance score > 0.7 (corresponding to metabolites with importance scores in the top 30 percentile) across all 20 iterations were selected to develop an algorithm for distinguishing early OvCa from benign disease. Seven models (including deep learning, random forest, ensemble learning, and gradient boosting method algorithms) incorporating seven metabolites were evaluated for distinguishing early OvCa from benign disease. The performance of the models was evaluated using 5-fold cross validation. To further assess the stability of the model, perturbations (e.g., random selection and replacement) were introduced into the training set and performance was re-evaluated.

[0134] A deep learning model (DLM) with 3 hidden layers and 3 nodes per layer was selected to model the 7-marker metabolite panel (7MetP) based on AUC and tested 7MetP for detecting OvCa in the MDACC cohort using fixed parameters.

[0135] To assess the contribution of 7MetP and ROMA, logistic regression was first fitted with 7MetP and ROMA as two independent predictors (Table 3). For ROMA, percentage hazards were used as described above. Initial modeling was performed using early OvCa cases and individuals with BPM from FHCRC, and the model was tested in the MDACC cohort.

[0136] To directly compare the performance of the combined 7MetP+ROMA model with ROMA, fixed risk thresholds (11.4% for premenopausal women and 29.9% for postmenopausal women) were used, and positive predictive value (PPV), negative predictive value (NPV), and sensitivity and specificity estimates were calculated.

[0137] The combined score of the logistic regression model was converted to risk by the following formula: exp(combined score) / (1+exp(combined score)).

[0138] The discrimination of the model was evaluated based on the receiver operating characteristic curve (ROC) and the sensitivity and specificity estimates. The 95% confidence interval (CI) of the AUC was estimated using the Delong method. The P values ​​for specificity and sensitivity were estimated by calculating the 2.5 and 97.5 percentiles of the δ values ​​of 1000 boot straps. All modeling was performed using the h2o package and the R statistical program. Table 3. Estimated coefficients of the combined model of 7MetP plus ROMA. intercept ROMA 7MetP coefficient -3.15 4.06 4.12 Example 4: Cancer-related metabolite database

[0139] An untargeted metabolomics study was performed on a serum training set of 101 OvCa cases (39 early and 62 late) and 134 subjects with BPM from the Fred Hutchinson Cancer Research Center (FHCRC) (Table 1). A total of 475 uniquely annotated metabolites were quantified (Table 4). To prioritize the metabolites, relative importance scores were calculated using the Gedeon method, and metabolites were selected based on importance scores that consistently showed above 0.7. This approach resulted in seven metabolites being selected for model building, each with prior evidence of association with cancer: diacetylspermidine (DAS), diacetylspermidine (DiAcSpmd), N-(3-acetamidopropyl)pyrrolidin-2-one (N3AP), N-acetylneuraminic acid (NANA), N-acetyl-mannosamine (NAcMan), N-acetyl-lactosamine (NAcLac), and hydroxyisobutyric acid (HBA). Individual classifier performance of these metabolites for distinguishing OvCa cases from individuals with BPM ranged from 0.55 to 0.82 ( Table 5 ; Figure 3 ). Table 5. Individual prediction performance of selected metabolites in the training set. Example 5: Model building and testing

[0140] An optimal combination rule incorporating seven metabolites was developed for distinguishing early OvCa from benign disease. For model construction, seven different machine learning algorithms were tested. Among them, a deep learning model (DLM) with 3 hidden layers and 3 nodes per layer achieved the highest predictive performance and was used to establish a 7-marker metabolite panel (7MetP), which had an AUC of 0.75 (95% CI: 0.66-0.85) in distinguishing early OvCa cases from benign disease (Tables 6-8). When OvCa cases were stratified into serous and non-serous, the AUCs of 7MetP were 0.85 (95% CI: 0.79-0.91) and 0.80 (95% CI: 0.71-0.89), respectively (Table 9).

[0141] 7MetP was validated using fixed parameters in an independent test set of 118 OvCa cases (20 early-stage, 98 late-stage) and 56 individuals with BPM at the MD Anderson Cancer Center (MDACC). 7MetP had an AUC of 0.88 (95% CI: 0.82-0.93) in distinguishing all OvCa cases from individuals with BPM (Table 7) and an AUC of 0.86 (95% CI: 0.76-0.95) in early-stage OvCa (Table 7). Figure 1 ; Table 7). Table 6. Performance of different learning algorithms in distinguishing early OvCa cases from BPM using 5-fold cross validation in the training set. 1 AUC: Area under the ROC curve 2 AUCpr: Area under the precision-recall curve 3 RMSE: Root Mean Square Deviation Table 7. Performance of the 7-marker metabolite panel (7MetP) in distinguishing OvCa cases from individuals with BPM in the training set, independent test set, and combined training+test sample sets. Table 8. Stability check of deep learning model (DLM) in the training set. Table 9.7 Predictive performance of MetP for distinguishing OvCa cases stratified into serous and non-serous and BPM in the training set. Contribution of metabolite groups using the ROMA algorithm

