Early diagnosis of refractory kawasaki disease by novel microrna biomarker

Abstract

The present invention addresses the problem of providing a biomarker containing microRNA for predicting refractory Kawasaki disease and a novel method for predicting refractory Kawasaki disease. Provided are a biomarker for predicting refractory Kawasaki disease which contains at least one microRNA belonging to the hsa-miR-548 family and a method for predicting refractory Kawasaki disease based on the expression level of the microRNA.

Description

Early diagnosis of refractory Kawasaki disease using novel microRNA biomarkers

[0001] This invention relates to a microRNA for predicting refractory Kawasaki disease, a prediction method, and a kit using the same.

[0002] The number of people suffering from Kawasaki disease is on the rise. According to the latest 27th National Kawasaki Disease Survey Report (September 2023), 11,597 people were diagnosed with Kawasaki disease in 2021 and 10,333 in 2022. Including all previous National Kawasaki Disease Survey Reports, the cumulative number of people diagnosed with Kawasaki disease in Japan has reached 445,688.

[0003] Kawasaki disease was discovered in 1967 by Tomisaku Kawasaki, and a diagnostic criterion was established for diagnosing Kawasaki disease when five of six symptoms were present: persistent fever, conjunctival congestion, redness of the lips or strawberry tongue, amorphous rash, changes in the extremities (hardened edema or membranous desquamation), and non-suppurative cervical lymphadenopathy (Figure 1). This diagnostic criterion places emphasis on clinical symptoms and remains fundamentally unchanged to this day. The general clinical course of Kawasaki disease is that the above six symptoms gradually begin to appear, and by the time all symptoms are present, 4 to 6 days have passed since the onset of fever. At that point, according to the Kawasaki disease treatment guidelines shown in Figure 2, the first dose of high-dose gamma globulin therapy (IVIG) is started. If there is no response to the treatment, a second dose of IVIG is generally administered. If there is still no response, treatment including immunosuppressants (infliximab, IFX) is finally started (Figure 2).

[0004] The pathogenesis of Kawasaki disease is a systemic vasculitis syndrome centered on medium and small arteries in childhood, particularly causing vascular endothelial damage to the coronary arteries. Even when various treatments are performed according to the guidelines for the acute-phase treatment of Kawasaki disease, there is a certain number of patients who do not respond to the initial treatment of IVIG and become refractory to treatment. Some of the patients refractory to IVIG develop Kawasaki disease coronary artery aneurysms (coronary artery lesions, CAL), which are cardiovascular sequelae of Kawasaki disease, and the cases account for about 10% of all Kawasaki disease patients. Among them, there are cases that form giant coronary artery aneurysms at a rate of 0.1%-0.2%. As shown in Figure 3, both coronary arteries expand to the same diameter as the aorta, and in the acute phase, it is accompanied by a risk of death due to coronary artery rupture. Thereafter, the risk of acute myocardial infarction persists for a long time due to stenosis around the giant coronary artery aneurysm. When myocardial infarction occurs, it is necessary to perform a coronary artery bypass surgery with a high degree of difficulty in early childhood. In addition, once a Kawasaki disease coronary artery aneurysm is formed, it is necessary to take antiplatelet drugs and anticoagulants throughout life, which requires a high medical cost burden for the patients and society as a whole. Thus, it is important to prevent the aggravation of Kawasaki disease, and for this purpose, it is necessary to identify IVIG-refractory patients in the acute phase of Kawasaki disease at an early stage.

[0005] In general clinical practice, the Kobayashi score (serum Na < 133 mmol / L (2 points), neutrophil ratio ≥ 80% (2 points), CRP > 10 mg / dL (1 point), AST ≥ 100 IU / L (2 points), platelet count ≤ 300,000 / μL (1 point), age ≤ 12 months (1 point), less than 4 days from the first day of fever to the start of treatment (2 points)) has been used to predict IVIG refractory patients in the acute phase of Kawasaki disease. A Kobayashi score of 5 or higher is considered to indicate a high risk of IVIG refractory. However, this score does not take into account regional differences, racial differences, or individual patient differences. Furthermore, its predictive accuracy is limited due to its sensitivity of 76% and specificity of 80%. Therefore, treatment plans are not significantly altered based on the Kobayashi score. Generally, following the acute Kawasaki disease treatment guidelines mentioned above, the first IVIG treatment is administered 4 to 6 days after the onset of fever, when all symptoms of Kawasaki disease are present. A second IVIG treatment is then given, and if the patient is refractory to IVIG, treatment with immunosuppressants such as infliximab is initiated. During this treatment procedure, in refractory Kawasaki disease that is refractory to IVIG, coronary artery vasculitis progresses, and in the most severe cases, giant coronary artery aneurysms form, increasing the risk of coronary artery rupture and myocardial infarction. Thus, the conventional Kobayashi score is insufficient for individualized diagnosis and rapid risk assessment, and there is a need for an effective and early method to predict refractory Kawasaki disease.

[0006] The blood contains extracellular vesicles (EVs) derived from various cells. Endothelial cells, in particular, are known to release microparticles (EMPs). The inventors have previously confirmed that large amounts of EMPs are released from endothelial cells in a vasculitis model using scanning electron microscopy (Figure 4, Non-Patent Literature 1).

[0007] Furthermore, biomarkers for rapid detection of Kawasaki disease using microRNA as a marker, rapid diagnostic kits thereof, methods for detecting Kawasaki disease, and kits for detecting Kawasaki disease have been reported to date (Patent Documents 1 and 2).

[0008] Patent No. 6097457, Patent No. 6816963

[0009] Nakaoka H, ​​Hirono K, Yamamoto S, Takasaki I, Takahashi K, Kinoshita K, Takasaki A, Nishida N, Okabe M, Ce W, Miyao N, Saito K, Ibuki K, Ozawa S, Adachi Y, Ichida F, MicroRNA-145-5p and microRNA-320a encapsulated in endothelial microparticles contribute to the progression of vasculitis in acute Kawasaki Disease, Sci Rep, 8: 1016. doi: 10.1038 / s41598-018-19310-4, 2018

[0010] The present invention aims to provide a biomarker containing microRNA for predicting refractory Kawasaki disease and a novel method for predicting refractory Kawasaki disease.

[0011] As a result of diligent research, the inventors found that the expression profile of microRNAs within extracellular vesicles, which are unique to Kawasaki disease, is clearly different between the Kawasaki disease group that responds well to IVIG treatment and the refractory Kawasaki disease group that does not respond to IVIG treatment. By combining dimensionality reduction analysis and random forest analysis, the inventors succeeded in extracting several microRNAs with a high contribution rate in predicting refractory Kawasaki disease, and found that at least one microRNA belonging to the hsa-miR-548 family is an important microRNA for predicting refractory Kawasaki disease, thus completing the present invention. That is, the present invention includes the following aspects.

[0012] Item 1. A biomarker for predicting refractory Kawasaki disease, comprising at least one microRNA belonging to the hsa-miR-548 family. Item 2. The biomarker for predicting refractory Kawasaki disease according to Item 1, wherein the microRNA belonging to the hsa-miR-548 family is at least one microRNA selected from the group consisting of hsa-miR-548d-3p, hsa-miR-548b-5p, hsa-miR-548at-3p, hsa-miR-548au-5p, and hsa-miR-548as-5p. Item 3. The biomarker according to Item 1 or 2 for predicting IVIG response, wherein the microRNA belonging to the hsa-miR-548 family is at least one microRNA selected from the group consisting of hsa-miR-548d-3p, hsa-miR-548b-5p, and hsa-miR-548at-3p. Item 4. In addition, hsa-miR-3649, hsa-miR-500b-3p, hsa-miR-4700-5p, hsa-miR-455-5p, hsa-miR-4738-5p, hsa-miR-4786-3p, hsa-miR-5091, hsa-miR -8060, hsa-miR-515-5p, hsa-miR-4650-3p, hsa-miR-553, hsa-miR-675-5p, hsa-miR-6833-5p, hsa-miR-148a-5p, hsa-miR-6782-3p, hsa-m A biomarker for predicting IVIG response group as described in any one of items 1 to 3, comprising at least one microRNA selected from the group consisting of iR-2114-3p, hsa-miR-6727-3p, hsa-miR-4427, hsa-miR-29c-3p, hsa-miR-4723-5p, hsa-miR-362-3p, hsa-miR-5010-5p, hsa-miR-4799-3p, hsa-miR-1178-3p, hsa-miR-6783-3p, and hsa-miR-1306-5p.Item 5. Biomarkers for predicting IVIG response group as described in any one of items 1 to 4, including hsa-miR-3649, hsa-miR-500b-3p, hsa-miR-4700-5p, hsa-miR-455-5p, hsa-miR-4738-5p, hsa-miR-4786-3p, hsa-miR-548d-3p, hsa-miR-5091, hsa-miR-515-5p, hsa-miR-553, and hsa-miR-148a-5p. Item 6. hsa-miR-3649, hsa-miR-500b-3p, hsa-miR-4700-5p, hsa-miR-455-5p, hsa-miR-4738-5p, hsa-miR-4786-3p, hsa-miR-548d-3p, hsa-miR-5091 , hsa-miR-8060, hsa-miR-515-5p, hsa-miR-4650-3p, hsa-miR-553, hsa-miR-675-5p, hsa-miR-6833-5p, hsa-miR-148a-5p, hsa-miR-548b-5p, h Biomarkers for predicting IVIG response groups as described in any one of items 1 to 5, including sa-miR-6782-3p, hsa-miR-2114-3p, hsa-miR-6727-3p, hsa-miR-4427, hsa-miR-29c-3p, hsa-miR-4723-5p, hsa-miR-362-3p, hsa-miR-5010-5p, hsa-miR-4799-3p, hsa-miR-1178-3p, hsa-miR-548at-3p, hsa-miR-6783-3p, and hsa-miR-1306-5p. Item 7. A biomarker for predicting IVIG response, comprising hsa-miR-3649, hsa-miR-500b-3p, hsa-miR-4700-5p, hsa-miR-455-5p, hsa-miR-4786-3p, hsa-miR-5091, hsa-miR-515-5p, hsa-miR-553, and hsa-miR-148a-5p. Item 8. The microRNA belonging to the hsa-miR-548 family is at least one microRNA selected from the group consisting of hsa-miR-548au-5p, hsa-miR-548as-5p, and hsa-miR-548b-5p, and is a biomarker for predicting IVIG unresponsiveness according to item 1 or 2.Item 9. In addition, hsa-miR-1292-3p, hsa-miR-4744, hsa-miR-550b-2-5p, hsa-miR-553, hsa-miR-99a-3p, hsa-miR- 301a-5p, hsa-miR-371a-5p, hsa-miR-2682-5p, hsa-miR-6892-3p, hsa-miR-3591-3p, hsa-miR-4660, h A biomarker for predicting IVIG refractory grouping as described in item 1, 2, or 8, comprising at least one microRNA selected from the group consisting of sa-miR-425-3p, hsa-miR-320a, hsa-miR-593-5p, hsa-miR-346, hsa-miR-6877-5p, hsa-miR-8060, hsa-miR-18b-5p, and hsa-miR-626. Item 10. Biomarkers for predicting IVIG refractory groups as described in item 1, 2, 8, or 9, including hsa-miR-548au-5p, hsa-miR-548as-5p, hsa-miR-1292-3p, hsa-miR-550b-2-5p, hsa-miR-553, hsa-miR-99a-3p, hsa-miR-301a-5p, hsa-miR-371a-5p, hsa-miR-3591-3p, and hsa-miR-4660. Item 11. hsa-miR-548au-5p, hsa-miR-548as-5p, hsa-miR-1292-3p, hsa-miR-4744, hsa-miR-550b-2-5p, hsa-miR-553, hsa-miR-99a-3p, hsa-miR-301a-5p, hsa-miR-371a-5p, hsa-miR-2682-5p, hsa-miR-6892-3p, hsa-miR-3591-3 A biomarker for predicting IVIG refractory groups, as described in any one of items 1, 2, 8-10, including p, hsa-miR-4660, hsa-miR-425-3p, hsa-miR-320a, hsa-miR-593-5p, hsa-miR-346, hsa-miR-6877-5p, hsa-miR-8060, hsa-miR-18b-5p, hsa-miR-626, and hsa-miR-548b-5p.Item 12. The microRNA belonging to the hsa-miR-548 family is hsa-miR-548b-5p, and is a biomarker for predicting IVIG refractory group as described in Item 1 or 2. Item 13. Furthermore, hsa-miR-550b-2-5p, hsa-miR-626, hsa-miR-553, hsa-miR-3145-5p, hsa-miR-3649, hsa-miR-3678-3p, hsa-miR-1185-1-3p, hsa-miR-6895-3p, hsa-miR-8060, hsa-miR-371a-5p, hsa-miR-6853 A biomarker for predicting IVIG refractory grouping as described in item 1, 2, or 12, comprising at least one microRNA selected from the group consisting of -3p, hsa-miR-663b, hsa-miR-4518, hsa-miR-494-3p, hsa-miR-346, hsa-miR-1915-5p, hsa-miR-4325, and hsa-miR-6772-3p. Item 14. Biomarkers for predicting IVIG refractory groups as described in item 1, 2, 12, or 13, including hsa-miR-550b-2-5p, hsa-miR-626, hsa-miR-553, hsa-miR-3145-5p, hsa-miR-3649, hsa-miR-3678-3p, hsa-miR-1185-1-3p, hsa-miR-6895-3p, hsa-miR-8060, hsa-miR-548b-5p, and hsa-miR-663b. Item 15. hsa-miR-550b-2-5p, hsa-miR-626, hsa-miR-553, hsa-miR-3145-5p, hsa-miR-3649, hsa-miR-36 78-3p, hsa-miR-1185-1-3p, hsa-miR-6895-3p, hsa-miR-8060, hsa-miR-371a-5p, hsa-miR-548b Biomarkers for predicting IVIG refractory groups, as described in any one of items 1, 2, 12-14, including -5p, hsa-miR-6853-3p, hsa-miR-663b, hsa-miR-4518, hsa-miR-494-3p, hsa-miR-346, hsa-miR-1915-5p, hsa-miR-4325, and hsa-miR-6772-3p.Item 16. A method for predicting refractory Kawasaki disease based on the expression level of a microRNA belonging to the hsa-miR-548 family in a biological sample taken from a subject. Item 17. The method according to Item 16, comprising the following steps: (1) Performing a logistic regression analysis in which a score with the expression level of the microRNA as the explanatory variable is set as a linear predictor according to the following formula: and [Formula 1] Score = Σ((coefficient) × microRNA expression level) + intercept (2) Predicting that the subject has refractory Kawasaki disease if the score is lower than a cutoff value. Item 18. The method according to Item 17, wherein the cutoff value is a cutoff value that gives 100% sensitivity and specificity. Item 19. The method according to Item 16 or 17, wherein the subject is a patient before IVIG treatment. Item 20. A kit for predicting refractory Kawasaki disease, comprising means for detecting the expression level of the microRNA according to Item 1. Item 21. A biomarker for predicting refractory Kawasaki disease, comprising at least one microRNA selected from the group consisting of hsa-miR-369-3p, hsa-miR-708-3p, hsa-miR-367-3p, hsa-miR-26a-5p, hsa-miR-126-5p, hsa-miR-302b-5p, hsa-miR-1225-3p, hsa-miR-194-5p, hsa-miR-579-3p, hsa-miR-133b, hsa-miR-296-5p, hsa-miR-558, hsa-miR-1914-3p, hsa-miR-224-5p, hsa-miR-31-5p, hsa-miR-548c-3p, and hsa-miR-3648. Item 22. The biomarker described in item 21, comprising at least three microRNAs selected from the group consisting of hsa-miR-369-3p, hsa-miR-708-3p, hsa-miR-367-3p, hsa-miR-26a-5p, hsa-miR-126-5p, hsa-miR-302b-5p, hsa-miR-1225-3p, hsa-miR-194-5p, hsa-miR-579-3p, hsa-miR-133b, hsa-miR-296-5p, hsa-miR-558, hsa-miR-1914-3p, hsa-miR-224-5p, hsa-miR-31-5p, hsa-miR-548c-3p, and hsa-miR-3648.Item 23. A biomarker according to item 21 or 22, comprising at least three microRNAs selected from the group consisting of hsa-miR-369-3p, hsa-miR-708-3p, hsa-miR-367-3p, hsa-miR-26a-5p, hsa-miR-126-5p, hsa-miR-302b-5p, hsa-miR-1225-3p, hsa-miR-194-5p, hsa-miR-579-3p, hsa-miR-133b, hsa-miR-296-5p, hsa-miR-558, and hsa-miR-1914-3p. Item 24. A biomarker according to any one of items 21 to 23, comprising at least three microRNAs selected from the group consisting of hsa-miR-369-3p, hsa-miR-708-3p, hsa-miR-367-3p, hsa-miR-126-5p, and hsa-miR-194-5p. Item 25. A biomarker according to any one of items 21 to 24, comprising the microRNAs hsa-miR-369-3p, hsa-miR-708-3p, hsa-miR-367-3p, hsa-miR-126-5p, and hsa-miR-194-5p. Item 26. A biomarker according to any one of claims 21 to 25, including hsa-miR-369-3p, hsa-miR-708-3p, hsa-miR-367-3p, hsa-miR-26a-5p, hsa-miR-126-5p, hsa-miR-302b-5p, hsa-miR-1225-3p, hsa-miR-194-5p, hsa-miR-579-3p, hsa-miR-133b, hsa-miR-296-5p, hsa-miR-558, hsa-miR-1914-3p, hsa-miR-224-5p, hsa-miR-31-5p, hsa-miR-548c-3p, and hsa-miR-3648.

[0013] The microRNA for predicting refractory Kawasaki disease according to the present invention can predict refractory Kawasaki disease with higher sensitivity and specificity compared to conventional risk scores.

