Gene panel for predicting responsiveness of renal cancer patients to radiotherapy

Abstract

The present invention relates to a gene panel for predicting the responsiveness of renal cancer patients to radiotherapy. In the present invention, the responsiveness of patients to stereotactic body radiotherapy can be predicted with high accuracy by only biopsy of a small amount of tissue, thereby maximizing the efficiency of radiotherapy for patients with difficulty in surgery and at the same time, providing the advantage of preserving renal function.

Description

A gene panel for predicting radiotherapy responsiveness in patients with renal cell carcinoma

[0001] The present invention relates to a gene panel for predicting radiotherapy responsiveness in renal cancer patients.

[0002] Renal cell carcinoma (RCC) is the ninth most common cancer in men worldwide, accounting for approximately 2.2% of newly diagnosed malignancies. The cumulative risk of renal cell carcinoma (RCC) is higher in developed countries and is closely associated with frequent abdominal imaging, advanced age, and exposure to metabolic risk factors. Surgical resection is the standard treatment for localized RCC, but the clinical benefits of aggressive treatment may be limited in elderly patients or those with multiple comorbidities and frailty.

[0003] Several alternative treatment options exist for the management of small renal cell carcinomas in elderly patients with comorbidities who are not suitable for surgery. Radiofrequency ablation (RFA) is a viable option if the tumor is cystic, centrally located, adjacent to large intrarenal vessels and / or the pelvic-calycicular system, or if the patient does not have uncorrected coagulopathy.

[0004] Stereotactic body radiotherapy (SBRT) is a well-established and effective treatment for lung, liver, and spine cancers. For renal cell carcinoma, hypofractionated radiotherapy (HFRT) has emerged as an effective approach to provide local control in metastatic cases, and evidence regarding its role in local disease is accumulating (Singh R, Ansinelli H, Sharma D, et al: Stereotactic body radiation therapy (SBRT) for metastatic renal cell carcinoma: a multi-institutional experience. J Radiosurg SBRT 7:29-37, 2020; Siva S, Ali M, Correa RJM, et al: 5-year outcomes after stereotactic ablative body radiotherapy for primary renal cell carcinoma: an individual patient data meta-analysis from IROCK (the International Radiosurgery Consortium of the Kidney). Lancet Oncol 23:1508-1516, 2022). However, prospective and definitive data on the outcomes of SBRT for localized renal cell carcinoma are still lacking.

[0005] Renal function and its maintenance are key treatment priorities in the management of small renal cell carcinoma (RCC). Maintaining adequate renal function is closely related to cardiovascular health, quality of life, and all-cause morbidity and mortality, especially in frail populations. Curative interventions for small, asymptomatic RCC should also aim to preserve adequate renal function. In this context, we designed a prospective clinical trial to evaluate tumor control and renal function preservation in patients with T1a RCC. Furthermore, we identified clinical and genomic predictors of response and established a system to preemptively assess whether a patient will exhibit a significant therapeutic response to SBRT, thereby completing the present invention.

[0006] One object of the present invention is to provide a gene panel for predicting radiation therapy responsiveness in renal cancer patients.

[0007] However, the technical problems to be solved by the present invention are not limited to the problems mentioned above, and other problems not mentioned can be clearly understood by those skilled in the art from the description below.

[0008] To achieve the above purpose, the present invention provides a composition for diagnosing the prognosis of stereotactic body radiation therapy (SBRT) of renal cell carcinoma, comprising an agent that specifically detects a combination of 30 genes or transcripts thereof as described in Table 1:

[0009] [Table 1]

[0010]

[0011] In the present invention, the combination of the genes or their transcripts may be a combination of the 220 genes or their transcripts described in Table 2:

[0012] [Table 2]

[0013]

[0014]

[0015]

[0016]

[0017]

[0018] In the present invention, the renal cell carcinoma may be clear cell type renal cell carcinoma.

[0019] In the present invention, the agent capable of specifically detecting a combination of the gene or its transcript may be at least one selected from the group consisting of an antibody, oligopeptide, ligand, PNA (peptide nucleic acid), and aptamer, primer, or probe that specifically binds to the gene or its transcript.

[0020] The present invention also provides a SBRT prognostic diagnostic kit comprising the composition.

[0021]

[0022] In the present invention, the kit may be an RT-PCR kit, a DNA chip kit, an ELISA kit, a protein chip kit, a rapid kit, or an MRM (Multiple reaction monitoring) kit.

[0023] The present invention also provides:

[0024] (a) detecting 30 genes or transcript combinations thereof listed in Table 1 in a biological sample isolated from a patient with renal cell carcinoma; and

[0025] (b) a step of analyzing the output result value by inputting the detected degree into an artificial intelligence model trained to predict the SBRT prognosis; a method for providing information for SBRT prognosis diagnosis is provided.

[0026] In the present invention, the gene or its transcript combination may be any of the 220 genes or their transcript combinations described in Table 2.

[0027] In the present invention, the biological sample may be a biopsy tissue isolated from a renal cancer patient.

[0028] The present invention also comprises: (a) a measuring unit for detecting 30 genes or transcript combinations thereof described in Table 1 in a biological sample isolated from a renal cell carcinoma patient; and

[0029] (b) Provided is an SBRT prognostic diagnostic device including a detection unit that outputs SBRT prognostic responsiveness from the expression level of a gene or its transcript measured in the above measurement unit.

[0030] In the present invention, the gene or its transcript combination may be any of the 220 genes or their transcript combinations described in Table 2.

[0031] The present invention can predict a patient's responsiveness to stereotactic external beam radiotherapy with high accuracy using only a small amount of tissue biopsy, thereby maximizing the effectiveness of radiotherapy for patients who are difficult to operate on, while providing the advantage of preserving renal function.

[0032]

[0033] Figure 1 shows cancer-related outcomes and changes in renal function after SBRT treatment for renal cell carcinoma (RCC).

[0034] Figure 1a is a waterfall plot depicting the percentage change in maximum tumor size of primary renal cell carcinoma after SBRT compared to baseline. At a median follow-up of 36 months, local control was achieved in 78 patients, of whom 8 achieved complete remission.

[0035] Figure 1b is a Kaplan-Meier curve showing the 3-year progression-free survival (PFS) rate in patients treated with SBRT. The 3-year PFS rate was 96%, with a 95% confidence interval of 89.8% to 99.9%.

[0036] Figure 1c is a Kaplan-Meier curve showing the 3-year cancer-specific survival (CSS) rate for patients treated with SBRT. The 3-year CSS rate was 96%, with a 95% confidence interval of 89.4–99.9%.

[0037] Figure 1d is a graph showing the temporal changes in the estimated glomerular filtration rate (eGFR) of kidneys treated with SBRT. The GFR of the treated kidney decreased by 9.8 ml / min / 1.73 m² from 12 to 18 months after treatment, but then stabilized and recovered to a level 5.2% (2.0 ml / min / 1.73 m²) lower than baseline at 24 to 36 months.

