Urine metabolome markers and their use in differentiating urinary tumor models, devices

Abstract

The present application relates to the technical field of in vitro diagnosis, in particular to urine metabolome markers and application thereof in identifying urinary system tumor models and devices. The present application provides a SALDI-MS urinary system tumor identification model and device. The multi-metal component synergistic effect, tunable energy band structure, excellent light-heat conversion efficiency and rich surface defect sites of the sample pretreatment material are used to significantly enhance the laser energy absorption and interface charge transfer efficiency, realize high-throughput, high-sensitivity and low-background ionization of small molecule metabolites in urine, and combine machine learning algorithms to construct a multi-cancer identification model, thereby realizing efficient, accurate and synchronous identification of bladder cancer, kidney cancer and prostate cancer. The method has the advantages of simple operation, low cost and large-scale promotion.

Description

Technical Field

[0001] This invention relates to the field of in vitro diagnostic technology, and in particular to the application of urinary metabolomics in the effective identification of urinary system tumors. Background Technology

[0002] Effective differentiation of urinary system tumors (renal cancer, prostate cancer, bladder cancer) is of core clinical significance in directly guiding precise individualized treatment, improving patient prognosis, and avoiding overtreatment or unnecessary invasive examinations due to misdiagnosis. However, current clinical practice still faces significant technical bottlenecks: among traditional methods, cystoscopy, while considered the "gold standard," is invasive; urine cytology is non-invasive but lacks sensitivity; and the widely used serum PSA screening has low specificity and is prone to overdiagnosis. While the emerging urine ctDNA testing shows great promise, it is limited by translational barriers such as high technical costs and cannot yet completely replace traditional methods. Therefore, overcoming existing bottlenecks and developing high-precision, non-invasive, and widely applicable differentiation techniques is key to improving the diagnosis and treatment of urinary system tumors.

[0003] Current technologies for the differential diagnosis of urinary system tumors mainly rely on imaging, tissue biopsy, and serum biomarker detection. Imaging examinations (such as CT and MRI) have limited ability to identify early, small lesions and are difficult to accurately distinguish tumor types; while tissue biopsy is the gold standard, it is an invasive procedure with risks of bleeding, infection, and tumor seeding and dissemination, making it unsuitable for large-scale screening and dynamic monitoring; serum biomarkers (such as PSA) have insufficient specificity, which can easily lead to false positives and overdiagnosis.

[0004] In recent years, liquid biopsy technology based on urinary metabolomics has attracted much attention due to its advantages such as non-invasiveness and repeatable sampling. Currently, the main research methods in omics include nuclear magnetic resonance (NMR), gas chromatography-mass spectrometry (GC-MS), and liquid chromatography-mass spectrometry (LC-MS). Chromatographic separation takes time, and NMR has lower sensitivity compared to MS technology, thus limiting the rapid, sensitive, and high-throughput detection of metabolites in urine. Urine, as a biological sample with a high salt content, requires cumbersome and time-consuming pretreatment. Therefore, constructing a simple pretreatment urinary metabolomics platform based on the high-throughput MALDI-TOF MS platform is essential for practical clinical applications.

[0005] Urinary metabolomics based on SALDI-TOF MS is an ideal technical approach for identifying urinary system tumors. This technology uses inorganic nanomaterials to replace traditional organic matrices, which can effectively eliminate background interference in the low molecular weight range, thereby capturing specific small molecule fingerprints in urine with high sensitivity and high reproducibility generated by tumor metabolic reprogramming.

[0006] Current SALDI-MS methods for urine diagnosis still face significant bottlenecks: First, traditional SALDI-MS matrix materials (such as organic matrices and single metal oxide nanomaterials) suffer from low laser energy absorption efficiency, insufficient charge transfer capacity, and strong background interference, resulting in weak signals from low-abundance metabolites and difficulty in meeting clinical needs in terms of detection sensitivity and repeatability. Second, existing metabolic profiling methods mostly focus on single cancer types and lack the ability to simultaneously differentiate between multiple urinary system tumors such as bladder cancer, kidney cancer, and prostate cancer. Third, data analysis strategies are still mainly based on univariate statistics, which makes it difficult to fully explore the subtle differences in high-dimensional metabolic profiles, thus limiting diagnostic accuracy. Summary of the Invention

[0007] In view of this, the purpose of this invention is to provide a urine metabolism diagnostic technology based on high-entropy metal oxide (HEO)-assisted SALDI-MS. Utilizing the synergistic effect of the multi-metal components, tunable band structure, excellent photothermal conversion efficiency, and abundant surface defect sites of HEO materials, the laser energy absorption and interfacial charge transfer efficiency are significantly enhanced, achieving high throughput, high sensitivity, and low background ionization of small molecule metabolites in urine. Combined with machine learning algorithms, a multi-cancer identification model is constructed to achieve efficient, accurate, and simultaneous identification of bladder cancer, kidney cancer, and prostate cancer. This method has advantages such as being non-invasive, easy to operate, low-cost, and scalable, aiming to provide a novel technological platform for the early screening, auxiliary diagnosis, and accurate classification of urinary system tumors.

[0008] To achieve the above-mentioned objectives, the present invention provides the following technical solution:

[0009] This invention provides the application of urinary metabolomics biomarkers in the preparation of products for identifying urinary system tumors. The urinary metabolomics biomarkers include: 4-hydroxy-L-glutamate, 8-hydroxyguanosine, arginine, ascorbic acid, cysteine-S-sulfate, cysteylglycine, cytidine, glycolic acid, guanosine, hippuric acid, hydroquinone, hypoxanthine, 3-(imidazol-4-yl)propionic acid, imidazolide, indoleacrylic acid, L-3-hydroxykynurenine, L-cystathionine, maleic acid, malic acid, N-formyl-L-methionine, nicotinamide, paraxanthine, phthalic acid, pimelic acid, pyruvic acid, S-cysteylsuccinic acid, guanidinotaurine, and uracil.

[0010] The urinary system tumors include bladder cancer, kidney cancer, and / or prostate cancer.

[0011] In some specific embodiments of the present invention, the product described above acquires the characteristic signal intensity of urinary metabolomics biomarkers in the sample to be tested based on SALDI-MS, and identifies urinary system tumors based on the characteristic signal intensity of urinary metabolomics biomarkers.

[0012] The SALDI-MS is based on a high-entropy metal oxide nanomatrix.

[0013] The high-entropy metal oxide nanomatrix includes high-entropy metal oxides with metal elements of Fe, Co, Ni, Cu, and Zn.

[0014] In some specific embodiments of the present invention, the identification described above includes:

[0015] Based on the intensity of the characteristic signal of the first biomarker group in the sample to be tested, the SVR model is used to determine whether the sample to be tested is a urinary system tumor sample or a non-urinary system tumor sample.