[0014] (A) This figure shows the general clinical symptoms of Kawasaki disease. 1. Fever, 2. Conjunctival hyperemia of both eyes, 3. Lip and oral findings: redness of the lips, strawberry tongue, redness of the oral and pharyngeal mucosa, 4. Atypical rash, 5. Changes in the extremities: (acute phase) indurated edema of the hands and feet, palmar erythema, (recovery phase) membranous desquamation from the fingertips, 6. Non-suppurative cervical lymphadenopathy in the acute phase. (B) This figure shows the general course of Kawasaki disease. The upper part of the graph shows the change in body temperature from before admission to discharge. The lower part of the graph shows the sequential appearance of symptoms such as cervical lymphadenopathy, atypical erythema, conjunctival hyperemia, redness of the lips / strawberry tongue, and edema of the hands and feet, followed by desquamation of the hands and feet. This figure shows the treatment guidelines for Kawasaki disease (including incomplete forms). * indicates that it is not covered by insurance. If the IVIG resistance prediction score is classified as low risk, a moderate dose of IVIG + ASA is administered as the first line. If the IVIG resistance prediction score is classified as high risk, a moderate dose of IVIG + ASA is standard as the first line, with prednisolone (PSL) and cytoplasmic analgesia (CsA) being recommended, and IVMP* and UTI* being considered. Typically, a low dose of ASA is administered after fever subsides. If fever persists or relapses 24 to 36 hours after discontinuation of IVIG, IVIG is recommended as the second line. Administration of PSL, IVMP*, and IFX is considered. The second line is administered alone or in combination. If fever persists or relapses thereafter, administration of IVIG, PSL, IVMP*, CsA, IFX, and PE is considered as the third line or later. The third line and later are administered alone or in combination. CsA in the third line can be upgraded to the second line, but IFX cannot be re-administered if used in the second line. If fever has already subsided after diagnosis of Kawasaki disease, consider IVIG or other medications in case of recurrent fever. When administering low doses of ASA, antiplatelet agents or anticoagulants should be added as appropriate in cases complicated by CAA. This figure shows a giant coronary artery aneurysm, a sequela of refractory Kawasaki disease. In the left figure, the thin arrows indicate that CAL is likely to form at the coronary artery origin. In the right figure, RCA indicates Right Coronary Artery, LCA indicates Left Coronary Artery, and Ao indicates Aorta. This figure shows EMPs released from vascular endothelial cells confirmed using a scanning electron microscope.(a) Before LPS stimulation, (b) 10 seconds after LPS stimulation, (c) 10 minutes after LPS stimulation, and (d) the appearance of EMPs in the culture medium. (A) A figure comparing the number of EMPs in the Kawasaki disease group and the control group. (B) A figure showing the changes in EMPs before IVIG treatment, immediately after IVIG treatment, and during the recovery period in the CAL group and the Non-CAL group. A schematic diagram showing the analysis method for extracting microRNAs for predicting refractory Kawasaki disease. Results of the study using dataset (A). (A) A figure showing the discrimination of Responder, Refractory, and non-Kawasaki disease group (non-KD group) in the acute phase of Kawasaki disease as a preliminary study. (B) A figure showing the discrimination of refractory Kawasaki disease patients using the novel microRNA biomarker of the present invention. A figure showing the discrimination of refractory Kawasaki disease patients using the novel microRNA biomarker of the present invention in the acute phase of Kawasaki disease using dataset (A). This shows the results of performing logistic regression using dataset (A) by removing one microRNA expression level at a time from the explanatory variables, each time minimizing the absolute value of the coefficients predicted by logistic regression, and repeating the logistic regression each time. The five microRNA expression levels that remained at the top and were not removed were then used as explanatory variables. This shows the results of performing logistic regression using dataset (A) by removing one microRNA expression level at a time from the explanatory variables, each time minimizing the absolute value of the coefficients predicted by logistic regression, and repeating the logistic regression each time. The thirteen microRNA expression levels that remained at the top and were not removed were then used as explanatory variables. This figure shows the distinction between the IVIG-responsive group (Figure 12-1) and the IVIG-unresponsive group (Figures 12-2, 12-3) using the novel microRNA biomarker of the present invention, using datasets (B) to (D).These figures show the distinction between Responders (Figure 13-1) and Refractories (Figures 13-2, 13-3) using the novel microRNA biomarkers of the present invention, using datasets (B) to (D). The results are shown by performing logistic regression using the microRNA expression levels of 23 types (Figure 14-1), 16 types (Figure 14-2), 11 types (Figure 14-3), or 9 types (Figure 14-4) that remained at the top of the list without being excluded, after each removal of one microRNA expression level from the explanatory variables using dataset (B). The results shown are as follows: Using dataset (C), one microRNA expression level that minimizes the absolute value of the coefficients predicted by logistic regression was removed from the explanatory variables one by one, and the logistic regression was repeated each time. The results shown are as follows: Logistic regression was performed using the microRNA expression levels of 17 types (Figure 15-1), 14 types (Figure 15-2), or 10 types (Figure 15-3) that remained at the top and were not removed as explanatory variables. Using dataset (D), one microRNA expression level that minimizes the absolute value of the coefficients predicted by logistic regression was removed from the explanatory variables one by one, and the logistic regression was repeated each time. The results shown are as follows: Logistic regression was performed using the microRNA expression levels of 16 types (Figure 16-1), 14 types (Figure 16-2), or 11 types (Figure 16-3) that remained at the top and were not removed as explanatory variables.

[0015] <Definition of Terms> In this specification, the inability to respond to high-dose gamma globulin therapy (IVIG treatment) based on the treatment guidelines for the acute phase of Kawasaki disease is referred to as "not responding to IVIG treatment" or "IVIG unresponsiveness." Conversely, a response to IVIG treatment is expressed as "responding to IVIG treatment" or "IVIG response." In this specification, patients who do not respond to the first course of IVIG treatment based on the treatment guidelines for the acute phase of Kawasaki disease are referred to as the Non-Responder group. The Non-Responder group includes the Delayed responder group, which responds to the second course of IVIG treatment even if it did not respond to the first course of IVIG treatment, and the Refractory group, which does not respond to either the first or second course of IVIG treatment. In this specification, the Refractory group, which does not respond to two courses of IVIG treatment, is referred to as refractory Kawasaki disease, and more specifically as the IVIG unresponsive group. Such a condition is also referred to as severe disease. In this invention, refractory Kawasaki disease also includes patients who do not respond to two or more courses of IVIG treatment. If IVIG treatment is unsuccessful and the treatment period is prolonged, coronary artery lesions (CALs) may form. Among the refractory group, patients who develop coronary artery lesions are specifically referred to as the CAL group. In this invention, patients who respond to the first IVIG treatment based on the treatment guidelines for the acute phase of Kawasaki disease are referred to as the Responder group, and are specifically referred to as the IVIG response group. The state of responding to the first IVIG treatment (Responder group) and the state of responding to the second IVIG treatment (Delayed responder group) may also be expressed as non-severe disease. Patients who respond to IVIG treatment do not develop CALs. Therefore, when the term Non-CAL group is used herein, it includes patients in the IVIG refractory group who do not develop coronary artery lesions, and patients who respond to the first or second IVIG treatment (Responder and Delayed responder).

[0016] 1. microRNA In one embodiment, the biomarker for predicting refractory Kawasaki disease according to the present invention includes at least one microRNA belonging to the hsa-miR-548 family. This microRNA can be identified as follows.

[0017] First, extracellular vesicles (EVs) specific to Kawasaki disease are collected from serum derived from Kawasaki disease patients. MicroRNAs are extracted from the EVs and subjected to a microRNA array capable of measuring human microRNAs to obtain microRNA expression levels (profiles). At this time, microRNA profiles are obtained for three groups of patients: those who respond to the first IVIG treatment (Responders), those who are unresponsive to the first IVIG treatment but respond to the second IVIG treatment (Delayed responders), and those who do not respond to either the first or second IVIG treatment (Refractory). Next, Analysis 1 and Analysis 2 below are performed separately to identify the microRNA sets obtained in common from both. Furthermore, a logistic regression model is constructed using the expression levels of the microRNA sets selected above as explanatory variables, and the intercept and coefficients are obtained. MicroRNAs with large absolute values ​​of coefficients are important microRNAs for diagnosing refractory Kawasaki disease.

[0018] In this invention, the prediction of refractory Kawasaki disease includes predicting patients who are refractory to IVIG and patients who are responders to IVIG. By identifying patients who respond to the first dose of IVIG (predicting the IVIG response group), it is possible to predict the refractory group, i.e., patients with refractory Kawasaki disease. Depending on which patient's microRNA is used to differentiate the groups, it is possible to extract microRNAs that are particularly important for predicting a specific group (e.g., refractory or responder). For example, when an analysis is performed to differentiate between responders and others (delayed responder and refractory) (dataset (B)), the microRNAs extracted are a set of microRNAs that predict whether or not a patient is a responder who responds to the first dose of IVIG. When an analysis is performed to differentiate between refractory and others (delayed responder and responder) (dataset (C)), the microRNAs extracted are a set of microRNAs that predict whether or not a patient is refractory to the second dose of IVIG. The microRNAs extracted when performing analysis to distinguish between refractory and responder (dataset (D)) characterize the differences between two groups with extremely different treatment outcomes, such as the group unresponsive to two IVIG treatments and the group that responded to the first IVIG treatment. This microRNA set is thought to be significantly involved in IVIG treatment responsiveness.

[0019] 1.1 Analysis 1. Using the obtained microRNA profiles, PLS-DA, a dimensionality reduction analysis method, is used to classify the datasets into the following categories: dataset (A): IVIG-unresponsive group (Refractory in Figure 7(A)) and IVIG-responsive group (Responder in Figure 7(A)); dataset (B): Responder and others (Delayed responder and Refractory); dataset (C): Refractory and others (Delayed responder and Responder); and dataset (D): Refractory and Responder. In particular, it is preferable to use the VIP score, which evaluates the magnitude of contribution to the first principal component, and select microRNAs that show a VIP score of a certain value or higher as biomarkers.

[0020] 1.2 Analysis 2. The obtained microRNA profiles are subjected to a random forest, a machine learning method. Nine random forest prediction models are constructed for dataset (A), eight for dataset (B), nine for dataset (C), and eight for dataset (D), each with different probabilities. MicroRNAs that fit the criteria of "whether they were selected as important in common by n or more models" are selected for classification into the target groups in datasets (A) to (D). In the datasets used in this invention, it is preferable that n be between 2 and 5 from the viewpoint of achieving 100% sensitivity and specificity. Among these, from the viewpoint of improving prediction accuracy, it is preferable that more types of microRNAs are included as explanatory variables, so it is preferable that n be 5, more preferable that n be 4, even more preferable that n be 3, and most preferable that n be 2. It is preferable that n be 2 or more in order to perform analysis under the condition that multiple models have selected microRNAs. For group discrimination, we use "CVPFI: Cross-validated Permutation Feature Importance" (Kaneko, 2022, Anal. Sci. Adv. https: / / chemistry-europe.onlinelibrary.wiley.com / doi / 10.1002 / ansa.202200018), a method developed in the field of chemical engineering. For example, we extract microRNAs with a positive CVPFI value. By extracting microRNAs based on CVPFI, we can perform as many cross-validations as possible and prevent model overfitting, given the smaller amount of data in clinical data compared to other machine learning subjects. MicroRNAs that show a certain CVPFI value can be said to contribute highly to group discrimination, and these microRNAs can be considered important for predicting IVIG unresponsiveness and IVIG response.

[0021] The microRNA sets extracted in common from both Analysis 1 and Analysis 2 are identified. The identified microRNA sets are important for the progression of Kawasaki disease and are likely to be targets for diagnosis and therapeutic intervention. For example, in dataset (A), Analysis 1 shows a high contribution to classification, and in Analysis 2, 6 types of microRNAs can be extracted when n is 5, 12 types when n is 4, 14 types when n is 3, and 22 (17+5) types when n is 2. In dataset (B), Analysis 1 shows a high contribution to classification, and in Analysis 2, 13 types of microRNAs can be extracted when n is 4 or greater, 19 types when n is 3, and 31 (29+2) types when n is 2. In dataset (C), Analysis 1 shows a high contribution to classification, and Analysis 2 shows a high contribution to classification. When n is 7 or greater, 8 types of microRNA can be extracted; when n is 6, 15 types of microRNA can be extracted; and when n is 5, 22 types of microRNA can be extracted. In dataset (D), analysis 1 shows a high contribution to classification, and analysis 2 shows that when n is 6 or greater, 8 types of microRNA can be extracted; when n is 5, 11 types of microRNA can be extracted; and when n is 4, 19 types of microRNA can be extracted. In this invention, by combining dimensionality reduction analysis and random forest analysis, prediction accuracy is further improved, making it possible to predict refractory Kawasaki disease with higher sensitivity and specificity compared to conventional risk scores.

[0022] 1.3 Logistic Regression Model Next, a logistic regression model is constructed using the expression levels of the microRNA sets selected as described above as explanatory variables to obtain the intercept and coefficients (regression coefficients). If the number of types of miRNAs used for classification is too large, correlations between variables tend to occur, the regression does not work well, and the prediction accuracy decreases. Conversely, if the number of microRNAs used for classification is too small, there is insufficient information and the prediction accuracy decreases. It is necessary to adjust both of these aspects. As a result, it is preferable that in dataset (A) of the present invention, 17 types of microRNAs are extracted by excluding 5 types of microRNAs that had a large correlation with other microRNAs when n was 2 in analysis 2; in dataset (B), 29 types of microRNAs are extracted by excluding 2 types of microRNAs that had a large correlation with other microRNAs when n was 2 in analysis 2; in dataset (C), 22 types of microRNAs are extracted when n was 5 in analysis 2; and in dataset (D), 19 types of microRNAs are extracted when n was 4 in analysis 2.

[0023] The microRNAs selected as described above include at least one microRNA belonging to the hsa-miR-548 family. Among the microRNAs belonging to the hsa-miR-548 family, it is preferable that the microRNAs include at least one microRNA selected from the group consisting of hsa-miR-548c-3p, hsa-miR-548d-3p, hsa-miR-548b-5p, hsa-miR-548at-3p, hsa-miR-548au-5p, and hsa-miR-548as-5p. When using dataset (A), it is preferable that the microRNAs belonging to the hsa-miR-548 family include hsa-miR-548c-3p. When using dataset (B), it is preferable that the microRNA belonging to the hsa-miR-548 family is at least one microRNA selected from the group consisting of hsa-miR-548d-3p, hsa-miR-548b-5p, and hsa-miR-548at-3p. When using dataset (C), it is preferable that the microRNA belonging to the hsa-miR-548 family is at least one microRNA selected from the group consisting of hsa-miR-548au-5p, hsa-miR-548as-5p, and hsa-miR-548b-5p. When using dataset (D), it is preferable that the microRNA belonging to the hsa-miR-548 family is hsa-miR-548b-5p.

[0024] 1.3.1 Dataset (A) In the present invention, from the 17 types of microRNAs obtained in Dataset (A), it is preferable to select at least one and three microRNAs as biomarkers from the group of five microRNAs that remain unexcluded due to the small absolute value of the coefficients predicted by logistic regression, by excluding one microRNA expression level from the explanatory variables one by one from the 17 types of microRNAs obtained in Dataset (A), and repeating the logistic regression each time. It is more preferable to select at least one and three microRNAs as biomarkers from the group of six microRNAs, even more preferable to select at least one and three microRNAs as biomarkers from the group of thirteen microRNAs, and most preferable to select all 17 types of microRNAs as biomarkers. Alternatively, one could select at least five microRNAs as biomarkers from the group of six microRNAs that remain after each removal of the microRNA expression level from the explanatory variables (the microRNA expression level that minimizes the absolute value of the coefficient predicted by logistic regression) from the 17 microRNAs mentioned above, and then repeating the logistic regression. Another method would be to select at least five microRNAs as biomarkers from the group of thirteen microRNAs.

[0025] The biomarker for predicting refractory Kawasaki disease obtained as described above preferably contains at least one microRNA selected from the group consisting of hsa-miR-369-3p, hsa-miR-708-3p, hsa-miR-367-3p, hsa-miR-26a-5p, hsa-miR-126-5p, hsa-miR-302b-5p, hsa-miR-1225-3p, hsa-miR-194-5p, hsa-miR-579-3p, hsa-miR-133b, hsa-miR-296-5p, hsa-miR-558, hsa-miR-1914-3p, hsa-miR-224-5p, hsa-miR-31-5p, hsa-miR-548c-3p, and hsa-miR-3648.

[0026] The biomarkers for predicting refractory Kawasaki disease obtained as described above are hsa-miR-369-3p, hsa-miR-708-3p, hsa-miR-367-3p, hsa-miR-26a-5p, hsa-miR-126-5p, hsa-miR-302b-5p, hsa-miR-1225-3p, hsa-miR-194-5p, hsa-miR-579-3p, hsa-miR-133b, hsa-miR-296-5p, and hsa-miR-5 Preferably, it contains at least three microRNAs selected from the group consisting of 58, hsa-miR-1914-3p, hsa-miR-224-5p, hsa-miR-31-5p, hsa-miR-548c-3p, and hsa-miR-3648; more preferably, it contains at least five microRNAs; even more preferably, it contains at least ten microRNAs; and even more preferably, it contains at least thirteen microRNAs.

