A vhl-related renal cell carcinoma treatment effect prediction model, system and device

Abstract

The application discloses a VHL-related renal cell carcinoma treatment effect prediction model, system and device. In the present study, liquid biopsy technology is used to integrate genomic features derived from blood shallow whole genome sequencing (sWGS) and metabolomics features. Through multi-omics joint analysis, the problem of insufficient information capture by a single biomarker is overcome, and a precise prediction model for HIF-2alpha inhibitor efficacy is constructed.

Description

Technical Field

[0001] This invention belongs to the field of intelligent medical care, specifically relating to a VHL-related renal cell carcinoma efficacy prediction model, system, and device. Background Technology

[0002] Von Hippel-Lindau syndrome (VHL syndrome) is a rare autosomal dominant inherited neoplastic syndrome in which patients are highly susceptible to benign and malignant tumors in multiple organs throughout the body, such as the kidneys, brain, and pancreas. VHL-associated renal cell carcinoma (VHL-RCC) is the leading cause of death among these patients. Unlike sporadic renal cell carcinoma, VHL-RCC is characterized by multifocality, bilaterality, and recurrent episodes.

[0003] For a long time, the clinical management of VHL-RCC has relied on repeated surgical procedures. This high frequency of invasive procedures not only causes immense physical and psychological suffering for patients but also significantly increases the risk of end-stage renal disease, severely impacting their quality of life. Although the advent of novel targeted drugs such as HIF-2α inhibitors (e.g., bezutefal) has provided patients with non-surgical treatment options, clinical data show that their objective response rate is only around 49%, and the onset of action is generally long.

[0004] Currently, there are no effective molecular markers in clinical practice to predict which patients will benefit from HIF-2α inhibitor treatment. This results in approximately half of the patients potentially not benefiting after long-term treatment, wasting valuable treatment windows and creating a significant economic burden on healthcare. Given the high tumor heterogeneity of VHL-RCC, there is an urgent clinical need for a non-invasive and accurate predictive method to identify the predisposing population that will benefit from HIF-2α inhibitor treatment before treatment begins. Summary of the Invention

[0005] To overcome the shortcomings of existing technologies, this invention provides a VHL-related renal cell carcinoma treatment efficacy prediction model, system, and device.

[0006] To achieve the above objectives, the present invention adopts the following technical solution: A first aspect of the present invention provides a method for predicting the efficacy of HIF-2α inhibitors in VHL-related renal cell carcinoma, the method being performed by a computer and comprising the following steps: Data Acquisition: Acquire any one or more of the following characteristics from the patient to be tested: 3β-hydroxy-5-cholenic acid, 10-(2,3-dihydroxypropoxy)-10-oxodecanoic acid, yeast amino acid, pyridoxamine phosphate, 12(S)-hydroxyheptadecanoic acid, cfDNA, chromosomal instability trait CX10, phosphatidic acid (15:0 / 20:2 (11Z,14Z)), 5(S)-hydroperoxyeicabalaenoic acid, and lysophosphatidylcholine 15:1 level; Data processing: Input the above features into the constructed VHL-related renal cell carcinoma efficacy prediction model; Output: Based on the prediction model, predict whether the VHL-related renal cell carcinoma patients under test will respond to HIF-2α inhibitor treatment.

[0007] Furthermore, the construction steps of the VHL-related renal cell carcinoma efficacy prediction model are as follows: One or more of the following are used as feature variables and input into a machine learning algorithm to construct a predictive model: 3β-hydroxy-5-cholenic acid, 10-(2,3-dihydroxypropoxy)-10-oxodecanoic acid, yeast amino acid, pyridoxamine phosphate, 12(S)-hydroxyheptadecanoic acid, cfDNA, chromosome instability feature CX10, phosphatidic acid (15:0 / 20:2(11Z,14Z)), 5(S)-hydroperoxyeicosapentaenoic acid, and lysophosphatidylcholine 15:1.

[0008] A second aspect of the present invention provides a predictive model for the efficacy of HIF-2α inhibitors in VHL-related renal cell carcinoma, wherein the indicators of the model include any one or more of the following: 3β-hydroxy-5-cholenic acid, 10-(2,3-dihydroxypropoxy)-10-oxodecanoic acid, yeast amino acid, pyridoxamine phosphate, 12(S)-hydroxyheptadecanoic acid, cfDNA, chromosomal instability trait CX10, phosphatidic acid (15:0 / 20:2 (11Z,14Z)), 5(S)-hydroperoxyeicadenosine monophosphate, and lysophosphatidylcholine 15:1.