[0016] Based on the intensity of the characteristic signal of the second biomarker group in the urinary system tumor sample, the SVR model is used to determine whether the sample to be tested is a non-bladder cancer urinary system tumor sample or a bladder cancer sample.

[0017] Based on the intensity of the characteristic signal of the third biomarker group in the non-bladder cancer urinary system tumor sample, the SVR model is used to determine whether the sample to be tested is a prostate cancer sample or a kidney cancer sample.

[0018] The first biomarker group includes hydroquinone, uracil, maleic acid, malic acid, hypoxanthine, pimelic acid, guanidine taurine, arginine, ascorbic acid, S-cysteine ​​succinate, cytidine, guanosine, and 8-hydroxyguanosine.

[0019] The second biomarker group includes glycolic acid, imidazoline, pyruvate, nicotinamide, 3-(imidazol-4-yl)propionic acid, 4-hydroxy-L-glutamic acid, phthalic acid, N-formyl-L-methionine, cysteylglycine, hippuric acid, paraxanthine, indoleacrylic acid, cysteine-S-sulfate, L-cystathionine, and L-3-hydroxykynurenine.

[0020] The third biomarker group includes glycolic acid, imidazoline, pyruvate, nicotinamide, 3-(imidazo-4-yl)propionic acid, 4-hydroxy-L-glutamic acid, phthalic acid, and indoleacrylic acid.

[0021] In some specific embodiments of the present invention, the SVR model described above is trained based on clinical samples with known diagnostic results;

[0022] The training methods include:

[0023] Multiple clinical samples with known diagnostic results are taken as training samples. The predictive variables in the training samples are standardized, and the support vector regression model is trained using the training samples.

[0024] The support vector regression model uses the following parameters:

[0025] (i) The kernel function is a second-order polynomial kernel with a penalty coefficient C of 0.3706, an insensitive band width of 0.0371, and a kernel scale parameter of 2.906; or

[0026] (ii) The kernel function is a second-order polynomial kernel with a penalty coefficient C of 1.4826, an insensitive band width of 0.1483, and a kernel scale parameter of 3.6030; or

[0027] (iii) The kernel function is a second-order polynomial kernel with a penalty coefficient C of 0.7413, an insensitive band width of 0.0741, and a kernel scale parameter of 3.2785.

[0028] In some specific embodiments of the present invention, the characteristic signal intensity of the above application includes the normalized mass spectrometry signal intensity.

[0029] In some specific embodiments of the present invention, the method for preparing the high-entropy metal oxide described above includes:

[0030] Will include Fe 3+ Co 2+ Ni 2+ Cu 2+ and Zn 2+ The solution was mixed evenly with urea and polyvinylpyrrolidone, and the pH value was adjusted to 10-11 with sodium hydroxide to obtain the precursor solution.

[0031] The precursor solution was hydrothermally treated at 180°C for 12 hours, then cooled, separated, washed, dried, and annealed at 500°C for 2 hours under normal pressure by heating at 5°C / min to 500°C to obtain the high-entropy metal oxide.

[0032] The Fe 3+ Co 2+ Ni 2+ Cu 2+ and Zn 2+ The molar ratio is 1:1:1:1:1, and the concentration is 0.02 M;

[0033] The concentration of the urea is 0.1 M;

[0034] The mass fraction of the polyvinylpyrrolidone is 1%.

[0035] The diagnostic module is used to determine, based on the characteristic signal intensity of urinary metabolomics markers in the sample, whether the sample is a kidney cancer sample, prostate cancer sample, bladder cancer sample, or non-urinary system tumor sample using an SVR model.

[0036] The urine metabolomics biomarkers include a first biomarker group, a second biomarker group, and a third biomarker group;

[0037] The first biomarker group includes hydroquinone, uracil, maleic acid, malic acid, hypoxanthine, pimelic acid, guanidine taurine, arginine, ascorbic acid, S-cysteine ​​succinate, cytidine, guanosine, and 8-hydroxyguanosine.

[0038] The second biomarker group includes glycolic acid, imidazoline, pyruvate, nicotinamide, 3-(imidazol-4-yl)propionic acid, 4-hydroxy-L-glutamic acid, phthalic acid, N-formyl-L-methionine, cysteylglycine, hippuric acid, paraxanthine, indoleacrylic acid, cysteine-S-sulfate, L-cystathionine, and L-3-hydroxykynurenine.

[0039] The third biomarker group includes glycolic acid, imidazoline, pyruvate, nicotinamide, 3-(imidazo-4-yl)propionic acid, 4-hydroxy-L-glutamic acid, phthalic acid, and indoleacrylic acid.

[0040] In some specific embodiments of the present invention, the diagnostic module of the above-described diagnostic system includes:

[0041] The first-layer diagnostic module is used to determine whether the sample to be tested is a urinary system tumor sample or a non-urinary system tumor sample based on the characteristic signal intensity of the first biomarker group in the sample to be tested and through the SVR model.

[0042] The second-layer diagnostic module is used to determine whether the sample to be tested is a non-bladder cancer urinary system tumor sample or a bladder cancer sample based on the characteristic signal intensity of the second biomarker group in the urinary system tumor sample through the SVR model.

[0043] The third-layer diagnostic module is used to determine whether the sample to be tested is a prostate cancer sample or a kidney cancer sample based on the characteristic signal intensity of the third biomarker group in the non-bladder cancer urinary system tumor sample and through the SVR algorithm model.

[0044] In some specific embodiments of the present invention, the kernel function of the SVR model in the first-layer diagnostic module of the above-mentioned diagnostic system is a second-order polynomial kernel, the penalty coefficient C is 0.3706, the insensitive band width is 0.0371, the kernel scale parameter is 2.906, and the predictor variables are standardized.

[0045] The kernel function of the SVR model in the second-layer diagnostic module is a second-order polynomial kernel, with a penalty coefficient C of 1.4826, an insensitive band width of 0.1483, a kernel scale parameter of 3.6030, and the predictor variables are standardized.

[0046] The kernel function of the SVR model in the third-layer diagnostic module is a second-order polynomial kernel with a penalty coefficient C of 0.7413, an insensitive band width of 0.0741, a kernel scale parameter of 3.2785, and the predictor variables are standardized.

[0047] In some specific embodiments of the present invention, the SVR model in the first-layer diagnostic module of the above-mentioned diagnostic system is trained based on clinical samples with known diagnostic results;

[0048] The training methods include:

[0049] Multiple clinical samples with known diagnostic results are taken as training samples. The predictive variables in the training samples are standardized, and the support vector regression model is trained using the training samples.

[0050] The support vector regression model uses a second-order polynomial kernel function with a penalty coefficient C of 0.3706, an insensitive band width of 0.0371, and a kernel scale parameter of 2.906.