[0027] The biomarkers for predicting refractory Kawasaki disease obtained as described above are at least three, at least five, and at least ten mi selected from the group consisting of hsa-miR-369-3p, hsa-miR-708-3p, hsa-miR-367-3p, hsa-miR-26a-5p, hsa-miR-126-5p, hsa-miR-302b-5p, hsa-miR-1225-3p, hsa-miR-194-5p, hsa-miR-579-3p, hsa-miR-133b, hsa-miR-296-5p, hsa-miR-558, and hsa-miR-1914-3p. It is preferable to include croRNA, and it is preferable to include at least three microRNAs selected from the group consisting of hsa-miR-369-3p, hsa-miR-708-3p, hsa-miR-367-3p, hsa-miR-26a-5p, hsa-miR-126-5p and hsa-miR-194-5p, and at least five microRNAs, and it is preferable to include at least three microRNAs selected from the group consisting of hsa-miR-369-3p, hsa-miR-708-3p, hsa-miR-367-3p, hsa-miR-126-5p and hsa-miR-194-5p. Furthermore, the biomarkers for predicting refractory Kawasaki disease obtained as described above are hsa-miR-369-3p, hsa-miR-708-3p, hsa-miR-367-3p, hsa-miR-26a-5p, hsa-miR-126-5p, hsa-miR-302b-5p, hsa-miR-1225-3p, hsa-miR-194-5p, hsa-miR-579-3p, hsa-miR-133b, hsa-miR-296- It is preferable to include at least five microRNAs selected from the group consisting of 5p, hsa-miR-558, and hsa-miR-1914-3p, and more preferably to include at least five microRNAs selected from the group consisting of hsa-miR-369-3p, hsa-miR-708-3p, hsa-miR-367-3p, hsa-miR-26a-5p, hsa-miR-126-5p, and hsa-miR-194-5p.

[0028] The biomarker for predicting refractory Kawasaki disease obtained as described above preferably contains the microRNAs of hsa-miR-369-3p, hsa-miR-708-3p, hsa-miR-367-3p, hsa-miR-126-5p and hsa-miR-194-5p, and more preferably contains the microRNAs of hsa-miR-369-3p, hsa-miR-708-3p, hsa-miR-367-3p, hsa-miR-26a-5p, hsa-miR-126-5p and hsa-miR-194-5p. It is more preferable to include the following microRNAs: hsa-miR-369-3p, hsa-miR-708-3p, hsa-miR-367-3p, hsa-miR-26a-5p, hsa-miR-126-5p, hsa-miR-302b-5p, hsa-miR-1225-3p, hsa-miR-194-5p, hsa-miR-579-3p, hsa-miR-133b, hsa-miR-296-5p, hsa-miR-558, and hsa-miR-1914-3p. Furthermore, it is most preferable to include all of the following: hsa-miR-369-3p, hsa-miR-708-3p, hsa-miR-367-3p, hsa-miR-26a-5p, hsa-miR-126-5p, hsa-miR-302b-5p, hsa-miR-1225-3p, hsa-miR-194-5p, hsa-miR-579-3p, hsa-miR-133b, hsa-miR-296-5p, hsa-miR-558, hsa-miR-1914-3p, hsa-miR-224-5p, hsa-miR-31-5p, hsa-miR-548c-3p, and hsa-miR-3648.

[0029] 1.3.2 Dataset (B) The microRNA set obtained using Dataset (B) is a biomarker for predicting the IVIG response group (Responder). In this invention, it is preferable to select at least one, at least three, at least five, or at least ten microRNAs as biomarkers from the groups consisting of 9, 11, 16, 23, or 29 microRNAs that remain unexcluded due to the small absolute value of the coefficients predicted by logistic regression, by removing one microRNA expression level at a time from the explanatory variables and repeating logistic regression each time. Furthermore, it is more preferable to select all nine microRNAs or all eleven microRNAs that remain unexcluded due to the small absolute value of the coefficients as biomarkers, and most preferable to select all 29 microRNAs as biomarkers.

[0030] The biomarkers for predicting IVIG response groups obtained as described above are hsa-miR-3649, hsa-miR-500b-3p, hsa-miR-4700-5p, hsa-miR-455-5p, hsa-miR-4738-5p, hsa-miR-4786-3p, hsa-miR-548d-3p, hsa-miR-5091, hsa-miR-8060, hsa-miR-515-5p, hsa-miR-4650-3p, hsa-miR-553, hsa-miR-675-5p, hsa-miR-6833-5p, hsa-miR-148a-5p, hsa-miR-548b-5p, and hsa-miR-67 Preferably, the microRNA contains at least one, at least three, at least five, at least ten, at least fifteen, at least twenty, or at least twenty-five microRNAs selected from the group consisting of 82-3p, hsa-miR-2114-3p, hsa-miR-6727-3p, hsa-miR-4427, hsa-miR-29c-3p, hsa-miR-4723-5p, hsa-miR-362-3p, hsa-miR-5010-5p, hsa-miR-4799-3p, hsa-miR-1178-3p, hsa-miR-548at-3p, hsa-miR-6783-3p, and hsa-miR-1306-5p.

[0031] The biomarkers for predicting IVIG response groups obtained as described above are hsa-miR-3649, hsa-miR-500b-3p, hsa-miR-4700-5p, hsa-miR-455-5p, hsa-miR-4738-5p, hsa-miR-4786-3p, hsa-miR-548d-3p, hsa-miR-5091, hsa-miR-8060, hsa-miR-515-5p, hsa-miR-4650-3p, hsa-miR-553, hsa-miR-675-5p, and hsa-m Preferably, the microRNAs include at least one, at least three, at least five, at least ten, at least fifteen, or at least twenty types selected from the group consisting of iR-6833-5p, hsa-miR-148a-5p, hsa-miR-548b-5p, hsa-miR-6782-3p, hsa-miR-2114-3p, hsa-miR-6727-3p, hsa-miR-4427, hsa-miR-29c-3p, hsa-miR-4723-5p, and hsa-miR-362-3p.

[0032] The biomarkers for predicting IVIG response groups obtained as described above are hsa-miR-3649, hsa-miR-500b-3p, hsa-miR-4700-5p, hsa-miR-455-5p, hsa-miR-4738-5p, hsa-miR-4786-3p, hsa-miR-548d-3p, hsa-miR-5091, hsa-miR-8060, and hsa-m Preferably, the microRNAs include at least one, at least three, at least five, at least ten, or at least fifteen microRNAs selected from the group consisting of iR-515-5p, hsa-miR-4650-3p, hsa-miR-553, hsa-miR-6833-5p, hsa-miR-148a-5p, hsa-miR-548b-5p, and hsa-miR-2114-3p.

[0033] The biomarker for predicting IVIG response group obtained as described above preferably contains at least one, at least three, at least five, or at least ten microRNAs selected from the group consisting of hsa-miR-3649, hsa-miR-500b-3p, hsa-miR-4700-5p, hsa-miR-455-5p, hsa-miR-4738-5p, hsa-miR-4786-3p, hsa-miR-548d-3p, hsa-miR-5091, hsa-miR-515-5p, hsa-miR-553, and hsa-miR-148a-5p.

[0034] The biomarker for predicting IVIG response group obtained as described above preferably contains at least one, at least three, at least five, or at least eight microRNAs selected from the group consisting of hsa-miR-3649, hsa-miR-500b-3p, hsa-miR-4700-5p, hsa-miR-455-5p, hsa-miR-4786-3p, hsa-miR-5091, hsa-miR-515-5p, hsa-miR-553, and hsa-miR-148a-5p.

[0035] The biomarker for predicting IVIG response group obtained as described above comprises at least one microRNA selected from the group consisting of hsa-miR-548d-3p, hsa-miR-548b-5p, and hsa-miR-548at-3p, and further comprises hsa-miR-3649, hsa-miR-500b-3p, hsa-miR-4700-5p, hsa-miR-455-5p, hsa-miR-4738-5p, hsa-miR-4786-3p, hsa-miR-5091, hsa-miR-8060, hsa-miR-515-5p, hsa-miR-4650-3p, hsa-miR-553, hsa-miR-675-5p, and hsa-miR Preferably, the microRNAs include at least one, at least three, at least five, at least ten, at least fifteen, at least twenty, or at least twenty-five microRNAs selected from the group consisting of -6833-5p, hsa-miR-148a-5p, hsa-miR-6782-3p, hsa-miR-2114-3p, hsa-miR-6727-3p, hsa-miR-4427, hsa-miR-29c-3p, hsa-miR-4723-5p, hsa-miR-362-3p, hsa-miR-5010-5p, hsa-miR-4799-3p, hsa-miR-1178-3p, hsa-miR-6783-3p, and hsa-miR-1306-5p.

[0036] The biomarker for predicting the IVIG response group obtained as described above may include hsa-miR-3649, hsa-miR-500b-3p, hsa-miR-4700-5p, hsa-miR-455-5p, hsa-miR-4786-3p, hsa-miR-5091, hsa-miR-515-5p, hsa-miR-553, and hsa-miR-148a-5p. As shown in Figure 14-4, even when using the above nine types of microRNAs as markers, the IVIG response group (responders) can be detected with 100% sensitivity and specificity.

[0037] The biomarker for predicting the IVIG response group obtained as described above may include hsa-miR-3649, hsa-miR-500b-3p, hsa-miR-4700-5p, hsa-miR-455-5p, hsa-miR-4738-5p, hsa-miR-4786-3p, hsa-miR-548d-3p, hsa-miR-5091, hsa-miR-515-5p, hsa-miR-553, and hsa-miR-148a-5p. As shown in Figure 14-3, even when using the above 11 types of microRNAs as markers, the IVIG response group (responders) can be detected with 100% sensitivity and specificity.

[0038] The biomarkers for predicting IVIG response groups obtained as described above may include hsa-miR-3649, hsa-miR-500b-3p, hsa-miR-4700-5p, hsa-miR-455-5p, hsa-miR-4738-5p, hsa-miR-4786-3p, hsa-miR-548d-3p, hsa-miR-5091, hsa-miR-8060, hsa-miR-515-5p, hsa-miR-4650-3p, hsa-miR-553, hsa-miR-6833-5p, hsa-miR-148a-5p, hsa-miR-548b-5p, and hsa-miR-2114-3p.

[0039] The biomarkers for predicting IVIG response groups obtained as described above are hsa-miR-3649, hsa-miR-500b-3p, hsa-miR-4700-5p, hsa-miR-455-5p, hsa-miR-4738-5p, hsa-miR-4786-3p, hsa-miR-548d-3p, hsa-miR-5091, hsa-miR-8060, hsa-miR-515-5p, hsa-miR-4650- It may also include 3p, hsa-miR-553, hsa-miR-675-5p, hsa-miR-6833-5p, hsa-miR-148a-5p, hsa-miR-548b-5p, hsa-miR-6782-3p, hsa-miR-2114-3p, hsa-miR-6727-3p, hsa-miR-4427, hsa-miR-29c-3p, hsa-miR-4723-5p, and hsa-miR-362-3p.

[0040] The biomarkers for predicting the IVIG response group of the present invention obtained as described above are hsa-miR-3649, hsa-miR-500b-3p, hsa-miR-4700-5p, hsa-miR-455-5p, hsa-miR-4738-5p, hsa-miR-4786-3p, hsa-miR-548d-3p, hsa-miR-5091, hsa-miR-8060, hsa-miR-515-5p, hsa-miR-4650-3p, hsa-miR-553, hsa-miR-675-5p, hsa-miR-6833-5p, hsa This may include -miR-148a-5p, hsa-miR-548b-5p, hsa-miR-6782-3p, hsa-miR-2114-3p, hsa-miR-6727-3p, hsa-miR-4427, hsa-miR-29c-3p, hsa-miR-4723-5p, hsa-miR-362-3p, hsa-miR-5010-5p, hsa-miR-4799-3p, hsa-miR-1178-3p, hsa-miR-548at-3p, hsa-miR-6783-3p, and hsa-miR-1306-5p.

[0041] 1.3.3 Dataset (C) The microRNA set obtained using Dataset (C) is a biomarker for predicting the refractory group that will not respond to a second IVIG treatment. In this invention, it is preferable to select at least one, at least three, at least five, or at least ten microRNAs as biomarkers from the groups consisting of 10, 14, 17, or 22 microRNAs that remain unexcluded due to the small absolute value of the coefficients predicted by logistic regression, by removing one microRNA expression level at a time from the explanatory variables from the 22 types of microRNAs obtained from Dataset (C), and repeating the logistic regression each time. It is also preferable to select all 10 microRNAs that remain unexcluded due to the small absolute value of the coefficients from the 22 types of microRNAs as biomarkers, and most preferably to select all 22 microRNAs as biomarkers.

[0042] The biomarkers for predicting IVIG refractory groups obtained as described above are hsa-miR-548au-5p, hsa-miR-548as-5p, hsa-miR-1292-3p, hsa-miR-4744, hsa-miR-550b-2-5p, hsa-miR-553, hsa-miR-99a-3p, hsa-miR-301a-5p, hsa-miR-371a-5p, hsa-miR-2682-5p, hsa-miR-6892-3p, and hsa-miR-3591-3p. Preferably, the microRNA contains at least one, at least three, at least five, at least ten, at least fifteen, or at least twenty microRNAs selected from the group consisting of hsa-miR-4660, hsa-miR-425-3p, hsa-miR-320a, hsa-miR-593-5p, hsa-miR-346, hsa-miR-6877-5p, hsa-miR-8060, hsa-miR-18b-5p, hsa-miR-626, and hsa-miR-548b-5p.

[0043] The biomarker for predicting the IVIG non-responding group of the present invention obtained as described above preferably contains at least one, at least three, at least five, at least ten, or at least fifteen microRNAs selected from the group consisting of hsa-miR-548au-5p, hsa-miR-548as-5p, hsa-miR-1292-3p, hsa-miR-4744, hsa-miR-550b-2-5p, hsa-miR-553, hsa-miR-99a-3p, hsa-miR-301a-5p, hsa-miR-371a-5p, hsa-miR-2682-5p, hsa-miR-6892-3p, hsa-miR-3591-3p, hsa-miR-4660, hsa-miR-425-3p, hsa-miR-320a, hsa-miR-593-5p, and hsa-miR-346.

[0044] The biomarker for predicting the IVIG non-responding group of the present invention obtained as described above preferably contains at least one, at least three, at least five, at least ten microRNAs selected from the group consisting of hsa-miR-548au-5p, hsa-miR-548as-5p, hsa-miR-1292-3p, hsa-miR-4744, hsa-miR-550b-2-5p, hsa-miR-553, hsa-miR-99a-3p, hsa-miR-301a-5p, hsa-miR-371a-5p, hsa-miR-2682-5p, hsa-miR-6892-3p, hsa-miR-3591-3p, hsa-miR-4660, and hsa-miR-320a.

[0045] The biomarker for predicting the IVIG non-responding group of the present invention obtained as described above preferably contains at least one, at least three, at least five, or at least eight microRNAs selected from the group consisting of hsa-miR-548au-5p, hsa-miR-548as-5p, hsa-miR-1292-3p, hsa-miR-550b-2-5p, hsa-miR-553, hsa-miR-99a-3p, hsa-miR-301a-5p, hsa-miR-371a-5p, hsa-miR-3591-3p, and hsa-miR-4660.

[0046] The biomarker for predicting IVIG-refractory groups obtained as described above comprises at least one microRNA selected from the group consisting of hsa-miR-548au-5p, hsa-miR-548as-5p, and hsa-miR-548b-5p, and further comprises hsa-miR-1292-3p, hsa-miR-4744, hsa-miR-550b-2-5p, hsa-miR-553, hsa-miR-99a-3p, hsa-miR-301a-5p, hsa-miR-371a-5p, hsa-miR-2682- Preferably, the microRNA contains at least one, at least three, at least five, at least ten, at least fifteen, or at least eighteen microRNAs selected from the group consisting of 5p, hsa-miR-6892-3p, hsa-miR-3591-3p, hsa-miR-4660, hsa-miR-425-3p, hsa-miR-320a, hsa-miR-593-5p, hsa-miR-346, hsa-miR-6877-5p, hsa-miR-8060, hsa-miR-18b-5p, and hsa-miR-626.

[0047] The biomarkers for predicting IVIG refractory groups obtained as described above may include hsa-miR-548au-5p, hsa-miR-548as-5p, hsa-miR-1292-3p, hsa-miR-550b-2-5p, hsa-miR-553, hsa-miR-99a-3p, hsa-miR-301a-5p, hsa-miR-371a-5p, hsa-miR-3591-3p, and hsa-miR-4660. As shown in Figure 15-3, even when using the above 10 types of microRNAs as markers, IVIG refractory groups can be detected with 100% sensitivity and specificity.

[0048] The biomarker for predicting the IVIG non-responsive group of the present invention obtained as described above may include hsa-miR-548au-5p, hsa-miR-548as-5p, hsa-miR-1292-3p, hsa-miR-4744, hsa-miR-550b-2-5p, hsa-miR-553, hsa-miR-99a-3p, hsa-miR-301a-5p, hsa-miR-371a-5p, hsa-miR-2682-5p, hsa-miR-6892-3p, hsa-miR-3591-3p, hsa-miR-4660, and hsa-miR-320a.

[0049] The biomarker for predicting the IVIG non-responsive group of the present invention obtained as described above may include hsa-miR-548au-5p, hsa-miR-548as-5p, hsa-miR-1292-3p, hsa-miR-4744, hsa-miR-550b-2-5p, hsa-miR-553, hsa-miR-99a-3p, hsa-miR-301a-5p, hsa-miR-371a-5p, hsa-miR-2682-5p, hsa-miR-6892-3p, hsa-miR-3591-3p, hsa-miR-4660, hsa-miR-425-3p, hsa-miR-320a, hsa-miR-593-5p, and hsa-miR-346.

[0050] The biomarkers for predicting IVIG refractory groups obtained as described above are hsa-miR-548au-5p, hsa-miR-548as-5p, hsa-miR-1292-3p, hsa-miR-4744, hsa-miR-550b-2-5p, hsa-miR-553, hsa-miR-99a-3p, hsa-miR-301a-5p, hsa-miR-371a-5p, hsa-miR-2682- It may also include 5p, hsa-miR-6892-3p, hsa-miR-3591-3p, hsa-miR-4660, hsa-miR-425-3p, hsa-miR-320a, hsa-miR-593-5p, hsa-miR-346, hsa-miR-6877-5p, hsa-miR-8060, hsa-miR-18b-5p, hsa-miR-626, and hsa-miR-548b-5p.