[0009] A third aspect of the present invention provides a predictive system for the efficacy of HIF-2α inhibitors in VHL-related renal cell carcinoma, the system comprising: Data acquisition unit: Acquire any one or more of the following characteristics of the patient to be tested: 3β-hydroxy-5-cholenic acid, 10-(2,3-dihydroxypropoxy)-10-oxodecanoic acid, yeast amino acid, pyridoxamine phosphate, 12(S)-hydroxyheptadecanoic acid, cfDNA, chromosomal instability trait CX10, phosphatidic acid (15:0 / 20:2 (11Z,14Z)), 5(S)-hydroperoxyeicosapentaenoic acid, and lysophosphatidylcholine 15:1 level; Data processing unit: Inputs the above features into the constructed VHL-related renal cell carcinoma efficacy prediction model; Results output unit: Based on the prediction model, predict whether the VHL-related renal cell carcinoma patient to be tested is effective for HIF-2α inhibitor treatment.

[0010] Furthermore, the construction steps of the VHL-related renal cell carcinoma efficacy prediction model are as follows: One or more of the following are used as feature variables and input into a machine learning algorithm to construct a predictive model: 3β-hydroxy-5-cholenic acid, 10-(2,3-dihydroxypropoxy)-10-oxodecanoic acid, yeast amino acid, pyridoxamine phosphate, 12(S)-hydroxyheptadecanoic acid, cfDNA, chromosome instability feature CX10, phosphatidic acid (15:0 / 20:2(11Z,14Z)), 5(S)-hydroperoxyeicosapentaenoic acid, and lysophosphatidylcholine 15:1.

[0011] A fourth aspect of the present invention provides a device for predicting the efficacy of HIF-2α inhibitors in VHL-related renal cell carcinoma, the device comprising a memory and a processor, the memory being used to store program instructions; the processor being used to invoke the program instructions, which, when executed, implement the method described in the first aspect of the present invention.

[0012] A fifth aspect of the present invention provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the method described in the first aspect of the present invention.

[0013] A sixth aspect of the present invention provides a computer program product comprising a computer program that, when executed by a processor, implements the method described in the first aspect of the present invention.

[0014] A seventh aspect of the invention provides the use of reagents for detecting biomarker levels in the preparation of products for predicting the efficacy of HIF-2α inhibitors in VHL-related renal cell carcinoma, said biomarkers including one or more of the following: 3β-hydroxy-5-cholenic acid, 10-(2,3-dihydroxypropoxy)-10-oxodecanoic acid, yeast amino acid, pyridoxamine phosphate, 12(S)-hydroxyheptadecanoic acid, cfDNA, chromosomal instability trait CX10, phosphatidic acid (15:0 / 20:2 (11Z,14Z)), 5(S)-hydroperoxyeicadenosine, and lysophosphatidylcholine 15:1.

[0015] The eighth aspect of the present invention provides a product for predicting the efficacy of HIF-2α inhibitors in VHL-related renal cell carcinoma, the product comprising a reagent for detecting biomarker levels, the biomarkers including one or more of the following: 3β-hydroxy-5-cholenic acid, 10-(2,3-dihydroxypropoxy)-10-oxodecanoic acid, yeast amino acid, pyridoxamine phosphate, 12(S)-hydroxyheptadecanoic acid, cfDNA, chromosomal instability trait CX10, phosphatidic acid (15:0 / 20:2 (11Z,14Z)), 5(S)-hydroperoxyeicadenosine monophosphate, and lysophosphatidylcholine 15:1.

[0016] Advantages and beneficial effects of the present invention: This study utilized liquid biopsy technology to integrate genomic features derived from superficial whole-genome sequencing (sWGS) of blood (such as chromosomal copy number variations and fragmentation characteristics) with metabolomics features. Through multi-omics joint analysis, the study overcame the problem of insufficient information capture by single biomarkers and constructed a precise predictive model for the efficacy of HIF-2α inhibitors. Attached Figure Description

[0017] Figure 1 This is a schematic diagram of the efficacy prediction method for HIF-2α inhibitors in VHL-related renal cell carcinoma provided in this application; Figure 2 This is a schematic diagram of the efficacy prediction system for HIF-2α inhibitors in VHL-related renal cell carcinoma provided in this application; Figure 3 This is a schematic diagram of the efficacy prediction device for HIF-2α inhibitors in VHL-related renal cell carcinoma provided in this application; Figure 4 These are single-omics feature screening diagrams for the response group and the non-responder group. Among them, 4A and 4B are metabolite difference analysis diagrams, 4C and 4D are metabolite analysis diagrams that distinguish the response group and the non-responder group by the OPLS-DA model, and 4E and 4F are diagrams of 23 metabolites screened using LOOCV. Figure 5 It is a horizontal graph verifying the central markers; Figure 6 These are the performance validation plots of the model. 6A is the MDA ranking plot of the top 10 variables, 6B is the ROC curve of the prediction model on the training set, and 6C is the ROC curve of the prediction model on the validation set. Detailed Implementation

[0018] To enable those skilled in the art to better understand the present invention, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings.

