Biomarker for predicting early mortality in cervical cancer patients

Abstract

The present invention relates to a biomarker for predicting early mortality in cervical cancer patients. As a result of analyzing changes in hematologic factors before and after treatment in cervical cancer patients who have undergone anticancer radiation treatment, it was confirmed that changes in compete blood count (CBC) representing absolute neutrophil count (ANC) level (cells / μL), platelet (PLT) level (cells / μL), hemoglobin (Hb) level (g / dL), or monocyte (Mo) count (cells / μL) were associated with early mortality, and plasma-derived exosome RNA associated with the changes in the CBC was found to be capable of predicting early mortality in a cervical cancer patient data cohort, thereby being provided as a biomarker associated with early mortality of cervical cancer patients.

Description

Biomarkers for Predicting Early Death in Cervical Cancer Patients

[0001] The present invention relates to a biomarker for predicting early death of cervical cancer patients.

[0002] Cervical cancer is a common malignant tumor in women, with more than 350,000 deaths worldwide each year. Cervical cancer is divided into stages 0 to 4, with stage 0 cancer being called intraepithelial carcinoma, and stages 1 to 4 being called invasive cervical cancer. Intraepithelial carcinoma develops about 10 years earlier than invasive cervical cancer, around the age of 35 to 40. Invasive cervical cancer begins to increase after age 30, peaks in the 50s, and then declines rapidly. In Korea, the prevalence of cervical cancer is approximately 31 per 100,000 women, and the mortality rate is 6.8 per 100,000 women, showing no significant change over the past decade. When looking at the frequency of incidence and prevalence by age, it is in the 50s, 60s, and 40s, in that order, and the incidence rate is three times that of the United States, 2.5 times that of Japan, and 1 / 3 that of Brazil. The mortality rate due to cervical cancer accounts for 6% of cancer mortality, which is higher than the 2% or so in the United States, Japan, and Switzerland. Currently, the sexually transmitted infection model is the most widely accepted for the occurrence of cervical cancer, and early onset of sexual activity, multiple sexual partners, male factors, human papillomavirus (HPV) infection, and human immunodeficiency virus infection are known to be risk factors for the occurrence of cervical cancer.

[0003] The recurrence rate after treatment for cervical cancer is approximately 50% within 1 year, 25% within 2 years, and 15% within 3 years, with approximately 85% of patients experiencing a recurrence within 3 years. The recurrence rate after radiation therapy for cervical cancer is also over 55%. Treatment for recurrent cervical cancer varies depending on the type of treatment previously received and the patient's condition. Generally, if surgery was the primary treatment, radiation therapy is considered in the event of a recurrence. If radiation therapy was the primary treatment, surgery is considered. If neither of these treatments is possible, chemotherapy is administered. However, there is no standard treatment method, and appropriate treatment is administered depending on the site of recurrence. Generally, the treatment method aims to alleviate symptoms rather than cure the disease.

[0004] While the treatment outcomes for patients undergoing chemoradiation therapy for cervical cancer can be predictable depending on the stage of the disease, sometimes the disease progresses much faster than anticipated, leading to poor outcomes. Such premature deaths have been reported in a variety of cancers, including breast, stomach, lung, head and neck, colon, and bladder cancers, and these patients are prone to resistance to the three major cancer treatments (surgery, chemotherapy, and radiation therapy). However, the precise cause of this is unknown. A possible explanation is that patients who are unable to maintain homeostasis in an environment of excessive inflammation and stress may experience accelerated cancer progression after treatment.

[0005] Meanwhile, exosomes are extracellular vesicles of 40-150 nm secreted by cancer cells or other normal cells. They perform various physiological functions through cell-to-cell signaling and are known to be involved in the regulation of cancer development, metastasis, immunity, inflammation, and stress responses. They contain specific non-coding / coding RNAs, and miRNAs, which regulate various coding RNAs, play a key role in regulating these functions. Therefore, exosome miRNAs can be considered as high-level regulatory factors that can be considered at this point in time. In particular, because blood is rich in exosomes originating from both cancer cells and normal cells, it is easy to understand the relationship between the physical responses of cancer patients and cancer cells, and at the same time, it is possible to examine the correlation between miRNAs and coding genes.

[0006] [Prior Art Literature]

[0007] [Patent Document]

[0008] Korean Patent No. 10-2170726 (announced on October 27, 2020)

[0009] Korean Patent No. 10-2280516 (announced on July 22, 2021)

[0010] Korean Patent No. 10-2534199 (announced on May 19, 2023)

[0011] The purpose of the present invention is to provide a biomarker associated with the early death of cervical cancer patients by analyzing hematological factors associated with early death within 15 months after the end of treatment from a data cohort of cervical cancer patients who received chemoradiation therapy, selecting plasma-derived exosomal RNAs that show significant changes before and after treatment, and analyzing the correlation between the hematological factors and exosomal RNAs to select and provide a biomarker associated with the early death of cervical cancer patients from exosomal RNAs.

[0012] The present invention provides a biomarker composition for predicting early death of a cervical cancer patient, comprising miR-574-3p (NCBI ID: NR_030300.1), a microRNA, and LINC01003 (NCBI ID: NR_027387.1), a long non-coding RNA (lncRNA).

[0013] In addition, the present invention provides a composition for predicting early death of a cervical cancer patient, comprising a preparation for measuring the expression level of miR-574-3p (NCBI ID: NR_030300.1), a microRNA, and LINC01003 (NCBI ID: NR_027387.1), a long non-coding RNA (lncRNA).

[0014] In addition, the present invention provides a kit for predicting early death of a cervical cancer patient, comprising a preparation for measuring the expression level of miR-574-3p (NCBI ID: NR_030300.1), a microRNA, and LINC01003 (NCBI ID: NR_027387.1), a long non-coding RNA (lncRNA).

[0015] In addition, the present invention provides a method for providing information for predicting early death of a cervical cancer patient, comprising the step of measuring the expression level of miR-574-3p (NCBI ID: NR_030300.1), a microRNA, and LINC01003 (NCBI ID: NR_027387.1), a long non-coding RNA (lncRNA), in a biological sample derived from a subject.

[0016] In addition, the present invention provides a method for selecting a biomarker for predicting early death in a patient with cervical cancer, comprising the steps of: collecting hematological factor data from a cohort of patients with cervical cancer (preparatory step); selecting hematological factors that show significant differences in patients with early death from the hematological factor data, and analyzing trends among the hematological factors (CBC analysis step); And a step of selecting exosomal RNA associated with the hematological factor and the patient who died early (RNA-CBC integrated analysis step); wherein the hematological factor is a CBC (compete blood count) measured in a blood sample, and the CBC includes an absolute neutrophil count (ANC) (cells / μL), a platelet (PLT) count (cells / μL), a hemoglobin (Hb) level (g / dL), an absolute lymphocyte count (ALC) (cells / μL), a monocyte (Mo) count (cells / μL), an NLR which is a value obtained by substituting the ANC and the ALC into ANC / ALC, a PLR which is a value obtained by substituting the PLT and the ALC into PLT / ALC, and a LMR which is a value obtained by substituting the ALC and the Mo into ALC / Mo. Provides a method for selecting biomarkers.

[0017] According to the present invention, as a result of analyzing changes in hematological factors before and after treatment in cervical cancer patients who received chemoradiation therapy, it was confirmed that changes in the CBC (compete blood count), which represents the absolute neutrophil count (ANC) level (cells / μL), the platelet (PLT) level (cells / μL), the hemoglobin (Hb) level (g / dL), or the monocyte (Mo) count (cells / μL), were associated with early death, and it was demonstrated that plasma-derived exosomal RNA associated with changes in the CBC (compete blood count) can predict early death in a cervical cancer patient data cohort, thereby being able to serve as a biomarker associated with early death in cervical cancer patients.

[0018] Figure 1 is a schematic diagram of the composition of the study cohort to evaluate whether quantitative changes in blood cells before and after anticancer radiotherapy are associated with early death.

[0019] Figure 2 shows the results of analyzing the degree of death and cancer progression in Cohort 1, which was created to evaluate whether quantitative changes in blood cells before and after chemoradiation therapy are associated with early death.

[0020] Figure 3 shows the results of analyzing the association between changes in hemoglobin (Hb) levels and early death (ED) in Cohort 1 and Cohort 2, which were designed to evaluate whether quantitative changes in blood cells before and after chemoradiation therapy are associated with early death.

[0021] Figure 4 shows the results of analyzing the correlation between hematological factors in Cohort 1 and Cohort 2, which were designed to evaluate whether quantitative changes in blood cells before and after chemoradiation therapy are associated with early death.

[0022] Figure 5 is a schematic diagram illustrating the treatment process, blood sampling time, MRI imaging time, and follow-up observation process for cervical cancer patients who underwent anticancer radiation therapy.

[0023] Figure 6 shows the results of a network analysis on the condition that the absolute value of the Pearson correlation coefficient between RNAs associated with early death and hematological factors in cervical cancer patients has a correlation of |R|>0.5 in the log2FC (log2Fold Change) values ​​of plasma exosomal RNAs showing quantitative changes before and after anticancer radiation therapy.

[0024] Figure 7 shows the results of a network analysis of the absolute value of the Pearson correlation coefficient between RNAs associated with early death and hematological factors in cervical cancer patients in plasma exosomal RNA showing quantitative changes before and after anticancer radiation therapy, with a correlation of |R|>0.6.

[0025] Figure 8 shows the adjusted R for a subset of RNAs in the network that have an association of |R|>0.5 (A) or |R|>0.6 (B) between RNAs associated with early death and hematological factors in plasma exosomal RNA showing quantitative changes before and after chemoradiation therapy. 2 This is the result of comparing values.

[0026] Figure 9 shows the results of analyzing the correlation between CBC3 (log(CBC2 / CBC0)) and |CBC0-CBC2| of RNAs associated with hematological factors among plasma exosomal RNAs showing quantitative changes before and after anticancer radiation therapy, by (A) relative ratio and (B) number.