[0038] Figure 2 shows differences in gene expression patterns between SBRT responders and non-responders.

[0039] Figure 2a is a volcano diagram depicting differentially expressed genes (DEGs) between responders and non-responders. Genes overexpressed in responders primarily belonged to functional ontologies associated with the apical region of cells, while genes upregulated in non-responders were associated with the mitochondrial matrix and flavin adenine dinucleotide (FAD) binding.

[0040] Figure 2b shows the results of Gene Set Enrichment Analysis (GSEA) using the Hallmark data set. Responders were enriched in gene sets related to cell surface and junctional pathways, while non-responders were enriched in gene sets related to oxidative phosphorylation, cell cycle, hypoxia, and DNA repair pathways. These results are consistent with existing evidence for metabolic reprogramming in cancers associated with radioresistance.

[0041] Figure 3 shows cases of complete remission and cases of progressive disease.

[0042] Figure 3A illustrates a case of complete remission. The tumor size was 20 mm at the start of treatment, decreased by 15% at 3 months, and 55% at 12 months. Complete remission was achieved at 30 months, and this status was maintained up to 60 months. Figure 3B illustrates a case of progressive disease. The tumor size was 28 mm at the start of treatment, and decreased by 45% at 3 months. However, it began to increase again at 12 months, and progressed to a larger mass than baseline at the subsequent follow-up (d, e).

[0043] Figure 4 shows the Weighted Gene Co-expression Network Analysis (WGCNA).

[0044] Figure 4a is a cluster dendrogram illustrating 22 gene modules formed based on gene expression patterns.

[0045] Figure 4b is a volcano diagram of module eigengenes (MEs) to identify differentially expressed functional modules between responders and non-responders. Modules showing statistically significant differences between responders and non-responders are presented.

[0046] Figure 4c depicts the ME1 module enriched in the responder group. The ME1 module is associated with pathways related to extracellular structural organization, normal kidney formation, and the cytoskeleton (e.g., GATA3), and the bar chart shows the degree of enrichment and statistical significance of these pathways.

[0047] Figure 4d depicts the ME18 module, which was enriched in the non-responder group. The ME18 module is associated with pathways related to DNA repair and DNA replication (e.g., SSBP1), and the bar chart shows the degree of enrichment and statistical significance of these pathways.

[0048] The bar charts in Figures 4c and 4d visualize the enrichment and statistical significance of each pathway and module. FC stands for Fold Change, GO stands for Gene Ontology, and KEGG stands for Kyoto Encyclopedia of Genes and Genomes.

[0049] Figure 5 shows the weighted gene co-expression network analysis (WGCNA) of the entire module before merging. This diagram illustrates the process of merging intersecting genes with the original WGCNA results to generate a total of 43 merged modules. The profiles of upregulated genes and associated pathways in the merged modules showed similar patterns to the entire cohort, and expression enrichment patterns were also confirmed. The bar charts indicate specific pathways, the degree of enrichment for each module, and statistical significance. ECM stands for extracellular matrix, FC stands for Fold Change, and ME stands for Module Eigengenes.

[0050] Figure 6 shows the reactivity prediction modeling.

[0051] Figure 6a illustrates the process of building a reactivity prediction model and selecting an optimal feature (gene) set.

[0052] Figure 6b presents the integrated confusion matrix, Receiver Operating Characteristic (ROC) curve, and evaluation metrics after five-fold cross-validation. The model's predictive performance was 0.9 (area under the curve), with a 95% confidence interval of 0.704–1.0. Evaluation metrics such as AUC, False Negative (FN), False Positive (FP), False Positive Rate (FPR), True Negative (TN), True Positive (TP), and True Positive Rate (TPR) are also shown.

[0053] Figure 7 depicts the study flow from enrollment to the end of 24-month follow-up after stereotactic extracorporeal radiotherapy (SBRT). This diagram is a study flow diagram depicting the follow-up protocol from patient enrollment to 24 months after SBRT treatment. AP stands for abdomen-pelvis, CT stands for computed tomography, DTPA stands for diethylenetriamine pentaacetic acid, and P / Ex stands for physical examination.

[0054] Figure 8 shows the RNA sequencing data analysis.

[0055] Figure 8a depicts one outlier from the initial 14 RNA sequencing data sets that was excluded from analysis. The data showed a good response but was judged to be an outlier and was therefore excluded.

[0056] Figure 8b shows the results of the analysis of the remaining 13 RNA sequencing data sets, excluding outliers. PC stands for Principal Component, and QC stands for Quality Control.

[0057]

[0058] The detailed description of the present invention, which follows, will be described with reference to specific drawings (where drawings exist) regarding specific embodiments in which the present invention may be practiced; however, the present invention is not limited thereto, but is defined solely by the appended claims to the full scope equivalent to or equivalent to what the claims describe. It should be understood that the various embodiments / embodiments of the present invention, while different from each other, are not necessarily mutually exclusive. For example, specific shapes, structures, and characteristics described herein may be changed from one embodiment / embodiment to another, or multiple embodiments / embodiments may be combined, without departing from the spirit and scope of the present invention. Technical and scientific terms used herein, unless otherwise defined, have the same meaning as commonly used in the art to which the present invention belongs. For the purpose of interpreting this specification, the following definitions will apply, and a term expressed in the singular should be construed to also refer to the plural (i.e., at least one), unless the context makes it inappropriate.

[0059] The term "about" means an amount, level, value, number, frequency, percentage, dimension, size, quantity, weight, or length that varies by 30, 25, 20, 15, 10, 9, 8, 7, 6, 5, 4, 3, 2, or 1% of the reference amount, level, value, number, frequency, percentage, dimension, size, quantity, weight, or length. For example, the term "about" when used in relation to a value x expressed as a number or numerical value can mean x ± 10%.

[0060]

[0061] The present invention relates, in one aspect, to a gene panel for diagnosing (or evaluating) the responsiveness to stereotactic body radiation therapy (SBRT).

[0062] The above gene panel may be used to analyze gene expression changes in biological samples isolated from renal cell carcinoma patients to assess responsiveness to stereotactic external beam radiotherapy.

[0063] Accordingly, the present invention provides, as one embodiment, a gene panel for evaluating the responsiveness of a patient with renal cancer to stereotactic external beam radiotherapy.

[0064] In the present invention, the renal cancer patient may be a patient with stage T1a.

[0065] In renal cell carcinoma, stage T1a refers to early-stage renal cell carcinoma with a tumor size of 4 cm or less and confined to the kidney.

[0066] In the present invention, the gene panel may include the genes listed in Table 1.

[0067] [Table 1]

[0068]

[0069] Preferably, the gene panel may include the genes listed in Table 2.

[0070] [Table 2]

[0071]

[0072]

[0073]

[0074]

[0075]

[0076] As another embodiment, the present invention provides a composition for diagnosing (or evaluating) the prognosis of stereotactic extracorporeal radiotherapy of renal cell carcinoma, comprising an agent that specifically detects a gene or transcript thereof listed in Table 1.