[0019] In some of the processes described in the specification, claims, and accompanying drawings of this invention, multiple operations appearing in a specific order are included. However, it should be clearly understood that these operations may not be executed in the order they appear herein, or may be executed in parallel. The operation numbers, such as 101, 102, etc., are merely used to distinguish different operations and do not represent any execution order. Furthermore, these processes may include more or fewer operations, and these operations may be executed sequentially or in parallel. It should be noted that the descriptions such as "first," "second," etc., in this document are used to distinguish different messages, devices, modules, etc., and do not represent a sequential order, nor do they limit "first" and "second" to different types.

[0020] Figure 1This is a schematic diagram of a method for predicting the efficacy of HIF-2α inhibitors in VHL-related renal cell carcinoma, as provided in this application, specifically including: 101 Data Acquisition: Acquire any one or more of the following characteristics from the patient to be tested: 3β-hydroxy-5-cholenic acid, 10-(2,3-dihydroxypropoxy)-10-oxodecanoic acid, yeast amino acid, pyridoxamine phosphate, 12(S)-hydroxyheptadecanoic acid, cfDNA, chromosomal instability trait CX10, phosphatidic acid (15:0 / 20:2 (11Z,14Z)), 5(S)-hydroperoxyeicadenosine monophosphate, and lysophosphatidylcholine 15:1 level.

[0021] 102 Data Processing: Input the above features into the constructed VHL-related renal cell carcinoma efficacy prediction model.

[0022] 103 Output Results: Based on the prediction model, predict whether the VHL-related renal cell carcinoma patients under test are effective for HIF-2α inhibitor treatment.

[0023] In some implementations, a VHL-related renal cell carcinoma patient is considered to have responded to HIF-2α inhibitor treatment (response) if any one or more of the following criteria are met, relative to the non-response group: 3β-hydroxy-5-cholenoic acid, 10-(2,3-dihydroxypropoxy)-10-oxodecanoic acid, Saccharopine, or Pyridoxamine. The levels of phosphate, 12(S)-hydroxyheptadecanoic acid (12S-HHT), phosphatidic acid (15:0 / 20:2(11Z,14Z)) (PA(15:0 / 20:2(11Z,14Z))), and 5(S)-hydroperoxyeicosapride (5(S)-HpEPE) were elevated, while the levels of lysophosphatidylcholine 15:1 (LySoPC 15-1) and cfDNA were decreased, and the proportion of the chromosomal instability trait CX10 (CX10) was high.

[0024] Figure 2 This is a schematic diagram of a HIF-2α inhibitor efficacy prediction system for VHL-related renal cell carcinoma provided in this application, specifically including: 201 Data Acquisition Unit: Acquires the levels of 3β-hydroxy-5-cholenic acid, 10-(2,3-dihydroxypropoxy)-10-oxodecanoic acid, yeast amino acid, pyridoxamine phosphate, 12(S)-hydroxyheptadecanoic acid, cfDNA, chromosomal instability trait CX10, phosphatidic acid (15:0 / 20:2 (11Z,14Z)), 5(S)-hydroperoxyeicosapentaenoic acid, and lysophosphatidylcholine 15:1 in the patients to be tested.

[0025] 202 Data Processing Unit: Input the above features into the constructed VHL-related renal cell carcinoma efficacy prediction model.

[0026] 203 Result Output Unit: Based on the prediction model, predict whether the VHL-related renal cell carcinoma patient to be tested is effective for HIF-2α inhibitor treatment.

[0027] In some implementations, the system includes a processor, which may be a single-core or multi-core processor or multiple processors for parallel processing. The system also includes memory (e.g., random access memory, read-only memory, flash memory), electronic storage units (e.g., hard disks), communication interfaces (e.g., network adapters) for communicating with one or more other systems, and peripheral devices such as cache memory, other memory, data storage, and / or electronic display adapters. The memory, electronic storage units, communication interfaces, and peripheral devices communicate with the processor via a communication bus (solid line), such as a motherboard. The storage units may be data storage units (or databases) for storing data. The system may be operatively coupled to a computer network via the communication interface. The network may be the Internet, an intranet and / or an extranet, or an intranet and / or extranet communicating with the Internet. In some cases, the network is a communication and / or data network. The network may include one or more computer servers, which may support distributed computing, such as cloud computing. In some cases, the network may enable a peer-to-peer network, allowing devices coupled to the system to operate as clients or servers.