[0051] In some specific embodiments of the present invention, the SVR model in the second-layer diagnostic module of the above-described diagnostic system is trained based on clinical samples with known diagnostic results;

[0052] The training methods include:

[0053] Multiple clinical samples with known diagnostic results are taken as training samples. The predictive variables in the training samples are standardized, and the support vector regression model is trained using the training samples.

[0054] The support vector regression model uses a second-order polynomial kernel function with a penalty coefficient C of 1.4826, an insensitive band width of 0.1483, and a kernel scale parameter of 3.6030.

[0055] In some specific embodiments of the present invention, the SVR model in the third-layer diagnostic module of the above-mentioned diagnostic system is trained based on clinical samples with known diagnostic results;

[0056] The training methods include:

[0057] Multiple clinical samples with known diagnostic results are taken as training samples. The predictive variables in the training samples are standardized, and the support vector regression model is trained using the training samples.

[0058] The support vector regression model uses a second-order polynomial kernel function with a penalty coefficient C of 0.7413, an insensitive band width of 0.0741, and a kernel scale parameter of 3.2785.

[0059] In some specific embodiments of the present invention, the characteristic signal intensity of the above application includes the normalized mass spectrometry signal intensity.

[0060] In some specific embodiments of the present invention, the diagnostic system further includes:

[0061] The data acquisition unit is used to acquire metabolomics information of the test samples via SALDI-MS.

[0062] In some specific embodiments of the present invention, the data acquisition unit of the above-mentioned diagnostic system includes a high-entropy metal oxide nanomatrix;

[0063] The high-entropy metal oxide nanomatrix includes high-entropy metal oxides with metal elements of Fe, Co, Ni, Cu, and Zn.

[0064] In some specific embodiments of the present invention, the method for preparing the high-entropy metal oxide of the above-mentioned diagnostic system includes:

[0065] Will include Fe 3+ Co 2+ Ni 2+ Cu 2+ and Zn 2+ The solution was mixed evenly with urea and polyvinylpyrrolidone, and the pH value was adjusted to 10-11 with sodium hydroxide to obtain the precursor solution.

[0066] The precursor solution was hydrothermally treated at 180°C for 12 hours, then cooled, separated, washed, dried, and annealed at 500°C for 2 hours under normal pressure by heating at 5°C / min to 500°C to obtain the high-entropy metal oxide.

[0067] The Fe 3+ Co 2+ Ni 2+ Cu 2+ and Zn 2+ The molar ratio is 1:1:1:1:1, and the concentration is 0.02 M;

[0068] The concentration of the urea is 0.1 M;

[0069] The mass fraction of the polyvinylpyrrolidone is 1%.

[0070] In some specific embodiments of the present invention, the high-entropy metal oxide nanomatrix of the above-mentioned diagnostic system further includes a matrix dispersion;

[0071] The matrix dispersion comprises water and ethanol in a volume ratio of 1:1.

[0072] The present invention also provides a method for diagnosing or identifying urinary system tumors, comprising:

[0073] S1. Determine whether the sample to be tested is a urinary system tumor sample based on the characteristic signal intensity of the first biomarker group in the sample to be tested;

[0074] S2. If the determination result of S1 is yes, then determine whether the sample to be tested is a bladder cancer sample based on the characteristic signal intensity of the second biomarker group in the sample to be tested;

[0075] S3. If the result of S2 is negative, then determine whether the sample to be tested is prostate cancer or kidney cancer based on the intensity of the characteristic signal of the third biomarker group in the sample to be tested;

[0076] The first biomarker group includes hydroquinone, uracil, maleic acid, malic acid, hypoxanthine, pimelic acid, guanidine taurine, arginine, ascorbic acid, S-cysteine ​​succinate, cytidine, guanosine, and 8-hydroxyguanosine.

[0077] in:

[0078] The determination in S1 is made by using a second-order polynomial kernel function, a penalty coefficient C of 0.3706, an insensitive band width of 0.0371, a kernel scale parameter of 2.906, and a standardized SVR model for the predicted variables.

[0079] The determination in S2 is made by using a second-order polynomial kernel function, a penalty coefficient C of 1.4826, an insensitive band width of 0.1483, a kernel scale parameter of 3.6030, and a standardized SVR model for the predictor variables.

[0080] The determination in S3 is made by using a second-order polynomial kernel function, a penalty coefficient C of 0.7413, an insensitive band width of 0.0741, a kernel scale parameter of 3.2785, and a standardized SVR model for the predictor variables.

[0081] The second biomarker group includes glycolic acid, imidazoline, pyruvate, nicotinamide, 3-(imidazol-4-yl)propionic acid, 4-hydroxy-L-glutamic acid, phthalic acid, N-formyl-L-methionine, cysteylglycine, hippuric acid, paraxanthine, indoleacrylic acid, cysteine-S-sulfate, L-cystathionine, and L-3-hydroxykynurenine.

[0082] The third biomarker group includes glycolic acid, imidazoline, pyruvate, nicotinamide, 3-(imidazo-4-yl)propionic acid, 4-hydroxy-L-glutamic acid, phthalic acid, and indoleacrylic acid.

[0083] In some specific embodiments of the present invention, the SVR model in S1 of the above-described method for diagnosing or identifying urinary system tumors is trained based on clinical samples with known diagnostic results;

[0084] The training methods include:

[0085] Multiple clinical samples with known diagnostic results are taken as training samples. The predictive variables in the training samples are standardized, and the support vector regression model is trained using the training samples.

[0086] The support vector regression model uses a second-order polynomial kernel function with a penalty coefficient C of 0.3706, an insensitive band width of 0.0371, and a kernel scale parameter of 2.906.

[0087] In some specific embodiments of the present invention, the SVR model in S2 of the above-described method for diagnosing or identifying urinary system tumors is trained based on clinical samples with known diagnostic results;

[0088] The training methods include:

[0089] Multiple clinical samples with known diagnostic results are taken as training samples. The predictive variables in the training samples are standardized, and the support vector regression model is trained using the training samples.

[0090] The support vector regression model uses a second-order polynomial kernel function with a penalty coefficient C of 1.4826, an insensitive band width of 0.1483, and a kernel scale parameter of 3.6030.

[0091] In some specific embodiments of the present invention, the SVR model in S3 of the above-described method for diagnosing or identifying urinary system tumors is trained based on clinical samples with known diagnostic results;

[0092] The training methods include:

[0093] Multiple clinical samples with known diagnostic results are taken as training samples. The predictive variables in the training samples are standardized, and the support vector regression model is trained using the training samples.

[0094] The support vector regression model uses a second-order polynomial kernel function with a penalty coefficient C of 0.7413, an insensitive band width of 0.0741, and a kernel scale parameter of 3.2785.