[0077] As another embodiment, the present invention provides a composition for prognostic diagnosis of stereotactic extracorporeal radiotherapy of renal cell carcinoma, comprising an agent that specifically detects a gene or transcript thereof listed in Table 2.

[0078] In the present invention, the renal cell carcinoma may be clear cell renal cell carcinoma.

[0079] In the present invention, the agent capable of specifically detecting the gene or its transcript may be at least one selected from the group consisting of an antibody, oligopeptide, ligand, PNA (peptide nucleic acid), aptamer, primer, probe or antisense nucleotide that specifically binds to each of the genes or their transcripts, but is not limited thereto.

[0080] The above composition is intended for application to a biological sample isolated from a patient diagnosed with renal cell carcinoma, including but not limited to a solid tissue sample, a tissue culture, a liquid tissue sample, a cell or a cell fragment. Also, non-limiting examples of biological samples include whole blood, leukocytes, peripheral blood mononuclear cells, buffy coat, plasma, serum, sputum, tears, mucus, nasal washes, nasal aspirate, breath, urine, semen, saliva, peritoneal washings, ascites, cystic fluid, meningeal fluid, amniotic fluid, glandular fluid, pancreatic fluid, lymph fluid, pleural fluid, nipple aspirate, bronchial aspirate, synovial fluid, joint fluid. The biological sample may include, but is not limited to, one or more selected from the group consisting of joint aspirate, organ secretions, cells, cell extracts, and cerebrospinal fluid. In a preferred embodiment, the biological sample may be a biopsy tissue isolated from a patient diagnosed with renal cell carcinoma.

[0081] According to the present invention, total RNA is isolated from a biopsy tissue isolated from a patient diagnosed with renal cell carcinoma, reverse transcribed into complementary DNA, and gene expression profiling is performed through gene expression analysis, and the expression levels of genes or transcripts thereof listed in Table 1 or Table 2 are compared with a control group to diagnose the prognosis for stereotactic extracorporeal radiotherapy of renal cell carcinoma.

[0082] In the present invention, "antibody" refers to a substance that specifically binds to an antigen and causes an antigen-antibody reaction. For the purposes of the present invention, an antibody refers to an antibody that specifically binds to the protein. Antibodies of the present invention include polyclonal antibodies, monoclonal antibodies, and recombinant antibodies. The antibodies can be readily produced using techniques well known in the art. For example, polyclonal antibodies can be produced by methods well known in the art, including a process of injecting an antigen of the protein into an animal and collecting blood from the animal to obtain serum containing the antibody. Such polyclonal antibodies can be produced from any animal, such as a goat, rabbit, sheep, monkey, horse, pig, cow, or dog. In addition, monoclonal antibodies can be produced using the hybridoma method well known in the art (see Kohler and Milstein (1976) European Journal of Immunology 6:511-519) or the phage antibody library technology (see Clackson et al, Nature, 352:624-628, 1991; Marks et al, J Mol Biol, 222:58, 1-597, 1991). Antibodies produced by the above methods can be separated and purified using methods such as gel electrophoresis, dialysis, salt precipitation, ion exchange chromatography, and affinity chromatography. In addition, the antibody of the present invention includes not only a complete form having two full-length light chains and two full-length heavy chains, but also functional fragments of antibody molecules. Functional fragments of antibody molecules mean fragments that have at least an antigen-binding function, and include Fab, F(ab'), F(ab')2, and Fv.

[0083] In the present invention, “oligopeptide” is a peptide composed of 2 to 20 amino acids and may include, but is not limited to, dipeptides, tripeptides, tetrapeptides, and pentapeptides.

[0084] In the present invention, "ligand" refers to a molecule capable of selectively and specifically binding to a specific target molecule (e.g., a protein, nucleic acid, receptor, sugar chain, etc.). Such ligands may be of natural origin or synthetically manufactured, and may include various chemical structures such as small-molecule compounds, peptides, sugars, lipids, and nucleic acid analogs. In one embodiment of the present invention, the ligand can be used to specifically bind to a target gene or its transcript and detect or isolate it.

[0085] In the present invention, "PNA (Peptide Nucleic Acid)" refers to an artificially synthesized polymer similar to DNA or RNA, and was first introduced in 1991 by Professors Nielsen, Egholm, Berg, and Buchardt of the University of Copenhagen, Denmark. While DNA has a phosphate-ribose sugar backbone, PNA has a repeated N-(2-aminoethyl)-glycine backbone linked by peptide bonds, which greatly increases its binding affinity and stability to DNA or RNA, and is used in molecular biology, diagnostic analysis, and antisense therapy. PNA is disclosed in detail in the literature [Nielsen PE, Egholm M, Berg RH, Buchardt O (December 1991) "Sequence-selective recognition of DNA by strand displacement with a thymine-substituted polyamide" Science 254 (5037): 1497-1500].

[0086] In the present invention, "aptamer" is an oligonucleotide or peptide molecule, and the general contents of aptamers are disclosed in detail in the literature [Bock LC et al, Nature 355(6360):5646(1992); Hoppe-Seyler F, Butz K "Peptide aptamers: powerful new tools for molecular medicine" J Mol Med 78(8):42630(2000); Cohen BA, Colas P, Brent R "An artificial cell cycle inhibitor isolated from a combinatorial library" Proc Natl Acad Sci USA 95(24): 142727(1998)].

[0087] In the present invention, a "primer" is a fragment that recognizes a target gene sequence, and includes a pair of forward and reverse primers, but is preferably a pair of primers that provide analysis results with specificity and sensitivity. When the nucleic acid sequence of the primer is a sequence that does not match the non-target sequence present in the sample, and thus only amplifies the target gene sequence containing the complementary primer binding site and does not cause non-specific amplification, high specificity can be imparted.

[0088] In the present invention, the term "probe" refers to a substance that can specifically bind to a target substance to be detected in a sample, and refers to a substance that can specifically confirm the presence of the target substance in the sample through said binding. The type of probe is not limited to a substance commonly used in the art, but is preferably PNA (peptide nucleic acid), LNA (locked nucleic acid), peptide, polypeptide, protein, RNA, or DNA, and most preferably PNA. More specifically, the probe includes a biomaterial derived from or similar to a living organism or manufactured in vitro, and may be, for example, an enzyme, a protein, an antibody, a microorganism, an animal or plant cell and organ, a nerve cell, DNA, and RNA. DNA includes cDNA, genomic DNA, and oligonucleotides, RNA includes genomic RNA, mRNA, and oligonucleotides, and examples of proteins include antibodies, antigens, enzymes, peptides, etc. "LNA (Locked nucleic acids)" refers to a nucleic acid analog containing a 2'-O, 4'-C methylene bridge [J Weiler, J Hunziker and J Hall Gene Therapy (2006) 13, 496502] LNA nucleosides contain common nucleic acid bases of DNA and RNA and can form base pairs according to the Watson-Crick base pairing rule. However, due to the 'locking' of the molecule caused by the methylene bridge, LNA does not form an ideal shape in Watson-Crick binding. When LNA is included in a DNA or RNA oligonucleotide, LNA can pair more quickly with a complementary nucleotide chain and increase the stability of the double helix.