[0142] Next, 7MetP was evaluated to determine whether it improved the predictive performance of the ROMA algorithm. Using model scores derived from the 7MetP and ROMA algorithms, a logistic regression model for distinguishing early OvCa from BPM was developed in the training set, and its performance was evaluated in the test set. In the test set, the AUC of the 7MetP+ROMA combination in early OvCa was 0.93 (95% CI: 0.86-1.00), while the AUC of ROMA alone was 0.91 (95% CI: 0.84-0.98) (likelihood ratio test p: 0.03). Compared with ROMA, the 7MetP+ROMA combination improved PPV by 21.0% (one-sided p<.001) and specificity by 14.0% (one-sided p<.001) in early OvCa (Table 10). When all OvCa cases were considered, the AUC of the 7MetP+ROMA combined model in the test set was 0.97 (95% CI: 0.94-0.99) (Table 11). Table 10. Performance estimates of ROMA and the combined 7MetP+ROMA model for early OvCa in the training set and independent test set. PPV: Positive Predictive Value NPV: Negative predictive value P: P value of the likelihood ratio test Table 11. Performance estimates of ROMA and the combined 7MetP+ROMA model for all OvCa in the training set and independent test set. PPV: Positive Predictive Value NPV: Negative predictive value P: P value of likelihood ratio test Performance of metabolite groups individually and in combination with ROMA in the combined training and test sets.

[0143] The predictive performance of 7MetP alone and in combination with ROMA was further evaluated in the entire sample set (n=219 OvCa cases (59 early and 160 late) and 190 BPM). The AUC of 7MetP in distinguishing all OvCa cases from individuals with BPM was 0.85 (95% CI: 0.81-0.88) and the AUC in early OvCa was 0.81 (95% CI: 0.76-0.86) (Figure 7). The final AUC of the 7MetP+ROMA combined model in early OvCa was 0.87 (95% CI: 0.85-0.93), which was a significant improvement over ROMA alone (AUC: 0.84 (95% CI: 0.81-0.90); likelihood ratio test p value: <0.001) (Tables 12 and 13). Importantly, the 7MetP+ROMA model produced statistically significantly (one-sided P < .001) higher PPV (0.68 vs 0.52) and specificity (0.89 vs 0.78) for early OvCa compared with ROMA alone (Table 5). Table 12. Performance estimates of ROMA and the combined 7MetP+ROMA model for early OvCa in the combined sample set. PPV: Positive Predictive Value NPV: Negative predictive value P: P value of the likelihood ratio test Table 13. Performance estimates of ROMA and the combined 7MetP+ROMA model for all OvCa in the combined sample set.

[0144] All references, patents, or applications cited in this application, whether U.S. or foreign, are incorporated herein by reference in their entirety, as if written herein. In the event of any inconsistency, the actual disclosure herein controls.

[0145] From the above description, those skilled in the art can easily ascertain the essential characteristics of the present invention and without departing from the spirit and scope of the invention, can make various changes and modifications to the invention to adapt it to different usages and conditions.

Claims

1. A method of treating ovarian cancer in a patient having elevated levels of diacetylspermine, diacetylspermidine, N-(3-acetamidopropyl)pyrrolidin-2-one, N-acetylneuraminate, N-acetyl-mannosamine, N-acetyl-lactosamine, and hydroxyisobutyric acid, and optionally elevated levels of human epididymis protein 4 (HE4) and mucin 16 (CA125), wherein the elevated levels classify the patient as having ovarian cancer, the method comprising administering to the patient a therapeutically effective amount of an ovarian cancer treatment.

2. A method for treating ovarian cancer, the method comprising: a) identifying a patient having elevated levels of diacetylspermine, diacetylspermidine, N-(3-acetamidopropyl)pyrrolidin-2-one, N-acetylneuraminate, N-acetyl-mannosamine, N-acetyl-lactosamine and hydroxyisobutyric acid and optionally elevated levels of human epididymis protein 4 (HE4) and mucin 16 (CA125), wherein these elevated levels classify the patient as having ovarian cancer; and b) administering to the patient a therapeutically effective amount of an ovarian cancer treatment.

3. A method of distinguishing between ovarian cancer and benign pelvic mass (BPM) in a subject, the method comprising, in a biological sample obtained from the subject: a) measuring the levels of diacetylspermine, diacetylspermidine, N-(3-acetamidopropyl)pyrrolidin-2-one, N-acetylneuraminate, N-acetyl-mannosamine, N-acetyl-lactosamine and hydroxyisobutyric acid, and optionally the levels of human epididymis protein 4 (HE4) and mucin 16 (CA125) in the biological sample; and b) classifying the subject as having ovarian cancer or BPM based on the measured levels.