[0027] Figure 10 shows the Pearson correlation coefficient (R) of plasma exosomal RNAs showing quantitative changes before and after chemoradiation treatment, with the CBC0 value indicating the CBC (compete blood count) value before treatment and the CBC1 ( indicating the overall CBC before and after treatment). ) values ​​and the Pearson correlation coefficient (R) of RNA, and the CBC3 (log (CBC2 / CBC0)) values ​​indicating the degree of blood cell recovery for two weeks after treatment and the Pearson correlation coefficient (R) of RNA were analyzed using a scatter plot.

[0028] Figure 11 shows the correlation coefficient between the pretreatment CBC (compete blood count) value of plasma exosomal RNAs showing quantitative changes before and after chemoradiation therapy associated with early death and the CBC0 value indicating the pretreatment CBC (compete blood count), and the CBC1 indicating the overall CBC before and after treatment. ) value and the correlation coefficient with the CBC3 (log(CBC2 / CBC0)) value, which indicates the degree of blood cell recovery for two weeks before and after treatment, were analyzed using a scatter plot.

[0029] Figure 12 shows the quantitative changes in plasma exosomal RNAs before and after chemoradiation therapy associated with early death, CBC0 indicating the pre-treatment CBC (compete blood count) value, and CBC1 indicating the overall CBC before and after treatment. ) values, and the CBC3 (log(CBC2 / CBC0)) values, which indicate the degree of blood cell recovery for two weeks before and after treatment.

[0030] Figure 13 shows the results of analyzing the number of RNAs (RNAs in Figure 6 above) in a network with a correlation of |R|>0.5 between RNAs associated with early death and hematological factors in plasma exosomal RNA showing quantitative changes before and after anticancer radiation therapy, by type of CBC (compete blood count), using a Venn diagram.

[0031] Figure 14 shows the results of verifying the possibility of predicting early death using a biomarker for predicting early death in cervical cancer patients in a research subject cohort diagnosed with cervical cancer and receiving chemoradiation therapy.

[0032] Figure 15 shows a Venn diagram of the number of mRNAs associated with biomarkers for predicting early death in cervical cancer patients, and the results of gene enrichment pathway analysis for the associated mRNAs.

[0033] Figure 16 shows the results of analyzing the biological ontology of RNAs (RNAs in Figure 6 above) in a network with a correlation of |R|>0.5 between RNAs associated with early death and hematological factors in plasma exosomal RNAs that show quantitative changes before and after anticancer radiation therapy.

[0034] The terms used in this specification have been selected from widely used, current terms, taking into account the functions of the present invention. However, these terms may vary depending on the intentions of those skilled in the art, precedents, the emergence of new technologies, etc. Furthermore, in certain cases, terms may be arbitrarily selected by the applicant, and in such cases, their meanings will be described in detail in the relevant description of the invention. Therefore, the terms used in this invention should not be defined simply as names, but rather based on their inherent meanings and the overall content of the present invention.

[0035] Unless otherwise defined, all terms used herein, including technical or scientific terms, have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains. Terms defined in commonly used dictionaries should be interpreted as having a meaning consistent with their meaning in the context of the relevant technology, and shall not be interpreted in an idealized or overly formal sense unless explicitly defined herein.

[0036] Hereinafter, the present invention will be described in more detail.

[0037] A method for selecting a biomarker for predicting early death in a cervical cancer patient according to the present invention comprises the steps of collecting hematological factor data from a cohort of cervical cancer patients (preparation step); selecting hematological factors showing significant differences in patients who died early from the hematological factor data and analyzing trends among the hematological factors (CBC analysis step); and selecting exosomal RNA correlated with the hematological factors and patients who died early (RNA-CBC integrated analysis step).

[0038] The above cohort includes two types of cohorts, cohort 1 and cohort 2, which are composed of patients with cervical cancer.

[0039] The above cohort 1 was composed of data from 402 patients with cervical cancer staged from Stage 2 to Stage 4A according to the criteria presented by the International Federation of Gynecology and Obstetrics in 2018, but did not include exosome RNA data. The above cohort 1 was divided into Group 1, an early death group consisting of patients who died within 15 months after the end of treatment due to a recurrence of cancer; Group 2, a progression group consisting of patients who died 30 months or more after the end of treatment due to a recurrence of cancer; and Group 3, a non-progression group consisting of patients whose cancer did not recur 60 months after the end of treatment.

[0040] The above cohort 2 was composed of data from 32 patients including exosome RNA data among 402 patients with cervical cancer staged from Stage 2 to Stage 4A according to the criteria presented by the International Federation of Gynecology and Obstetrics in 2018, and data from 10 patients including exosome RNA data among patients with cervical cancer staged from Stage 1B or Stage 4B. The above cohort 2 was divided into an early death group consisting of patients who died due to cancer progression within 12 months after the end of treatment, and a non-early death group.

[0041] In Cohort 2, early death was clinically determined based on the treatment progress among patients whose disease progressed within 12 months from the end of chemoradiation therapy, as described in Table 5.

[0042] In the step (preparatory step) of collecting hematological factor data from the cohort of cervical cancer patients, the hematological factor is a CBC (compete blood count) value measured from a blood sample separated from the patient, and the CBC includes an absolute neutrophil count (ANC) (cells / μL), a platelet (PLT) count (cells / μL), a hemoglobin (Hb) level (g / dL), an absolute lymphocyte count (ALC) (cells / μL), and a monocyte (Mo) count (cells / μL), and includes an NLR, which is an ANC / ALC value, a PLR, which is a PLT / ALC value, and a LMR, which is an ALC / Mo value, for comparing each blood cell. That is, the CBC components are ANC, PLT, Hb, ALC, Mo, NLR, PLR, and LMR.

[0043] Specifically, the hematological factor data includes CBC0, which is a CBC (pretreatment CBC) value measured in a blood sample isolated from a patient before chemoradiation treatment; min CBC, which is a minimum CBC (compete blood count) value measured in a blood sample isolated from a patient during chemoradiation treatment; and the CBC0 value and the min CBC value. CBC1, which is a value obtained by substituting the value into log(CBC2 / CBC0); CBC2, which is a CBC (compete blood count) value measured from a blood sample isolated from a patient two weeks after starting chemoradiation therapy; and CBC3, which is a value obtained by substituting the CBC0 value and the CBC2 value into log(CBC2 / CBC0).

[0044] The above CBC0 includes ANC0, PLT0, Hb0, ALC0, and Mo0, which refer to the absolute neutrophil count (ANC) (cells / μL), platelet (PLT) count (cells / μL), hemoglobin (Hb) level (g / dL), absolute lymphocyte count (ALC) (cells / μL), and monocyte (Mo) count (cells / μL) measured in a blood sample isolated from a patient before chemoradiation treatment, and includes NLR0 (ANC0 / ALC0), PLT0 (PLT0 / ALC0), and LMR0 (ALC0 / Mo0).

[0045] The above min CBC includes min ANC, min PLT, min Hb, min ALC, and min Mo, which refer to the absolute neutrophil count (ANC) (cells / μL), platelet (PLT) count (cells / μL), hemoglobin (Hb) level (g / dL), absolute lymphocyte count (ALC) (cells / μL), and monocyte (Mo) count (cells / μL) measured in blood samples isolated from patients during chemoradiation therapy.

[0046] The above CBC1 is The value is ANC1, The value is PLT1, The value of Hb1, The value is ALC1, It includes the value Mo1, the value NLR1 of ANC1 / ALC1, the value PLR1 of PLT1 / ALC1, and the value LMR1 of ALC1 / Mo1.

[0047] The above CBC2 includes ANC2, PLT2, Hb2, ALC2, and Mo2, which refer to the absolute neutrophil count (ANC) (cells / μL), platelet (PLT) count (cells / μL), hemoglobin (Hb) level (g / dL), absolute lymphocyte count (ALC) (cells / μL), and monocyte (Mo) count (cells / μL) measured in a blood sample isolated from a patient 2 weeks after the start of chemoradiation therapy.

[0048] The above CBC3 includes ANC3 which is a log(ANC2 / ANC0) value, PLT3 which is a log(PLT2 / PLT0) value, Hb3 which is a log(Hb2 / Hb0) value, ALC3 which is a log(ALC2 / ALC0) value, or Mo3 which is a log(Mo2 / Mo0) value.

[0049] The step of selecting hematological factors showing significant differences in patients with early death from the above hematological factor data and analyzing the tendency between hematological factors (CBC analysis step) is to compare the hematological factor data, CBC0, CBC1, CBC3, and min CBC, between patients with early death and non-early death among cervical cancer patients, to select CBC components showing significant differences, and to analyze the tendency of the CBC components between patients with early death and non-early death.

[0050] In one embodiment of the present invention, the CBC analysis step is performed in each of the cohort 1 and the cohort 2, and specifically, in the cohort 1, propensity score matching is performed between the early death group Group 1, the progression group Group 2, and the non-progression group Group 3 to balance the groups, and the hematological factor data CBC0, CBC1, CBC3, and min CBC are compared to select CBC components showing significant differences (Table 1, Table 2, Table 3, and Figure 2) (cohort 1 analysis); A step of performing propensity score matching between the early death group and the non-early death group in Cohort 2 to balance the groups and comparing the hematological factor data, CBC0, CBC1, CBC3, and min CBC, to select CBC components showing significant differences (Table 4, Table 5, Table 6, and Fig. 3) (Cohort 2 analysis); and a step of analyzing the correlation between the hematological factor data, CBC0 (excluding NLR0, PLR0, and LMR0), CBC1 (excluding NLR1, PLR1, and LMR1), CBC3, and min CBC, using a correlation coefficient matrix (Pearson's correlation matrix) in Cohort 1 and Cohort 2 (comparison of Cohort 1 and Cohort 2).