[0051] 1.3.4 Dataset (D) The microRNA set obtained using Dataset (D) is a biomarker for predicting the distinction between two groups with extreme differences in treatment outcomes: a group that is refractory to the second IVIG treatment (Refractory) and a group that responds to the first treatment (Responder). In this invention, it is preferable to select at least one, at least three, at least five, or at least ten microRNAs as biomarkers from the groups consisting of 11, 14, 16, and 19 microRNAs that remain unexcluded due to the small absolute value of the coefficients predicted by logistic regression, by excluding one microRNA expression level at a time from the explanatory variables, and repeating logistic regression each time. Furthermore, it is more preferable to select all 11 microRNAs that remain unexcluded due to the small absolute value of the coefficients as biomarkers, and most preferable to select all 19 microRNAs as biomarkers.

[0052] The biomarkers for predicting IVIG refractory groups obtained as described above are hsa-miR-550b-2-5p, hsa-miR-626, hsa-miR-553, hsa-miR-3145-5p, hsa-miR-3649, hsa-miR-3678-3p, hsa-miR-1185-1-3p, hsa-miR-6895-3p, hsa-miR-8060, hsa-miR-371a-5p, hsa-miR-548b- Preferably, the microRNA contains at least one, at least three, at least five, at least ten, at least fifteen, or at least eighteen microRNAs selected from the group consisting of 5p, hsa-miR-6853-3p, hsa-miR-663b, hsa-miR-4518, hsa-miR-494-3p, hsa-miR-346, hsa-miR-1915-5p, hsa-miR-4325, and hsa-miR-6772-3p.

[0053] The biomarkers for predicting IVIG refractory groups obtained as described above are hsa-miR-550b-2-5p, hsa-miR-626, hsa-miR-553, hsa-miR-3145-5p, hsa-miR-3649, hsa-miR-3678-3p, hsa-miR-1185-1-3p, hsa-miR-6895-3p, hsa-miR-8060, hs Preferably, the microRNAs include at least one, at least three, at least five, at least ten, or at least fifteen types selected from the group consisting of a-miR-371a-5p, hsa-miR-548b-5p, hsa-miR-6853-3p, hsa-miR-663b, hsa-miR-4518, hsa-miR-494-3p, and hsa-miR-346.

[0054] The biomarker for predicting IVIG-refractory groups obtained as described above preferably contains at least one, at least three, at least five, or at least ten microRNAs selected from the group consisting of hsa-miR-550b-2-5p, hsa-miR-626, hsa-miR-553, hsa-miR-3145-5p, hsa-miR-3649, hsa-miR-3678-3p, hsa-miR-1185-1-3p, hsa-miR-6895-3p, hsa-miR-8060, hsa-miR-371a-5p, hsa-miR-548b-5p, hsa-miR-6853-3p, hsa-miR-663b, and hsa-miR-494-3p.

[0055] The biomarker for predicting IVIG-refractory groups obtained as described above preferably includes at least one, at least three, at least five, or at least eight microRNAs selected from the group consisting of hsa-miR-550b-2-5p, hsa-miR-626, hsa-miR-553, hsa-miR-3145-5p, hsa-miR-3649, hsa-miR-3678-3p, hsa-miR-1185-1-3p, hsa-miR-6895-3p, hsa-miR-8060, hsa-miR-548b-5p, and hsa-miR-663b.

[0056] The biomarker for predicting IVIG refractory groups obtained as described above includes hsa-miR-548b-5p, and further includes hsa-miR-550b-2-5p, hsa-miR-626, hsa-miR-553, hsa-miR-3145-5p, hsa-miR-3649, hsa-miR-3678-3p, hsa-miR-1185-1-3p, hsa-miR-6895-3p, hsa-miR-8060, and hsa-mi Preferably, the microRNAs include at least one, at least three, at least five, at least ten, or at least fifteen types selected from the group consisting of R-371a-5p, hsa-miR-6853-3p, hsa-miR-663b, hsa-miR-4518, hsa-miR-494-3p, hsa-miR-346, hsa-miR-1915-5p, hsa-miR-4325, and hsa-miR-6772-3p.

[0057] The biomarkers for predicting IVIG refractory groups obtained as described above may include hsa-miR-550b-2-5p, hsa-miR-626, hsa-miR-553, hsa-miR-3145-5p, hsa-miR-3649, hsa-miR-3678-3p, hsa-miR-1185-1-3p, hsa-miR-6895-3p, hsa-miR-8060, hsa-miR-548b-5p, and hsa-miR-663b. As shown in Figure 16-3, even when using the above 11 types of microRNAs as markers, IVIG refractory groups can be detected with 100% sensitivity and specificity.

[0058] The biomarkers for predicting IVIG-refractory groups obtained as described above may include hsa-miR-550b-2-5p, hsa-miR-626, hsa-miR-553, hsa-miR-3145-5p, hsa-miR-3649, hsa-miR-3678-3p, hsa-miR-1185-1-3p, hsa-miR-6895-3p, hsa-miR-8060, hsa-miR-371a-5p, hsa-miR-548b-5p, hsa-miR-6853-3p, hsa-miR-663b, and hsa-miR-494-3p.

[0059] The biomarkers for predicting IVIG-refractory groups obtained as described above may include hsa-miR-550b-2-5p, hsa-miR-626, hsa-miR-553, hsa-miR-3145-5p, hsa-miR-3649, hsa-miR-3678-3p, hsa-miR-1185-1-3p, hsa-miR-6895-3p, hsa-miR-8060, hsa-miR-371a-5p, hsa-miR-548b-5p, hsa-miR-6853-3p, hsa-miR-663b, hsa-miR-4518, hsa-miR-494-3p, and hsa-miR-346.

[0060] The biomarkers for predicting IVIG refractory groups obtained as described above are hsa-miR-550b-2-5p, hsa-miR-626, hsa-miR-553, hsa-miR-3145-5p, hsa-miR-3649, hsa-miR-3678-3p, hsa-miR-1185-1-3p, hsa-miR-6895-3p, hsa-miR- It may also include 8060, hsa-miR-371a-5p, hsa-miR-548b-5p, hsa-miR-6853-3p, hsa-miR-663b, hsa-miR-4518, hsa-miR-494-3p, hsa-miR-346, hsa-miR-1915-5p, hsa-miR-4325, and hsa-miR-6772-3p.

[0061] 2. Prediction Method The prediction method for refractory Kawasaki disease according to the present invention can predict refractory Kawasaki disease based on the expression level of microRNA, particularly the expression level of microRNA belonging to the hsa-miR-548 family in a biological sample collected from a subject. The prediction method of the present invention sets a score for predicting Refractory (severe) and Responder (non-severe) using the expression profile of the target microRNA and the coefficients and intercept obtained from the above logistic regression model. This score is adopted as a linear predictor in the logistic regression model and used to distinguish between the IVIG-refractory group (severe group, Refractory) and the IVIG-responsive group (non-severe group, Responder). By setting a cutoff value so that the sensitivity and specificity for refractory Kawasaki disease are at the desired values, it is possible to predict whether a subject has refractory Kawasaki disease (IVIG-refractory group) or is in the IVIG-responsive group by comparing the expression level of microRNA in a biological sample collected from a subject with the cutoff value.

[0062] 2.1 Dataset (A) The target microRNA is at least one microRNA selected from the group consisting of hsa-miR-369-3p, hsa-miR-708-3p, hsa-miR-367-3p, hsa-miR-26a-5p, hsa-miR-126-5p, hsa-miR-302b-5p, hsa-miR-1225-3p, hsa-miR-194-5p, hsa-miR-579-3p, hsa-miR-133b, hsa-miR-296-5p, hsa-miR-558, hsa-miR-1914-3p, hsa-miR-224-5p, hsa-miR-31-5p, hsa-miR-548c-3p, and hsa-miR-3648. In other words, the method for predicting refractory Kawasaki disease according to the present invention can be predicted based on the expression level of at least one microRNA selected from the group consisting of hsa-miR-369-3p, hsa-miR-708-3p, hsa-miR-367-3p, hsa-miR-26a-5p, hsa-miR-126-5p, hsa-miR-302b-5p, hsa-miR-1225-3p, hsa-miR-194-5p, hsa-miR-579-3p, hsa-miR-133b, hsa-miR-296-5p, hsa-miR-558, hsa-miR-1914-3p, hsa-miR-224-5p, hsa-miR-31-5p, hsa-miR-548c-3p, and hsa-miR-3648.

[0063] The method for predicting refractory Kawasaki disease according to the present invention is as follows: hsa-miR-369-3p, hsa-miR-708-3p, hsa-miR-367-3p, hsa-miR-26a-5p, hsa-miR-126-5p, hsa-miR-302b-5p, hsa-miR-1225-3p, hsa-miR-194-5p, hsa-miR-579-3p, hsa-miR-133b, hsa-miR-296-5p, hsa-miR-558, hsa-miR-1914-3p, hsa-miR- It is preferable to make predictions based on the expression levels of at least three microRNAs selected from the group consisting of 224-5p, hsa-miR-31-5p, hsa-miR-548c-3p, and hsa-miR-3648, more preferably based on the expression levels of at least five microRNAs, even more preferably based on the expression levels of at least ten microRNAs, and even more preferably based on the expression levels of at least thirteen microRNAs.

[0064] The method for predicting refractory Kawasaki disease according to the present invention is preferably based on the expression levels of at least three, at least five, and at least ten microRNAs selected from the group consisting of hsa-miR-369-3p, hsa-miR-708-3p, hsa-miR-367-3p, hsa-miR-26a-5p, hsa-miR-126-5p, hsa-miR-302b-5p, hsa-miR-1225-3p, hsa-miR-194-5p, hsa-miR-579-3p, hsa-miR-133b, hsa-miR-296-5p, hsa-miR-558, and hsa-miR-1914-3p. It is preferable to make predictions based on the expression levels of at least three and at least five microRNAs selected from the group consisting of hsa-miR-369-3p, hsa-miR-708-3p, hsa-miR-367-3p, hsa-miR-26a-5p, hsa-miR-126-5p, and hsa-miR-194-5p. Furthermore, the biomarkers for predicting refractory Kawasaki disease obtained as described above are hsa-miR-369-3p, hsa-miR-708-3p, hsa-miR-367-3p, hsa-miR-26a-5p, hsa-miR-126-5p, hsa-miR-302b-5p, hsa-miR-1225-3p, hsa-miR-194-5p, hsa-miR-579-3p, hsa-miR-133b, hsa-miR-296-5p, hsa-miR It is preferable to make predictions based on the expression levels of at least five microRNAs selected from the group consisting of -558 and hsa-miR-1914-3p, and more preferably to make predictions based on the expression levels of at least five microRNAs selected from the group consisting of hsa-miR-369-3p, hsa-miR-708-3p, hsa-miR-367-3p, hsa-miR-26a-5p, hsa-miR-126-5p and hsa-miR-194-5p.

[0065] The method for predicting refractory Kawasaki disease according to the present invention is preferably based on the expression levels of the microRNAs hsa-miR-369-3p, hsa-miR-708-3p, hsa-miR-367-3p, hsa-miR-126-5p and hsa-miR-194-5p, and the method for predicting refractory Kawasaki disease according to the present invention is preferably based on the expression levels of hsa-miR-369-3p, hsa-miR-708-3p, hsa-miR-367-3p, hsa-miR-26a-5p, hsa It is more preferable to predict based on the expression levels of the microRNAs -miR-126-5p and hsa-miR-194-5p, including hsa-miR-369-3p, hsa-miR-708-3p, hsa-miR-367-3p, hsa-miR-26a-5p, hsa-miR-126-5p, hsa-miR-302b-5p, hsa-miR-1225-3p, hsa-miR-194-5p, and hsa-miR It is even more preferable to predict based on the expression levels of the microRNAs -579-3p, hsa-miR-133b, hsa-miR-296-5p, hsa-miR-558 and hsa-miR-1914-3p, as well as hsa-miR-369-3p, hsa-miR-708-3p, hsa-miR-367-3p, hsa-miR-26a-5p, hsa-miR-126-5p, and hsa-miR-302b-5 It is most preferable to make predictions based on the expression levels of all hsa-miR-1225-3p, hsa-miR-194-5p, hsa-miR-579-3p, hsa-miR-133b, hsa-miR-296-5p, hsa-miR-558, hsa-miR-1914-3p, hsa-miR-224-5p, hsa-miR-31-5p, hsa-miR-548c-3p, and hsa-miR-3648.

[0066] 2.2 Dataset (B) The microRNA set obtained using Dataset (B) is a biomarker for predicting the IVIG response group (Responder). The target microRNAs are hsa-miR-3649, hsa-miR-500b-3p, hsa-miR-4700-5p, hsa-miR-455-5p, hsa-miR-4738-5p, hsa-miR-4786-3p, hsa-miR-548d-3p, hsa-miR-5091, hsa-miR-8060, hsa-miR-515-5p, hsa-miR-4650-3p, hsa-miR-553, hsa-miR-675-5p, hsa-miR-6833-5p, hsa-miR-148a-5p, hsa-miR It is at least one microRNA selected from the group consisting of -548b-5p, hsa-miR-6782-3p, hsa-miR-2114-3p, hsa-miR-6727-3p, hsa-miR-4427, hsa-miR-29c-3p, hsa-miR-4723-5p, hsa-miR-362-3p, hsa-miR-5010-5p, hsa-miR-4799-3p, hsa-miR-1178-3p, hsa-miR-548at-3p, hsa-miR-6783-3p, and hsa-miR-1306-5p.In other words, the IVIG response group prediction method according to the present invention is hsa-miR-3649, hsa-miR-500b-3p, hsa-miR-4700-5p, hsa-miR-455-5p, hsa-miR-4738-5p, hsa-miR-4786-3p, hsa-miR-548d-3p, hsa-miR-5091, hsa-miR-8060, hsa-miR-515-5p, hsa-miR-4650-3p, hsa-miR-553, hsa-miR-675-5p, hsa-miR-6833-5p, hsa-miR-148a-5p, hsa-miR-5 This can be predicted based on the expression level of at least one microRNA selected from the group consisting of 48b-5p, hsa-miR-6782-3p, hsa-miR-2114-3p, hsa-miR-6727-3p, hsa-miR-4427, hsa-miR-29c-3p, hsa-miR-4723-5p, hsa-miR-362-3p, hsa-miR-5010-5p, hsa-miR-4799-3p, hsa-miR-1178-3p, hsa-miR-548at-3p, hsa-miR-6783-3p, and hsa-miR-1306-5p.

[0067] The IVIG response group prediction method according to the present invention includes hsa-miR-3649, hsa-miR-500b-3p, hsa-miR-4700-5p, hsa-miR-455-5p, hsa-miR-4738-5p, hsa-miR-4786-3p, hsa-miR-548d-3p, hsa-miR-5091, hsa-miR-8060, hsa-miR-515-5p, hsa-miR-4650-3p, hsa-miR-553, hsa-miR-675-5p, hsa-miR-6833-5p, hsa-miR-148a-5p, hsa-miR-548b-5p, hsa-miR-6782-3p, hsa-mi The results can be predicted based on the expression levels of at least one, at least three, at least five, at least ten, at least fifteen, at least twenty, or at least twenty-five microRNAs selected from the group consisting of R-2114-3p, hsa-miR-6727-3p, hsa-miR-4427, hsa-miR-29c-3p, hsa-miR-4723-5p, hsa-miR-362-3p, hsa-miR-5010-5p, hsa-miR-4799-3p, hsa-miR-1178-3p, hsa-miR-548at-3p, hsa-miR-6783-3p, and hsa-miR-1306-5p.

[0068] The IVIG response group prediction method according to the present invention includes hsa-miR-3649, hsa-miR-500b-3p, hsa-miR-4700-5p, hsa-miR-455-5p, hsa-miR-4738-5p, hsa-miR-4786-3p, hsa-miR-548d-3p, hsa-miR-5091, hsa-miR-8060, hsa-miR-515-5p, hsa-miR-4650-3p, hsa-miR-553, hsa-miR-675-5p, hsa-miR-6833-5p, h It is preferable to make predictions based on the expression levels of at least one, at least three, at least five, at least ten, at least fifteen, or at least twenty microRNAs selected from the group consisting of sa-miR-148a-5p, hsa-miR-548b-5p, hsa-miR-6782-3p, hsa-miR-2114-3p, hsa-miR-6727-3p, hsa-miR-4427, hsa-miR-29c-3p, hsa-miR-4723-5p, and hsa-miR-362-3p.

[0069] The IVIG response group prediction method according to the present invention includes hsa-miR-3649, hsa-miR-500b-3p, hsa-miR-4700-5p, hsa-miR-455-5p, hsa-miR-4738-5p, hsa-miR-4786-3p, hsa-miR-548d-3p, hsa-miR-5091, hsa-miR-8060, hsa-miR-515-5p, hsa It is preferable to predict based on the expression levels of at least one, at least three, at least five, at least ten, or at least fifteen microRNAs selected from the group consisting of -miR-4650-3p, hsa-miR-553, hsa-miR-6833-5p, hsa-miR-148a-5p, hsa-miR-548b-5p, and hsa-miR-2114-3p.

[0070] The method for predicting the IVIG response group according to the present invention preferably predicts based on the expression levels of at least one, at least three, at least five, or at least ten microRNAs selected from the group consisting of hsa-miR-3649, hsa-miR-500b-3p, hsa-miR-4700-5p, hsa-miR-455-5p, hsa-miR-4738-5p, hsa-miR-4786-3p, hsa-miR-548d-3p, hsa-miR-5091, hsa-miR-515-5p, hsa-miR-553, and hsa-miR-148a-5p.

[0071] The method for predicting the IVIG response group according to the present invention preferably predicts based on the expression levels of at least one, at least three, at least five, or at least eight microRNAs selected from the group consisting of hsa-miR-3649, hsa-miR-500b-3p, hsa-miR-4700-5p, hsa-miR-455-5p, hsa-miR-4786-3p, hsa-miR-5091, hsa-miR-515-5p, hsa-miR-553, and hsa-miR-148a-5p.