[0095] In some specific embodiments of the present invention, the characteristic signal intensity of the first biomarker group, the characteristic signal intensity of the second biomarker group, or the characteristic signal intensity of the third biomarker group in the above-described method for diagnosing or identifying urinary system tumors is obtained by SALDI-MS.

[0096] The SALDI-MS is based on a high-entropy metal oxide nanomatrix.

[0097] The high-entropy metal oxide nanomatrix includes high-entropy metal oxides with metal elements of Fe, Co, Ni, Cu, and Zn.

[0098] The method for preparing the high-entropy metal oxide includes:

[0099] Will include Fe 3+ Co 2+ Ni 2+ Cu 2+ and Zn 2+ The solution was mixed evenly with urea and polyvinylpyrrolidone, and the pH value was adjusted to 10-11 with sodium hydroxide to obtain the precursor solution.

[0100] The precursor solution was hydrothermally treated at 180°C for 12 hours, then cooled, separated, washed, dried, and annealed at 500°C for 2 hours under normal pressure by heating at 5°C / min to 500°C to obtain the high-entropy metal oxide.

[0101] The Fe 3+ Co 2+ Ni 2+ Cu 2+ and Zn 2+ The molar ratio is 1:1:1:1:1, and the concentration is 0.02 M;

[0102] The concentration of the urea is 0.1 M;

[0103] The mass fraction of the polyvinylpyrrolidone is 1%.

[0104] The present invention also provides a device for identifying urinary system tumors, comprising at least one processor and a memory connected to said processor, wherein:

[0105] The memory is used to store computer programs;

[0106] The processor is used to execute the computer program so that the device can implement the above-described methods for diagnosing or identifying urinary system tumors.

[0107] In some specific embodiments of the present invention, in the above-described applications, diagnostic systems, or methods for diagnosing or identifying urinary system tumors, the sample to be tested is a urine sample.

[0108] This invention constructs a urinary metabolomics platform that can rapidly detect metabolites in urine samples and utilize differentially characterized molecules to effectively identify urinary system tumors (renal cancer, prostate cancer, bladder cancer). Specifically, the beneficial effects include:

[0109] This invention significantly improves the sensitivity and reproducibility of urine detection. Traditional SALDI matrices (such as gold and silver nanoparticles) have single catalytic active sites. However, the high-entropy oxides of Fe, Co, Ni, Cu, and Zn used in this invention, due to the "cocktail effect" of multiple metal elements and uniform mixing at the atomic scale, produce abundant lattice distortion and electronic structure changes. This unique structure provides a large number of unsaturated coordination sites, resulting in stronger adsorption capacity and photothermal conversion efficiency for polar metabolites in urine (such as amino acids, nucleosides, and organic acids). During laser irradiation, energy transfer is more efficient, resulting in stronger mass spectrometry signal peaks (high sensitivity). Furthermore, due to the stable material structure, signal fluctuations between points and between samples are small (good reproducibility).

[0110] This invention enables effective identification of three types of urinary system tumors. Fe, Co, Ni, Cu, and Zn, as transition metals, act not only as energy transfer mediators but also as "reaction centers" during the SALDI process. Under laser irradiation, they catalyze specific and mild desorption / ionization behaviors of metabolites. This "soft ionization" characteristic results in highly specific fragmentation patterns of the metabolites. The microenvironmental differences among different tumors (renal cancer, prostate cancer, and bladder cancer) lead to variations in the composition of metabolites in urine. The matrix of this invention amplifies these subtle metabolite differences, generating unique "fingerprint" fragmentation peaks for each, thereby achieving high-resolution identification capabilities that are difficult to achieve with traditional methods.

[0111] This invention is non-invasive and highly time-efficient. Urine testing itself has the advantages of being non-invasive and readily available. The high sensitivity of high-entropy metal oxides allows for testing using only a very small amount of urine (microliter), eliminating the need for complex sample pretreatment (such as centrifugation, extraction, and derivatization). This invention establishes a rapid "sample drop-detection-result" testing mode, which, compared to traditional tissue biopsies or serological indicators such as PSA, greatly reduces patient discomfort and testing costs, while avoiding the risk of cross-infection. It can complete large-scale population screening and typing in a short time. Attached Figure Description

[0112] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the accompanying drawings used in the description of the embodiments or the prior art are briefly introduced below.

[0113] Figure 1 This is a flowchart illustrating the urine metabolomics detection process based on high-entropy metal oxide nanomatrix.

[0114] Figure 2 Urinary metabolic fingerprints of healthy controls, patients with kidney cancer, prostate cancer, and bladder cancer were displayed.

[0115] Figure 3A. Distribution of intra-batch and inter-batch relative standard deviations in the metabolic profiles of quality control urine samples; B. Distribution of relative standard deviations (RSD) for intra-batch detections; C. Distribution of RSD for inter-batch detections.

[0116] Figure 4 The text describes the performance evaluation of different machine learning algorithms. A represents binary classification of healthy controls and urinary system tumors; B represents binary classification of bladder cancer and non-bladder cancer urinary system tumors; C represents binary classification of prostate cancer and kidney cancer; DT represents decision tree; KNN represents nearest neighbor classification; NB represents Naive Bayes classification; LDA represents linear regression; Logi represents logistic regression; and SVR represents support vector regression.

[0117] Figure 5 Comparison of the discrimination results of high-entropy metal oxides (HEO) and single-component metal oxides (α-Fe2O3, Co3O4, NiO, CuO, ZnO) on healthy controls and urinary system tumors;

[0118] Figure 6 A shows the discrimination results of the first-level diagnosis; B shows the prediction scores of the modeling set and the validation set; C shows the confusion matrix of the modeling set; D shows the confusion matrix of the validation set.

[0119] Figure 7 A shows the discrimination results of the second-level diagnosis; B shows the predicted scores of the modeling set and the validation set; C shows the confusion matrix of the modeling set; D shows the confusion matrix of the validation set.

[0120] Figure 8 A shows the discrimination results of the third-level diagnosis; B shows the predicted scores of the modeling set and the validation set; C shows the confusion matrix of the modeling set; D shows the confusion matrix of the validation set.

[0121] Figure 9 The validation set shows the discrimination results; A~C are the ROC curves for the first-level diagnosis, the second-level diagnosis, and the third-level diagnosis, respectively; D shows the overall prediction accuracy of each group of samples in the three-step urine discrimination model.

[0122] Figure 10 The diagram shows a comparison of the discrimination results between three-class and two-class classification; A shows the prediction score of stepwise binary classification; B~C show the ROC curves of stepwise binary classification; D shows the prediction score of three-class classification; E~G show the ROC curves of three-class classification; H shows the overall prediction accuracy of each group of samples in the three-class classification model; I shows the overall prediction accuracy of each group of samples in the stepwise binary classification model.