[0051] In one embodiment of the present invention, in the analysis of Cohort 1, Group 1, which is the early death group, showed a tendency for ANC0, min ANC, ANC1, and PLR0 to increase, and Group 3, which is the non-progression group, showed a tendency for Hb0 and ALC0 to increase. In addition, in the analysis of Cohort 2, ANC showed a tendency for increase in the early death group, and Hb3 showed a tendency for decrease. In the comparison of Cohort 1 and Cohort 2, CBC1 showed a positive correlation with both CBC0 and min CBC, CBC3 showed a negative correlation with CBC0, and it was shown that only ANC, Hb, and Mo showed a negative correlation between CBC3 and CBC1.

[0052] That is, in the CBC analysis step, the hematological factor most highly related to early death of cervical cancer patients was selected as ANC, and when the correlation between hematological factor data was analyzed, it was confirmed that if the hematological factors were at a high level before treatment, the degree of decrease tended to be high as the treatment progressed, and if the hematological factors were at a low level before treatment, the degree of decrease tended to be low, indicating that the hematological factors tend to maintain internal homeostasis.

[0053] The step of selecting exosomal RNA associated with the above hematological factors and patients with early death (RNA-CBC integrated analysis step) includes the steps of calculating a log2Fold Change (log2FC) value of the exosomal RNA; calculating an absolute value |R| of a Pearson correlation coefficient for the log2FC value and early death (ED) to select candidate RNAs having significant log2FC values; and selecting RNAs associated with hematological factor data and early death from cervical cancer from the candidate RNAs.

[0054] The above exosomal RNA includes miRNA (microRNA), piRNA (piwi-interacting RNA), snoRNA (small nucleaolar RNA), snRNA (small nuclear RNA), tRNA (transfer RNA), yRNA (Y RNA), lncRNA (long non-coding RNA), and mRNA (messenger RNA). The amount of change in RNA expression is specifically calculated as the log2Fold Change (log2FC) value by comparing the read count of the exosomal RNA before chemoradiation treatment to the read count of the exosomal RNA after treatment using NGS (Next Generation Sequencing) analysis of exosomal RNA extracted from a blood sample isolated from a patient.

[0055] In one embodiment of the present invention, the RNA-CBC integrated analysis step was performed on cohort 2 containing exosome RNA data.

[0056] The step of calculating the log2Fold Change (log2FC) value of the above exosomal RNA was performed by isolating exosomal RNA from blood samples isolated from patients before and after chemoradiation treatment, and calculating the change in the read count value of the exosomal RNA after treatment compared to the read count value of the exosomal RNA before treatment using NGS (Next Generation Sequencing) analysis, thereby calculating the log2Fold Change (log2FC) value.

[0057] The step of selecting candidate RNAs by calculating the absolute value |R| of the Pearson correlation coefficient for the above log2FC value and early death (ED) includes a step of selecting a first candidate RNA and a step of selecting a second candidate RNA, wherein the candidate RNAs include a first candidate RNA and a second candidate RNA.

[0058] The absolute value |R| of the above Pearson correlation coefficient was calculated using Pearson's correlation matrix.

[0059] The step of selecting the above first candidate RNA is a step of selecting RNAs whose log2FC values ​​are correlated with three or more CBC components, while the absolute value of the Pearson correlation coefficient for the log2FC value and early death (ED) satisfies |R|>0. The CBC components are ANC, PLT, Hb, ALC, Mo, NLR, PLR, and LMR, and the RNA correlated with the CBC components refers to RNA associated with ANC, PLT, Hb, ALC, Mo, NLR, PLR, or LMR. More specifically, the RNA associated with the ANC is an RNA associated with at least one of ANC0, min ANC or ANC1, the RNA associated with PLT is an RNA associated with at least one of PLT0, min PLT or PLT1, the RNA associated with Hb is an RNA associated with at least one of Hb0, min Hb or Hb1, the RNA associated with ALC is an RNA associated with at least one of ALC0, min ALC or ALC1, the RNA associated with Mo is an RNA associated with at least one of Mo0, min Mo or Mo1, the RNA associated with NLR is an RNA associated with at least one of NLR0 or NLR1, the RNA associated with PLR is an RNA associated with at least one of PLR0 or PLR1, and the RNA associated with LMR is an RNA associated with at least one of LMR0 or LMR1.

[0060] The step of selecting the second candidate RNA is a step of selecting RNA whose absolute value of the Pearson correlation coefficient for the log2FC value and early death (ED) satisfies |R|>0.5.

[0061] The steps of selecting RNAs associated with hematological factor data and early death from cervical cancer from the above candidate RNAs include network analysis, subset analysis, association analysis with hematological factors, and RNA distribution analysis.

[0062] Specifically, a network showing a correlation of |R|>0.5 or |R|>0.6 or higher is constructed for the candidate RNAs, and multiple regression analysis is performed on a statistically possible subset to select the network-centered RNA. In addition, among the candidate RNAs, RNAs associated with the condition that CBC0 (excluding NLR0, PLR0, and LMR0) and CBC3 have a negative correlation or CBC1 (excluding NLR1, PLR1, and LMR1) and CBC3 have a negative correlation are selected as conditional RNAs associated with premature death among RNAs satisfying the above conditions using a correlation matrix. RNAs corresponding to the network-centered RNA and simultaneously corresponding to the conditional RNA are selected.

[0063] In one embodiment of the present invention, a network showing a correlation of |R|>0.5 or |R|>0.6 or higher was constructed for the candidate RNAs, and seven RNAs, URS00001034C4 (piRNA), LINC01003 (lncRNA), ,STSRBP, LMBRD2, miR-574-3p, MOGS, and NIT1, were statistically selected (Fig. 8). Among the seven RNAs, LINC01003 (Hb0-Hb3) and miR-574-3p (ANC1-ANC3) were finally selected as ncRNAs associated with the condition in which CBC0 (excluding NLR0, PLR0, and LMR0) and CBC3 have a negative correlation or the condition in which CBC1 (excluding NLR1, PLR1, and LMR1) and CBC3 have a negative correlation.

[0064] Additionally, the types of CBC components associated with each RNA on the above-described network were analyzed, and RNAs associated with ANC, PLT, Hb, ALC, and Mo were selected. Specifically, in one embodiment of the present invention, the mRNA of the ACOT9 gene was selected as an RNA associated with all CBC components. In one embodiment of the present invention, the prediction of premature death using the selected RNA was evaluated (Fig. 14), and gene enrichment analysis was performed using the mRNAs associated with the log2FC values ​​of the selected RNAs and the RNAs on the network (Figs. 15 and 16).

[0065] Finally, the present invention has proven that miR-574-3p (NCBI ID: NR_030300.1), which is an RNA (microRNA), and LINC01003 (NCBI ID: NR_027387.1), which is an lncRNA (long non-coding RNA), are related to the early death of cervical cancer patients, and provides a method for predicting the early death of cervical cancer through miR-574-3p (NCBI ID: NR_030300.1), which is an RNA (microRNA), and LINC01003 (NCBI ID: NR_027387.1), which is an lncRNA (long non-coding RNA).

[0066] The present invention provides a biomarker composition for predicting early death of a cervical cancer patient, comprising miR-574-3p (NCBI ID: NR_030300.1), a microRNA, and LINC01003 (NCBI ID: NR_027387.1), a long non-coding RNA (lncRNA).

[0067] The above miR-574-3p is composed of a base sequence represented by sequence number 1, and the above LINC01003 is composed of a base sequence represented by sequence number 2.

[0068] The above biomarker composition may additionally include mRNA of the ACOT9 gene (NCBI ID: NG_016780.2). The mRNA of the ACOT9 gene is composed of a base sequence represented by SEQ ID NO: 4 (NCBI ID: NM_001037171.2), a base sequence represented by SEQ ID NO: 5 (NCBI ID: NM_001033583.3), or a base sequence represented by SEQ ID NO: 6 (NCBI ID: NM_001330259.2).

[0069] In addition, the present invention provides a composition for predicting early death of a cervical cancer patient, comprising a preparation for measuring the expression level of miR-574-3p (NCBI ID: NR_030300.1), a microRNA, and LINC01003 (NCBI ID: NR_027387.1), a long non-coding RNA (lncRNA).

[0070] The agent for measuring the expression level may be a sense and antisense primer, or a probe, that complementarily binds to the microRNA.

[0071] The above primer refers to a short genetic sequence that serves as the starting point for DNA synthesis, and is an oligonucleotide synthesized for the purpose of diagnosis, DNA sequencing, etc. The above primers are typically synthesized and used with a length of 15 to 30 base pairs, but this may vary depending on the intended use, and can be modified by methylation, capping, etc. using known methods.

[0072] The above probe refers to a nucleic acid capable of specifically binding to mRNA, ranging from a few bases to several hundred bases in length, produced through enzymatic, chemical, or synthetic processes. The presence or absence of mRNA can be confirmed by labeling with radioactive isotopes or enzymes, and can be designed and modified using known methods.

[0073] In addition, the present invention provides a kit for predicting early death of a cervical cancer patient, comprising a preparation for measuring the expression level of miR-574-3p (NCBI ID: NR_030300.1), a microRNA, and LINC01003 (NCBI ID: NR_027387.1), a long non-coding RNA (lncRNA).

[0074] The kit comprises one or more different component compositions, solutions, or devices suitable for the analysis method. For example, the kit may be a kit comprising genomic DNA derived from a sample to be analyzed, a primer set specific for the marker gene of the present invention, an appropriate amount of DNA polymerase, a dNTP mixture, a PCR buffer, and water for performing PCR. The PCR buffer may contain KCl, Tris-HCl, and MgCl2. In addition, components necessary for performing electrophoresis to confirm whether a PCR product has been amplified may be additionally included in the kit.

[0075] Additionally, the kit may be a kit containing the essential components required for performing RT-PCR. In addition to each primer pair specific for a marker gene, the RT-PCR kit may include a test tube or other appropriate container, reaction buffer, deoxynucleotides (dNTPs), enzymes such as Taq polymerase and reverse transcriptase, DNase and RNase inhibitors, DEPC water, sterile water, etc. It may also contain a primer pair specific for a gene used as a quantitative control.