[0072] The IVIG response group prediction method according to the present invention includes at least one microRNA selected from the group consisting of hsa-miR-548d-3p, hsa-miR-548b-5p, and hsa-miR-548at-3p, and further includes hsa-miR-3649, hsa-miR-500b-3p, hsa-miR-4700-5p, hsa-miR-455-5p, hsa-miR-4738-5p, hsa-miR-4786-3p, hsa-miR-5091, hsa-miR-8060, hsa-miR-515-5p, hsa-miR-4650-3p, hsa-miR-553, hsa-miR-675-5p, hsa-miR-6833-5p, hsa It is preferable to make predictions based on the expression levels of at least one, at least three, at least five, at least ten, at least fifteen, at least twenty, or at least twenty-five microRNAs selected from the group consisting of -miR-148a-5p, hsa-miR-6782-3p, hsa-miR-2114-3p, hsa-miR-6727-3p, hsa-miR-4427, hsa-miR-29c-3p, hsa-miR-4723-5p, hsa-miR-362-3p, hsa-miR-5010-5p, hsa-miR-4799-3p, hsa-miR-1178-3p, hsa-miR-6783-3p, and hsa-miR-1306-5p.

[0073] The method for predicting the IVIG response group according to the present invention can be used to predict based on the expression levels of the microRNAs hsa-miR-3649, hsa-miR-500b-3p, hsa-miR-4700-5p, hsa-miR-455-5p, hsa-miR-4786-3p, hsa-miR-5091, hsa-miR-515-5p, hsa-miR-553, and hsa-miR-148a-5p.

[0074] The IVIG response group prediction method according to the present invention can be predicted based on the expression levels of the following microRNAs: hsa-miR-3649, hsa-miR-500b-3p, hsa-miR-4700-5p, hsa-miR-455-5p, hsa-miR-4738-5p, hsa-miR-4786-3p, hsa-miR-548d-3p, hsa-miR-5091, hsa-miR-515-5p, hsa-miR-553, and hsa-miR-148a-5p.

[0075] The IVIG response group prediction method according to the present invention can be predicted based on the expression levels of the following microRNAs: hsa-miR-3649, hsa-miR-500b-3p, hsa-miR-4700-5p, hsa-miR-455-5p, hsa-miR-4738-5p, hsa-miR-4786-3p, hsa-miR-548d-3p, hsa-miR-5091, hsa-miR-8060, hsa-miR-515-5p, hsa-miR-4650-3p, hsa-miR-553, hsa-miR-6833-5p, hsa-miR-148a-5p, hsa-miR-548b-5p, and hsa-miR-2114-3p.

[0076] The IVIG response group prediction method according to the present invention includes hsa-miR-3649, hsa-miR-500b-3p, hsa-miR-4700-5p, hsa-miR-455-5p, hsa-miR-4738-5p, hsa-miR-4786-3p, hsa-miR-548d-3p, hsa-miR-5091, hsa-miR-8060, hsa-miR-515-5p, hsa-miR-4650-3p, hsa-miR-553, h This can be predicted based on the expression levels of the following microRNAs: sa-miR-675-5p, hsa-miR-6833-5p, hsa-miR-148a-5p, hsa-miR-548b-5p, hsa-miR-6782-3p, hsa-miR-2114-3p, hsa-miR-6727-3p, hsa-miR-4427, hsa-miR-29c-3p, hsa-miR-4723-5p, and hsa-miR-362-3p.

[0077] The IVIG response group prediction method according to the present invention includes hsa-miR-3649, hsa-miR-500b-3p, hsa-miR-4700-5p, hsa-miR-455-5p, hsa-miR-4738-5p, hsa-miR-4786-3p, hsa-miR-548d-3p, hsa-miR-5091, hsa-miR-8060, hsa-miR-515-5p, hsa-miR-4650-3p, hsa-miR-553, hsa-miR-675-5p, hsa-miR-6833-5p, hsa-miR-148a-5p, hs This can be predicted based on the expression levels of the following microRNAs: a-miR-548b-5p, hsa-miR-6782-3p, hsa-miR-2114-3p, hsa-miR-6727-3p, hsa-miR-4427, hsa-miR-29c-3p, hsa-miR-4723-5p, hsa-miR-362-3p, hsa-miR-5010-5p, hsa-miR-4799-3p, hsa-miR-1178-3p, hsa-miR-548at-3p, hsa-miR-6783-3p, and hsa-miR-1306-5p.

[0078] 2.3 Dataset (C) The microRNA set obtained using Dataset (C) is a biomarker for predicting the refractory group that will not respond to a second IVIG treatment. The target microRNAs are hsa-miR-548au-5p, hsa-miR-548as-5p, hsa-miR-1292-3p, hsa-miR-4744, hsa-miR-550b-2-5p, hsa-miR-553, hsa-miR-99a-3p, hsa-miR-301a-5p, hsa-miR-371a-5p, hsa-miR-2682-5p, hsa-miR-6892-3p, h It is at least one microRNA selected from the group consisting of sa-miR-3591-3p, hsa-miR-4660, hsa-miR-425-3p, hsa-miR-320a, hsa-miR-593-5p, hsa-miR-346, hsa-miR-6877-5p, hsa-miR-8060, hsa-miR-18b-5p, hsa-miR-626, and hsa-miR-548b-5p. In other words, the IVIG-refractory group prediction method according to the present invention is hsa-miR-548au-5p, hsa-miR-548as-5p, hsa-miR-1292-3p, hsa-miR-4744, hsa-miR-550b-2-5p, hsa-miR-553, hsa-miR-99a-3p, hsa-miR-301a-5p, hsa-miR-371a-5p, hsa-miR-2682-5p, hsa-miR-6892-3p, hsa This can be predicted based on the expression level of at least one microRNA selected from the group consisting of -miR-3591-3p, hsa-miR-4660, hsa-miR-425-3p, hsa-miR-320a, hsa-miR-593-5p, hsa-miR-346, hsa-miR-6877-5p, hsa-miR-8060, hsa-miR-18b-5p, hsa-miR-626, and hsa-miR-548b-5p.

[0079] The IVIG-refractory group prediction method according to the present invention includes hsa-miR-548au-5p, hsa-miR-548as-5p, hsa-miR-1292-3p, hsa-miR-4744, hsa-miR-550b-2-5p, hsa-miR-553, hsa-miR-99a-3p, hsa-miR-301a-5p, hsa-miR-371a-5p, hsa-miR-2682-5p, hsa-miR-6892-3p, hsa-miR-3591-3p, and hsa-miR-4660. It is preferable to predict based on the expression levels of at least one, at least three, at least five, at least ten, at least fifteen, or at least twenty microRNAs selected from the group consisting of hsa-miR-425-3p, hsa-miR-320a, hsa-miR-593-5p, hsa-miR-346, hsa-miR-6877-5p, hsa-miR-8060, hsa-miR-18b-5p, hsa-miR-626, and hsa-miR-548b-5p.

[0080] The IVIG-refractory group prediction method according to the present invention includes hsa-miR-548au-5p, hsa-miR-548as-5p, hsa-miR-1292-3p, hsa-miR-4744, hsa-miR-550b-2-5p, hsa-miR-553, hsa-miR-99a-3p, hsa-miR-301a-5p, hsa-miR-371a-5p, hsa-miR-2682-5p, hsa It is preferable to predict based on the expression levels of at least one, at least three, at least five, at least ten, or at least fifteen microRNAs selected from the group consisting of -miR-6892-3p, hsa-miR-3591-3p, hsa-miR-4660, hsa-miR-425-3p, hsa-miR-320a, hsa-miR-593-5p, and hsa-miR-346.

[0081] The method for predicting IVIG-refractory groups according to the present invention preferably predicts based on the expression levels of at least one, at least three, at least five, or at least ten microRNAs selected from the group consisting of hsa-miR-548au-5p, hsa-miR-548as-5p, hsa-miR-1292-3p, hsa-miR-4744, hsa-miR-550b-2-5p, hsa-miR-553, hsa-miR-99a-3p, hsa-miR-301a-5p, hsa-miR-371a-5p, hsa-miR-2682-5p, hsa-miR-6892-3p, hsa-miR-3591-3p, hsa-miR-4660, and hsa-miR-320a.

[0082] The method for predicting IVIG-refractory groups according to the present invention is preferably based on the expression levels of at least one, at least three, at least five, or at least eight microRNAs selected from the group consisting of hsa-miR-548au-5p, hsa-miR-548as-5p, hsa-miR-1292-3p, hsa-miR-550b-2-5p, hsa-miR-553, hsa-miR-99a-3p, hsa-miR-301a-5p, hsa-miR-371a-5p, hsa-miR-3591-3p, and hsa-miR-4660.

[0083] The method for predicting IVIG-refractory groups according to the present invention includes at least one microRNA selected from the group consisting of hsa-miR-548au-5p, hsa-miR-548as-5p, and hsa-miR-548b-5p, and further includes hsa-miR-1292-3p, hsa-miR-4744, hsa-miR-550b-2-5p, hsa-miR-553, hsa-miR-99a-3p, hsa-miR-301a-5p, hsa-miR-371a-5p, hsa-miR-2682-5p, and hsa-miR-68 It is preferable to make predictions based on the expression levels of at least one, at least three, at least five, at least ten, at least fifteen, or at least eighteen microRNAs selected from the group consisting of 92-3p, hsa-miR-3591-3p, hsa-miR-4660, hsa-miR-425-3p, hsa-miR-320a, hsa-miR-593-5p, hsa-miR-346, hsa-miR-6877-5p, hsa-miR-8060, hsa-miR-18b-5p, and hsa-miR-626.

[0084] The method for predicting IVIG-refractory groups according to the present invention can be used to predict based on the expression levels of the following microRNAs: hsa-miR-548au-5p, hsa-miR-548as-5p, hsa-miR-1292-3p, hsa-miR-550b-2-5p, hsa-miR-553, hsa-miR-99a-3p, hsa-miR-301a-5p, hsa-miR-371a-5p, hsa-miR-3591-3p, and hsa-miR-4660.

[0085] The method for predicting IVIG-refractory groups according to the present invention can be used to predict based on the expression levels of the following microRNAs: hsa-miR-548au-5p, hsa-miR-548as-5p, hsa-miR-1292-3p, hsa-miR-4744, hsa-miR-550b-2-5p, hsa-miR-553, hsa-miR-99a-3p, hsa-miR-301a-5p, hsa-miR-371a-5p, hsa-miR-2682-5p, hsa-miR-6892-3p, hsa-miR-3591-3p, hsa-miR-4660, and hsa-miR-320a.

[0086] The method for predicting IVIG-refractory groups according to the present invention can be used to predict based on the expression levels of the following microRNAs: hsa-miR-548au-5p, hsa-miR-548as-5p, hsa-miR-1292-3p, hsa-miR-4744, hsa-miR-550b-2-5p, hsa-miR-553, hsa-miR-99a-3p, hsa-miR-301a-5p, hsa-miR-371a-5p, hsa-miR-2682-5p, hsa-miR-6892-3p, hsa-miR-3591-3p, hsa-miR-4660, hsa-miR-425-3p, hsa-miR-320a, hsa-miR-593-5p, and hsa-miR-346.

[0087] The IVIG-refractory group prediction method according to the present invention includes hsa-miR-548au-5p, hsa-miR-548as-5p, hsa-miR-1292-3p, hsa-miR-4744, hsa-miR-550b-2-5p, hsa-miR-553, hsa-miR-99a-3p, hsa-miR-301a-5p, hsa-miR-371a-5p, hsa-miR-2682-5p, and hsa-miR-6892. This can be predicted based on the expression levels of the following microRNAs: -3p, hsa-miR-3591-3p, hsa-miR-4660, hsa-miR-425-3p, hsa-miR-320a, hsa-miR-593-5p, hsa-miR-346, hsa-miR-6877-5p, hsa-miR-8060, hsa-miR-18b-5p, hsa-miR-626, and hsa-miR-548b-5p.

[0088] 2.4 Dataset (D) The microRNA set obtained using Dataset (D) is a biomarker for predicting the distinction between two groups with extreme differences in treatment outcomes: the group that is refractory to the second IVIG treatment (Refractory) and the group that responds to the first treatment (Responder). The target microRNAs are hsa-miR-550b-2-5p, hsa-miR-626, hsa-miR-553, hsa-miR-3145-5p, hsa-miR-3649, hsa-miR-3678-3p, hsa-miR-1185-1-3p, hsa-miR-6895-3p, hsa-miR-8060, and hsa-miR-371a-5p. It is at least one microRNA selected from the group consisting of hsa-miR-548b-5p, hsa-miR-6853-3p, hsa-miR-663b, hsa-miR-4518, hsa-miR-494-3p, hsa-miR-346, hsa-miR-1915-5p, hsa-miR-4325, and hsa-miR-6772-3p. In other words, the IVIG-refractory group prediction method according to the present invention is hsa-miR-550b-2-5p, hsa-miR-626, hsa-miR-553, hsa-miR-3145-5p, hsa-miR-3649, hsa-miR-3678-3p, hsa-miR-1185-1-3p, hsa-miR-6895-3p, hsa-miR-8060, hsa-miR-371a-5p, h This can be predicted based on the expression level of at least one microRNA selected from the group consisting of sa-miR-548b-5p, hsa-miR-6853-3p, hsa-miR-663b, hsa-miR-4518, hsa-miR-494-3p, hsa-miR-346, hsa-miR-1915-5p, hsa-miR-4325, and hsa-miR-6772-3p.

[0089] The IVIG-refractory group prediction method according to the present invention includes hsa-miR-550b-2-5p, hsa-miR-626, hsa-miR-553, hsa-miR-3145-5p, hsa-miR-3649, hsa-miR-3678-3p, hsa-miR-1185-1-3p, hsa-miR-6895-3p, hsa-miR-8060, hsa-miR-371a-5p, hsa-miR-548b-5p, and hsa-miR-68 It is preferable to make predictions based on the expression levels of at least one, at least three, at least five, at least ten, at least fifteen, or at least eighteen microRNAs selected from the group consisting of 53-3p, hsa-miR-663b, hsa-miR-4518, hsa-miR-494-3p, hsa-miR-346, hsa-miR-1915-5p, hsa-miR-4325, and hsa-miR-6772-3p.

[0090] The IVIG-refractory group prediction method according to the present invention includes hsa-miR-550b-2-5p, hsa-miR-626, hsa-miR-553, hsa-miR-3145-5p, hsa-miR-3649, hsa-miR-3678-3p, hsa-miR-1185-1-3p, hsa-miR-6895-3p, hsa-miR-8060, hsa-miR-371a-5p, hsa-miR-548b-5p, hsa-miR-6853-3p, hsa-miR-663b, hsa-miR-4518, hsa It is preferable to predict based on the expression levels of at least one, at least three, at least five, at least ten, or at least fifteen microRNAs selected from the group consisting of -miR-494-3p and hsa-miR-346.

[0091] The method for predicting IVIG-refractory groups according to the present invention is preferably based on the expression levels of at least one, at least three, at least five, or at least ten microRNAs selected from the group consisting of hsa-miR-550b-2-5p, hsa-miR-626, hsa-miR-553, hsa-miR-3145-5p, hsa-miR-3649, hsa-miR-3678-3p, hsa-miR-1185-1-3p, hsa-miR-6895-3p, hsa-miR-8060, hsa-miR-371a-5p, hsa-miR-548b-5p, hsa-miR-6853-3p, hsa-miR-663b, and hsa-miR-494-3p.

[0092] The method for predicting IVIG-refractory groups according to the present invention is preferably based on the expression levels of at least one, at least three, at least five, or at least eight microRNAs selected from the group consisting of hsa-miR-550b-2-5p, hsa-miR-626, hsa-miR-553, hsa-miR-3145-5p, hsa-miR-3649, hsa-miR-3678-3p, hsa-miR-1185-1-3p, hsa-miR-6895-3p, hsa-miR-8060, hsa-miR-548b-5p, and hsa-miR-663b.

[0093] The method for predicting IVIG-refractory groups according to the present invention includes hsa-miR-548b-5p, and further includes hsa-miR-550b-2-5p, hsa-miR-626, hsa-miR-553, hsa-miR-3145-5p, hsa-miR-3649, hsa-miR-3678-3p, hsa-miR-1185-1-3p, hsa-miR-6895-3p, hsa-miR-8060, hsa-miR-371a-5p, hs It is preferable to make predictions based on the expression levels of at least one, at least three, at least five, at least ten, or at least fifteen microRNAs selected from the group consisting of a-miR-6853-3p, hsa-miR-663b, hsa-miR-4518, hsa-miR-494-3p, hsa-miR-346, hsa-miR-1915-5p, hsa-miR-4325, and hsa-miR-6772-3p.

[0094] The method for predicting IVIG-refractory groups according to the present invention can be used to predict based on the expression levels of the following microRNAs: hsa-miR-550b-2-5p, hsa-miR-626, hsa-miR-553, hsa-miR-3145-5p, hsa-miR-3649, hsa-miR-3678-3p, hsa-miR-1185-1-3p, hsa-miR-6895-3p, hsa-miR-8060, hsa-miR-548b-5p, and hsa-miR-663b.

[0095] The method for predicting IVIG-refractory groups according to the present invention can be used to predict based on the expression levels of the following microRNAs: hsa-miR-550b-2-5p, hsa-miR-626, hsa-miR-553, hsa-miR-3145-5p, hsa-miR-3649, hsa-miR-3678-3p, hsa-miR-1185-1-3p, hsa-miR-6895-3p, hsa-miR-8060, hsa-miR-371a-5p, hsa-miR-548b-5p, hsa-miR-6853-3p, hsa-miR-663b, and hsa-miR-494-3p.

[0096] The method for predicting IVIG-refractory groups according to the present invention can be used to predict based on the expression levels of the following microRNAs: hsa-miR-550b-2-5p, hsa-miR-626, hsa-miR-553, hsa-miR-3145-5p, hsa-miR-3649, hsa-miR-3678-3p, hsa-miR-1185-1-3p, hsa-miR-6895-3p, hsa-miR-8060, hsa-miR-371a-5p, hsa-miR-548b-5p, hsa-miR-6853-3p, hsa-miR-663b, hsa-miR-4518, hsa-miR-494-3p, and hsa-miR-346.