[0089] As used herein, "antisense nucleotide" refers to an oligomer having a sequence of nucleotide bases and an intersubunit backbone that allows the antisense oligomer to hybridize with a target sequence within RNA by Watson-Crick base pairing, typically allowing the formation of an mRNA and RNA:oligomer heteroduplex within the target sequence. The oligomer may have exact sequence complementarity or approximate sequence complementarity to the target sequence.

[0090] As another embodiment of the present invention, the present invention provides a kit for prognostic diagnosis of stereotactic extracorporeal radiotherapy for a patient with renal cancer.

[0091] The above kit may be, but is not limited to, an RT-PCR kit, a DNA chip kit, an ELISA kit, a protein chip kit, a rapid kit, or an MRM (Multiple reaction monitoring) kit.

[0092] The "kit" of the present invention refers to a tool capable of assessing the expression level of a biomarker by labeling a probe or antibody that specifically binds to the biomarker component with a detectable label. Specifically, the kit may refer to a collection of diagnostic tools comprising reagents, containers, and accessory components.

[0093] In addition to direct labeling of a detectable substance related to a probe or antibody by reaction with a substrate, indirect labeling is also included in which a chromogenic label is conjugated by reactivity with another directly labeled reagent. The label may include a chromogenic substrate solution, a washing solution, and other solutions that will react with the label for chromogenic reaction, and may be manufactured by including the reagent components used.

[0094] The kit of the present invention may be a kit including essential elements required for performing RT-PCR, and may include, in addition to each primer pair specific for a marker gene, a test tube, a reaction buffer, deoxynucleotides (dNTPs), Taq polymerase, reverse transcriptase, DNase, RNase inhibitor, sterile water, etc.

[0095] Additionally, the kit of the present invention includes the essential elements necessary for performing a DNA chip, and may be a kit for detecting the gene. The DNA chip kit includes a substrate to which cDNA corresponding to a gene or a fragment thereof is attached as a probe, and the substrate may include cDNA corresponding to a quantitative control gene or a fragment thereof. The kit of the present invention is not limited to any known technology in the art.

[0096] In the present invention, the kit may be an RT-PCR kit, a DNA chip kit, an ELISA kit, a protein chip kit, a rapid kit, or an MRM (Multiple reaction monitoring) kit.

[0097] The kit of the present invention may further comprise one or more other component compositions, solutions, or devices suitable for the analysis method. For example, the kit of the present invention may further comprise essential elements necessary for performing a reverse transcription polymerase reaction. The reverse transcription polymerase reaction kit comprises a pair of primers specific for a gene encoding a marker protein. The primers are nucleotides having a sequence specific for the nucleic acid sequence of the gene, and may have a length of about 7 bp to 50 bp, more preferably about 10 bp to 30 bp. It may also comprise a primer specific for the nucleic acid sequence of a control gene. In addition, the reverse transcription polymerase reaction kit may comprise a test tube or other appropriate container, a reaction buffer (with various pH and magnesium concentrations), deoxynucleotides (dNTPs), an enzyme such as Taq polymerase and reverse transcriptase, DNase, RNase inhibitor DEPC-water, sterile water, etc. A DNA chip kit may include a substrate to which cDNA or oligonucleotides corresponding to a gene or a fragment thereof are attached, and reagents, preparations, enzymes, etc. for producing a fluorescently labeled probe. The substrate may also include cDNA or oligonucleotides corresponding to a control gene or a fragment thereof.

[0098] Additionally, the kit of the present invention may include essential components necessary for performing an ELISA. The ELISA kit includes an antibody specific for a protein expressed by the gene of the present invention. The antibody has high specificity and affinity for the marker protein and little cross-reactivity with other proteins, and may be a monoclonal antibody, polyclonal antibody, or recombinant antibody. The ELISA kit may also include an antibody specific for a control protein. In addition, the ELISA kit may include reagents capable of detecting bound antibodies, such as labeled secondary antibodies, chromophores, enzymes (e.g., conjugated to antibodies), and their substrates or other substances capable of binding to antibodies.

[0099] In the present invention, a fixative for the antigen-antibody binding reaction may be a nitrocellulose membrane, a PVDF membrane, a well plate synthesized from polyvinyl resin or polystyrene resin, a glass slide glass, etc., but is not limited thereto.

[0100] In addition, in the present invention, the label of the secondary antibody is preferably a conventional chromogen that undergoes a color development reaction, and labels such as fluorescein and dyes such as HRP (horseradish peroxidase), alkaline phosphatase, colloid gold, FITC (poly L-lysine-fluorescein isothiocyanate), and RITC (rhodamine-B-isothiocyanate) can be used, but are not limited thereto.

[0101] In addition, in the present invention, it is preferable to use a chromogenic substrate for inducing color development according to a marker that undergoes a color development reaction, and TMB (3,3',5,5'-tetramethyl bezidine), ABTS [2,2'-azino-bis(3-ethylbenzothiazoline-6-sulfonic acid)], OPD (o-phenylene diamine), etc. can be used. At this time, it is more preferable that the chromogenic substrate is provided in a state dissolved in a buffer solution (0.1 M NaAc, pH 5.5). A chromogenic substrate such as TMB is decomposed by HRP used as a marker of a secondary antibody conjugate to generate a chromogenic precipitate, and the presence or absence of the marker proteins is detected by visually confirming the degree of deposition of this chromogenic precipitate.

[0102] In the present invention, the washing solution preferably contains a phosphate buffer solution, NaCl, and Tween 20, and a buffer solution (PBST) composed of 0.02 M phosphate buffer solution, 0.13 M NaCl, and 0.05% Tween 20 is more preferred. After the antigen-antibody binding reaction, the washing solution reacts the secondary antibody with the antigen-antibody complex, and then an appropriate amount is added to the fixative and washed 3 to 6 times. The reaction stopping solution can preferably be a sulfuric acid solution (H2SO4).

[0103] In the present invention, the measurement of the expression level of the protein or biomarker can be performed by protein chip analysis, immunoassay, ligand binding assay, MALDI-TOF (Matrix Assisted Laser Desorption / Ionization Time of Flight Mass Spectrometry) analysis, SELDI-TOF (Sulface Enhanced Laser Desorption / Ionization Time of Flight Mass Spectrometry) analysis, radioimmunoassay, radioimmunodiffusion, aukteroni immunodiffusion, rocket immunoelectrophoresis, tissue immunostaining, complement fixation assay, two-dimensional electrophoresis analysis, liquid chromatography-mass spectrometry (LCMS), liquid chromatography-mass spectrometry / mass spectrometry (LC-MS / MS), western blotting, or enzyme linked immunosorbent assay (ELISA).

[0104] Additionally, in the present invention, the measurement of the expression level of the protein or biomarker can be performed using a multiple reaction monitoring (MRM) method.