[0076] In addition, the present invention provides a method for providing information for predicting early death of a cervical cancer patient, comprising the step of measuring the expression level of miR-574-3p (NCBI ID: NR_030300.1), a microRNA, and LINC01003 (NCBI ID: NR_027387.1), a long non-coding RNA (lncRNA), in a biological sample derived from a subject.

[0077] The above biological sample is an exosome derived from blood or plasma.

[0078] The above expression level can be measured by one or more methods selected from the group consisting of next generation sequencing (NGS), polymerase chain reaction (PCR), reverse transcription polymerase chain reaction (RT-PCR), real-time polymerase chain reaction (Real-time PCR), RNase protection assay (RPA), microarray, and northern blotting.

[0079] The above-mentioned providing method further comprises a step of measuring the expression level of mRNA of ACOT9 gene (NCBI ID: NG_016780.2) in a biological sample derived from a subject.

[0080] The subject may be a patient receiving radiation therapy to treat cervical cancer.

[0081] Specifically, the expression level may be the expression amount of the RNA or the change in the expression amount, and the change in the expression amount may be the relative change in the expression amount 1 to 3 weeks, preferably 2 weeks, after the start of the anticancer radiation treatment compared to the RNA expression amount before the anticancer radiation treatment as the reference value, and the relative change in the expression amount may be calculated as log2FC (fold change). More preferably, the change in the RNA expression amount 2 weeks after the start of the anticancer radiation treatment compared to the RNA expression amount before the anticancer radiation treatment may be calculated as a log2FC (fold change) value that takes the log value with 2 as the base.

[0082] The method for providing information for predicting early death of the above cervical cancer patient is as follows: If the value of miR-574-3p-LINC01003-ACOT9 obtained by subtracting the log2FC (fold change) value of LINC01003 and the log2FC (fold change) value of ACOT9 mRNA from the log2FC (fold change) value of miR-574-3p meets the condition of >3, it can be determined that the cervical cancer patient is likely to die early.

[0083] Hereinafter, to aid understanding of the present invention, examples will be given in detail. However, the following examples are intended only to illustrate the scope of the present invention and are not intended to limit its scope. These examples are provided to more fully explain the present invention to those of average skill in the art.

[0084]

[0085] [Preparation]

[0086] 1. Patient

[0087] This study was conducted under the approval of the Institutional Review Board of Ajou University Hospital (Approval Number: DB-2023-424). From April 2001 to April 2023, a total of 460 patients diagnosed with cervical cancer stages I to IVB by the 2018 International Federation of Gynecology and Obstetrics were enrolled. The subjects received weekly cisplatin-based concurrent chemoradiotherapy (CCRT) at the Department of Radiation Oncology, Ajou University Hospital. Among the subjects, 42 patients have next-generation sequencing data on plasma exosomal RNA from blood samples collected before RT (radiation therapy) and two weeks after RT. For the matching analysis, 370 patients were included as subjects, excluding 42 patients with RNA data and 48 patients diagnosed with stage 1B or 4B and with a total dose (TD) of less than 50 Gy, as shown in Figure 1.

[0088] All patients had pathologic findings confirmed by cervical cancer biopsy. Local staging, regional lymph nodes (LNs), and distant metastases were assessed using magnetic resonance imaging (MRI), computed tomography (CT), or positron emission tomography-CT. Suspected bladder or rectal metastases were confirmed by cystoscopy or sigmoidoscopy. External beam radiotherapy (EBRT) was administered to the pelvis using 10-15 MV photons, regardless of the presence of para-aortic lymph nodes. The pelvic RT dose was 45 Gy delivered in 25 fractions, and nodal lesions received up to 55-60 Gy. In six patients with stage IVB cervical cancer, the RT dose to the metastatic site was 30–55 Gy in 10–25 fractions. One patient with a single lesion in the left lung was treated with stereotactic body RT using 48 Gy in four fractions. 352 patients, excluding 10 who were replaced with EBRT, received high-dose intracavitary brachytherapy. All patients received weekly cisplatin (30–70 mg / m) during RT. 2 ) were administered for 4 to 6 cycles. Patients were followed up every 1 to 3 months after completion of treatment.

[0089] 2. Clinical and hematological factor analysis

[0090] Information on age at diagnosis, pathologic status, 2018 FIGO stage, total dose (TD), overall treatment time (OT), and extent of RT were collected. In addition, information on tumor markers such as pretretment Sqcc antigen level (SCC0), Sqcc antigen during CCRT (SCC2), pretreatment carcinoembryonic antigen 21-1 level (Cyfra0), and carcinoembryonic antigen 21-1 during CCRT (Cyfra2) were also collected. TD was calculated as the sum of EBRT and brachytherapy doses to the central lesion (equivalent dose in 2 Gy fractions). Follow-up time was measured as the period between the end of concurrent chemoradiation (CCRT) and the occurrence of an event (progression or disease-specific death).

[0091] The compete blood count (CBC) value includes the absolute neutrophil count (ANC) (cells / μL), platelet (PLT) count (cells / μL), hemoglobin (Hb) level (g / dL), absolute lymphocyte count (ALC) (cells / μL), and monocyte (Mo) count (cells / μL), and includes the NLR (ANC / ALC value), PLR (PLT / ALC value), and LMR (ALC / Mo value) values ​​for comparing each blood cell. CBC0 is the value measured in a CBC (pretreatment CBC) from a blood sample isolated from a patient before chemoradiation treatment, and includes the values ​​of ANC0, PLT0, Hb0, ALC0, Mo0, NLR0, PLR0, or NLR0. min CBC is the minimum CBC during concurrent chemo-radiotherpy (CCRT) measured in a blood sample from a patient and includes min ANC, min PLT, min Hb, min ALC, or min Mo. CBC1 is the CBC0 value indicating the overall CBC before and after treatment, and the min CBC value is This is the value obtained by substituting it into The value is ANC1, The value is PLT1, The value of Hb1, The value is ALC1, It includes Mo1 which is a value, NLR1 which is an ANC1 / ALC1 value, PLR1 which is a PLT1 / ALC1 value, and LMR1 which is an ALC1 / Mo1 value. CBC2 is a CBC value measured in a blood sample isolated from a patient 2 weeks after the start of chemoradiation treatment and includes ANC2, PLT2, Hb2, ALC2, and Mo2. CBC3 is a value indicating the degree of blood cell recovery before and after treatment and is a value obtained by substituting the CBC0 value and the CBC2 value into log(CBC2 / CBC0), and includes ANC3 which is a value of log(ANC2 / ANC0), PLT3 which is a value of log(PLT2 / PLT0), Hb3 which is a value of log(Hb2 / Hb0), ALC3 which is a value of log(ALC2 / ALC0), or Mo3 which is a value of log(Mo2 / Mo0).

[0092] 3. Comparison of CBC intervals by time of death by disease

[0093] Among the 370 subjects, patients with advanced cervical cancer or disease-specific death (DSD) were selected for comparison. The timing of DSD, the frequency of DSD occurrence, the frequency of disease progression, and the timing of DSD between patients who experienced recurrence within 1 year and those who did not were compared. Based on these comparisons, patients were classified into the early death (ED) group (group 1), progression group (group 2), or non-progression group (group 3). One-to-one propensity score matching (PSM) was performed between groups 1 and 2, 2 and 3, or 1 and 3 based on pathology, stage, and RT extent. The differences in CBC values ​​between each group were then compared. Patients included in this comparison were classified as cohort 1. Patients in Cohort 2 were propensity score matched (PSM) based on age, pathology, and stage between patients who died early (ED) and those who did not. The differences in CBC values ​​between these two groups were then compared. Categorical data were analyzed using the Chi-Square test, while continuous data were first tested for normality, followed by the t-test or Wilcoxon test.

[0094] 4. Calculating log2FC (fold change) values ​​for plasma exosomal RNA

[0095] The log2FC (log2-fold change) values ​​of plasma exosomal RNA were investigated from next-generation sequencing (NGS) data obtained from plasma exosomes, which contain information on non-coding RNA (ncRNA) and mRNA. Among 84 samples, RNAs that were not detected in more than 50% of the samples were excluded, and 508 miRNAs, 352 piRNAs, 629 snoRNAs, 2000 snRNAs, 2523 tRNAs, 700 yRNAs, 3151 lncRNAs, and 14,908 mRNAs were analyzed. The log2FC values ​​were collected from 5-10 cc of blood before treatment (control group) and during the second week of CCRT treatment (treatment group). Plasma was separated from the collected blood, and 3-5 cc of plasma samples from 29 individuals stored in a human material repository were additionally used in the experiment. RNA read counts were compared using edgeR with M-value normalization for 42 patients in cohort 2. Plasma exosomal RNA sequencing and profiling were performed at Macrogen and ROKIT Genomics (www.macrogen.com and www.rokitgenomics.com, respectively).

[0096] 4-1. RNA library construction and sequencing

[0097] Exosomes were isolated from plasma samples and subjected to next-generation sequencing (NGS) analysis. Samples with low expression levels were excluded from the analysis. Exosomes were isolated by mixing plasma with Exo2D RNA solution (Exosomeplus). Exosomes were extracted using an aqueous two-phase system (ATPS) tailored to the surface properties of exosomes. Plasma-derived exosomal RNA was extracted using the miRNeasy Serum / Plasma Kit (Qiagen, Valencia, CA) according to the manufacturer's instructions. The concentration of the extracted RNA was calculated by Quant-IT RiboGreen (Invitrogen). RNA size was confirmed using the Agilent RNA 6000 Pico Kit and Small RNA Kit on an Agilent 2100 Bioanalyzer (Agilent Technologies, Boblingen, Germany).