[0123] Figure 11 The external validation results are shown; A to C are the ROC curves for the first-level diagnosis, the second-level diagnosis, and the third-level diagnosis, respectively; D shows the overall prediction accuracy of each group of samples in the three-step urine discrimination model.

[0124] In the attached diagram, HC represents healthy controls; UC represents urinary system tumors; BC represents bladder cancer; non-BC represents non-bladder cancer urinary system tumors; PC represents prostate cancer; and KC represents kidney cancer. Detailed Implementation

[0125] This invention discloses an effective method for the identification of urinary system tumors using urinary metabolomics. Those skilled in the art can refer to the content of this document and appropriately modify the process parameters to achieve the desired result. It is particularly important to note that all similar substitutions and modifications are obvious to those skilled in the art and are considered to be included in this invention. The methods and applications of this invention have been described through preferred embodiments. Those skilled in the art can clearly modify or appropriately change and combine the methods and applications described herein without departing from the content, spirit, and scope of this invention to realize and apply the technology of this invention.

[0126] It should be understood that the expression “one or more of…” individually includes each of the objects described after the expression, as well as various different combinations of two or more of the described objects, unless otherwise understood from the context and usage. The expression “and / or” combined with three or more described objects should be understood to have the same meaning, unless otherwise understood from the context.

[0127] The terms “including,” “having,” or “containing,” including the use of their grammatical synonyms, should generally be understood as open-ended and non-restrictive, for example, not excluding other unstated elements or steps, unless otherwise specifically stated or understood from the context.

[0128] It should be understood that the order of the steps or the order in which certain actions are performed is not important as long as the invention remains operational. Furthermore, two or more steps or actions can be performed simultaneously.

[0129] The use of any and all instances or exemplary language such as “e.g.” or “including” in this document is merely intended to better illustrate the invention and does not constitute a limitation on the scope of the invention. No language in this specification should be construed as indicating that any unclaimed element is essential to the practice of the invention.

[0130] Furthermore, the numerical ranges and parameters used to define the present invention are approximate values, and the relevant values ​​in the specific embodiments have been presented as precisely as possible. However, any value inevitably contains standard deviations due to individual test methods. Therefore, unless explicitly stated otherwise, it should be understood that all ranges, quantities, values, and percentages used in this disclosure are modified with the word "approximately". Here, "approximately" generally means that the actual value is within plus or minus 10%, 5%, 1%, or 0.5% of a specific value or range.

[0131] Unless otherwise specified, the raw materials, reagents, consumables and instruments involved in this invention are all commercially available products and can be purchased from the market.

[0132] This invention provides a set of biomarkers composed of urinary metabolites for the effective identification of bladder cancer, kidney cancer, and prostate cancer, enabling health management of high-risk populations.

[0133] In practice, the combination of multiple biomarkers for the diagnosis of urinary system tumors has advantages in terms of diagnostic sensitivity and specificity compared to a single biomarker.

[0134] In practice, this invention constructs a pre-processing-free metabolomics detection technology for urine samples, which can collect urine metabolic fingerprints in high throughput using a high-entropy metal oxide nanomatrix.

[0135] In its specific implementation, this invention uses accuracy, F-metric value, Kappa coefficient, and precision as indicators to evaluate model performance, selects the optimal machine learning algorithm, and constructs a urinary system tumor diagnostic model based on a combination of urine metabolic biomarkers.

[0136] The present invention will be further illustrated below with reference to the embodiments.

[0137] Example 1: Mass spectrometry detection of urinary metabolites and establishment of a diagnostic model for urinary system tumors

[0138] 1. Samples and Instruments

[0139] A total of 320 urine samples were collected from Sir Run Run Shaw Hospital, Zhejiang University (details are shown in Table 1). Among them, 80 were from healthy controls, 80 from patients with renal cell carcinoma, 80 from patients with prostate cancer, and 80 from patients with bladder cancer. All urine samples were midstream morning urine collected on an empty stomach. Patients avoided eating, drinking alcohol, and taking medications within 8 hours prior to sample collection. Age, sex ratio, smoking status, and alcohol consumption were kept consistent across all groups.

[0140] Table 1

[0141]

[0142] Urine metabolic profiles were acquired using a matrix-assisted laser desorption / ionization time-of-flight mass spectrometer (MALDI-TOF MS) from Bruker Daltonics. The analysis and processing of metabolic data were performed using software such as Clinprotools provided by Bruker Daltonics. Data normalization was performed using the affy algorithm package in the statistical analysis software R 3.5.2.

[0143] 2. Technical Approach

[0144] (1) Preparation of high-entropy metal oxide nanomatrix

[0145] Step 1: Preparation of the precursor solution;

[0146] Preparation of Fe with equimolar concentration 3+ Co 2+ Ni 2+ Cu 2+ and Zn 2+ The solution (all ion concentrations were 0.02 M) was prepared using deionized water as the solvent. Urea (0.1 M in the precursor) and polyvinylpyrrolidone (1% by mass in the precursor) were then dissolved separately in this solution, and the mixture was magnetically stirred (800 rpm) to promote homogeneous precipitation and prevent particle aggregation. The pH was then adjusted to 10–11 by adding 1 M sodium hydroxide solution dropwise.

[0147] Step two: High-temperature calcination of the material;

[0148] The mixture was transferred to a 50 mL PTFE-lined autoclave (70%–80% filling) and hydrothermally treated at 180 °C for 12 hours, then allowed to cool naturally to room temperature. The resulting precipitate was separated by vacuum filtration, repeatedly washed with deionized water, and dried overnight at 80 °C. The collected powder was finally annealed at 500 °C for 2 hours under normal pressure with a heating rate of 5 °C / min.

[0149] (2) Preparation of urine samples

[0150] Step 1: Pretreatment of urine samples;

[0151] After collection, urine samples were centrifuged (8000 rpm, 10 min, 4℃) to obtain a clear supernatant, which was then aliquoted and stored at -80℃ for later use. Before spotting, the urine samples were thawed on crushed ice at 4℃.

[0152] Step two, LDI-MS sample preparation follows a two-step "drop-drying" process;

[0153] The specific steps are as follows: 1 μL of urine was dropped onto an MTP 384 stainless steel target plate and allowed to dry. Then, 1 μL of a nanomaterial matrix solution (1 mg / mL) was deposited. The nanomaterial matrix solution consisted of a high-entropy metal oxide or a single-component metal oxide uniformly dispersed in a water / ethanol (1:1, v / v) mixture. The target plate after spotting was stored in a dry vacuum environment until mass spectrometry analysis.