[0105] In the present invention, the internal standard material in the multiple reaction monitoring method may be a synthetic peptide in which a specific amino acid constituting the target peptide is substituted with an isotope or E. coli beta-galactosidase.

[0106] In the present invention, the measurement of the expression level of the gene encoding the protein or biomarker can be performed by reverse transcription polymerase chain reaction (RT-PCR), competitive reverse transcription polymerase reaction (Competitive RT-PCR), real-time reverse transcription polymerase reaction (Real-time RT-PCR), RNase protection assay (RPA), Northern blotting, or DNA chip.

[0107]

[0108] As another embodiment, the present invention provides a method for providing information capable of diagnosing (or evaluating) the prognosis of stereotactic extracorporeal radiotherapy.

[0109] The method of providing the above information may include the following steps:

[0110] (a) a step of detecting the expression level of a gene or its transcript listed in Table 1 in a biological sample isolated from a patient with renal cell carcinoma; and

[0111] (b) A step of analyzing the output result value by inputting the detected degree into an artificial intelligence model trained to predict the prognosis of stereotactic extracorporeal radiotherapy.

[0112] The above step (a) may be a step of specifically detecting the expression level of a gene or its transcript listed in Table 2.

[0113] Additionally, the biological sample may be a biopsy tissue isolated from a renal cancer patient. The renal cancer patient may be a T1a stage clear cell renal cell carcinoma patient.

[0114] In the present invention, predicting the prognosis of stereotactic external beam radiotherapy (ESRT) refers to detecting the expression level of a patient's genes or their transcripts prior to treatment and, based on this, assessing the patient's likelihood of responding to ESRT. If responsiveness is predicted, ESRT can be performed. If non-responsiveness is predicted, alternative treatments can be considered or ESRT can be postponed.

[0115] The above information provision method can proceed as follows. Prepare a biopsy tissue sample collected from a patient considering stereotactic external beam radiotherapy. Extract total RNA from the collected biopsy tissue sample. The quality of the extracted RNA can be assessed (e.g., DV200 value) using Quant-IT RiboGreen and TapeStation RNA Screentape to determine if it is suitable for analysis. Create a library for RNA sequencing (e.g., using the SureSelectXT RNA Direct Library Preparation Kit) and obtain gene reads (e.g., using the 101 bp paired-end method on the Illumina platform). Map these reads to the reference genome and quantify transcript abundance to generate gene expression data (e.g., using programs such as HISAT2 and StringTie). Select the expression data of genes corresponding to the gene panel listed in Table 1 or Table 2 and convert them to fit the input format of a pre-trained artificial intelligence model. Preprocessed gene expression data is input into a pre-trained AI predictive model (built using machine learning algorithms such as Random Forest, Earth, Adaboost, and svmRadial). The input gene expression data is analyzed to classify patients into "responders" or "non-responders," determining their likelihood of responding to stereotactic external beam radiotherapy.

[0116] As used herein, “diagnosis” means determining a medical condition, such as the presence or absence of a specific disease, the degree of progression, or the responsiveness to treatment, by detecting and analyzing the expression level of a patient’s gene or its transcript.

[0117] As used herein, “evaluation” means estimating or determining in advance the patient’s prognosis, likelihood of treatment response, or treatment effect based on the obtained biological and clinical information, including the above diagnostic results.

[0118] In this specification, the terms diagnosis and assessment may be used interchangeably. Furthermore, the diagnosis and / or assessment may be based on an artificial intelligence model, excluding the subjective judgment of a clinician.

[0119] Responsiveness can be predicted by analyzing the expression levels of genes or their transcripts listed in Table 1 or Table 2, and this can be predicted by detecting the expression levels of genes or their transcripts of patients before treatment, inputting them into a pre-trained artificial intelligence model, and analyzing the output results.

[0120] In one aspect, artificial intelligence model learning according to the present invention can be performed in the following manner.

[0121] First, patient gene expression data is acquired. The data may include a matrix of expression levels for each sample and a label indicating whether the data was responsive or non-responsive. Preprocessing may be performed, as needed, through logarithmic transformation, normalization, or other preprocessing methods. Preprocessing may be performed using log2 transformation, Z-score normalization, or any similar method commonly used in the art. Furthermore, the dataset may be partitioned into a training set and a validation set, using k-fold cross-validation or a predetermined ratio partition (e.g., 80:20 partition). For example, the entire patient dataset can be partitioned into a training set and a validation set using R's createDataPartition function. In this case, the training set may be allocated approximately 80%, and the validation set approximately 20%.

[0122] Next, a recursive feature elimination (RFE) procedure based on repeated cross-validation is performed on the training set to derive the optimal set of features contributing to prediction. Specifically, the initial set of approximately 460 features defined through DEG analysis is used to calculate the importance of variables using the random forest-based RFE function built into the R caret package (v6.0-94), and variables with low importance are sequentially removed to find the optimal number of features. This process is performed through 10 rounds of repeated cross-validation, ultimately selecting 220 feature genes (see Table 2). These selected feature genes are used as input variables in the subsequent training and validation steps.

[0123] Afterwards, the classification model can be constructed using algorithms such as Random Forest, Support Vector Machine (SVM, radial kernel), AdaBoost, and Earth regression. In a specific embodiment of the present invention, a Random Forest model was constructed using R's randomForest package (v4.7-1.2), and a total of 500 trees were applied to secure the stability of the prediction. The learning of the above model can be performed through cross-validation, and in the embodiment, 5-fold cross-validation was performed to estimate the generalization performance of the model.

[0124] Finally, the model's output for the validation set is compared with the actual response results to evaluate model performance by calculating performance indicators such as AUC, accuracy, sensitivity, and specificity. In one example, a model based on 220 feature genes demonstrated excellent performance, with an AUC value of approximately 0.9. Furthermore, by analyzing the importance of the feature genes included in the model, the biological significance of the genes contributing to the prediction can be interpreted.

[0125] The predictive model constructed as described above can be applied to predicting the response to stereotactic external beam radiotherapy (ESRT) in renal cancer patients. Specifically, gene expression data is acquired from the patient's biological samples (e.g., biopsy tissue) and then preprocessed to fit the model's input format. The transformed data is then input into a pre-trained predictive model, which outputs the patient's likelihood of responding to ESRT.

[0126] The above output can be categorized into a responder or non-responder group, and in some cases, it can be provided in the form of a probability value (e.g., a score between 0 and 1). Based on these results, stereotactic external beam radiotherapy can be administered, or alternative treatment methods can be considered.

[0127] As another embodiment of the present invention, a diagnostic device capable of diagnosing (or evaluating) the prognosis of stereotactic extracorporeal radiotherapy of the present invention is provided.

[0128] The above diagnostic device may include (a) a measuring unit that detects the expression level of a gene or its transcript listed in Table 1 in a biological sample isolated from a patient with renal cell carcinoma; and (b) a detection unit that outputs the prognostic response to stereotactic extracorporeal radiotherapy from the expression level of the gene or its transcript measured by the measuring unit.