[0098] Ten nanograms of RNA isolated from each sample was used to construct sequencing libraries using the SMARTer smRNA-Seq kit for Illumina according to the manufacturer's protocol. Input RNA was first polyadenylated to provide priming sequences for the oligo(dT) primers. cDNA synthesis was primed by a 3'smRNA dT Primer, which incorporates an adapter sequence at the 5' end of each RNA template, followed by the addition of non-template nucleotides bound by SMRT smRNA Oligos enriched with locked nucleic acid (LNA) technology. In the template switching step, PrimeScript RT used the SMART smRNA Oligos as templates to add a second adapter sequence to the 3' end of each first-strand cDNA molecule. Subsequently, full-length Illumina adapters containing index sequences for sample multiplexing were added during PCR amplification. The forward PCR primer bound to the sequence added by the SMART smRNA Oligo, while the reverse PCR primer bound to the sequence added by the 3'smRNA dT primer. The amplified library was purified on a 6% Novex TBE-PAGE gel (Thermo Fisher, MA), and a fraction longer than 138 bp (≥18 bp of cDNA + ≥120 bp of adapter) was excised. The resulting library cDNA molecules contained the sequences required for clustering on an Illumina flow cell. The library was gel-purified and verified for size, purity, and concentration using an Agilent Bioanalyzer.Libraries were quantified using qPCR according to the qPCR Quantification Protocol Guide (KAPA Library Quantification Kit for Illumina Sequence Platforms) and verified using TapeStation D1000 ScreenTape (Agilent Technologies, Waldbronn, Germany). Libraries were pooled in equimolar amounts and sequenced on an Illumina HiSeq 2500 (Illumina, San Diego, USA) instrument to generate 51-base reads. Image decomposition and quality value calculation were performed using modules in the Illumina pipeline.

[0099] 4-2. Adapter trimming

[0100] Raw sequence reads of small RNAs obtained from different experimental samples were preprocessed and analyzed with miRDeep2. Adapter trimming was performed using the cutadapt program to remove adapter sequences attached to miRNAs during the construction of the smRNA library. The first 3 nt of each read was trimmed to remove additional bases introduced during the SMART template-switching process. The adapter sequence and all 3' nt of the adapter were also removed. If a read matched the first 5 bp or more of the 3' adapter sequence, the sequence was considered to be the adapter sequence and then trimmed from the read. For analysis, only trimmed reads longer than 18 bp were considered reliable. The remaining adapter sequences were classified as non-adapter reads. In this analysis, trimmed and non-adapter reads were combined and considered as reads processed for downstream analysis.

[0101] 4-3. Clustering

[0102] To minimize sequence uniqueness and computational complexity, reads processed with adapter sequences were collected and clustered. These clusters contained reads with 100% concordance in sequence ID and read length, and a provisional cluster ID and number of retained reads were provided.

[0103] 4-4. Ribosomal RNA filtering

[0104] Most RNA is known to be rRNA. To remove the effect of abundant rRNA, reads were aligned and matched to the 45S pre-rRNA and mitochondrial rRNA of Homo sapiens.

[0105] 4-5. mRNA expression profiling

[0106] Homo sapience (GRCh38; release 109.20190607) was searched in NCBI using the reference gene. Since the generated reads included both small RNA and mRNA, mRNA expression profiling was performed using RSEM (v1.3.1) with the options (--estimate-rspd --seed-length 15 --strandedness forward).

[0107] 4-6. Identification of known miRNAs

[0108] Sequence alignment and detection of microRNAs were performed using the miRDeep2 software algorithm. Before performing sequence alignment, the Homo sapiens reference genome release hg19 was retrieved from the UCSC Genome Browser and indexed using Bowtie (1.1.2) to align sequence reads to the reference sequence. Precursor miRNAs obtained from miRBase v21 were aligned to the reference sequence. The miRDeep2 algorithm is based on a model of miRNA biogenesis, aligns reads to potential hairpin structures in a manner consistent with Dicer processing, and assigns a score indicating the likelihood that the hairpin is a true miRNA precursor.

[0109] 4-7. miRNA and other RNA category ratios

[0110] Clustered reads were sequentially aligned to the reference genome, miRBase v21, and the noncoding RNA database RNAcentral release 10.0 to identify known miRNAs and other types of RNA.

[0111] All data and visualizations were performed using R version 4.3.0 (www.r-project.org), and all comparative analyses assumed nonparametric models. Propensity score matching was performed using the matchit package. The analysis included NGS (Next Generation Sequencing) miRNA, piRNA, snoRNA, snRNA, tRNA, yRNA, lncRNA, and mRNA obtained through the above process. The change in read count value 2 weeks after radiation treatment was calculated as the log2FC (log2Fold Change) value, using the read count value before radiation treatment as the reference (control). When calculating the log2FC (log2Fold Change) value, TMM normalization was performed using edgeR.

[0112] 5. Correlation Analysis between CBC Changes and RNA

[0113] Correlations between CBC0 (excluding NLR0, PLR0, and LMR0), CBC1 (excluding NLR1 PLR1, and LMR1), CBC2, CBC3, and min CBC were analyzed using a correlation matrix for Cohort 1 and Cohort 2, respectively. After identifying RNAs significantly associated with CBC0, CBC1, CBC2, CBC3, or min CBC in Cohort 2, a Venn diagram was constructed based on ANC, PLT, Hb, ALC, and Mo. The number of RNAs associated with CBC0, CBC1, CBC2, CBC3, and min CBC was determined from the Venn diagram. In addition, the correlation between |CBC0-CBC2| or |CBC1-CBC2| and CBC3 was analyzed. Secondarily, RNAs associated with early death (ED) were extracted. Pearson correlation coefficient (R) matrices related to CBC, RNA, and early death (ED) were calculated using the recorr function in the Hmisc package.

[0114] Correlation analysis was conducted in three steps:

[0115] Step 1: Hmisc package - rcorr function - pearson's correlation

[0116] Step 2: Leaps package - regsubsets function - exhaustive algorithm (maximum variable set to 10)

[0117] Step 3: Mean comparison (Wincoxon rank sum test or Kruskal-Wallis test) and boxplots

[0118] 6. RNA screening for predicting early death (ED)

[0119] In the network, RNAs were defined as being associated with at least one of ANC0, ANC1, ANC2, ANC3, or min ANC. The same definition was also applied to PLT, Hb, ALC, and Mo. Venn diagrams for NC, PLT, Hb, ALC, and Mo were constructed based on RNAs associated with CBC. All types of CBC-associated RNAs were selected. The regsubsets function in the leap package in R was used to conduct an exhaustive search to identify major components among the RNAs in the network. Among the RNAs, ncRNAs associated with CBC0 and CBC3, or with CBC1 and CBC3, were selected.

[0120] 7. Survival Analysis

[0121] The Wilcoxon rank sum test was used to compare the sum and difference of log2FC values ​​of selected RNAs between patients with early death (ED) and those without ED, and the two groups were compared using Kaplan-Meier plots and log-rank tests based on these values.

[0122] 8. Biological functions of selected RNAs

[0123] Enrichment analysis and visualization of pathway and gene ontology annotations were performed in Enrichr (https: / maayanlab.cloud / Enrichr) using the log2FC values ​​of selected RNAs and significantly altered mRNAs based on the RNAs in the network.

[0124] Network and subset analyses were performed using the Igraph package - prim algorithm. In all networks, positive directions are indicated by red edges, and negative directions are indicated by blue edges. Functional analysis of RNA was performed using biological pathways or ontology in the enrichR database (https: / maayanlab.cloud / Enrichr / ). Venn diagrams were created using the ggvenndiagram package. Additionally, graphics, grouping, and calculations were performed using the tidyverse package. Survival analysis was performed using the survival and survminer packages. All tables were created using mytable from the 1.1 moonBook package.

[0125]

[0126] [Example 1] Analysis of hematological factors associated with premature death

[0127] To identify biomarkers associated with early mortality in cervical cancer patients who underwent chemoradiation, we examined quantitative changes in blood cells before and after chemoradiation in the study cohort and selected hematological factors associated with early mortality in cervical cancer patients. Cervical cancer staging was determined based on FIGO (The International Federation of Gynecology and Obstetrics) stages. The patients in this study were analyzed from previously collected data, and this study was conducted with the approval of the Institutional Review Board of Ajou University Hospital (Approval No. DB-2023-424). As shown in Figure 1, the study cohort was selected from the data of 460 patients diagnosed with Stage I to IVB cervical cancer and received primary concurrent chemoradiotherapy from April 2001 to April 2023. Patients received weekly cisplatin-based CCRT at the Department of Radiation Oncology, Ajou University Hospital.

[0128] 1-1. Cohort 1 Analysis

[0129] Cohort 1 was comprised of data from 402 patients with cervical cancer stages 2 to 4A, but did not include exosome RNA data. Cohort 1 was divided into three subgroups by analyzing the degree of death and cancer progression, as shown in Figure 2. Among the patients, 74 patients who died were divided into patients who died within 15 months of treatment and patients who died after 30 months regardless of the time of cancer progression by creating a scatter plot (Figure 2A) and a density plot (Figure 2B) of the death date versus the progression date. As a result, the patients were divided into patients who died within 15 months of treatment and patients who died after 30 months of treatment regardless of the time of cancer progression. As a result of creating a density plot (Figure 2C) of the progression date of 119 patients whose cancer recurred after treatment, it was confirmed that most patients had a recurrence within 12 months, and that there were almost no recurrences after 60 months. Based on the above results, a boxplot (Fig. 2D) was created to compare the death dates of patients whose cancer progressed within 12 months and those whose cancer progressed after 12 months based on the data of patients who died due to cancer progression. As a result, it was confirmed that most patients whose cancer progressed within 12 months died within 30 months, and patients whose cancer progressed after 12 months died after 30 months. Based on the above results, Cohort 1 was divided into the following three subgroups.

[0130] Group 1: Early death. Patients whose cancer recurred and died within 15 months after the end of treatment.

[0131] Group 2: Progression. Patients whose cancer recurred and died 30 months or more after the end of treatment.

[0132] Group 3: Non-progression. Patients whose cancer has not recurred 60 months after the end of treatment.

[0133] For each group, clinical information such as age, pathological characteristics, FIGO stage, RT field, treatment time, cancer progression type, and disease-specific death (DSD) is summarized in Table 1 below.