[0154] 3. Bioinformatics methods

[0155] (a) Mass spectrometry data acquisition

[0156] Metabolic fingerprints of urine were acquired using an ultrafleXtreme MALDI-TOF / TOF instrument (Bruker Daltonics) equipped with a 355 nm Nd:YAG laser beam. Data acquisition was performed in reflective negative ion mode, with a molecular weight range of 20–350 Da and a relative laser pulse energy of 68% of the total energy for mass spectrometry acquisition. The lens voltage was set to 8.50 kV, and the voltages of ion source 1, ion source 2, reflector 1, and reflector 2 were set to 20.00 kV, 17.75 kV, 21.10 kV, and 10.70 kV, respectively. The ion extraction time was 120 ns, the laser parameters were set to "4_large", and the total number of mass spectrometry shots per sample was "2500 shots". Figure 1 This is a flowchart of the urine metabolomics detection process in this invention. Figure 2 This study investigated the urinary metabolic fingerprint profiles of healthy controls, patients with kidney cancer, prostate cancer, and bladder cancer.

[0157] Experimental quality control: (1) For each acquired raw spectrum, the number of peaks with S / N >= 3 was set as the standard for judging the quality of the spectrum; only spectra with more than 100 peaks were retained, and spectra with less than 100 peaks were discarded. (2) For the entire experimental operation, the relative standard deviation (relative standard deviation = standard deviation / average) of the mixed quality control urine samples (all enrolled urine samples were mixed in equal volumes) was used to ensure the consistency of the experiment. In this embodiment, the intra-batch relative standard deviation of the metabolic spectrum of the quality control urine samples was 10.17%, and the inter-batch relative standard deviation was 18.74%; which met the consistency range (<30%), indicating that the experimental consistency was good. See Figure 3 .

[0158] (II) Preprocessing of raw data

[0159] Raw metabolomic data were processed using FlexAnalysis (Bruker Daltonics), and peaks with S / N >= 3 were selected for subsequent statistical analysis. Normalization was performed using the cubic spline method from the affy package in R 3.5.2 software.

[0160] (III) Selection of characteristic metabolites of urinary system tumors

[0161] Differentially differentiated metabolites were screened using t-tests in MATLAB software, and p-values ​​were corrected using the p.adjust function in R 3.5.2. Metabolites with p < 0.05 after correction based on the Benjamini-Hochberg method were defined as potential biomarkers for urinary system tumors. One-way ANOVA was used to screen for metabolites showing significant differences among the renal cell carcinoma, prostate cancer, and bladder cancer groups.

[0162] (iv) Machine Learning Algorithms

[0163] Machine learning is a field of research and algorithms that focuses on finding patterns in data and using these patterns to make predictions. Common machine learning algorithms include linear regression, logistic regression, decision trees, support vector regression, Naive Bayes, and nearest neighbor. In this embodiment, a 10-fold cross-validation process is introduced when training the machine learning model to avoid overlearning. Accuracy, F-metric, kappa coefficient, and precision are used to evaluate the model performance of different classification methods.

[0164] (v) Identification of characteristic metabolites

[0165] Identification of Characteristic Metabolites: The structure of differentially identified metabolites was determined by matching secondary mass spectrometry data with the Human Metabolome Database (http: / / www.hmdb.ca / ) and metabolic standards. First, UPLC-MS / MS analysis provided the accurate molecular weights and secondary fragment peaks of metabolites in urine samples, which were then used to identify metabolites through database searches. The relative error of the accurate molecular weights was controlled within 30 ppm to obtain a preliminary list of identified differentially identified metabolites. Subsequently, the preliminarily identified metabolites were validated using MALDI-TOF / TOF tandem mass spectrometry, matching metabolite standards with metabolites in urine samples, including matching the accurate molecular weights and secondary mass spectrometry fragment peaks of the metabolites. The results are shown in Table 2.

[0166] Table 2: Results of Metabolite Identification

[0167]

[0168] Example 2: Blind Selection Test of a Diagnostic Model for Urinary System Tumors

[0169] Forty healthy controls, 40 patients with renal cell carcinoma, 40 patients with prostate cancer, and 40 patients with bladder cancer were randomly selected from 240 urine samples to form the training set for model building. An additional 80 samples from the remaining pool were selected as validation samples for blind testing; these included 20 samples from healthy controls, 20 from patients with renal cell carcinoma, 20 from patients with prostate cancer, and 20 from patients with bladder cancer. Furthermore, 20 additional samples from healthy controls, 20 from patients with renal cell carcinoma, 20 from patients with prostate cancer, and 20 from patients with bladder cancer were collected for external validation.

[0170] 1. Performance evaluation of different machine learning algorithms

[0171] Accuracy, F-metric, Kappa coefficient, and precision were used as performance metrics to compare the diagnostic performance of metabolomics models constructed using different machine learning algorithms for binary classification. The machine learning algorithms evaluated included Support Vector Regression (SVR), Decision Tree (DT), Naive Bayes (NB), Logistic Regression (Logi), Linear Regression (LDA), and Nearest Neighbor Classification (KNN). Binary classification included distinguishing between healthy controls (HC) and urinary tract tumors (UC), bladder cancer (BC) and non-bladder urinary tract tumors (non-BC), and prostate cancer (PC) and kidney cancer (KC). Figure 4 As can be seen, the models built based on the SVR algorithm all exhibit the best performance in binary classification.

[0172] 2. Comparison of diagnostic results between high-entropy metal oxides and single-component metal oxides

[0173] To demonstrate the advantages of high-entropy metal oxides compared to single-component metal oxides (Fe2O3, Co3O4, NiO, CuO, ZnO), the diagnostic performance of prediction models built using high-entropy or single-component metal oxide data was evaluated in a binary classification test on the validation set. Support vector regression (SVR) was used for modeling, and the diagnostic accuracy was calculated for healthy controls and urinary system tumor groups. Compared to single-component metal oxide models, the diagnostic model based on high-entropy metal oxides significantly improved the diagnostic accuracy in the validation cohort. Figure 5 ).

[0174] 3. A model for identifying urinary system tumors based on the SVR algorithm

[0175] Based on the selected SVR algorithm, a three-step urine discrimination model was established to identify urinary system tumors. In the first-level diagnosis, a biomarker group of 13 metabolites was used for urine tumor screening to distinguish between healthy controls and patients with urinary system tumors (kidney cancer, prostate cancer, and bladder cancer). The biomarker group specifically included hydroquinone, uracil, maleic acid, malic acid, hypoxanthine, pimelic acid, taurocyamine, arginine, ascorbic acid, S-cysteinosuccinic acid, cytidine, guanosine, and 8-hydroxyguanosine. An output value below 0.54 was considered a healthy control, while an output value above 0.54 was considered a urinary system tumor.

[0176] The kernel function of SVR is polynomial, and the order of the polynomial kernel is 2.