4. A method for determining a subject's risk of developing ovarian cancer, the method comprising, in a biological sample obtained from the subject: a) measuring the levels of diacetylspermine, diacetylspermidine, N-(3-acetamidopropyl)pyrrolidin-2-one, N-acetylneuraminate, N-acetyl-mannosamine, N-acetyl-lactosamine and hydroxyisobutyric acid, and optionally the levels of human epididymis protein 4 (HE4) and mucin 16 (CA125) in the biological sample; and b) classifying the subject as being at risk for ovarian cancer or not at risk for ovarian cancer based on the measured level.

5. A method for generating a risk profile for ovarian cancer in a subject, the method comprising, in a biological sample obtained from the subject: a) measuring the levels of diacetylspermine, diacetylspermidine, N-(3-acetamidopropyl)pyrrolidin-2-one, N-acetylneuraminate, N-acetyl-mannosamine, N-acetyl-lactosamine and hydroxyisobutyric acid, and optionally the levels of human epididymis protein 4 (HE4) and mucin 16 (CA125) in the biological sample; and b) classifying the subject as being at risk for ovarian cancer or not at risk for ovarian cancer based on the measured level.

6. A method for risk stratifying a patient at risk for developing ovarian cancer, the method comprising, in a biological sample obtained from the patient: (a) measuring the levels of diacetylspermine, diacetylspermidine, N-(3-acetamidopropyl)pyrrolidin-2-one, N-acetylneuraminate, N-acetyl-mannosamine, N-acetyl-lactosamine and hydroxyisobutyric acid, and optionally the levels of human epididymis protein 4 (HE4) and mucin 16 (CA125) in the biological sample; and (b) determining, by a processor circuit, a risk score for the patient, wherein the risk score is determined by a scoring function derived from metabolite profiles of biological samples collected from a plurality of individuals undergoing ovarian cancer monitoring.

7. A method for calculating a biomarker score or a risk score for ovarian cancer in a patient, the method comprising, in a biological sample obtained from the patient: (a) measuring the levels of diacetylspermine, diacetylspermidine, N-(3-acetamidopropyl)pyrrolidin-2-one, N-acetylneuraminate, N-acetyl-mannosamine, N-acetyl-lactosamine and hydroxyisobutyric acid, and optionally the levels of human epididymis protein 4 (HE4) and mucin 16 (CA125) in the biological sample; and (b) Using the numerical values ​​of the measured levels in a deep learning model (DLM) to calculate the biomarker score or risk score.

8. The method of claim 7, wherein the DLM comprises an artificial neural network having three hidden layers and three nodes per layer.

9. The method of any one of claims 1-7, further comprising measuring the levels of human epididymis protein 4 (HE4) and mucin 16 (CA125) or identifying a patient with elevated levels of human epididymis protein 4 (HE4) and mucin 16 (CA125).

10. The method of any one of claims 1-9, wherein the levels of HE4 and CA125 are determined by immunoassay.

11. The method of claim 10, wherein the levels of HE4 and CA125 are used to calculate the prediction index (PI) for premenopausal women using the formula: PI=-12.0+2.38*ln[HE4]+0.0626*ln[CA125].

12. The method of claim 10, wherein the levels of HE4 and CA125 are used to calculate the prediction index (PI) for postmenopausal women using the formula: PI=-8.09+1.04*ln[HE4]+0.732*ln[CA125].

13. The method of claim 11 or 12, wherein the predictive index (PI) is used to calculate the ovarian malignancy risk algorithm (ROMA) score, the formula being:

14. The method of claim 13, wherein the combined model score is calculated using logistic regression of the ROMA score and the biomarker score.

15. The method of any one of claims 1-14, wherein the ovarian cancer is early stage (eg, stage I or stage II).

16. The method of any one of claims 1-14, wherein the ovarian cancer is advanced (eg, stage III or stage IV).

17. The method of any one of claims 1-7, wherein the levels of diacetylspermine, diacetylspermidine, N-(3-acetamidopropyl)pyrrolidin-2-one, N-acetylneuraminate, N-acetyl-mannosamine, N-acetyl-lactosamine and hydroxyisobutyric acid, and optionally HE4 and CA125 are elevated relative to a reference patient or group not suffering from ovarian cancer.

18. The method of any one of claims 1-7, wherein the levels of diacetylspermine, diacetylspermidine, N-(3-acetamidopropyl)pyrrolidin-2-one, N-acetylneuraminate, N-acetyl-mannosamine, N-acetyl-lactosamine and hydroxyisobutyric acid, and optionally HE4 and CA125 are elevated relative to a reference patient or group with benign pelvic mass (BPM).