[0134] [Table 1]

[0135]

[0136] Pathological characteristics included adenocarcinoma, adenosquamous cell carcinoma, unknown carcinoma, and squamous cell carcinoma. FIGO stage was analyzed by the proportion (%) of patients belonging to Stage 2A to Stage 3C1 or Stage 3C2 to Stage 4A. RT field was analyzed by the proportion (%) of cases where only the pelvis was treated and cases where the pelvis and para-aortic region were treated together. Cancer progression type was analyzed by cases where cancer did not develop (No), cases with local progression (LP), and cases with distant metastasis (DM).

[0137] Based on the above analysis data, propensity score matching was performed on the hematological characteristics of 174 patients belonging to groups 1, 2, and 3. Comparisons were performed between groups 1 and 2, between groups 2 and 3, or between groups 1 and 3. Data from patients in each group were extracted equally based on pathological characteristics, FIGO stage, and RT field, and changes in hematological characteristics were compared.

[0138] As shown in Tables 2 and 3 below, Group 1, patients with early death, showed a tendency for ANC to increase compared to the other groups, and this tendency was consistent before treatment (CBC0), during treatment (min CBC), and overall (CBC1). Group 3, patients without recurrence, showed a tendency for Hb and ALC to increase compared to the other groups.

[0139] [Table 2]

[0140]

[0141] [Table 3]

[0142]

[0143] 1-2. Cohort 2 Analysis

[0144] As shown in Fig. 1, Cohort 2 was established by including data from 32 patients with exosome RNA data among 402 patients with cervical cancer stages 2 to 4A, and data from 10 patients with cervical cancer stages 1B or 4B, including exosome RNA data. Clinical data was analyzed on the patient data in the same manner as Cohort 1, and the clinical data is as shown in Table 4, and specific information on some patients is as shown in Table 5 below.

[0145] [Table 4]

[0146]

[0147] [Table 5]

[0148]

[0149] In Table 5, patients 6 and 8, whose deaths could not be confirmed due to loss of follow-up, were excluded from the data. As shown in Tables 4 and 5 above, among the 40 patient data, early deaths were confirmed within 15 months, and some patients were confirmed to have died around 30 months. Bevacizumab was used as a second-line treatment starting in 2015, but in 2018 and 2019, with the development of targeted therapies and the use of pembrolizumab, the survival period of patients showed a gradual trend of increasing. Therefore, among patients treated after 2018 and 2019, patients who died between 25 and 34 months were also included in the early death group. Based on the above analysis data, propensity score matching for hematological characteristics according to early death was performed in the same manner as in the cohort 1 analysis, based on age, pathological characteristics, and FIGO stage of 5 patients with early death and 10 patients who did not fall into the early death group. As shown in Table 6 below, ANC tended to increase in the early death group, and Hb3 tended to decrease.

[0150] [Table 6]

[0151]

[0152] 1-3. Comparative Analysis of Cohort 1 and Cohort 2

[0153] Based on the results of Cohort 1 and Cohort 2, hematological factors associated with early death in cervical cancer patients were identified. First, to determine whether there were differences in Hb3 between Cohort 1 and Cohort 2, the associations between Hb0 and Hb3 were compared and analyzed in Cohort 1 and Cohort 2.

[0154] As shown in Fig. 3, the correlation between Hb0 and Hb3 according to whether or not early death (ED) occurred was analyzed. In Cohort 1 (Fig. 3A), as the Hb0 value increased, Hb3, which indicates the degree of hemoglobin recovery before and after treatment, tended to decrease, and as Hb0 decreased, Hb3 tended to increase. In particular, in Cohort 2 (Fig. 3B), cases where Hb3 had relatively low values ​​were identified, which corresponded to early death.

[0155] When examining the correlations between hematological factors in Cohort 1 and Cohort 2, as shown in Figure 4, CBC1 showed a positive correlation (red) with both CBC0 and min CBC, and CBC3 showed a negative correlation (blue) with CBC0. In particular, in Cohort 2 (Figure 4B), ANC was the only factor showing a negative correlation between CBC3 and CBC1.

[0156] The above results generally indicate that if hematologic factors were at high pretreatment levels, the rate of decrease was higher as treatment progressed. If pretreatment levels were low, the rate of decrease was lower. This suggests that hematologic factors tend to maintain internal homeostasis. Considering the results of Cohort 1 and Cohort 2 comprehensively, a correlation was confirmed between early death and changes in ANC in cervical cancer.

[0157]

[0158] [Example 2] Detection of markers for predicting early death associated with hematological factors

[0159] Using the dataset of exosomal ncRNA (non-coding RNA) and mRNA from Cohort 2 of Example 1, we analyzed ncRNAs associated with early mortality and hematological factors in cervical cancer patients. As shown in Figure 5, the RNA data from Cohort 2 were collected from 28 cervical cancer patients who underwent chemoradiation therapy, including the treatment process, blood sampling time, MRI imaging time, and follow-up observation.

[0160] 2-1. Network construction

[0161] RNA sequencing of miRNA (microRNA), piRNA (piwi-interacting RNA), snoRNA (small nuclear RNA), snRNA (small nuclear RNA), tRNA (transfer RNA), yRNA (Y RNA), lncRNA (long non-coding RNA), and mRNA (messenger RNA) was performed on plasma exosomes of 42 patients before and 2 weeks after radiotherapy, and NGS data were obtained using the method described above. For each patient, the read count value of the plasma exosome before radiotherapy was used as a control, and the change in the read count value 2 weeks after radiotherapy was calculated using edgeR for TMM normalization and the log2FC (Fold Change) value.

[0162] A total of 24,772 RNAs were analyzed, and a network was constructed by extracting 55 RNAs with log2FC values ​​that met either of the following two conditions. Network analysis and visualization were performed using the igraph package.

[0163] Condition 1. RNAs (35) that are associated with at least three of the eight types of CBC below and have an absolute value of the Pearson correlation coefficient of |R|>0 for early death (ED).

[0164] 1) RNA associated with ANC: RNA associated with at least one of ANC0, min ANC, or ANC1.

[0165] 2) RNA associated with PLT: RNA associated with at least one of PLT0, min PLT, or PLT1.

[0166] 3) RNA associated with Hb: RNA associated with at least one of Hb0, min Hb, or Hb1.

[0167] 4) RNA associated with ALC: RNA associated with at least one of ALC0, min ALC, or ALC1.

[0168] 5) RNA associated with Mo: RNA associated with at least one of Mo0, min Mo, or Mo1.

[0169] 6) RNA associated with NLR: RNA associated with at least one of NLR0 or NLR1

[0170] 7) RNA associated with PLR: RNA associated with at least one of PLR0 or PLR1

[0171] 8) RNA associated with LMR: RNA associated with at least one of LMR0 or LMR1

[0172] Condition 2. RNAs (20) with a high correlation with early death (ED) with an absolute value of the Pearson correlation coefficient of |R|>0.5

[0173] As a result, 4 out of 508 miRNAs (microRNAs) were selected, 2 out of 353 piRNAs (piwi-interacting RNAs) were selected, 3 out of 629 snoRNAs (small nuclear RNAs) were selected, none out of 2000 snRNAs (small nuclear RNAs) were selected, 1 out of 2523 tRNAs (transfer RNAs) was selected, 1 out of 700 yRNAs (Y RNAs) was selected, 11 out of 3151 lncRNAs (long non-coding RNAs) were selected, and 33 out of 14908 mRNAs (messenger RNAs) were selected. Network analysis and subset analysis were performed using the 55 RNAs selected above. Network analysis extracted ncRNAs that were highly correlated with early death (ED) (|R|>0.5) or significantly associated with three or more of ANC, PLT, Hb, ALC, Mo, NLR, PLR, and LMR (|R|>0) while being associated with early death (ED).

[0174] Using the igraph package in R, a network showing a correlation of |R|>0.5 or |R|>0.6 or higher was constructed for the extracted RNAs. Subset analysis extracted major RNAs that were mostly included in subsets that could well predict early death (ED). Network and subset analyses were performed using the igraph package and the prim algorithm, and in all networks, red edges indicate positive directions, and blue edges indicate negative directions.

[0175] Among the 55 RNAs above, a network with a correlation of |R|>0.5 was constructed, and 47 RNAs formed the network, as shown in Fig. 6. In addition, among the 55 RNAs above, a network with a correlation of |R|>0.6 was constructed, and 20 RNAs formed the network, as shown in Fig. 7. All of the 20 RNAs were found to be included in the 47 RNAs of Fig. 6. In Fig. 6, the CBC to which they are associated is indicated above the RNA name.

[0176] When the adjusted R value was calculated for a subset of RNAs constituting a network composed of 47 RNAs with a correlation of |R|>0.5 and a subset analysis was performed to form a subgroup associated with early death (ED), as shown in Fig. 8A, the piRNA (piwi-interacting RNA) URS00001034C4, the lncRNA (long non-coding RNA) LINC01003, the mRNA (messenger RNA) STSRBP, and LMBRD2 were selected as major RNAs (indicated by ** at the top). When the adjusted R value was calculated for a subset of RNAs forming a network composed of 20 RNAs with a correlation of |R|R>0.6 and a subset analysis was performed to form a subgroup associated with early death (ED), as shown in Figure 8B, miR-574-3p, a microRNA (miRNA), and MOGS and NIT1, messenger RNAs (mRNAs), were selected as major RNAs (indicated by ** at the top).