[0177] The BoxConstraint constraint (penalty coefficient C) is 0.3706;

[0178] The Epsilon insensitive bandwidth (ε) is 0.0371;

[0179] The kernel scale parameter is 2.906;

[0180] "StandardizeData = 1" (i.e., standardize the predictor variables);

[0181] The decision function is (Equation I):

[0182] ;

[0183] The kernel function is (Equation II):

[0184] ;

[0185] γ is the kernel scale parameter.

[0186] like Figure 6The figure shows the prediction scores and confusion matrix of the modeling set and validation set in the first-level diagnosis. The results for the validation samples are as follows: all 20 healthy controls were correctly identified, with a specificity of 100%; 58 out of 60 patients with urinary system tumors were correctly identified, with a sensitivity of 96.7%.

[0187] Samples identified as urinary system tumors were then input into a second-level diagnostic panel, which used a biomarker panel including 15 metabolites to effectively differentiate between bladder cancer and non-bladder cancer urinary system tumors. The biomarker group specifically includes glycolic acid, imidazolone, pyruvic acid, nicotinamide, 3-(imidazol-4-yl)propionic acid, 4-hydroxy-L-glutamic acid, phthalic acid, N-formyl-L-methionine, cysteinylglycine, hippuric acid, paraxanthine, indoleacrylic acid, cysteine-S-sulfate, L-cystathionine, and L-3-hydroxykynurenine. An output value below 0.67 indicates bladder cancer, while a value above 0.67 indicates a non-bladder cancer urinary system tumor.

[0188] The kernel function of SVR is polynomial, and the order of the polynomial kernel is 2.

[0189] The BoxConstraint constraint (penalty coefficient C) is 1.4826;

[0190] The Epsilon insensitive bandwidth (ε) is 0.1483;

[0191] The kernel scale parameter is 3.6030;

[0192] "StandardizeData = 1" (i.e., standardize the predictor variables);

[0193] The decision function is shown in Equation I above, and the kernel function is shown in Equation II above.

[0194] like Figure 7The figure shows the prediction scores and confusion matrix of the modeling set and validation set in the second-level diagnosis. For the 58 validation samples input into the second-level prediction model: 18 out of 20 bladder cancer patients were correctly identified, and 34 out of 38 non-bladder cancer urinary system tumor patients were correctly identified, with an overall accuracy of 89.7%.

[0195] Samples diagnosed as urinary system tumors were then input into the third-level diagnostic panel, where a biomarker group, including eight metabolites, was used to effectively differentiate between prostate and kidney cancer. The biomarker group specifically included glycolic acid, imidazolone, pyruvic acid, nicotinamide, 3-(imidazol-4-yl)propionic acid, 4-hydroxy-L-glutamic acid, phthalic acid, and indoleacrylic acid. Output values ​​below 1.50 indicated kidney cancer, while values ​​above 1.50 indicated prostate cancer.

[0196] The kernel function of SVR is polynomial, and the order of the polynomial kernel is 2.

[0197] The BoxConstraint constraint (penalty coefficient C) is 0.7413;

[0198] The Epsilon insensitive bandwidth (ε) is 0.0741;

[0199] The kernel scale parameter is 3.2785;

[0200] "StandardizeData = 1" (i.e., standardize the predictor variables);

[0201] The decision function is shown in Equation I above, and the kernel function is shown in Equation II above.

[0202] like Figure 8 The figure shows the prediction scores and confusion matrix of the modeling set and validation set in the third-level diagnosis. For the 34 validation samples input into the third-level prediction model, 17 cases of renal cell carcinoma and 17 cases of prostate cancer were correctly identified, with an overall accuracy of 100%.

[0203] Overall, for the 80 urine samples in the validation set, the AUC values ​​for each step of the three-step urine discrimination model were 0.997, 0.978, and 1.000, respectively. All samples in the healthy control group were correctly identified; 85.0% of the samples in the renal cell carcinoma group and the prostate cancer group were correctly identified; and 90.0% of the samples in the bladder cancer group were correctly identified. Figure 9 ).

[0204] Comparison of Three-Part and Two-Part Classification: For 40 patients with urinary system tumors (40 cases each of bladder cancer, kidney cancer, and prostate cancer), a three-part classification model and a stepwise two-part classification model were established. The stepwise two-part classification model first distinguishes between bladder cancer and non-bladder cancer urinary system tumors, and then distinguishes between prostate cancer and kidney cancer. The results showed that the stepwise two-part classification model had significantly better discriminative performance than the three-part model, successfully identifying 92.5% of kidney cancer cases, 87.5% of prostate cancer cases, and 95.0% of bladder cancer cases. Figure 10 External validation results are as follows: Figure 11 As shown.

[0205] The above description is only a preferred embodiment of the present invention. It should be noted that for those skilled in the art, several improvements and modifications can be made without departing from the principle of the present invention, and these improvements and modifications should also be considered within the scope of protection of the present invention.

Claims

1. The application of urinary metabolomics biomarkers in the preparation of products for identifying urinary system tumors, characterized in that, The urinary metabolomics markers include: 4-hydroxy-L-glutamate, 8-hydroxyguanosine, arginine, ascorbic acid, cysteine-S-sulfate, cysteylglycine, cytidine, glycolic acid, guanosine, hippuric acid, hydroquinone, hypoxanthine, 3-(imidazol-4-yl)propionic acid, imidazolidinone, indoleacrylic acid, L-3-hydroxykynurenine, L-cystathionine, maleic acid, malic acid, N-formyl-L-methionine, nicotinamide, paraxanthine, phthalic acid, pimelic acid, pyruvic acid, S-cysteylsuccinic acid, guanidinotaurine, and uracil. The urinary system tumors include bladder cancer, kidney cancer, and / or prostate cancer.

2. The application as described in claim 1, characterized in that, The product uses SALDI-MS to acquire the characteristic signal intensity of urinary metabolomics biomarkers in the test sample, and identifies urinary system tumors based on the characteristic signal intensity of urinary metabolomics biomarkers. The SALDI-MS is based on a high-entropy metal oxide nanomatrix. The high-entropy metal oxide nanomatrix includes high-entropy metal oxides with metal elements of Fe, Co, Ni, Cu, and Zn.

3. The application as described in claim 2, characterized in that, The method for preparing the high-entropy metal oxide includes: Will include Fe 3+ Co 2+ Ni 2+ Cu 2+ and Zn 2+ The solution was mixed evenly with urea and polyvinylpyrrolidone, and the pH value was adjusted to 10-11 with sodium hydroxide to obtain the precursor solution. The precursor solution was hydrothermally treated at 180°C for 12 hours, then cooled, separated, washed, dried, and annealed at 500°C for 2 hours under normal pressure by heating at 5°C / min to 500°C to obtain the high-entropy metal oxide. The Fe 3+ Co 2+ Ni 2+ Cu 2+ and Zn 2+ The molar ratio is 1:1:1:1:1, and the concentration is 0.02 M; The concentration of the urea is 0.1 M; The mass fraction of the polyvinylpyrrolidone is 1%.