[0097] The IVIG-refractory group prediction method according to the present invention includes hsa-miR-550b-2-5p, hsa-miR-626, hsa-miR-553, hsa-miR-3145-5p, hsa-miR-3649, hsa-miR-3678-3p, hsa-miR-1185-1-3p, hsa-miR-6895-3p, hsa-miR-8060, and hsa-miR-37 This can be predicted based on the expression levels of the following microRNAs: 1a-5p, hsa-miR-548b-5p, hsa-miR-6853-3p, hsa-miR-663b, hsa-miR-4518, hsa-miR-494-3p, hsa-miR-346, hsa-miR-1915-5p, hsa-miR-4325, and hsa-miR-6772-3p.

[0098] Any method that can specifically detect some or all of the microRNAs can be used to detect the expression levels of these microRNAs. For example, one method involves extracting microRNAs from the EVs of a subject according to existing methods and subjecting them to a microRNA array capable of measuring human microRNAs (a microarray consisting of complementary base sequences that hybridize to the microRNAs). Other methods that can be used for measurement include quantitative PCR methods such as competitive PCR, real-time PCR, and (droplet) digital PCR; molecular beacon methods and their simplified versions that enable highly sensitive detection using microRNA-specific molecular probes; semiconductor sensors that can electronically measure the hybridization of the microRNA or reverse-transcribed amplified DNA with probe molecules with high sensitivity and rapid accuracy; and methods using nanopore sensing or CRISPR / Cas systems.

[0099] Logistic regression analysis is performed using the microRNAs obtained in Analysis 1 and Analysis 2 as explanatory variables to determine the coefficients and intercepts. Next, the obtained coefficients and intercepts are applied to a dataset of microRNA profiles obtained from samples including both the severe and non-severe groups (as described later, in this invention, a dataset of 30 samples (Dataset (A)) or 59 samples (Datasets (B) to (D)) owned by Toyama University is used), and a score represented by the following formula 1 is determined for each sample. This score can be adopted as a linear predictor in the logistic regression model and used for group discrimination. The method for predicting refractory Kawasaki disease according to the present invention may include the step of determining a score represented by the formula 1.

[0100] [Math 1] Score = Σ((coefficient) × microRNA expression level) + intercept

[0101] By plotting the above scores for each sample, a desired cutoff value can be set so that the sensitivity and specificity are at specific values. For subjects whose status as refractory Kawasaki disease is unknown, the score calculated using the subject's microRNA profile is compared with the cutoff value. If the score is lower than the cutoff value, it is predicted to be refractory Kawasaki disease; if it is higher, it is predicted not to be refractory Kawasaki disease. In the present invention, it is preferable to set a cutoff value where the sensitivity and specificity are 80%, more preferably 90%, and most preferably 100%. The method for predicting refractory Kawasaki disease of the present invention may further include the step of setting a cutoff value so that the sensitivity and specificity are at specific values, and if the subject's score is lower than the cutoff value, it can be predicted to be refractory Kawasaki disease.

[0102] The present invention's method for predicting refractory Kawasaki disease can use serum from patients (subjects) either before or after IVIG treatment, but it is preferable to use serum from patients (subjects) before IVIG treatment. The present invention's method for predicting refractory Kawasaki disease can predict refractory Kawasaki disease even when using serum collected before IVIG treatment (3 to 7 days after the onset of fever, which is called the acute phase of Kawasaki disease). Therefore, compared to conventional techniques that perform two IVIG treatments according to guidelines even for IVIG-refractory patients, this method is advantageous because it allows for earlier diagnosis of refractory Kawasaki disease and early initiation of appropriate treatment. Furthermore, patients immediately after IVIG treatment have more time elapsed compared to those before IVIG treatment, and their symptoms are thought to be more advanced. Therefore, it is presumed that microRNAs in patient serum after IVIG treatment reflect the characteristics of refractory Kawasaki disease more strongly than those in patient serum before IVIG treatment. From this, when using patient serum immediately after IVIG treatment, refractory Kawasaki disease can be predicted by using expression data of fewer microRNAs, specifically, at least five types of microRNAs when using dataset (A).

[0103] 3. Kit The kit for predicting refractory Kawasaki disease according to the present invention includes means for detecting the expression level of the microRNA disclosed herein. Such means may be any method that can specifically detect part or all of the microRNA, such as a microarray consisting of a complementary base sequence that hybridizes with the microRNA, quantitative PCR, molecular beacon method and its simplified method, a method using an electronically measurable semiconductor sensor, nanopore sensing, or a method using a CRISPR / Cas system. In addition to means for detecting the expression level of microRNA, the kit may further include buffers, pH adjusters, reaction vessels, and instructions for predicting refractory Kawasaki disease according to the method disclosed herein, as needed and for the purpose.

[0104] 3.1 Dataset (A) In one embodiment of the present invention, the above kit preferably includes means for detecting the expression level of at least one microRNA selected from the group consisting of hsa-miR-369-3p, hsa-miR-708-3p, hsa-miR-367-3p, hsa-miR-26a-5p, hsa-miR-126-5p, hsa-miR-302b-5p, hsa-miR-1225-3p, hsa-miR-194-5p, hsa-miR-579-3p, hsa-miR-133b, hsa-miR-296-5p, hsa-miR-558, hsa-miR-1914-3p, hsa-miR-224-5p, hsa-miR-31-5p, hsa-miR-548c-3p, and hsa-miR-3648.

[0105] The above kits include hsa-miR-369-3p, hsa-miR-708-3p, hsa-miR-367-3p, hsa-miR-26a-5p, hsa-miR-126-5p, hsa-miR-302b-5p, hsa-miR-1 225-3p, hsa-miR-194-5p, hsa-miR-579-3p, hsa-miR-133b, hsa-miR-296-5p, hsa-miR-558, hsa-miR-1914-3p, hsa-miR-224-5p, Preferably, the system includes means for detecting the expression levels of at least three microRNAs selected from the group consisting of hsa-miR-31-5p, hsa-miR-548c-3p, and hsa-miR-3648; preferably, means for detecting the expression levels of at least five microRNAs; more preferably, means for detecting the expression levels of at least ten microRNAs; and even more preferably, means for detecting the expression levels of at least thirteen microRNAs.

[0106] The above kit preferably includes means for detecting the expression levels of at least three, at least five, and at least ten microRNAs selected from the group consisting of hsa-miR-369-3p, hsa-miR-708-3p, hsa-miR-367-3p, hsa-miR-26a-5p, hsa-miR-126-5p, hsa-miR-302b-5p, hsa-miR-1225-3p, hsa-miR-194-5p, hsa-miR-579-3p, hsa-miR-133b, hsa-miR-296-5p, hsa-miR-558, and hsa-miR-1914-3p. Preferably, the system includes means for detecting the expression levels of at least three and at least five microRNAs selected from the group consisting of a-miR-369-3p, hsa-miR-708-3p, hsa-miR-367-3p, hsa-miR-26a-5p, hsa-miR-126-5p, and hsa-miR-194-5p. Furthermore, the biomarkers for predicting refractory Kawasaki disease obtained as described above are hsa-miR-369-3p, hsa-miR-708-3p, hsa-miR-367-3p, hsa-miR-26a-5p, hsa-miR-126-5p, hsa-miR-302b-5p, hsa-miR-1225-3p, hsa-miR-194-5p, hsa-miR-579-3p, hsa-miR-133b, hsa-miR-296-5p, hsa-miR- It is preferable to include means for detecting the expression levels of at least five microRNAs selected from the group consisting of 558 and hsa-miR-1914-3p, and more preferably to include means for detecting the expression levels of at least five microRNAs selected from the group consisting of hsa-miR-369-3p, hsa-miR-708-3p, hsa-miR-367-3p, hsa-miR-26a-5p, hsa-miR-126-5p and hsa-miR-194-5p.

[0107] The above kit preferably includes means for detecting the expression levels of the microRNAs hsa-miR-369-3p, hsa-miR-708-3p, hsa-miR-367-3p, hsa-miR-126-5p, and hsa-miR-194-5p, and more preferably includes means for detecting the expression levels of the microRNAs hsa-miR-369-3p, hsa-miR-708-3p, hsa-miR-367-3p, hsa-miR-26a-5p, hsa-miR-126-5p, and hsa-miR-194-5p. Furthermore, it is even more preferable to include means for detecting the expression levels of the microRNAs hsa-miR-369-3p, hsa-miR-708-3p, hsa-miR-367-3p, hsa-miR-26a-5p, hsa-miR-126-5p, hsa-miR-302b-5p, hsa-miR-1225-3p, hsa-miR-194-5p, hsa-miR-579-3p, hsa-miR-133b, hsa-miR-296-5p, hsa-miR-558, and hsa-miR-1914-3p. Furthermore, it is most preferable to include means for detecting the expression levels of all of the following: hsa-miR-369-3p, hsa-miR-708-3p, hsa-miR-367-3p, hsa-miR-26a-5p, hsa-miR-126-5p, hsa-miR-302b-5p, hsa-miR-1225-3p, hsa-miR-194-5p, hsa-miR-579-3p, hsa-miR-133b, hsa-miR-296-5p, hsa-miR-558, hsa-miR-1914-3p, hsa-miR-224-5p, hsa-miR-31-5p, hsa-miR-548c-3p, and hsa-miR-3648.

[0108] 3.2 Dataset (B) In another embodiment of the present invention, the above kit includes hsa-miR-3649, hsa-miR-500b-3p, hsa-miR-4700-5p, hsa-miR-455-5p, hsa-miR-4738-5p, hsa-miR-4786-3p, hsa-miR-548d-3p, hsa-miR-5091, hsa-miR-8060, hsa-miR-515-5p, hsa-miR-4650-3p, hsa-miR-553, hsa-miR-675-5p, hsa-miR-6833-5p, hsa-miR-148a-5p, hsa-miR-548b-5p, hsa-miR-6782-3p, hsa-miR Preferably, the method includes means for detecting the expression levels of at least one, at least three, at least five, at least ten, at least fifteen, at least twenty, or at least twenty-five microRNAs selected from the group consisting of -2114-3p, hsa-miR-6727-3p, hsa-miR-4427, hsa-miR-29c-3p, hsa-miR-4723-5p, hsa-miR-362-3p, hsa-miR-5010-5p, hsa-miR-4799-3p, hsa-miR-1178-3p, hsa-miR-548at-3p, hsa-miR-6783-3p, and hsa-miR-1306-5p.

[0109] The above kits include hsa-miR-3649, hsa-miR-500b-3p, hsa-miR-4700-5p, hsa-miR-455-5p, hsa-miR-4738-5p, hsa-miR-4786-3p, hsa-miR-548d -3p, hsa-miR-5091, hsa-miR-8060, hsa-miR-515-5p, hsa-miR-4650-3p, hsa-miR-553, hsa-miR-675-5p, hsa-miR-6833-5p, hsa-miR- Preferably, the method includes means for detecting the expression levels of at least one, at least three, at least five, at least ten, at least fifteen, or at least twenty microRNAs selected from the group consisting of 148a-5p, hsa-miR-548b-5p, hsa-miR-6782-3p, hsa-miR-2114-3p, hsa-miR-6727-3p, hsa-miR-4427, hsa-miR-29c-3p, hsa-miR-4723-5p, and hsa-miR-362-3p.

[0110] The above kits are: hsa-miR-3649, hsa-miR-500b-3p, hsa-miR-4700-5p, hsa-miR-455-5p, hsa-miR-4738-5p, hsa-miR-4786-3p, hsa-miR-548d-3p, hsa-miR-5091, hsa-miR-8060, hsa-miR-515-5p, hsa-miR-46 Preferably, the method includes means for detecting the expression levels of at least one, at least three, at least five, at least ten, or at least fifteen microRNAs selected from the group consisting of 50-3p, hsa-miR-553, hsa-miR-6833-5p, hsa-miR-148a-5p, hsa-miR-548b-5p, and hsa-miR-2114-3p.

[0111] The above kit preferably includes means for detecting the expression levels of at least one, at least three, at least five, or at least ten microRNAs selected from the group consisting of hsa-miR-3649, hsa-miR-500b-3p, hsa-miR-4700-5p, hsa-miR-455-5p, hsa-miR-4738-5p, hsa-miR-4786-3p, hsa-miR-548d-3p, hsa-miR-5091, hsa-miR-515-5p, hsa-miR-553, and hsa-miR-148a-5p.

[0112] The above kit preferably includes means for detecting the expression levels of at least one, at least three, at least five, or at least eight microRNAs selected from the group consisting of hsa-miR-3649, hsa-miR-500b-3p, hsa-miR-4700-5p, hsa-miR-455-5p, hsa-miR-4786-3p, hsa-miR-5091, hsa-miR-515-5p, hsa-miR-553, and hsa-miR-148a-5p.

[0113] The above kit contains at least one microRNA selected from the group consisting of hsa-miR-548d-3p, hsa-miR-548b-5p, and hsa-miR-548at-3p, and further includes hsa-miR-3649, hsa-miR-500b-3p, hsa-miR-4700-5p, hsa-miR-455-5p, hsa-miR-4738-5p, hsa-miR-4786-3p, hsa-miR-5091, hsa-miR-8060, hsa-miR-515-5p, hsa-miR-4650-3p, hsa-miR-553, hsa-miR-675-5p, hsa-miR-6833-5p, and hsa-miR-14 Preferably, the method includes means for detecting the expression levels of at least one, at least three, at least five, at least ten, at least fifteen, at least twenty, or at least twenty-five microRNAs selected from the group consisting of 8a-5p, hsa-miR-6782-3p, hsa-miR-2114-3p, hsa-miR-6727-3p, hsa-miR-4427, hsa-miR-29c-3p, hsa-miR-4723-5p, hsa-miR-362-3p, hsa-miR-5010-5p, hsa-miR-4799-3p, hsa-miR-1178-3p, hsa-miR-6783-3p, and hsa-miR-1306-5p.

[0114] The above kit preferably includes means for detecting the expression levels of the microRNAs hsa-miR-3649, hsa-miR-500b-3p, hsa-miR-4700-5p, hsa-miR-455-5p, hsa-miR-4786-3p, hsa-miR-5091, hsa-miR-515-5p, hsa-miR-553, and hsa-miR-148a-5p.

[0115] The above kit preferably includes means for detecting the expression levels of the microRNAs hsa-miR-3649, hsa-miR-500b-3p, hsa-miR-4700-5p, hsa-miR-455-5p, hsa-miR-4738-5p, hsa-miR-4786-3p, hsa-miR-548d-3p, hsa-miR-5091, hsa-miR-515-5p, hsa-miR-553, and hsa-miR-148a-5p.

[0116] The above kit preferably includes means for detecting the expression levels of the microRNAs hsa-miR-3649, hsa-miR-500b-3p, hsa-miR-4700-5p, hsa-miR-455-5p, hsa-miR-4738-5p, hsa-miR-4786-3p, hsa-miR-548d-3p, hsa-miR-5091, hsa-miR-8060, hsa-miR-515-5p, hsa-miR-4650-3p, hsa-miR-553, hsa-miR-6833-5p, hsa-miR-148a-5p, hsa-miR-548b-5p, and hsa-miR-2114-3p.

[0117] The above kits include hsa-miR-3649, hsa-miR-500b-3p, hsa-miR-4700-5p, hsa-miR-455-5p, hsa-miR-4738-5p, hsa-miR-4786 -3p, hsa-miR-548d-3p, hsa-miR-5091, hsa-miR-8060, hsa-miR-515-5p, hsa-miR-4650-3p, hsa-miR-553, hsa-miR- Preferably, the method includes means for detecting the expression levels of the microRNAs 675-5p, hsa-miR-6833-5p, hsa-miR-148a-5p, hsa-miR-548b-5p, hsa-miR-6782-3p, hsa-miR-2114-3p, hsa-miR-6727-3p, hsa-miR-4427, hsa-miR-29c-3p, hsa-miR-4723-5p, and hsa-miR-362-3p.

[0118] The above kits include hsa-miR-3649, hsa-miR-500b-3p, hsa-miR-4700-5p, hsa-miR-455-5p, hsa-miR-4738-5p, hsa-miR-4786-3p, hsa-miR-548d-3p, hsa-m iR-5091, hsa-miR-8060, hsa-miR-515-5p, hsa-miR-4650-3p, hsa-miR-553, hsa-miR-675-5p, hsa-miR-6833-5p, hsa-miR-148a-5p, hsa-miR-54 Preferably, the method includes means for detecting the expression levels of the microRNAs 8b-5p, hsa-miR-6782-3p, hsa-miR-2114-3p, hsa-miR-6727-3p, hsa-miR-4427, hsa-miR-29c-3p, hsa-miR-4723-5p, hsa-miR-362-3p, hsa-miR-5010-5p, hsa-miR-4799-3p, hsa-miR-1178-3p, hsa-miR-548at-3p, hsa-miR-6783-3p, and hsa-miR-1306-5p.

[0119] 3.3 Dataset (C)

[0120] In another embodiment of the present invention, the above kit includes hsa-miR-548au-5p, hsa-miR-548as-5p, hsa-miR-1292-3p, hsa-miR-4744, hsa-miR-550b-2-5p, hsa-miR-553, hsa-miR-99a-3p, hsa-miR-301a-5p, hsa-miR-371a-5p, hsa-miR-2682-5p, hsa-miR-6892-3p, hsa-miR-3591-3p, and hsa-miR-4660. Preferably, the system includes means for detecting the expression levels of at least one, at least three, at least five, at least ten, at least fifteen, or at least twenty microRNAs selected from the group consisting of hsa-miR-425-3p, hsa-miR-320a, hsa-miR-593-5p, hsa-miR-346, hsa-miR-6877-5p, hsa-miR-8060, hsa-miR-18b-5p, hsa-miR-626, and hsa-miR-548b-5p.