[0129] The above measurement unit may be a measurement unit that specifically detects the expression level of a gene or transcript thereof listed in Table 2.

[0130]

[0131] As another embodiment of the present invention, software capable of diagnosing (or evaluating) the prognosis of stereotactic extracorporeal radiotherapy is provided.

[0132] The above software,

[0133] (a) an input module that receives gene expression data (e.g., data including the expression levels of genes or their transcripts listed in Table 1 or Table 2) for a biological sample isolated from a patient with renal cell carcinoma; and

[0134] (b) an analysis module that calculates the patient's responsiveness to stereotactic extracorporeal radiotherapy by applying a pre-learned prediction model based on the gene expression data collected from the input module; and

[0135] (c) It may include an output module that classifies the result values ​​produced by the above analysis module into a response group or a non-response group or provides the result values ​​to the user in the form of a response probability value.

[0136] The above input module may be configured to specifically receive expression level data of genes or transcripts thereof listed in Table 1 or Table 2, and the analysis module may derive results using a prediction model learned based on a machine learning algorithm such as random forest, SVM, or Adaboost.

[0137] As another embodiment of the present invention, a computer-readable recording medium storing the software is provided.

[0138] As another embodiment of the present invention, the use of a gene panel comprising a gene or transcript thereof listed in Table 1 or Table 2 for the prognosis diagnosis (or evaluation) of stereotactic extracorporeal radiotherapy in a patient with renal cell carcinoma is provided.

[0139] As another embodiment of the present invention, there is provided a use of a composition comprising an agent that specifically detects a gene or transcript thereof listed in Table 1 or Table 2, for use in the prognosis diagnosis (or evaluation) of stereotactic extracorporeal radiotherapy in a patient with renal cell carcinoma.

[0140] As another embodiment of the present invention, a gene panel comprising the genes listed in Table 1 or Table 2 or their transcripts, which is used for prognostic diagnosis (or evaluation) of stereotactic extracorporeal radiotherapy in patients with renal cell carcinoma, is provided.

[0141] As another embodiment of the present invention, a composition comprising an agent that specifically detects a gene or transcript thereof listed in Table 1 or Table 2, which is used for prognostic diagnosis (or evaluation) of stereotactic extracorporeal radiotherapy in a patient with renal cell carcinoma is provided.

[0142] As another embodiment of the present invention, the use of a gene panel comprising a gene or transcript thereof listed in Table 1 or Table 2 is provided for the manufacture of a composition for prognostic diagnosis (or evaluation) of stereotactic extracorporeal radiotherapy in a patient with renal cell carcinoma.

[0143] As another embodiment of the present invention, a composition comprising an agent that specifically detects a gene or transcript thereof listed in Table 1 or Table 2 is provided for use in the manufacture of a composition for prognostic diagnosis (or evaluation) of stereotactic extracorporeal radiotherapy in a patient with renal cell carcinoma.

[0144] In this study, we demonstrated that stereotactic external beam radiotherapy (ESRT) provides an acceptable 3-year progression-free survival rate without renal toxicity in patients with T1a RCC. Furthermore, we identified differential gene expression patterns between responders and non-responders, suggesting that these differential gene expression patterns could serve as potential biomarkers for developing personalized treatment strategies and identifying optimal candidates for ESRT.

[0145] Percutaneous resection is highly operator-dependent, and treatment outcomes vary significantly depending on tumor location. Anterior tumors or tumors near major blood vessels or the renal pelvis are particularly challenging to treat. In contrast, stereotactic external beam radiotherapy (ESRT) delivers a uniform and sufficient dose and is suitable for patients with advanced coagulopathy or cystic tumors who are ineligible for percutaneous resection. Furthermore, ESRT does not require anesthesia, allowing even patients who are not candidates for surgery to complete treatment without significant toxicity.

[0146] Genetic analysis revealed that non-responders were enriched in pathways related to the cell cycle, DNA repair, oxidative phosphorylation, and hypoxia. Related genes include SSBP1, RFC1, RFC4, and MBD4, whose protein products play a key role in maintaining genome stability through single-strand protection, DNA repair checkpoints, and prevention of CpG mutations. Conversely, responders were primarily enriched in genes associated with kidney development, including GATA3, a key factor in endothelial cell biology and a key regulator of the ionizing radiation response of human keratinocytes.

[0147] Preservation of renal function is a key goal in RCC treatment, as surgery inevitably results in nephron loss. Indeed, a recent study of young, healthy patients found that robot-assisted partial nephrectomy for tumors averaging 2.5 cm in size resulted in a greater decline in GFR at 3 months postoperatively compared with stereotactic external beam radiotherapy (22%, 9.9 ml / min / 1.73 m² vs. 6.8%, 2.5 ml / min / 1.73 m²). This suggests that stereotactic external beam radiotherapy may be superior to surgery in preserving renal function. Furthermore, the slight decline in total GFR may be attributable to comorbidities in the patient population, consistent with the observation that renal function decline is generally more influenced by systemic disease than by locoregional therapy.

[0148] The model according to the present invention can provide important information that enables stereotactic extracorporeal radiotherapy to be performed on a patient group with a high probability of response, and alternative treatments (e.g., RFA, immunotherapy) to a patient group with a low probability of response, by predicting in advance whether or not a patient will respond to stereotactic extracorporeal radiotherapy.

[0149]

[0150] Hereinafter, the present invention will be described in more detail through examples. These examples are intended solely to illustrate the present invention more specifically, and it will be apparent to those skilled in the art that the scope of the present invention is not limited by these examples, in accordance with the gist of the present invention.

[0151]

[0152] Example

[0153]

[0154] Example 1. Patient selection

[0155] The phase 2 clinical trial was approved by the institutional ethics committee. Eligible patients were those with cT1a renal masses, aged 75 years or older, with an Eastern Cooperative Oncology Group (ECOG) performance score of 2 or higher, a Charlson comorbidity score of 3 or higher, or who refused surgery.

[0156] Exclusion criteria included multiple or bilateral tumors, tumors larger than 4 cm, lymph node or distant metastases, or prior systemic treatment. All patients, except those with Bosniak IV cyst-type complex cystic masses, underwent computed tomography (CT)-guided biopsy and complete metastasis evaluation.

[0157] A total of 87 patients were screened for eligibility, and four were excluded (two refused to participate and two were excluded due to worsening of their general condition). Ultimately, 83 patients were assigned to the treatment group; however, three of these were lost to follow-up after 3 months, leaving 80 patients in the final analysis.

[0158]

[0159] [Table 3] Baseline characteristics of 80 patients with T1a RCC who received SBRT

[0160]

[0161]

[0162] Example 2. Radiation treatment

[0163] All patients were treated after administration of three fiducial markers for the Synchrony tracking system using the CyberKnife version 9.5 system (Accuray, Sunnyvale, CA, USA).