4. A diagnostic system for identifying urinary system tumors, characterized in that, include: The diagnostic module is used to determine, based on the characteristic signal intensity of urinary metabolomics markers in the sample, whether the sample is a kidney cancer sample, prostate cancer sample, bladder cancer sample, or non-urinary system tumor sample using an SVR model. The urine metabolomics biomarkers include a first biomarker group, a second biomarker group, and a third biomarker group; The first biomarker group includes hydroquinone, uracil, maleic acid, malic acid, hypoxanthine, pimelic acid, guanidine taurine, arginine, ascorbic acid, S-cysteine ​​succinate, cytidine, guanosine, and 8-hydroxyguanosine. The second biomarker group includes glycolic acid, imidazoline, pyruvate, nicotinamide, 3-(imidazol-4-yl)propionic acid, 4-hydroxy-L-glutamic acid, phthalic acid, N-formyl-L-methionine, cysteylglycine, hippuric acid, paraxanthine, indoleacrylic acid, cysteine-S-sulfate, L-cystathionine, and L-3-hydroxykynurenine. The third biomarker group includes glycolic acid, imidazoline, pyruvate, nicotinamide, 3-(imidazo-4-yl)propionic acid, 4-hydroxy-L-glutamic acid, phthalic acid, and indoleacrylic acid; The SVR model was trained on clinical samples with known diagnostic results.

5. The diagnostic system as described in claim 4, characterized in that, The diagnostic module includes: The first-layer diagnostic module is used to determine whether the sample to be tested is a urinary system tumor sample or a non-urinary system tumor sample based on the characteristic signal intensity of the first biomarker group in the sample to be tested and through the SVR model. The second-layer diagnostic module is used to determine whether the sample to be tested is a non-bladder cancer urinary system tumor sample or a bladder cancer sample based on the characteristic signal intensity of the second biomarker group in the urinary system tumor sample through the SVR model. The third-layer diagnostic module is used to determine whether the sample to be tested is a prostate cancer sample or a kidney cancer sample based on the characteristic signal intensity of the third biomarker group in the non-bladder cancer urinary system tumor sample through the SVR model.

6. The diagnostic system as described in claim 5, characterized in that, The characteristic signal intensity includes the normalized mass spectrometry signal intensity; The kernel function of the SVR model in the first-layer diagnostic module is a second-order polynomial kernel, with a penalty coefficient C of 0.3706, an insensitive band width of 0.0371, a kernel scale parameter of 2.906, and the predictor variables are standardized. The kernel function of the SVR model in the second-layer diagnostic module is a second-order polynomial kernel, with a penalty coefficient C of 1.4826, an insensitive band width of 0.1483, a kernel scale parameter of 3.6030, and the predictor variables are standardized. The kernel function of the SVR model in the third-layer diagnostic module is a second-order polynomial kernel with a penalty coefficient C of 0.7413, an insensitive band width of 0.0741, a kernel scale parameter of 3.2785, and the predictor variables are standardized.

7. The diagnostic system as described in claim 4, characterized in that, Also includes: The data acquisition module is used to acquire metabolomics information of the test samples via SALDI-MS.

8. The diagnostic system as described in claim 7, characterized in that, The data acquisition module includes a high-entropy metal oxide nanomatrix; The high-entropy metal oxide nanomatrix includes high-entropy metal oxides with metal elements of Fe, Co, Ni, Cu, and Zn. The method for preparing the high-entropy metal oxide includes: Will include Fe 3+ Co 2+ Ni 2+ Cu 2+ and Zn 2+ The solution was mixed evenly with urea and polyvinylpyrrolidone, and the pH value was adjusted to 10-11 with sodium hydroxide to obtain the precursor solution. The precursor solution was hydrothermally treated at 180°C for 12 hours, then cooled, separated, washed, dried, and annealed at 500°C for 2 hours under normal pressure by heating at 5°C / min to 500°C to obtain the high-entropy metal oxide. The Fe 3+ Co 2+ Ni 2+ Cu 2+ and Zn 2+ The molar ratio is 1:1:1:1:1, and the concentration is 0.02 M; The concentration of the urea is 0.1 M; The mass fraction of the polyvinylpyrrolidone is 1%.

9. A device for identifying tumors of the urinary system, characterized in that, It includes at least one processor and a memory connected to the processor, wherein: The memory is used to store computer programs; The processor is used to execute the computer program to enable the device to perform the following steps: S1. Determine whether the sample to be tested is a urinary system tumor sample based on the characteristic signal intensity of the first biomarker group in the sample. If it is determined to be yes, proceed to the next step; if it is determined to be no, output a negative result for urinary system tumor. S2. Determine whether the sample is a bladder cancer sample based on the intensity of the characteristic signal of the second biomarker group in the sample. If it is determined to be bladder cancer, output a positive result for bladder cancer; if it is determined to be bladder cancer, proceed to the next step. S3. Determine whether the sample is prostate cancer or kidney cancer based on the intensity of the characteristic signal of the third biomarker group in the sample. If it is determined to be prostate cancer, output a positive result for prostate cancer; if it is determined to be kidney cancer, output a positive result for kidney cancer. The first biomarker group includes hydroquinone, uracil, maleic acid, malic acid, hypoxanthine, pimelic acid, guanidine taurine, arginine, ascorbic acid, S-cysteine ​​succinate, cytidine, guanosine, and 8-hydroxyguanosine. The second biomarker group includes glycolic acid, imidazoline, pyruvate, nicotinamide, 3-(imidazol-4-yl)propionic acid, 4-hydroxy-L-glutamic acid, phthalic acid, N-formyl-L-methionine, cysteylglycine, hippuric acid, paraxanthine, indoleacrylic acid, cysteine-S-sulfate, L-cystathionine, and L-3-hydroxykynurenine. The third biomarker group includes glycolic acid, imidazoline, pyruvate, nicotinamide, 3-(imidazo-4-yl)propionic acid, 4-hydroxy-L-glutamic acid, phthalic acid, and indoleacrylic acid.

10. The device as claimed in claim 9, characterized in that, The determination in S1 is made by using a second-order polynomial kernel function, a penalty coefficient C of 0.3706, an insensitive band width of 0.0371, a kernel scale parameter of 2.906, and a standardized SVR model for the predicted variables. The determination in S2 is made by using a second-order polynomial kernel function, a penalty coefficient C of 1.4826, an insensitive band width of 0.1483, a kernel scale parameter of 3.6030, and a standardized SVR model for the predictor variables. The determination in S3 is based on a second-order polynomial kernel with a penalty coefficient C of 0.7413, an insensitive band width of 0.0741, and a kernel scale parameter of 3.2785, and is performed using a standardized SVR model to determine the predictor variables.

Images