[0028] In some embodiments, the processor can execute a series of machine-readable instructions, which can be embodied in a program or software. The instructions can be stored in a memory location, such as memory. The instructions can be directed to the processor, which can then be programmed or otherwise configured to implement the methods of the present invention. Examples of operations performed by the processor can include reading, decoding, executing, and writing back.

[0029] In some implementations, the processor may be part of a circuit such as an integrated circuit, and one or more other components of the system may be included in the circuit, which in some cases is an application-specific integrated circuit (ASIC).

[0030] In some implementations, the electronic storage unit can store files such as drivers, libraries, and saved programs. The electronic storage unit can also store user data, such as user preferences and user programs. In some cases, the system may include one or more additional data storage units located outside the computer system, such as on a remote server that communicates with the system via an intranet or the Internet.

[0031] In some implementations, the system can communicate with one or more remote computer systems via a network. For example, the system can communicate with a user's (e.g., a physician's) remote computer system. Examples of remote computer systems include personal computers, tablet PCs, telephones, smartphones, or personal digital assistants. Users can access the system via the network.

[0032] Figure 3 This is a schematic diagram of a device for predicting the efficacy of HIF-2α inhibitors in VHL-related renal cell carcinoma, as provided in this application. The device includes a memory and a processor. The memory is used to store program instructions, and the processor is used to call the program instructions. When the program instructions are executed, the above method is implemented.

[0033] The present invention also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the above-described method.

[0034] In some implementations, the computer-readable storage medium may be FRAM, ROM, PROM, EPROM, EEPROM, flash memory, magnetic surface memory, optical disc, or CD. ROM and other memory; or it can be any device that includes one or any combination of the above-mentioned memory.

[0035] In some implementations, executable instructions may take the form of programs, software, software modules, scripts, or code, written in any form of programming language (including compiled or interpreted languages, or declarative or procedural languages), and may be deployed in any form, including as stand-alone programs or as modules, components, subroutines, or other units suitable for use in a computing environment.

[0036] As an example, executable instructions may, but do not necessarily, correspond to files in a file system. They may be stored as part of a file that holds other programs or data, for example, in one or more scripts in a Hyper Text Markup Language (HTML) document, in a single file dedicated to the program in question, or in multiple collaborating files (e.g., a file that stores one or more modules, subroutines, or code sections).

[0037] As an example, executable instructions can be deployed to execute on a single computing device, or on multiple computing devices located in one location, or on multiple computing devices distributed across multiple locations and interconnected via a communication network.

[0038] In some embodiments, executable instructions can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium. When executed, the computer program can include the processes described in the embodiments of the methods above. Any references to memory, storage, databases, or other media used in the embodiments provided in this application can include non-volatile and / or volatile memory. Non-volatile memory can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory. Volatile memory can include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in various forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), dual data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link DRAM (SLDRAM), Rambus direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM), etc.

[0039] This invention provides a product for predicting the efficacy of HIF-2α inhibitors in VHL-related renal cell carcinoma. The product includes reagents for detecting biomarker levels, wherein the biomarkers include one or more of the following: 3β-hydroxy-5-cholenic acid, 10-(2,3-dihydroxypropoxy)-10-oxodecanoic acid, yeast amino acid, pyridoxamine phosphate, 12(S)-hydroxyheptadecanoic acid, cfDNA, chromosomal instability trait CX10, phosphatidic acid (15:0 / 20:2 (11Z,14Z)), 5(S)-hydroperoxyeicadenosine monophosphate, and lysophosphatidylcholine 15:1.

[0040] In some embodiments, the level or expression level refers to the absolute or relative amount of the marker in this application. The expression level of the marker in this application can be determined by various techniques; in particular, the absolute or relative amount of the marker in this application can be detected using methods well known to those skilled in the art.

[0041] In some implementations, methods for detecting metabolic biomarkers include, but are not limited to, chromatography, mass spectrometry, spectroscopy, chromatography-mass spectrometry, enzyme-linked immunosorbent assay (ELISA), radiochemical analysis, nuclear magnetic resonance spectroscopy, light scattering analysis, and turbidimetry.

[0042] The reagent also includes a detectable marker.

[0043] In some embodiments, a detectable label refers to any composition that can be detected by fluorescent, spectroscopic, photochemical, biochemical, immunological, electrical, optical, or chemical means. This includes, but is not limited to, UV-Vis labels, near-infrared labels, fluorescent groups, phosphorescent groups, magnetic spin resonance labels, photosensitizers, photolytically cleavable moieties, chelating centers, heavy atoms, radioactive isotopes, isotope-detectable spin resonance labels, paramagnetic moieties, chromophores, and luminescent organisms.