[0121] The above kits include hsa-miR-548au-5p, hsa-miR-548as-5p, hsa-miR-1292-3p, hsa-miR-4744, hsa-miR-550b-2-5 p, hsa-miR-553, hsa-miR-99a-3p, hsa-miR-301a-5p, hsa-miR-371a-5p, hsa-miR-2682-5p, hsa-miR-68 Preferably, the method includes means for detecting the expression levels of at least one, at least three, at least five, at least ten, or at least fifteen microRNAs selected from the group consisting of 92-3p, hsa-miR-3591-3p, hsa-miR-4660, hsa-miR-425-3p, hsa-miR-320a, hsa-miR-593-5p, and hsa-miR-346.

[0122] The above kit preferably includes means for detecting the expression levels of at least one, at least three, at least five, or at least ten microRNAs selected from the group consisting of hsa-miR-548au-5p, hsa-miR-548as-5p, hsa-miR-1292-3p, hsa-miR-4744, hsa-miR-550b-2-5p, hsa-miR-553, hsa-miR-99a-3p, hsa-miR-301a-5p, hsa-miR-371a-5p, hsa-miR-2682-5p, hsa-miR-6892-3p, hsa-miR-3591-3p, hsa-miR-4660, and hsa-miR-320a.

[0123] The above kit preferably includes means for detecting the expression levels of at least one, at least three, at least five, or at least eight microRNAs selected from the group consisting of hsa-miR-548au-5p, hsa-miR-548as-5p, hsa-miR-1292-3p, hsa-miR-550b-2-5p, hsa-miR-553, hsa-miR-99a-3p, hsa-miR-301a-5p, hsa-miR-371a-5p, hsa-miR-3591-3p, and hsa-miR-4660.

[0124] The above kit contains at least one microRNA selected from the group consisting of hsa-miR-548au-5p, hsa-miR-548as-5p, and hsa-miR-548b-5p, and further includes hsa-miR-1292-3p, hsa-miR-4744, hsa-miR-550b-2-5p, hsa-miR-553, hsa-miR-99a-3p, hsa-miR-301a-5p, hsa-miR-371a-5p, hsa-miR-2682-5p, hsa-miR-6892-3p, h Preferably, the method includes means for detecting the expression levels of at least one, at least three, at least five, at least ten, at least fifteen, or at least eighteen microRNAs selected from the group consisting of sa-miR-3591-3p, hsa-miR-4660, hsa-miR-425-3p, hsa-miR-320a, hsa-miR-593-5p, hsa-miR-346, hsa-miR-6877-5p, hsa-miR-8060, hsa-miR-18b-5p, and hsa-miR-626.

[0125] The above kit preferably includes means for detecting the expression levels of the microRNAs hsa-miR-548au-5p, hsa-miR-548as-5p, hsa-miR-1292-3p, hsa-miR-550b-2-5p, hsa-miR-553, hsa-miR-99a-3p, hsa-miR-301a-5p, hsa-miR-371a-5p, hsa-miR-3591-3p, and hsa-miR-4660.

[0126] The above kit preferably includes means for detecting the expression levels of the microRNAs hsa-miR-548au-5p, hsa-miR-548as-5p, hsa-miR-1292-3p, hsa-miR-4744, hsa-miR-550b-2-5p, hsa-miR-553, hsa-miR-99a-3p, hsa-miR-301a-5p, hsa-miR-371a-5p, hsa-miR-2682-5p, hsa-miR-6892-3p, hsa-miR-3591-3p, hsa-miR-4660, and hsa-miR-320a.

[0127] The above kit preferably includes means for detecting the expression levels of the microRNAs hsa-miR-548au-5p, hsa-miR-548as-5p, hsa-miR-1292-3p, hsa-miR-4744, hsa-miR-550b-2-5p, hsa-miR-553, hsa-miR-99a-3p, hsa-miR-301a-5p, hsa-miR-371a-5p, hsa-miR-2682-5p, hsa-miR-6892-3p, hsa-miR-3591-3p, hsa-miR-4660, hsa-miR-425-3p, hsa-miR-320a, hsa-miR-593-5p, and hsa-miR-346.

[0128] The above kits include hsa-miR-548au-5p, hsa-miR-548as-5p, hsa-miR-1292-3p, hsa-miR-4744, hsa-miR-550b-2-5p, hsa -miR-553, hsa-miR-99a-3p, hsa-miR-301a-5p, hsa-miR-371a-5p, hsa-miR-2682-5p, hsa-miR-6892-3p, hsa- Preferably, the method includes means for detecting the expression levels of the microRNAs miR-3591-3p, hsa-miR-4660, hsa-miR-425-3p, hsa-miR-320a, hsa-miR-593-5p, hsa-miR-346, hsa-miR-6877-5p, hsa-miR-8060, hsa-miR-18b-5p, hsa-miR-626, and hsa-miR-548b-5p.

[0129] 3.4 Dataset (D)

[0130] In another embodiment of the present invention, the above kit includes hsa-miR-550b-2-5p, hsa-miR-626, hsa-miR-553, hsa-miR-3145-5p, hsa-miR-3649, hsa-miR-3678-3p, hsa-miR-1185-1-3p, hsa-miR-6895-3p, hsa-miR-8060, hsa-miR-371a-5p, hsa-miR-548b-5p, hsa-miR-68 Preferably, the method includes means for detecting the expression levels of at least one, at least three, at least five, at least ten, at least fifteen, or at least eighteen microRNAs selected from the group consisting of 53-3p, hsa-miR-663b, hsa-miR-4518, hsa-miR-494-3p, hsa-miR-346, hsa-miR-1915-5p, hsa-miR-4325, and hsa-miR-6772-3p.

[0131] The above kit preferably includes means for detecting the expression levels of at least one, at least three, at least five, at least ten, or at least fifteen microRNAs selected from the group consisting of hsa-miR-550b-2-5p, hsa-miR-626, hsa-miR-553, hsa-miR-3145-5p, hsa-miR-3649, hsa-miR-3678-3p, hsa-miR-1185-1-3p, hsa-miR-6895-3p, hsa-miR-8060, hsa-miR-371a-5p, hsa-miR-548b-5p, hsa-miR-6853-3p, hsa-miR-663b, hsa-miR-4518, hsa-miR-494-3p, and hsa-miR-346.

[0132] The above kit preferably includes means for detecting the expression levels of at least one, at least three, at least five, or at least ten microRNAs selected from the group consisting of hsa-miR-550b-2-5p, hsa-miR-626, hsa-miR-553, hsa-miR-3145-5p, hsa-miR-3649, hsa-miR-3678-3p, hsa-miR-1185-1-3p, hsa-miR-6895-3p, hsa-miR-8060, hsa-miR-371a-5p, hsa-miR-548b-5p, hsa-miR-6853-3p, hsa-miR-663b, and hsa-miR-494-3p.

[0133] The above kit preferably includes means for detecting the expression levels of at least one, at least three, at least five, or at least eight microRNAs selected from the group consisting of hsa-miR-550b-2-5p, hsa-miR-626, hsa-miR-553, hsa-miR-3145-5p, hsa-miR-3649, hsa-miR-3678-3p, hsa-miR-1185-1-3p, hsa-miR-6895-3p, hsa-miR-8060, hsa-miR-548b-5p, and hsa-miR-663b.

[0134] The above kit includes hsa-miR-548b-5p, and also hsa-miR-550b-2-5p, hsa-miR-626, hsa-miR-553, hsa-miR-3145-5p, hsa-miR-3649, hsa-miR-3678-3p, hsa-miR-1185-1-3p, hsa-miR-6895-3p, hsa-miR-8060, hsa-miR-371a-5p, and hsa-miR-6 Preferably, the method includes means for detecting the expression levels of at least one, at least three, at least five, at least ten, or at least fifteen microRNAs selected from the group consisting of 853-3p, hsa-miR-663b, hsa-miR-4518, hsa-miR-494-3p, hsa-miR-346, hsa-miR-1915-5p, hsa-miR-4325, and hsa-miR-6772-3p.

[0135] The above kit preferably includes means for detecting the expression levels of the following microRNAs: hsa-miR-550b-2-5p, hsa-miR-626, hsa-miR-553, hsa-miR-3145-5p, hsa-miR-3649, hsa-miR-3678-3p, hsa-miR-1185-1-3p, hsa-miR-6895-3p, hsa-miR-8060, hsa-miR-548b-5p, and hsa-miR-663b.

[0136] The above kit preferably includes means for detecting the expression levels of the microRNAs hsa-miR-550b-2-5p, hsa-miR-626, hsa-miR-553, hsa-miR-3145-5p, hsa-miR-3649, hsa-miR-3678-3p, hsa-miR-1185-1-3p, hsa-miR-6895-3p, hsa-miR-8060, hsa-miR-371a-5p, hsa-miR-548b-5p, hsa-miR-6853-3p, hsa-miR-663b, and hsa-miR-494-3p.

[0137] The above kit preferably includes means for detecting the expression levels of the microRNAs hsa-miR-550b-2-5p, hsa-miR-626, hsa-miR-553, hsa-miR-3145-5p, hsa-miR-3649, hsa-miR-3678-3p, hsa-miR-1185-1-3p, hsa-miR-6895-3p, hsa-miR-8060, hsa-miR-371a-5p, hsa-miR-548b-5p, hsa-miR-6853-3p, hsa-miR-663b, hsa-miR-4518, hsa-miR-494-3p, and hsa-miR-346.

[0138] The above kits are: hsa-miR-550b-2-5p, hsa-miR-626, hsa-miR-553, hsa-miR-3145-5p, hsa-miR-3649, hsa-miR-3678-3p, hsa-miR-1185-1-3p, hsa-miR-6895-3p, hsa-miR-8060, hsa-miR-371a-5p, hs Preferably, the method includes means for detecting the expression levels of microRNAs a-miR-548b-5p, hsa-miR-6853-3p, hsa-miR-663b, hsa-miR-4518, hsa-miR-494-3p, hsa-miR-346, hsa-miR-1915-5p, hsa-miR-4325, and hsa-miR-6772-3p.

[0139] This invention relates to a specific novel microRNA biomarker that enables predictive diagnosis of patients with refractory Kawasaki disease. The aim of this invention is to identify patients with refractory Kawasaki disease from the early stages of the disease and to prevent the development of coronary artery aneurysms, a serious sequela, by providing appropriate treatment interventions from an early stage. Furthermore, this invention is adaptable to biomarker diagnostic systems, enabling rapid and precise diagnosis. Using this invention, it is possible to develop a simple, rapid kit for highly sensitive and highly specific microRNA for patients with refractory Kawasaki disease. It is expected that this simple, rapid kit will enable accurate risk classification of patients with refractory Kawasaki disease from the early stages. Moreover, by using the novel microRNA biomarker of this invention, it will be possible to initiate aggressive treatment, such as immunosuppressants, from an early stage for patients predicted to develop IVIG-refractory refractory Kawasaki disease, aiming for early recovery within 10 days of Kawasaki disease onset and preventing the development of giant coronary artery aneurysms, the most serious sequela of refractory Kawasaki disease.

[0140] The present invention is further illustrated by the following embodiments, which should not be construed as further limitations.

[0141] 4.1 Examination using dataset (A) Experiment 1. Verification of EMPs The patient groups shown in Table 1 were used for analysis. Of the 50 acute Kawasaki disease (KD) patients, 5 (10%) were in the CAL group and 45 (90%) were in the Non-CAL group. Serum samples were collected before IVIG treatment, immediately after IVIG treatment, and 2 to 4 weeks after IVIG treatment. 100 μL of serum from acute Kawasaki disease (KD) patients and control groups consisting of febrile patients (Febrile, 25 cases) and healthy individuals (Healthy, 25 cases) was centrifuged (12,000 g, 1 minute) to remove coarse particles, and the number of extracellular vesicles (EVs) contained in the supernatant was compared. As a result, it was revealed that endothelial microparticles (EMPs), particularly those released from vascular endothelial cells during the acute phase of Kawasaki disease, were significantly elevated compared to the control group (feverish patients) and the healthy control group (p<0.0001) (Figure 5(A)). Furthermore, it was found that EMPs normalized immediately after treatment in Kawasaki disease patients who did not develop coronary artery lesions (labeled as the Non-CAL group in the figure), while elevated EMPs persisted after treatment in the group with refractory Kawasaki disease and coronary artery lesions (CAL group) (p<0.001) (Figure 5(B)).

[0142]

[0143] Experiment 2. Preliminary Study Next, we examined the microRNAs contained within EVs. In acute Kawasaki disease patients (KD) and control groups consisting of febrile patients (Febrile) and healthy individuals (Healthy) as shown in Table 1, 500 μL of serum collected before and immediately after IVIG treatment was centrifuged (12,000 g, 1 minute) to remove coarse particles, and microRNA was extracted using 400 μL of the supernatant. MicroRNA extraction was performed according to existing methods. 130 ng of the obtained microRNA was subjected to a microarray experiment capable of measuring 2578 types of human microRNAs at once, and the data was analyzed. Based on this comprehensive microRNA expression data, an analysis was performed to distinguish each patient group (Figure 7).

[0144] Using the microRNA data obtained above before and after IVIG treatment, dimensionality reduction analysis (PLS-DA) was performed (Analysis 1 in Figure 6) to conduct a preliminary study to distinguish between the IVIG response group (Responder, circles in Figure 7(A)), the IVIG non-responder group (Refractory, triangles in Figure 7(A)), and the non-Kawasaki disease group (squares in Figure 7(A)) (Figure 7(A)). As a result, it was determined that it is possible to distinguish the clinical characteristics of patients according to the obtained microRNA profiles. In this analysis, the IVIG non-responder group includes Refractory and CAL groups that do not form CAL.

[0145] Experiment 3. Extraction of microRNAs by combining dimensionality reduction analysis and random forest analysis (1) Based on the results of the preliminary study, we investigated the extraction of microRNAs by combining dimensionality reduction analysis (Analysis 1) and random forest analysis (Analysis 2) from the perspective of predicting the response to IVIG treatment.

[0146] Experiment 3-1. Dimensionality Reduction Analysis First, dimensionality reduction analysis (Analysis 1 in Figure 6) was performed using microRNA data obtained before IVIG treatment, immediately after treatment, and during the recovery period, to determine the treatment prognosis. The results for the IVIG response group (Responder, circles in Figure 7(B)) and the IVIG refractory group (Refractory, triangles in Figure 7(B)) are shown in Figure 7(B). To distinguish between the IVIG refractory and IVIG response groups, the VIP score, which evaluates the magnitude of the contribution in the direction of the first principal component, was used. MicroRNAs with a VIP score of 1.0000 or higher were extracted. Table 2 shows an example of the obtained VIP scores (VIP scores of 17 types of microRNAs extracted after random forest analysis and logistic regression analysis, which will be shown later).

[0147]

[0148] Experiment 3-2. Random Forest Analysis Furthermore, in order to perform classification more accurately, nine random forest prediction models with different algorithms were constructed, and random forest analysis (Analysis 2 in Figure 6) was performed. In the dataset in Table 1 used in this invention, from the viewpoint of achieving 100% sensitivity and specificity, it is preferable that n is 4 or less in the criterion of "whether it was selected as important in common by n or more random forest models," and since the prediction accuracy can be improved by extracting more candidate microRNAs as explanatory variables, microRNAs were extracted under the condition that n = 2. When n = 5, the number of candidate microRNAs was too small, resulting in prediction errors, so it was determined that n = 5 is undesirable. Generally, sample sizes of clinical data ranging from a few to several tens of cases are often insufficient for prediction by machine learning. In the random forest analysis of this study, a variable importance index called CVPFI (Cross-validated Permutation Feature Importance (Kaneko 2022 Anal. Sci. Adv.)), developed in the field of chemical engineering, was adopted to address the decrease in prediction accuracy due to insufficient sample size. miRNAs with a positive CVPFI rating were extracted.

[0149] Experiment 3-3. MicroRNAs that were commonly extracted in both logistic regression dimensionality reduction analysis and random forest analysis were selected. Analysis 1 yielded a set of 22 microRNAs that contributed highly to classification, and Analysis 2 yielded 22 microRNAs that were important for distinguishing between the severe and non-severe disease groups when n was set to 2.

[0150] Analysis 2. The 22 microRNAs extracted when n was set to 2 were hsa-miR-369-3p, hsa-miR-708-3p, hsa-miR-367-3p, hsa-miR-26a-5p, hsa-miR-126-5p, hsa-miR-302b-5p, hsa-miR-1225-3p, hsa-miR-194-5p, hsa-miR-579-3p, hsa-miR-133b, hsa-miR-296-5p, hsa-miR-558, hsa-miR-1914-3p, hsa-miR-224-5p, hsa-miR-31-5p, hsa-miR-548c-3p, hs a-miR-3648, hsa-miR-296-3p, hsa-miR-762, hsa-miR-1909-5p, hsa-miR-1225-5p and hsa-miR-1237-3p.

[0151] A logistic regression model was constructed using the expression levels of the 22 microRNA sets selected above, before and after IVIG treatment, as explanatory variables. As a result, five microRNA sets with high correlations between variables were removed, leaving 17 microRNA sets. Intercepts and coefficients were also obtained for these sets (Table 3). The microRNA sets obtained above are biomarkers for diagnosing refractory Kawasaki disease.

[0152]

[0153] Experiment 3-4. Prediction of refractory Kawasaki disease Next, we investigated whether it is possible to predict refractory Kawasaki disease using the microRNA expression profiles obtained above in the present invention. For the investigation, we used a microRNA dataset of 24 samples from the 30 samples owned by Toyama University shown in Table 4, excluding 3 fever patients other than Kawasaki disease patients (fever1, fever2, and fever3) and 3 healthy individuals (healthy1, healthy2, and healthy3). The breakdown of these samples was as follows. To determine the IVIG response group, we used a total of 9 samples from three patients (Res1, Res2, and Res3): 3 samples each from before IVIG administration (pre), after IVIG administration (post), and during the recovery period (con), from patients who recovered completely with treatment, approximately one month after the onset of Kawasaki disease. In addition, we used 3 samples from the recovery period (con) of three refractory patients (Nonres1, Nonres2, and Nonres3) and 3 samples from the recovery period (con) of four CAL patients (CAL1, CAL2, CAL3, and CAL4), for a total of 15 samples. To determine the IVIG non-responder group, we similarly used 9 samples from three refractory patients (Nonres1, Nonres2, and Nonres3) and four CAL patients (CAL1, CAL2, CAL3, and CAL4): 2-3 samples each from before IVIG administration (pre) and after IVIG administration (post). Using the expression levels of these patient-derived microRNAs, as well as the coefficients and intercepts obtained above, a score for predicting severe and non-severe cases was calculated, as shown in Equation 1 below.