[0164] Unenhanced 3D CT simulations were performed with a 1.25-mm slice thickness and under end-expiratory breath-hold conditions, and were used to delineate the gross tumor volume (GTV) in conjunction with contrast-enhanced CT and / or magnetic resonance imaging (MRI).

[0165] The planning target volume (PTV) was generated with a 3-mm margin. Treatment plans were developed using Multiplan version 4.5 (Accuray) with a ray-tracing algorithm and a fixed cone diameter of 5–60 mm. A standardized dose of 42 Gy was administered in three fractions every other day. The dose was adjusted so that at least 95% of the PTV was within the 80–95% isodose line. Dose restrictions for organs at risk were based on the Timmerman report.

[0166]

[0167] Example 3. Follow-up

[0168] After treatment, patients were evaluated for radiation-related toxicity according to the Common Terminology Criteria for Adverse Events (CTCAE) version 5.0.

[0169] Fifteen patients (6.2%) experienced Grade 1 treatment-related nausea, and six patients (7.5%) experienced Grade 1 anorexia. All symptoms resolved spontaneously and improved within 1 month. There were no treatment-related deaths, and no Grade 2-4 toxicities were observed.

[0170]

[0171] [Table 4] Toxicity observed after SBRT according to Common Terminology Criteria for Adverse Events version 5.0

[0172]

[0173]

[0174] Recurrence was assessed by CT, and renal function was measured by diethylenetriamine pentaacetic acid (DTPA) renal scan. Total estimated glomerular filtration rate (GFR) was calculated according to the Chronic Kidney Disease-Epidemiology (CKD-EPI) formula. The 2-year follow-up protocol is presented in Figure 7.

[0175] Tumor response was assessed by an experienced radiologist according to the Response Evaluation Criteria in Solid Tumors (RECIST) version 1.1.

[0176] Tumor size began to decrease 3 months after SBRT. The mean reduction was 10.0% without any change in the contrast enhancement pattern. During a median follow-up period of 31.8 months (interquartile range, 19.9, 46.8), 78 patients (97.5%) achieved local control, including 8 complete remissions (CRs) (10.0%). The 3-year progression-free survival (PFS) and cancer-specific survival (CSS) rates were 96.4% and 96.3%, respectively. Local progression occurred in two patients, one with distant metastases and the other without distant metastases. Representative CT scan results of responders and non-responders are presented in Figure 3. Sixteen patients died of unrelated causes during the follow-up period. Among clinicopathologic factors potentially associated with CR, age (odds ratio (OR), 0.920; 95% confidence interval (CI), 0.855–0.990; p=0.025) and tumor size reduction at 3 months (OR, 1.084; 95% CI, 1.028–1.144; p=0.003) were significant. Tumor size reduction at 3 months after SBRT significantly predicted CR, with an area under the curve (AUROC) of 0.821 (95% CI, 0.600–1.000). The optimal threshold was a 20% reduction from baseline (sensitivity=0.875, specificity=0.847).

[0177]

[0178] Example 4. RNA sequencing and data preprocessing

[0179] Among patients with clear cell renal cell carcinoma (RCC) who were followed up for more than 24 months, 6 patients with a good response (2 complete remissions, 4 partial remissions) and 8 patients with a poor response (2 progressive disease, 6 stable disease) were selected as subjects for gene expression signature evaluation.

[0180] Total RNA was extracted from formalin-fixed, paraffin-embedded (FFPE) tumor biopsy blocks (5 mm × 5 mm, 10 μm thick), and sequencing was performed at Macrogen (Seoul, Korea). Total RNA concentration was measured using Quant-IT RiboGreen (Invitrogen), and RNA quality was assessed by the DV200 value (the proportion of RNA fragments ≥200 bp) using TapeStation RNA Screentape (Agilent).

[0181] Sequencing libraries were prepared using the SureSelectXT RNA Direct Library Preparation Kit, and sequenced using the 101 bp paired-end method on the Illumina platform. The acquired reads were mapped to the GRCh38 reference genome using the HISAT2 program, and the reads were then assembled using the StringTie program to quantify transcript abundance. This resulted in read counts and normalized values ​​[e.g., fragments per kilobase of transcript per million mapped reads (FPKM), transcripts per million (TPM)].

[0182] During the initial analysis of RNA sequencing data from 14 patients, one patient in the good response group was identified as an outlier, so the final analysis was performed based on data from 13 patients.

[0183]

[0184] Example 5. Differentially expressed genes and pathway analysis

[0185] In the normalized gene expression data, genes with expression values ​​less than 10 in more than 90% of the samples and 22 outlier genes were excluded (Fig. 8). Differentially expressed genes (DEGs) between responders and non-responders were identified using DESeq2 (version 1.38.3), and selection was performed based on a false discovery rate (FDR) q-value < 0.25 and |log2 fold change| > 0.5. Pathway analysis was performed using the clusterProfiler package in R (version 3.12.0) with reference to the Reactome and Kyoto Encyclopedia of Genes and Genomes (KEGG) databases. In addition, gene set enrichment analysis (GSEA) was performed to identify enriched pathways using the hallmark gene set of the Molecular Signatures Database (version 7.2).

[0186] First, we used volcano plots to search for differentially expressed genes (DEGs) between the five responders and eight non-responders. Based on an FDR q<0.25 and |Log2FC|>0.5 criteria, 511 DEGs were identified. Functional classification of the 511 DEGs was analyzed using the BP, CC, and MF databases. Genes overexpressed in the responders were primarily involved in the apical part functional ontology. Conversely, genes upregulated in the non-responders were primarily related to the mitochondrial matrix and flavin adenine dinucleotide binding. GSEA pathway analysis using the Hallmark database revealed that pathways such as the apical surface and apical junction were enriched in the responders. In contrast, non-responders were enriched in oxidative phosphorylation, hypoxia, DNA repair, cell cycle pathways, and glycolysis.

[0187]

[0188] Example 6. Weighted gene coexpression network analysis (WGCNA)

[0189] Weighted gene co-expression network analysis involved preprocessing raw count data in the base R environment using DESeq2 (v1.38.3), followed by WGCNA (v1.72-5) for network construction and module identification, limma (v3.54.2) for statistical analysis, ggplot2 (v3.5.0) for data visualization, and dplyr (v1.1.4) for data processing.

[0190] Next, we performed WGCNA analysis focusing on co-expressed gene modules to provide a comprehensive functional pathway associated with radiotherapy response. After constructing a scale-free network, hierarchical clustering revealed 22 distinct gene modules. Each module was examined for differential expression of member genes between responders and non-responders. The ME1 module, which was most strongly associated with responders, contained 4,625 genes, including GATA3, associated with extracellular structure organization, renal system development, the cytoskeleton, and focal adhesion. In contrast, the ME18 module, associated with non-responders and including SSBP1, contained genes involved in mismatch repair, DNA replication, and base excision repair. The DEG analysis results were merged with the WGCNA results to generate 43 merged modules. The profiles of upregulated genes and associated pathways in the merged modules were similar to those in the overall cohort analysis. Modules enriched in responders were associated with normal cellular structure and function, whereas modules enriched in non-responders were involved in DNA repair processes. Next, we reconstructed a transcriptional network by generating 1,300 regulons from a comprehensive list of regulators available in the tfsData dataset. Using FPKM-based RTN analysis, we identified 11 significant transcription factors (TFs) targeted by more than 15 genes, the two most prominent of which were PURG and TSHZ2. Among the 1,300 regulons, we found 27 genes targeted by PURG, which were associated with kidney development, the cytoskeleton, the PI3K-Akt signaling pathway, and focal adhesion. Furthermore, we found 18 genes targeted by TSHZ2, which were associated with cell-cell adhesion, Wnt signaling, and the actin cytoskeleton.