[0044] In some embodiments, the fluorescent groups include, but are not limited to, FAM (Carboxyfluorescein, green fluorescent), FITC (Fluorescein isothiocyanate), TET (Tetrachlorofluorescein), HEX (Hexachlorofluorescein), JOE (2,7-dimethyl-4,5-dichloro-6-carboxyfluorescein), rhodamine dyes (such as R110, TAMRA, Texas Red, etc.), ROX, and AlexaFluor dyes (such as Alexa 350, Alexa 405, Alexa 430, Alexa 488, Alexa 500, Alexa 514, Alexa 532, Alexa 546, Alexa 555, Alexa 568, Alexa 594, Alexa 610, Alexa 633, Alexa 635, Alexa 647, Alexa 633, Alexa 635, Alexa 647, Alexa 633, Alexa 634, Alexa 645, Alexa 635, Alexa 636, Alexa 637, Alexa 638, Alexa 639 ... 660, Alexa 680, Alexa 700, Alexa 750, Alexa 790), ATTO dyes (such as ATTO 390, ATTO 425, ATTO 465, ATTO 488, ATTO 495, ATTO 514, ATTO 520, ATTO 532, ATTO Rho6G, ATTO 542, ATTO 550, ATTO 565, ATTO Rho3B, ATTO Rho11, ATTO Rho12, ATTO Thio12, ATTO Rho101, ATTO 590, ATTO 594, ATTO Rho13, ATTO 610, ATTO 620, ATTO Rho14, ATTO 633, ATTO 647, ATTO 647N, ATTO 655, ATTO Oxa12, ATTO 665, ATTO 680, ATTO 700, ATTO 725, ATTO 740), DyLight dyes, cyanine dyes (such as Cy2, Cy3, Cy3.5, Cy3b, Cy5, Cy5.5, Cy7, Cy7.5), FluoProbes dyes, SulfoCy dyes, Seta dyes, IRIS dyes, SeTau dyes, SRfluor dyes, Square dyes.

[0045] The products include reagent kits, chips, and test strips.

[0046] In some embodiments, the kit may also include a fluorescent dye, and a variety of known fluorescent dyes may be used. Examples include methods using an intercalator with a marking function, and methods using probes that bind fluorescent substances to nucleotides that specifically hybridize to the relatively amplified DNA sequence. Examples of intercalators include ethidium bromide and SYBR Green I as unsaturated fluorescent dyes, and Resolight and EvaGreen as saturated fluorescent dyes. The dosage should be as recommended by the manufacturer or distributor of the fluorescent dye used.

[0047] The kit also includes an instruction manual, which may include guidance on obtaining and processing samples.

[0048] In some embodiments, a kit refers to any delivery system used to deliver materials. The components of the kit may be packaged in an aqueous medium or in a lyophilized form. Suitable containers in the kit typically include at least one vial, test tube, long-necked flask, PET bottle, syringe, or other container in which one component can be placed, and preferably, appropriately aliquoted. When more than one component is present in the kit, the kit will also typically include a second, third, or other additional container in which the additional components are placed separately. However, different combinations of components may be contained in a single vial. The kit of this application will also typically include a container for containing the reactants, sealed for commercial sale. Such a container may include injection-molded or blow-molded plastic containers in which the desired vials can be held.

[0049] The process of building and validating the prediction model is as follows: 1. Experimental Methods 1) Clinical cohort and sample collection and cfDNA extraction A total of 36 patients with VHL-RCC who received HIF-2α inhibitor therapy were included, collected at Peking University First Hospital from April 2023 to October 2025. The inclusion criteria were as follows: (1) confirmed germline VHL mutation by genetic testing; (2) clinically diagnosed with VHL syndrome, and at least one family member had undergone genetic testing; (3) imaging evidence showing a renal mass suspected of being renal cell carcinoma. The exclusion criteria included: (1) other active malignancies in addition to VHL-related tumors; (2) recent surgery, radiotherapy, ablation or other systemic antitumor therapy before the start of treatment; (3) severe active infection or inflammatory disease; (4) peripheral blood samples that did not meet quality control standards; (5) insufficient imaging follow-up data for efficacy assessment. Ten samples were randomly selected as the training set, and the remaining samples were used as the independent validation set. The training set included 5 responding patients and 5 non-responding patients, and the validation set included 12 responding patients and 14 non-responding patients. Fasting venous blood was collected from patients prior to the initiation of the first dose of drug treatment. Cell-free plasma DNA (cfDNA) was extracted using the standardized CONCERT cell-free DNA extraction kit (RC1006). The concentration of the extracted cfDNA was precisely quantified using a Qubit 4.0 fluorometer, and the fragment size distribution was assessed using the Qsep400 biological fragment analysis system to verify nucleic acid yield and exclude large genomic DNA contamination.