[0177] 2-2. Analysis of RNA distribution and selection of marker RNAs associated with hematological factors

[0178] Among a total of 24,772 target RNAs, regardless of early death (ED), we analyzed the distribution of RNAs associated with hematologic factors CBC0, CBC1, CBC2, CBC3, and min CBC. RNAs associated with Hb were found to have the highest distribution, while RNAs associated with PLT, ANC, ALC, or Mo were found to have similar levels. Approximately 20-50% of the RNAs were found to be associated with two or more hematologic factors. The decrease in RNA associated with CBC3 is shown in Figure 9 to be associated with the absolute difference in the number of RNAs associated with CBC0 and CBC2 (|CBC0-CBC2|) or the absolute difference in the number of RNAs associated with CBC1 and CBC2 (|CBC1-CBC2|). Considering the results showing a negative correlation between CBC0 and CBC1 and CBC3 in Figure 4B and the fact that many RNAs are associated with changes in the number and distribution of RNAs before and after radiation treatment in ANC and Hb in Figure 9, it can be inferred that RNAs mediating the negative correlation between CBC0-CBC3 or CBC1-CBC3 exist. Specifically, by confirming the changes in blood cells before and after radiation therapy, it was found that a decrease in CBC due to chemoradiation in a state where CBC0 was increased led to a decrease in CBC recovery, whereas a decrease due to chemoradiation in a state where CBC0 was decreased further stimulated the recovery of CBC, affecting the homeostasis of CBC maintained within a certain range, and it is proven that the exosomal RNAs associated with the changes in blood cells are associated with the maintenance of blood cell homeostasis before and after treatment. Based on the above results, RNAs related to CBCO or CBC1 and CBC3, while having a negative correlation with CBCO or CBC1 and CBC3, were selected.

[0179] As shown in Fig. 10, a scatter plot of RNAs associated with both CBC0-CBC3 or CBC1-CBC3 (CBC0(1)-CBC3) and associated RNAs revealed that there were 2825 RNAs mediating this, and depending on the blood cell type, there were 2087 RNAs in Hb0-Hb3 and Hb1-Hb3, 262 RNAs in ANC0-ANC3 and ANC1-ANC3, 201 RNAs in Mo0-Mo3 and Mo1-Mo3, 145 RNAs in ALC0-ALC3, and 130 RNAs in PLTO-PLT3. The results demonstrate that exosomal RNAs involved in maintaining the appropriate amount of CBC in the blood are abundant, and among these RNAs, RNAs associated with early death (ED) were selected again.

[0180] The scatter plot between the CBC0(1)-RNA3 or CBC3-RNA correlation coefficient and the early death (ED)-RNA correlation coefficient is shown in Fig. 11. It consists of 81 RNAs associated with both CBC0(1) and CBC3 ED: 6 PLT, 10 ANC, 60 Hb, and 5 Mo. If the increase in RNA change amount and the occurrence of ED have a negative relationship, it is expressed as a blue square (type 1, type 3), and if it has a positive relationship, it is expressed as a red square (type 2, type 4). According to the results of Cohort 1, the occurrence of ED is associated with the increase in ANC and PLT, and is expressed as a solid line (type 1, type 2), while Hb and Mo, which are not, are expressed as dotted lines (type 3, type 4). In other words, four types of associations can be described depending on the relationship between ED and RNA and whether the associated CBC is ANC or PLT, as shown in Fig. 12. Through this, we identified RNAs associated with CBC homeostasis and early death.

[0181] In the network of Fig. 6, there were 6 RNAs (red squares) associated with both CBC0(1) and CBC3. Specifically, LINC01003 and SRBD1 were associated with Hb0-Hb3, miR-574-3p was associated with ANC1-ANC3, LOC105375521 was associated with PLT0-PLT3, and miR-1229-3p and LRG1 were associated with Mo0-Mo3. In addition, there were 10 RNAs (blue squares) associated only with CBC3, URS000064525E, ANKRD55, MOGS, CCDC113, RUBCNL, STRBP, ZZZ3, LINC02513, LMBRD2 were associated with Hb3, and ACOT9 was associated with Mo3. A Venn diagram was created for each hematological variable using the RNAs in the network shown in Figure 6. As a result of analyzing the distribution of RNAs associated with ANC, PLT, Hb, ALC, and Mo, as shown in Figure 13, RNAs associated with three or more hematological variables and premature death were marked in blue, RNAs showing a high correlation with premature death were marked in red, and statistically selected RNAs in Figure 8 were marked with **. In addition, ACOT9 was confirmed to be the only RNA associated with five hematological factors in the Venn diagram. This suggests that ACOT9 is an essential RNA for all CBCs. Among the RNAs extracted from the above results, three RNAs satisfying the following conditions, LINC01003, miR-574-3p, and ACOT9, were selected as biomarkers for predicting premature death in cervical cancer patients.

[0182] (1) CBC0(1)-CBC3 associated RNA (Fig. 11, Fig. 12)

[0183] (2) ncRNAs, which are among the major RNAs that are always included in the subset combined with other types of RNAs in the network (Figure 8)

[0184] RNA satisfying both conditions (1) and (2) (LINC01003 and miR-574-3p)

[0185] (3) RNA associated with all types of CBC (Figure 13) (ACOT9)

[0186] 2-3. Validation and Functional Analysis of Biomarker RNA

[0187] To verify the predictive ability of early death in cervical cancer using LINC01003 and miR-574-3p, which are biomarkers for predicting early death in the selected cervical cancer patients, and ACOT9, which is associated with five hematological factors, the sum and difference of the log2FC values ​​of LINC01003, miR-574-3p, and ACOT9 were used to test with data from 42 patients in Cohort 2. Specifically, the log2FC value for LINC01003 and the log2FC value for ACOT9 were subtracted from the log2FC value for miR-574-3p to calculate the log2(Fold Change) value of miR-574-3p-LINC01003-ACOT9. As a result of analyzing the difference in the log2 (Fold Change) value of miR-574-3p-LINC01003-ACOT9 according to survival and the increase in the log2FC value of miR-574-3p-LINC01003-ACOT9, the increase in miR-574-3p-LINC01003-ACOT9 (log2FC) was found to have a significant association with early death (ED), as shown in Figure 14. Specifically,

[0188] When the values ​​of miR-574-3p-LINC01003-ACOT9, which were obtained by subtracting the log2FC (fold change) value of LINC01003 and the log2FC (fold change) value of ACOT9 mRNA from the log2FC (fold change) value of miR-574-3p, were sorted in descending order, it was confirmed that they were sorted sequentially according to the time of early death and disease progression, as shown in Table 7 below, and when the value of miR-574-3p-LINC01003-ACOT9 met the condition of >3.2, it was shown that the cervical cancer patient was likely to die early.

[0189] [Table 7]

[0190] [Correction pursuant to Rule 91, 04.12.2024]

[0191] The above validation demonstrated that three biomarkers, miR-574-3p, LINC01003, and ACOT9, can predict early death in cervical cancer patients. To analyze the biological functions of miR-574-3p, LINC01003, and ACOT9, a Venn diagram was created regarding the number of associated RNAs, and information on related mRNAs was collected. As shown in Figure 15, the Venn diagram illustrating the number of RNAs associated with the log2FC of RNA showed that the associated RNAs were in the order of miR-574-3p, LINC01003, and ACOT9. A total of 219 RNAs were identified that were associated with all three RNAs, and these 219 RNAs were identified as being related to platelet activation and hemostasis. As a result of selecting the 100 most correlated mRNAs for each RNA, it was confirmed that there were no common RNAs among miR-574-3p, LINC01003, and ACOT9. When selecting the 100 most correlated mRNAs for each RNA, miR-574-3p was confirmed to be related to neutrophil innate immunity, LINC01003 was confirmed to be related to the innate immune system, and ACOT9 was confirmed to be related to heme production.

[0192] The above results suggest that LINC01003 is associated with Hb0-Hb3, and considering the functions of mRNAs associated with ACOT9, ncRNAs involved in innate immunity associated with red blood cells may serve as biomarkers for predicting early death in cervical cancer patients. In addition, as a result of analyzing the biological functions (biological ontology) of 47 RNAs associated with hematological variables and early death (ED) in Fig. 6, it was confirmed that they were related to the toll-like receptor 9 signaling pathway (toll-like), as shown in Fig. 16, suggesting that miR-574-3p, an exosomal RNA, and LINC01003 are associated with innate immunity.

[0193] Considering the above results comprehensively, in order to discover a biomarker for predicting the early death of cervical cancer patients, RNAs associated with changes in hematological factors related to innate immunity before and after treatment can be selected from exosomal RNA as biomarkers for predicting the early death of cervical cancer patients. Through the above process, miR-574-3p and LINC01003, which are plasma exosomal RNAs selected as biomarkers for predicting the early death of cervical cancer patients, were proven to be involved in maintaining the homeostasis of hematological factors ANC and Hb, and can predict the early death of cervical cancer patients together with the mRNA of ACOT9.

[0194] 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 merely represent preferred embodiments and are not intended to limit the scope of the present invention. In other words, the substantial scope of the present invention is defined by the appended claims and their equivalents.

[0195] Numerical ranges are inclusive of the values ​​defined in the ranges above. Any maximum numerical limitation given throughout this specification includes any lower numerical limitation, as if that lower numerical limitation were explicitly stated. Any minimum numerical limitation given throughout this specification includes any higher numerical limitation, as if that higher numerical limitation were explicitly stated. Any numerical limitation given throughout this specification will include any better numerical range within the broader numerical range, as if that narrower numerical limitation were explicitly stated.