19. The method of any one of claims 1-7, wherein the subject presents with a pelvic mass.

20. The method of any preceding claim, wherein each of diacetylspermine, diacetylspermidine, N-(3-acetamidopropyl)pyrrolidin-2-one, N-acetylneuraminate, N-acetyl-mannosamine, N-acetyl-lactosamine and hydroxyisobutyric acid, and optionally HE4 and CA125, produces a detectable signal.

21. The method of claim 20, wherein the detectable signal is detectable by spectroscopy.

22. The method of claim 21, wherein the spectroscopy is selected from the group consisting of UV-visible spectroscopy, mass spectrometry, nuclear magnetic resonance (NMR) spectroscopy, proton NMR spectroscopy, nuclear magnetic resonance (NMR) spectroscopy, gas chromatography-mass spectrometry (GC-MS), liquid chromatography-mass spectrometry (LC-MS), correlation spectroscopy (COSY), nuclear Overhauser effect spectroscopy (NOESY), rotating coordinate nuclear Overhauser effect spectroscopy (ROESY), time-of-flight LC-MS (LC-TOF-MS), liquid chromatography-tandem mass spectrometry (LC-MS / MS), and capillary electrophoresis-mass spectrometry.

23. The method of claim 22, wherein the spectroscopic method is mass spectrometry.

24. The method of claim 23, wherein the mass spectrometry is LC-TOF-MS.

25. The method of claim 1, 2, 15 or 16, wherein the treatment is selected from surgery, chemotherapy, immunotherapy, radiation therapy, targeted therapy or a combination thereof.

26. The method of any one of claims 1-7, wherein the measured levels are used to calculate a biomarker score or risk profile based on sensitivity and specificity values ​​corresponding to the subject's risk of developing ovarian cancer.

27. The method of claim 26, wherein the sensitivity and specificity values ​​are not significantly different from the curves in Figure 1.

28. The method of claim 27, wherein the sensitivity and specificity values ​​differ by less than 10%.

29. The method of claim 28, wherein the difference in the sensitivity and specificity values ​​is less than 5%.

30. The method of claim 29, wherein the sensitivity and specificity values ​​differ by less than 1%.

31. The method of any one of claims 1-7, wherein the cutoff value comprises an AUC (95% CI) of at least 0.

76.

32. The method of any preceding claim, further comprising assigning the patient to an appropriate risk group based on the calculated risk score.

33. The method of claim 32, wherein there are at least two risk groups.

34. The method of any one of claims 1-7, wherein the AUC of the method is greater than the AUC of a different single biomarker, multiple biomarkers, panel, assay, or algorithm incorporating a combination thereof.

35. The method of claim 34, wherein the AUC is greater than 0.

76.

36. The method of claim 35, wherein the AUC is between 0.76 and 0.

95.

37. The method of claim 36, wherein the AUC is about 0.

88.

38. The method of claim 36, wherein the AUC is about 0.

86.

39. The method of claim 35, wherein the AUC is between 0.82 and 0.

93.

40. The method of claim 39, wherein the AUC is about 0.

87.

41. The method of any one of claims 1-7, wherein the positive predictive value (PPV) of the method is greater than the PPV of a different single biomarker, multiple biomarkers, panel, assay, or algorithm incorporating a combination thereof.

42. The method of claim 41, wherein the PPV is greater than 0.

67.

43. The method of claim 42, wherein the PPV is between 0.67 and 0.

87.

44. The method of claim 43, wherein the PPV is about 0.

79.

45. The method of any one of claims 34-44, wherein the algorithm is the Risk of Ovarian Malignancy Algorithm (ROMA).

46. ​​The method of any one of claims 34-44, wherein the biomarkers are only HE4 and CA125.

47. The method of claim 45 or 46, wherein the cut-off point of each method is used for classification.

48. The method of claim 45 or 46, analyzed by the same statistical method.

49. The method of any preceding claim, wherein the levels of diacetylspermine, diacetylspermidine, N-(3-acetamidopropyl)pyrrolidin-2-one, N-acetylneuraminate, N-acetyl-mannosamine, N-acetyl-lactosamine and hydroxyisobutyric acid are measured according to a given threshold value or thresholds.

50. The method of claim 49, wherein the values ​​exceed the one or more thresholds, classifying the patient as being at risk for ovarian cancer.

51. The method of claim 50, wherein the values ​​are below the one or more thresholds, classifying the patient as not at risk for ovarian cancer.

52. The method of claim 50, wherein the values ​​are below the one or more thresholds, classifying the patient as having BPM.

53. The method of claim 51, wherein the patient is subsequently designated to receive further ovarian cancer screening or treatment.

54. The method of claim 53, wherein the screening is selected from endoscopic ultrasound, magnetic resonance imaging (MRI), and computed tomography (CT).

55. The method of claim 54, wherein the screening is performed annually.

56. The method of claim 54, wherein the screening is performed semi-annually.

Images