[0050] 2) Shallow whole-genome sequencing (sWGS) and genome feature extraction Sequencing libraries were constructed following standard next-generation sequencing (NGS) procedures, including end repair, adapter ligation, and PCR amplification. After passing Qsep400 quality control, sWGS whole-genome shallow sequencing was performed on the Illumina platform. The following genomic dimensions were then used for deep feature extraction from the sequenced data: Tumor fraction (TF) estimation: After quality control and trimming, the raw sequencing sequences were aligned to the human reference genome (hg19) using the Burrows-WheelerAligner (BWA) algorithm. The generated BAM files were then sorted, indexed, and deduplicated by PCR. The absolute copy number was estimated using the R package ichorCNA to calculate the tumor fraction (TF) in peripheral blood.

[0051] cfDNA fragmentation feature analysis: High-quality paired-end alignment sequences (MAPQ≥30) were extracted from the sequencing data. Sequencing fragments were precisely divided into short fragments (100-150 bp) and long fragments (151-220 bp). The reference genome was divided into continuous, non-overlapping 100-kb bins, and "blacklist regions" and bins with an effective base ratio <50% were filtered out to ensure data robustness. A locally weighted scatter plot smoothing (LOESS) regression equation (span = 0.75) was introduced to correct the long / short fragment read counts in different bins based on GC content. Finally, the GC-corrected data were aggregated into a 5-Mb window (each window must contain at least 25 effective 100-kb sub-bins), and the fragmentation ratio (the ratio of GC-corrected short fragments to long fragments) within this 5-Mb window was calculated to characterize the whole-genome fragmentation pattern.

[0052] Quantitative analysis of chromosomal instability variants (CX1-17): Based on absolute copy number profiles generated by sWGS, five core genomic structural variation indicators were assessed at the whole-genome level (including copy number fragment size, absolute deviation of copy state between adjacent fragments, number of breakpoints per 10 Mb and per chromosome arm, and oscillating copy number chain length). The observed vector points were mapped to 43 established mixture model elements using a probabilistic model, deriving the posterior probability and vector for each sample. Subsequently, using the linear combination decomposition algorithm within the R language's YAPSA package and the standardized definition matrix, the dimensionality reduction was used to derive the exact copy number feature activity levels (i.e., chromosomal instability variants CX1-17).

[0053] 3) Non-targeted metabolomics analysis and metabolic feature mining Metabolomics profiling of serum samples was performed using a Waters Acquity I-Class PLUS ultra-high performance liquid chromatography (UPLC) system coupled with a Xevo G2-XS QTOF high-resolution mass spectrometer. Chromatographic separation was performed using a Waters Acquity UPLC HSS T3 column (1.8 μm, 2.1 × 100 mm), with the injection volume precisely controlled at 2 μL. The mobile phase consisted of solution A (an aqueous solution containing 0.1% formic acid) and solution B (an acetonitrile solution containing 0.1% formic acid). Mass spectrometry was performed in MSe full-scan acquisition mode (MassLynx V4.2 software system). The ESI source was configured as follows: capillary voltage 2500V (+) / -2000V (-); desolventizing temperature 500℃; and desolventizing gas flow rate 800 L / h.

[0054] The raw mass spectrometry data were preprocessed using Progenesis QI software for peak extraction and alignment, and then accurately identified by comparison with the built-in METLIN and proprietary metabolic databases. Subsequently, peak area data were normalized to the total peak area and then statistically screened. In data modeling, the Mann-Whitney U test (with p < 0.05) was used, and orthogonal partial least squares discriminant analysis (OPLS-DA) was performed using the R package ropls (verified to be overfit after 200 permutation tests). Based on multiple cross-validation, variable projection importance was calculated, and candidate metabolites with a VIP value > 1 and significant differences between groups were rigorously selected.

[0055] 4) Predictive model construction algorithm Rigorous initial screening of metabolites: First, leave-one-out cross-validation (LOOCV) was introduced to evaluate the candidate differential metabolites extracted in the early stage by univariate prediction using receiver operating characteristic (ROC). The feature group with area under the curve (AUC) > 0.7 and strong anti-interference ability was rigorously screened, and 23 metabolites were successfully extracted.

[0056] Composite Feature Dimensionality Reduction and Optimization (MDA Algorithm Scoring): The 23 metabolites obtained from the initial screening are combined with the extracted genomic variables (TF, cfDNA, fragmentation, and the complete set of CX1-17 chromosome instability features) in a matrix to construct a comprehensive input variable set. Then, the Random Forest (RF) machine learning algorithm framework is introduced for classification modeling. Within a randomly selected training set of 10 examples, the Mean Decrease Accuracy (MDA) of all the above metabolic and gene-level variables is calculated, and importance gradient ranking is performed.