[0196] [Explanation of symbols]

[0197] ED: Early death

[0198] CCRT: concurrent chemoradiotherapy

[0199] adenosqcc: adenosquamous cell carcinoma

[0200] Sqcc: squamous cell carcinoma

[0201] FIGO: The International Federation of Gynecology and Obstetrics

[0202] RT: Radiation therapy

[0203] PA: Para-aortic region

[0204] TD: total dose

[0205] EQD2: biologically equivalent dose in 2Gy fractions

[0206] LP: local progression DM: distant metastasis

[0207] fu: follow-up

[0208] DSD: disease-specific death

[0209] PSM: Propensity Score Matching

[0210] PD: disease progression

[0211] SCC0: pretretment Sqcc antigen

[0212] SCC2: Sqcc antigen during CCRT

[0213] Cyfra0: pretreatment carcinoembryonic antigen 21-1

[0214] Cyfra2: carcinoembryonic antigen 21-1 during CCRT

[0215] ANC: absolute neutrophil count

[0216] PLT: Platelet

[0217] Hb: hemoglobin

[0218] ALC: absolute lymphocyte count

[0219] Mo: monocyte

[0220] NLR :ANC / ALCPLR :PLT / ALCLMR :ALC / Mo

[0221] CBC: Compete blood count (ANC, PLT, Hb, ALC, or Mo)

[0222] CBC0: Pretreatment CBC (ANC0, PLT0, Hb0, ALC0, or Mo0)

[0223] min CBC: Minimum CBC during concurrent chemoradiotherpy (min ANC, min PLT, min Hb, min ALC, or min Mo)

[0224] CBC1: (ANC1, PLT1, Hb1, ALC1, or Mo1)

[0225] NLR0 :ANC0 / ALC0PLR0 :PLT0 / ALC0LMR0 :ALC0 / Mo0

[0226] NLR1 :minANC / minALCPLR1 :minPLT / minALCLMR1 :minALC / minMo

[0227] CBC2: CBC 2 weeks after CCRT (ANC2, PLT2, Hb2, ALC2, or Mo2)

[0228] CBC3: log(CBC2 / CBC0) (ANC3, PLT3, Hb3, ALC3 or Mo3)

Claims

1. A biomarker composition for predicting early death in cervical cancer patients, comprising miR-574-3p (NCBI ID: NR_030300.1), a microRNA, and LINC01003 (NCBI ID: NR_027387.1), a long non-coding RNA (lncRNA).

2. A biomarker composition for predicting early death of a cervical cancer patient, characterized in that in paragraph 1, the miR-574-3p is composed of a base sequence represented by sequence number 1, and the LINC01003 is composed of a base sequence represented by sequence number 2.

3. A biomarker composition for predicting early death of a cervical cancer patient, characterized in that in the first paragraph, the biomarker composition additionally comprises mRNA of the ACOT9 gene (NCBI ID: NG_016780.2).

4. A biomarker composition for predicting early death of a cervical cancer patient, characterized in that the mRNA of the ACOT9 gene in the third paragraph consists of a base sequence represented by sequence number 4 (NCBI ID: NM_001037171.2), a base sequence represented by sequence number 5 (NCBI ID: NM_001033583.3), or a base sequence represented by sequence number 6 (NCBI ID: NM_001330259.2).

5. A composition for predicting early death in a patient with cervical cancer, comprising a preparation for measuring the expression level of miR-574-3p (NCBI ID: NR_030300.1), a microRNA, and LINC01003 (NCBI ID: NR_027387.1), a long non-coding RNA (lncRNA).

6. A composition for predicting early death of a cervical cancer patient, characterized in that the agent for measuring the expression level in paragraph 5 is a sense and antisense primer or probe that complementarily binds to the microRNA.

7. A kit for predicting early death in cervical cancer patients, comprising a preparation measuring the expression level of miR-574-3p (NCBI ID: NR_030300.1), a microRNA, and LINC01003 (NCBI ID: NR_027387.1), a long non-coding RNA (lncRNA).

8. A method for providing information for predicting early death of a patient with cervical cancer, comprising the step of measuring the expression level of miR-574-3p (NCBI ID: NR_030300.1), a microRNA, and LINC01003 (NCBI ID: NR_027387.1), a long non-coding RNA (lncRNA), in a biological sample derived from a subject.

9. A method for providing information for predicting early death of a cervical cancer patient, characterized in that the biological sample is an exosome derived from blood or plasma in the 8th paragraph.

10. A method for providing information for predicting early death of a cervical cancer patient, characterized in that the expression level is measured by at least one method selected from the group consisting of next generation sequencing (NGS), polymerase chain reaction (PCR), reverse transcription polymerase chain reaction (RT-PCR), real-time polymerase chain reaction (Real-time PCR), RNase protection assay (RPA), microarray, and northern blotting.

11. A method for providing information for predicting early death of a cervical cancer patient, characterized in that the method further comprises a step of measuring the expression level of mRNA of the ACOT9 gene (NCBI ID: NG_016780.2) in a biological sample derived from the subject, in accordance with paragraph 8.

12. A method for providing information for predicting early death of a cervical cancer patient, characterized in that the subject is a patient receiving radiation therapy to treat cervical cancer, in accordance with paragraph 8.

13. A method for providing information for predicting early death of a cervical cancer patient, characterized in that the early death in paragraph 8 is death within 15 months from the end of chemoradiation treatment.

14. A method for selecting a biomarker to predict early death in patients with cervical cancer. A stage (preparatory stage) of collecting hematological factor data from a cohort of patients with cervical cancer; A step of selecting hematological factors that show significant differences in patients who died early from the above hematological factor data and analyzing the trends among the hematological factors (CBC analysis step); and A step of selecting exosomal RNA associated with the above hematological factors and patients with early death (RNA-CBC integrated analysis step); A method for selecting a biomarker for predicting early death of a patient with cervical cancer, characterized in that the hematological factor is a competitive blood count (CBC) measured in a blood sample, and the CBC includes an absolute neutrophil count (ANC) (cells / μL), a platelet (PLT) count (cells / μL), a hemoglobin (Hb) level (g / dL), an absolute lymphocyte count (ALC) (cells / μL), a monocyte (Mo) count (cells / μL), an NLR which is a value obtained by substituting the ANC and the ALC into ANC / ALC, a PLR which is a value obtained by substituting the PLT and the ALC into PLT / ALC, and a LMR which is a value obtained by substituting the ALC and the Mo into ALC / Mo.

15. In paragraph 14, The above hematological factor data is CBC0, which is a CBC value measured in a blood sample isolated from a patient before chemoradiation treatment (pretreatment CBC); min CBC, which is a minimum CBC (compete blood count) value measured in a blood sample isolated from a patient during chemoradiation treatment; the CBC0 value and the min CBC value. A method for selecting a biomarker for predicting early death of a patient with cervical cancer, characterized in that it includes: CBC1, which is a value obtained by substituting the value into log(CBC2 / CBC0); CBC2, which is a CBC (compete blood count) value measured in a blood sample isolated from a patient two weeks after starting chemoradiation treatment; and CBC3, which is a value obtained by substituting the CBC0 value and the CBC2 value into log(CBC2 / CBC0).

16. In paragraph 15, The above CBC0 includes ANC0, PLT0, Hb0, ALC0, and Mo0, which refer to the absolute neutrophil count (ANC) (cells / μL), platelet (PLT) count (cells / μL), hemoglobin (Hb) level (g / dL), absolute lymphocyte count (ALC) (cells / μL), and monocyte (Mo) count (cells / μL) measured in a blood sample isolated from a patient before chemoradiation treatment, and includes NLR0 (ANC0 / ALC0), PLT0 (PLT0 / ALC0), and LMR0 (ALC0 / Mo0). The above min CBC includes min ANC, min PLT, min Hb, min ALC, and min Mo, which refer to the absolute neutrophil count (ANC) (cells / μL), platelet (PLT) count (cells / μL), hemoglobin (Hb) level (g / dL), absolute lymphocyte count (ALC) (cells / μL), and monocyte (Mo) count (cells / μL) measured in blood samples isolated from patients during chemoradiation therapy. The above CBC1 is The value is ANC1, The value is PLT1, The value of Hb1, The value is ALC1, It includes the value Mo1, the value NLR1 of ANC1 / ALC1, the value PLR1 of PLT1 / ALC1, and the value LMR1 of ALC1 / Mo1, The above CBC2 includes ANC2, PLT2, Hb2, ALC2, and Mo2, which refer to the absolute neutrophil count (ANC) (cells / μL), platelet (PLT) count (cells / μL), hemoglobin (Hb) level (g / dL), absolute lymphocyte count (ALC) (cells / μL), and monocyte (Mo) count (cells / μL) measured in a blood sample isolated from a patient 2 weeks after the start of chemoradiation therapy. A method for selecting a biomarker for predicting early death of a patient with cervical cancer, characterized in that the CBC3 includes ANC3, which is a log(ANC2 / ANC0) value, PLT3, which is a log(PLT2 / PLT0) value, Hb3, which is a log(Hb2 / Hb0) value, ALC3, which is a log(ALC2 / ALC0) value, or Mo3, which is a log(Mo2 / Mo0) value.

17. In paragraph 14, The step of selecting exosomal RNA associated with the above hematological factors and patients with early death (RNA-CBC integrated analysis step) A step of calculating the log2Fold Change (log2FC) value of the above exosomal RNA; A step of calculating the absolute value |R| of the Pearson correlation coefficient for the above log2FC value and early death (ED) to select candidate RNAs having a significant log2FC value; and A method for selecting a biomarker for predicting early death of a cervical cancer patient, comprising the steps of: selecting RNA associated with hematological factor data and early death from cervical cancer from the candidate RNA; 18. In paragraph 17, A method for selecting a biomarker for predicting early death of a cervical cancer patient, characterized in that the exosomal RNA is at least one selected from the group consisting of miRNA (microRNA), piRNA (piwi-interacting RNA), snoRNA (small nucleaolar RNA), snRNA (small nuclear RNA), tRNA (transfer RNA), yRNA (Y RNA), lncRNA (long non-coding RNA), and mRNA (messenger RNA).

19. In paragraph 17, A method for selecting a biomarker for predicting early death of a patient with cervical cancer, characterized in that the step of calculating the log2Fold Change (log2FC) value of the exosomal RNA comprises isolating exosomal RNA from a blood sample isolated from a patient before and after chemoradiation treatment and calculating the change in the read count value of the exosomal RNA after treatment compared to the read count value of the exosomal RNA before treatment as a log2 (Fold Change) value through NGS (Next Generation Sequencing) analysis.

20. In paragraph 17, A method for selecting a biomarker for predicting early death of a cervical cancer patient, characterized in that the step of selecting RNA associated with hematological factor data and early death from cervical cancer from the above candidate RNA comprises calculating the absolute value |R| of the Pearson correlation coefficient for the log2FC value and early death (ED), and selecting RNA that shows a correlation between the CBC0 value, the CBC1 value, or the CBC3 value and the log2 (Fold Change) value while satisfying |R|>0.

Images