[0191] In cell type enrichment analysis for immune microenvironment profiling using the xCell score, there were no differences in most immune or stromal cells, including NK cells, macrophages, dendritic cells, monocytes, memory T cells, CD8+ T cells, Th1, Th2, Tregs, and B cells. However, CD4+ naïve T cells and eosinophils tended to be higher in responders. Composite indices, including ImmuneScore, MicroenvironmentScore, and StromaScore, were also similar between groups.

[0192]

[0193] Example 7. Predictive modeling using machine learning

[0194]

[0195] 7-1. Feature Selection

[0196] In this invention, the random forest-based Recursive Feature Elimination (RFE) function built into the caret package (v6.0-94) in R was used to select features that provide optimal predictive performance from training data. First, a set of 460 initial features was defined through differential equation (DEG) analysis. The number of optimal candidate features was set to be 1 to 99 consecutively, and then increased by 5 from 100 to 460, to search for the optimal number of variables for a total of 460 variables. The search process was performed using only the 11-person training set after partitioning the entire patient dataset into a training set of 11 individuals and a testing set of 2 individuals using the createDataPartition function built into the caret package. In the RFE process, cross-validation-based performance evaluation was applied. To this end, the repeated cross-validation method was used to reliably estimate the model's generalization performance. Cross-validation was set to repeat 10 times, and at each stage, a random forest model was used to calculate variable importance, sequentially removing variables with low importance. After training the model for all candidate variables and evaluating its predictive performance, the variable set with the highest performance was selected as the optimal feature set. In the present embodiment, 220 features were ultimately selected from a total of 460 initial variable sets, and this variable set was utilized in the subsequent predictive modeling stage.

[0197]

[0198] 7-2. Building and Validating Predictive Modeling

[0199] A machine learning model was constructed to predict SBRT treatment response using 220 selected features as input variables. In this study, the model was trained using the random forest algorithm, which can be implemented with the R randomForest package (v4.7-1.2), and a total of 500 trees were used to enhance prediction stability. No other parameters were changed. To prevent overfitting and evaluate the model's generalization performance, 5-fold cross-validation was performed. This means that the entire data was divided into five equally sized folds, and in each iteration, four folds were used for training and the remaining one fold for validation. This process was repeated five times to generate predicted values ​​for each sample, and the predicted values ​​and actual response data were integrated to produce a single performance index.

[0200] A machine learning model based on the expression of the 220 genes listed in Table 2 demonstrated excellent predictive performance, with an area under the curve (AUC) of 0.9. SBRT for T1a RCC was acceptable in terms of intermediate-stage tumor control and preservation of renal function (Figure 6). This is the first prospective clinical trial to evaluate the efficacy of hypofractionated radiotherapy as a curative treatment for T1a renal cell carcinoma in a large cohort with systematic follow-up of renal function. This novel machine learning-based model predicts genomic profile differences and outcomes and may be clinically useful for patient selection.

[0201]

[0202] Example 8. Statistical analysis

[0203] Progression-free survival (PFS) and cancer-specific survival (CSS) were estimated using the Kaplan-Meier method. Cox proportional hazards regression analysis was performed to identify factors predicting complete remission, and the area under the curve (AUC) of the receiver operating characteristic (ROC) curve was calculated to determine the optimal cutoff point. Time 0 was defined as the end of treatment, and patients without radiographic evidence of disease progression or death were censored at the time of the last radiographic assessment. Statistical analyses were performed using Python (v3.9.0). Statistical significance was set at p < 0.05, and all reported p values ​​are two-sided.

[0204]

[0205] [Table 5] Cox regression analysis results for factors predicting complete response after SBRT for T1a renal cell carcinoma.

[0206]

[0207]

[0208] While specific aspects of the present invention have been described in detail above, it should be apparent to those skilled in the art that these specific descriptions are merely preferred embodiments and do not limit the scope of the present invention. Therefore, the substantial scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A composition for diagnosing the prognosis of stereotactic body radiation therapy (SBRT) of renal cell carcinoma, comprising a preparation that specifically detects a combination of 30 genes or their transcripts listed in Table 1: [Table 1] 2. In the first paragraph, the combination of the genes or their transcripts is a combination of 220 genes or their transcripts described in Table 2, a composition for SBRT prognosis diagnosis: [Table 2] .

3. A composition for SBRT prognosis diagnosis in the first paragraph, wherein the renal cell carcinoma is clear cell type renal cell carcinoma.

4. In the first paragraph, the agent capable of specifically detecting a combination of the gene or its transcripts is at least one selected from the group consisting of antibodies, oligopeptides, ligands, PNA (peptide nucleic acid), and aptamers, primers, or probes that specifically bind to the gene or its transcripts, a composition for SBRT prognosis diagnosis.

5. A kit for SBRT prognosis diagnosis, comprising a composition of any one of claims 1 to 4.

6. In paragraph 5, the kit is an SBRT prognostic diagnostic kit that is an RT-PCR kit, a DNA chip kit, an ELISA kit, a protein chip kit, a rapid kit, or an MRM (Multiple reaction monitoring) kit. 7.(a) Step of detecting the 30 genes or their transcript combinations listed in Table 1 in a biological sample isolated from a patient with renal cell carcinoma: [Table 1] ; and (b) A method for providing information for SBRT prognosis diagnosis, comprising: a step of analyzing the output result value by inputting the detected degree into an artificial intelligence model trained to predict SBRT prognosis; 8. In the 7th paragraph, the gene or its transcript combination is a method for providing information for SBRT prognosis diagnosis, wherein the gene or its transcript combination is 220 genes or its transcript combinations described in Table 2: [Table 2] 9. A method for providing information for SBRT prognosis diagnosis, wherein the biological sample is a biopsy tissue isolated from a renal cancer patient in paragraph 7. 10.(a) A measuring unit for detecting the 30 genes or transcript combinations thereof listed in Table 1 in a biological sample isolated from a patient with renal cell carcinoma: [Table 1] ; and (b) An SBRT prognostic diagnostic device comprising a detection unit that outputs SBRT prognostic responsiveness from the expression level of a gene or its transcript measured in the above measurement unit.

11. In the 10th paragraph, the gene or its transcript combination is 220 genes or its transcript combinations described in Table 2, SBRT prognostic diagnostic device: [Table 2]

Images