[0057] Optimal rule selection and independent set evaluation: Based on the MDA decay importance descending score pool, the top 10 core multi-omics variables were selected to construct the final version of the random forest prediction model. The top 10 variables were input into a machine classifier to establish classification prediction rules on the training set. Finally, the model was directly applied to an independent validation set, and the generalization prediction performance and robustness against overfitting were independently evaluated using the area under the receiver operating characteristic (ROC) curve (AUC).

[0058] 2. Experimental Results 1) A panoramic view of single-omics feature screening and differential expression 264 differentially expressed metabolites were identified between the response and non-responder groups through differential analysis (p<0.05), followed by further screening. Figure 4 A and B). The overall metabolic profile shows a clear classification trend in the OPLS-DA model ( Figure 4 (C and D), 130 metabolites were obtained through VIP>1 screening. Leave-one-out cross-validation (LOOCV) was used to evaluate all metabolites individually, retaining metabolites with AUC>0.7 as candidate variables while removing interfering metabolites, ultimately yielding 23 core metabolites, including 5,9,11-trihydroxyprostaglandane-6E,14Z-diene-1-ester, D-urobilin, 3β-hydroxy-5-cholenic acid, glycocholic acid, pyridoxamine phosphate, 10-(2,3-dihydroxypropoxy)-10-oxodecanoic acid, prolyl-aspartic-arginine, hydrogenated cinnamic acid, 5(S)-Hydroperoxyeicosapride, galactosylceramide (d18:1 / 18:0), diglyceride (18:0 / LTE4 / 0:0), phosphatidic acid (15:0 / 20:2 (11Z,14Z)), yeast amino acid, 12(S)-hydroxyheptadecanoic acid, octadecanoic acid, octadecanoic acid, lysophosphatidylcholine 15:1, lysophosphatidylethanolamine (0:0 / 20:1), lysophosphatidylcholine (20:4 (5Z,8Z,11Z,14Z) / 0:0), phosphatidylcholine (17:0 / 20:4 (6Z,8E,10E,14Z)-2OH (5S,12R)), cysteyllysine, chenodeoxycholic acid sulfate, lysophosphatidylcholine (18:0 / 0:0) Figure 4 E and F).

[0059] 2) Core Feature Importance Ranking (MDA) and Determination of the Best Subset The metabolites from the initial screening were summarized along with genomic tumor scores (TF), chromosomal instability features, cfDNA, and fragmentation rates. A Random Forest (RF) machine learning algorithm was then introduced. The importance of all variables in the training set was scored based on Mean Decrease Accuracy (MDA).

[0060] The top 10 biomarker combinations include: 3β-hydroxy-5-cholenoic acid, 10-(2,3-dihydroxypropoxy)-10-oxodecanoic acid, Saccharopine, Pyridoxamine phosphate, 12S-HHT, cfDNA, CX10 (a chromosomal instability marker), phosphatidic acid (15:0 / 20:2 (11Z,14Z)) (PA (15:0 / 20:2 (11Z,14Z))), 5S-HpEPE (5(S)-HpEPE), and Lysophosphatidylcholine 15:1 (LySoPC 15-1). Figure 5 , Figure 6 A).

[0061] 3) Performance verification of multi-omics joint models The final classification model was built based on the found optimal Top 10 feature subset and data partitioning scheme. Results show that the prediction model achieves an area under the receiver operating characteristic (AUC) of 1.000 on the training set and an AUC of 0.982 on the independent validation set. Figure 6 (B and C).

[0062] The above description of the embodiments is only for understanding the method and core ideas of the present invention. It should be noted that those skilled in the art can make various improvements and modifications to the present invention without departing from the principles of the invention, and these improvements and modifications will also fall within the protection scope of the claims of the present invention.

Claims

1. A method for predicting the efficacy of HIF-2α inhibitors in VHL-related renal cell carcinoma, characterized in that, The method is performed by a computer and includes the following steps: Data Acquisition: Acquire any one or more of the following characteristics from the patient to be tested: 3β-hydroxy-5-cholenic acid, 10-(2,3-dihydroxypropoxy)-10-oxodecanoic acid, yeast amino acid, pyridoxamine phosphate, 12(S)-hydroxyheptadecanoic acid, cfDNA, chromosomal instability trait CX10, phosphatidic acid (15:0 / 20:2 (11Z,14Z)), 5(S)-hydroperoxyeicabalaenoic acid, and lysophosphatidylcholine 15:1 level; Data processing: Input the above features into the constructed VHL-related renal cell carcinoma efficacy prediction model; Output: Based on the prediction model, predict whether the VHL-related renal cell carcinoma patients under test will respond to HIF-2α inhibitor treatment.

2. The method according to claim 1, characterized in that, The steps for constructing the VHL-related renal cell carcinoma treatment efficacy prediction model are as follows: One or more of the following are used as feature variables and input into a machine learning algorithm to construct a predictive model: 3β-hydroxy-5-cholenic acid, 10-(2,3-dihydroxypropoxy)-10-oxodecanoic acid, yeast amino acid, pyridoxamine phosphate, 12(S)-hydroxyheptadecanoic acid, cfDNA, chromosome instability feature CX10, phosphatidic acid (15:0 / 20:2(11Z,14Z)), 5(S)-hydroperoxyeicosapentaenoic acid, and lysophosphatidylcholine 15:

1.

3. A predictive model for the efficacy of HIF-2α inhibitors in VHL-related renal cell carcinoma, characterized in that, The model's indices include any one or more of the following: 3β-hydroxy-5-cholenic acid, 10-(2,3-dihydroxypropoxy)-10-oxodecanoic acid, yeast amino acid, pyridoxamine phosphate, 12(S)-hydroxyheptadecanoic acid, cfDNA, chromosome instability trait CX10, phosphatidic acid (15:0 / 20:2 (11Z,14Z)), 5(S)-hydroperoxyeicosapentaenoic acid, and lysophosphatidylcholine 15:

1.

4. A predictive system for the efficacy of HIF-2α inhibitors in VHL-related renal cell carcinoma, characterized in that, The system includes: Data acquisition unit: Acquire any one or more of the following characteristics of the patient to be tested: 3β-hydroxy-5-cholenic acid, 10-(2,3-dihydroxypropoxy)-10-oxodecanoic acid, yeast amino acid, pyridoxamine phosphate, 12(S)-hydroxyheptadecanoic acid, cfDNA, chromosomal instability trait CX10, phosphatidic acid (15:0 / 20:2 (11Z,14Z)), 5(S)-hydroperoxyeicosapentaenoic acid, and lysophosphatidylcholine 15:1 level; Data processing unit: Inputs the above features into the constructed VHL-related renal cell carcinoma efficacy prediction model; Results output unit: Based on the prediction model, predict whether the VHL-related renal cell carcinoma patient to be tested is effective for HIF-2α inhibitor treatment.

5. The system according to claim 4, characterized in that, The steps for constructing the VHL-related renal cell carcinoma treatment efficacy prediction model are as follows: One or more of the following are used as feature variables and input into a machine learning algorithm to construct a predictive model: 3β-hydroxy-5-cholenic acid, 10-(2,3-dihydroxypropoxy)-10-oxodecanoic acid, yeast amino acid, pyridoxamine phosphate, 12(S)-hydroxyheptadecanoic acid, cfDNA, chromosome instability feature CX10, phosphatidic acid (15:0 / 20:2(11Z,14Z)), 5(S)-hydroperoxyeicosapentaenoic acid, and lysophosphatidylcholine 15:

1.

6. A device for predicting the efficacy of HIF-2α inhibitors in VHL-related renal cell carcinoma, characterized in that, The device includes a memory and a processor, the memory being used to store program instructions; the processor being used to invoke the program instructions, which, when executed, implement the method according to any one of claims 1-2.

7. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the method described in any one of claims 1-2.

8. A computer program product, characterized in that, Includes a computer program that, when executed by a processor, implements the method described in any one of claims 1-2.

9. The application of reagents for detecting biomarker levels in the preparation of efficacy prediction products for HIF-2α inhibitors in VHL-related renal cell carcinoma, wherein the biomarkers include one or more of the following: 3β-hydroxy-5-cholenic acid, 10-(2,3-dihydroxypropoxy)-10-oxodecanoic acid, yeast amino acid, pyridoxamine phosphate, 12(S)-hydroxyheptadecanoic acid, cfDNA, chromosomal instability trait CX10, phosphatidic acid (15:0 / 20:2 (11Z,14Z)), 5(S)-hydroperoxyeicadenosine monophosphate, and lysophosphatidylcholine 15:

1.

10. A product for predicting the efficacy of HIF-2α inhibitors in VHL-related renal cell carcinoma, characterized in that, The product includes reagents for detecting biomarker levels, the biomarkers including one or more of the following: 3β-hydroxy-5-cholenic acid, 10-(2,3-dihydroxypropoxy)-10-oxodecanoic acid, yeast amino acid, pyridoxamine phosphate, 12(S)-hydroxyheptadecanoic acid, cfDNA, chromosomal instability trait CX10, phosphatidic acid (15:0 / 20:2 (11Z,14Z)), 5(S)-hydroperoxyeicosapentaenoic acid, and lysophosphatidylcholine 15:1.

Images