[0154] [Math 1] Score = Σ((coefficient) × microRNA expression level) + intercept

[0155]

[0156] Figure 8 plots the scores for each patient. Using these scores as linear predictors in a logistic regression model, we searched for a score value that maximized both sensitivity and specificity in the logistic regression model. We were able to set a cutoff value of -1.018 that resulted in 100% sensitivity and specificity (the straight line in Figure 8). The method for predicting refractory Kawasaki disease according to the present invention was able to predict refractory Kawasaki disease even when patient samples were included before IVIG administration (labeled "pre" in Figure 8).

[0157] From Experiment 3, it is considered that by using the method for predicting refractory Kawasaki disease of the present invention, the expression level of microRNA in a subject before IVIG treatment can be measured, and the subject's score shown in Formula 1 above can be determined using the intercept and coefficients shown in Table 3. If the score is lower than -1.018, it is considered that the patient has refractory Kawasaki disease. Therefore, it is considered that by using the method for predicting refractory Kawasaki disease of the present invention, it is possible to predict IVIG-refractory Kawasaki disease with higher sensitivity and specificity than the conventional Kobayashi score, even before the start of IVIG treatment.

[0158] In Experiments 3-4 above, it was shown that, as an example, adopting a cutoff value of -1.018 could predict refractory Kawasaki disease with 100% sensitivity and specificity. However, in actual clinical practice, a test with a sensitivity and specificity of 80% or higher is generally considered reliable. Therefore, in the method for predicting refractory Kawasaki disease of the present invention, the cutoff value does not necessarily have to be -1.018. It is also possible to predict refractory Kawasaki disease by setting the cutoff value to have the desired sensitivity and specificity using the scores obtained from the expression levels of the 17 types of microRNAs shown in Table 3. Furthermore, the types of microRNAs to be input as explanatory variables may differ depending on the size and type of patient sample used. For each patient sample, a logistic regression model can be constructed using the microRNA sets extracted as important for distinguishing between the severe and non-severe disease groups in both Analysis 1 and 2 as explanatory variables, and the microRNA sets that predict refractory Kawasaki disease, as well as their coefficients and intercepts, can be determined.

[0159] Furthermore, it is possible to perform regression by gradually changing the number of microRNA types used for scoring from the extracted microRNA set according to their contribution, and then select the resulting coefficients, microRNA sets, and cutoff values. Figures 9-11 show the five types (hsa-miR-369-3p, hsa-miR-708-3p, hsa-miR-367-3p, hsa-miR-126-5p, and hsa-miR-194-5p) and six types (hsa-miR-369-3p, hsa-miR-708-3p, hsa-miR-367-3p, hsa-miR-26a-5p, hsa-miR-126-5p, and hsa-miR-194-5p) that remained at the top after each removal of one microRNA expression level from the explanatory variable, as shown in Table 3. Alternatively, logistic regression was performed using the expression levels of 13 types of microRNAs (hsa-miR-369-3p, hsa-miR-708-3p, hsa-miR-367-3p, hsa-miR-26a-5p, hsa-miR-126-5p, hsa-miR-302b-5p, hsa-miR-1225-3p, hsa-miR-194-5p, hsa-miR-579-3p, hsa-miR-133b, hsa-miR-296-5p, hsa-miR-558, and hsa-miR-1914-3p) as explanatory variables, and the score represented by Equation 1 was calculated and plotted. Actual severe cases are plotted in the area below the straight line in the figure, and non-severe cases are plotted above the dotted line. As is clear from Figure 9, even when logistic regression was performed using the five microRNAs with large absolute values ​​of coefficients shown in Table 3 as explanatory variables, the sensitivity and specificity were 100%, and refractory Kawasaki disease could be predicted. It is thought that increasing the number of microRNAs input into the logistic regression from five to six or thirteen will widen the difference in scores between the two groups, severe and non-severe cases, and thus improve diagnostic accuracy. Therefore, it can be said that a larger number of microRNAs used in the diagnosis of refractory Kawasaki disease is preferable.

[0160] 4.2 Analysis using datasets (B) to (D) A total of 59 microRNA datasets from 19 Responders, 11 Delayed Responders, and 29 Refractory Participants were analyzed. For each participant, two samples were used in the analysis: one before IVIG administration (pre) and one after a single dose of IVIG (post).

[0161] Experiment 4. Extraction of microRNAs by combining dimensionality reduction analysis and random forest analysis (2) Similar to Experiment 3 in 4.1 above, microRNAs were extracted by combining dimensionality reduction analysis (Analysis 1) and random forest analysis (Analysis 2), and logistic regression and prediction of refractory Kawasaki cosmetic surgery were investigated.

[0162] Experiment 4-1. Dimensionality Reduction Analysis First, dimensionality reduction analysis was performed using microRNA data obtained before IVIG treatment and immediately after the first IVIG treatment to determine the treatment outcome. The results of the classification for each dataset (B): 19 responders and 40 others (delayed responders and refractory), dataset (C): 29 refractory and 30 others (delayed responders and responders), and dataset (D): 29 refractory and 19 responders are shown in Figures 12-1 to 12-3. To distinguish between the IVIG-unresponsive and IVIG-responsive groups, the VIP score, which evaluates the magnitude of the contribution in the direction of the first principal component, was used. MicroRNAs with a VIP score greater than 2 were extracted (data not shown).

[0163] Experiment 4-2. Random Forest Analysis Next, we constructed eight random forest prediction models with different probabilities for dataset (B), nine for dataset (C), and eight for dataset (D). MicroRNAs were extracted under the condition that n, defined as "selected as important in common by n or more random forest models," was 2 for dataset (B), 5 for dataset (C), and 4 for dataset (D). When a number greater than the selected n was chosen, prediction errors occurred and these were not used. As in Experiment 3-2, CVPFI was used to address the decrease in prediction accuracy due to insufficient sample size.

[0164] Experiment 4-3. MicroRNAs that were commonly extracted in both logistic regression dimensionality reduction analysis and random forest analysis were selected. In dataset (B), 31 microRNA sets were obtained that contributed highly to classification in analysis 1 and were important for distinguishing between responders and others (delayed responders and refractories) when n was set to 2 in analysis 2. In dataset (C), 22 microRNA sets were obtained that contributed highly to classification in analysis 1 and were important for distinguishing between refractories and others (delayed responders and responders) when n was set to 5 in analysis 2. In dataset (D), 19 microRNA sets were obtained that contributed highly to classification in analysis 1 and were important for distinguishing between refractories and responders when n was set to 4 in analysis 2.

[0165] A logistic regression model was constructed using the expression levels of the microRNA sets selected above before and after IVIG treatment as explanatory variables. As a result, in dataset (B), two microRNAs with high correlations between variables were removed, and 29 microRNA sets were extracted (Table 5-1). Models based on the expression levels of 22 and 19 microRNAs were obtained for datasets (C) and (D), respectively (Tables 5-2 and 5-3). Furthermore, coefficients and intercepts for the extracted microRNAs were obtained for each model (Table 5). The microRNA sets obtained in dataset (B) are biomarkers that predict the IVIG response group. The microRNA sets obtained in datasets (C) and (D) are biomarkers that predict the IVIG non-responder group (refractory Kawasaki disease).

[0166]

[0167] Experiment 4-4. Prediction of IVIG response (Responder) and IVIG refractory. Furthermore, we investigated whether it is possible to predict IVIG response (Responder) or IVIG refractory using the microRNA expression profiles obtained from the above datasets (B) to (D). For the investigation, we used 19 responder samples, 11 delayed responder samples, and 29 refractory samples. Using the expression levels of microRNAs derived from these patients, as well as the coefficients and intercepts obtained above, we calculated a score expressed by Formula 1, similar to that in Experiment 3-4.

[0168] Figure 13-1 plots the scores of each patient obtained from dataset (B), Figure 13-2 plots the scores from dataset (C), and Figure 13-3 plots the scores from dataset (D). Using these scores as linear predictors in a logistic regression model, we searched for score values ​​that maximized both sensitivity and specificity in the logistic regression model. We were able to set cutoff values ​​of -2.08, -0.57, and -0.85, respectively, for sensitivity and specificity of 100% (straight lines in Figure 12). The prediction method for refractory Kawasaki disease according to the present invention was able to predict refractory or responder even when patient samples were included before IVIG administration.

[0169] From Experiment 4, it was found that by using the method for predicting refractory Kawasaki disease of the present invention, the microRNA expression level of a subject before IVIG treatment is measured, and the subject's score shown in Formula 1 above is determined using the intercepts and coefficients shown in Table 5-1. If the score is higher than -2.08, it can be predicted that the subject is an IVIG responder. Similarly, by determining the subject's score shown in Formula 1 above using the intercepts and coefficients shown in Table 5-2, and if the score is lower than -0.57, it can be predicted that the subject is an IVIG refractory, and in particular, it can predict the group that will be refractory to a second IVIG treatment. By determining the subject's score shown in Formula 1 above using the intercepts and coefficients shown in Table 5-3, and if the score is lower than -0.85, it can be predicted that the subject is an IVIG refractory, and in particular, it can predict the distinction between the group that will be refractory to a second IVIG treatment and the group that will respond to the first treatment, which are two groups with extreme differences in treatment outcomes. Therefore, it is believed that the present invention's method for predicting refractory Kawasaki disease can be used to predict IVIG-refractory Kawasaki disease with higher sensitivity and specificity than the conventional Kobayashi score, even before the initiation of IVIG treatment.

[0170] Similar to Experiment 3-4 above, Experiment 4-4 showed that, as an example, adopting cutoff values ​​of -2.08, -0.57, or -0.85 could predict Responder or Refractory with 100% sensitivity and specificity. However, in actual clinical practice, a test with a sensitivity and specificity of 80% or higher is generally considered reliable. Therefore, in the method for predicting refractory Kawasaki disease of the present invention, the cutoff value does not necessarily have to be one of the above values. It is also possible to predict refractory Kawasaki disease by setting the cutoff value to have the desired sensitivity and specificity using the scores obtained from the microRNA expression levels shown in Tables 5-1 to 5-3.

[0171] Furthermore, it is possible to perform regression by gradually changing the number of microRNA types used for scoring from the extracted microRNA set according to their contribution, and then select the resulting coefficients, microRNA sets, and cutoff values. Figures 14-1 to 14-4 show the results when the microRNA expression level that minimizes the absolute value of the coefficients shown in Table 5-1 is excluded one by one from the explanatory variables, and logistic regression is repeated each time, the 23 types that were not excluded and remained at the top (hsa-miR-3649, hsa-miR-500b-3p, hsa-miR-4700-5p, hsa-miR-455-5p, hsa-miR-4738-5p, hsa-miR-4786-3p, hsa-miR-548d-3p, hsa-miR-5091, hsa -miR-8060, hsa-miR-515-5p, hsa-miR-4650-3p, hsa-miR-553, hsa-miR-675-5p, hsa-miR-6833-5p, hsa-miR-148a-5p, hsa-miR-548 b-5p, hsa-miR-6782-3p, hsa-miR-2114-3p, hsa-miR-6727-3p, hsa-miR-4427, hsa-miR-29c-3p, hsa-miR-4723-5p, and hsa-miR-362- 3p), 16 types (hsa-miR-3649, hsa-miR-500b-3p, hsa-miR-4700-5p, hsa-miR-455-5p, hsa-miR-4738-5p, hsa-miR-4786-3p, hsa-miR-54 8d-3p, hsa-miR-5091, hsa-miR-8060, hsa-miR-515-5p, hsa-miR-4650-3p, hsa-miR-553, hsa-miR-6833-5p, hsa-miR-148a-5p, hsa- miR-548b-5p and hsa-miR-2114-3p), 11 types (hsa-miR-3649, hsa-miR-500b-3p, hsa-miR-4700-5p, hsa-miR-455-5p, hsa-miR-4738-5p, hsa-miR-4786-3p, hsa-miR-548d-3p, hsa-miR-5091, hsa-miR-515-5p, hsa-miR-553, and hsa-miR-148a-5p), or 9 types (hsa-miR-3649,Logistic regression was performed with the expression levels of microRNAs (hsa-miR-500b-3p, hsa-miR-4700-5p, hsa-miR-455-5p, hsa-miR-4786-3p, hsa-miR-5091, hsa-miR-515-5p, hsa-miR-553, and hsa-miR-148a-5p) as explanatory variables, and the scores represented by Equation 1 were calculated and plotted. In the figure, delayed responders and refractories are plotted in the area below the straight line, and responders are plotted above the dotted line. Figure 14-3 shows the regression results for 11 types of microRNAs: hsa-miR-3649, hsa-miR-500b-3p, hsa-miR-4700-5p, hsa-miR-455-5p, hsa-miR-4738-5p, hsa-miR-4786-3p, hsa-miR-548d-3p, hsa-miR-5091, hsa-miR-515-5p, hsa-miR-553, and hsa-miR-148a-5p. Figure 14-4 shows the regression results for nine types of microRNAs (hsa-miR-3649, hsa-miR-500b-3p, hsa-miR-4700-5p, hsa-miR-455-5p, hsa-miR-4786-3p, hsa-miR-5091, hsa-miR-515-5p, hsa-miR-553, and hsa-miR-148a-5p). As is clear from Figure 14-4, even when logistic regression was performed using the nine microRNAs with large absolute values ​​of coefficients shown in Table 5-1 as explanatory variables, the IVIG response group (Responders) could be detected with 100% sensitivity and specificity.

[0172] Similarly, in Figures 15-1 to 15-3, the microRNA expression level that minimizes the absolute value of the coefficients shown in Table 5-2 was excluded one by one from the explanatory variables, and logistic regression was repeated each time. The 17 types that were not excluded and remained at the top were (hsa-miR-548au-5p, hsa-miR-548as-5p, hsa-miR-1292-3p, hsa-miR-4744, hsa-miR-550b-2-5p, hsa-miR-553, hsa-miR-99a-3p, hsa-miR-301a-5p, hsa-miR-371a-5p, hsa-mi R-2682-5p, hsa-miR-6892-3p, hsa-miR-3591-3p, hsa-miR-4660, hsa-miR-425-3p, hsa-miR-320a, hsa-miR-593-5p, and hsa-miR-346); R-548au-5p, hsa-miR-548as-5p, hsa-miR-1292-3p, hsa-miR-4744, hsa-miR-550b-2-5p, hsa-miR-553, hsa-miR-99a-3p, hsa-miR-301a-5p, hsa-mi R-371a-5p, hsa-miR-2682-5p, hsa-miR-6892-3p, hsa-miR-3591-3p, hsa-miR-4660, and hsa-miR-320a) or 10 types (hsa-miR-548au-5p, hsa-miR-548a) s-5p, hsa-miR-1292-3p, hsa-miR-550b-2-5p, hsa-miR-553, hsa-miR-99a-3p, hsa-miR-301a-5p, hsa-miR-371a-5p, hsa-miR-3591-3p, and hsa-miR- The expression levels of microRNAs (4660) (dataset (C)), and the microRNA expression levels that minimize the absolute value of the coefficients shown in Table 5-3 are shown in Figures 16-1 to 16-3. When logistic regression was repeated by removing one type of microRNA from the explanatory variables each time, the 16 types that remained at the top without being removed (hsa-miR-550b-2-5p, hsa-miR-626, hsa-miR-553, hsa-miR-3145-5p, hsa-miR-3649, hsa-miR-3678-3p, hsa-miR-1185-1-3p, hsa-miR-6895-3p,hsa-miR-8060, hsa-miR-371a-5p, hsa-miR-548b-5p, hsa-miR-6853-3p, hsa-miR-663 b, hsa-miR-4518, hsa-miR-494-3p, and hsa-miR-346), 14 types (hsa-miR-550b-2-5p, hsa- miR-626, hsa-miR-553, hsa-miR-3145-5p, hsa-miR-3649, hsa-miR-3678-3p, hsa-miR -1185-1-3p, hsa-miR-6895-3p, hsa-miR-8060, hsa-miR-371a-5p, hsa-miR-548b-5p, Logistic regression was performed using the expression levels (Dataset (D)) of microRNAs (hsa-miR-6853-3p, hsa-miR-663b, and hsa-miR-494-3p), or 11 types (hsa-miR-550b-2-5p, hsa-miR-626, hsa-miR-553, hsa-miR-3145-5p, hsa-miR-3649, hsa-miR-3678-3p, hsa-miR-1185-1-3p, hsa-miR-6895-3p, hsa-miR-8060, hsa-miR-548b-5p, and hsa-miR-663b) as explanatory variables, and the scores represented by Equation 1 were calculated and plotted. Figure 15-3 shows the regression results for 10 types of microRNAs (hsa-miR-548au-5p, hsa-miR-548as-5p, hsa-miR-1292-3p, hsa-miR-550b-2-5p, hsa-miR-553, hsa-miR-99a-3p, hsa-miR-301a-5p, hsa-miR-371a-5p, hsa-miR-3591-3p, and hsa-miR-4660). Figure 16-3 shows the regression results for 11 types of microRNA (hsa-miR-550b-2-5p, hsa-miR-626, hsa-miR-553, hsa-miR-3145-5p, hsa-miR-3649, hsa-miR-3678-3p, hsa-miR-1185-1-3p, hsa-miR-6895-3p, hsa-miR-8060, hsa-miR-548b-5p, and hsa-miR-663b). The area below the line in the figure is Refractory.Responders (or Delayed responders and Responders) were plotted above the dotted line. Even when performing logistic regression using 10 or 11 microRNAs with large absolute coefficient values ​​(shown in Tables 5-2 and 6-3) as explanatory variables, refractory prediction was possible with 100% sensitivity and specificity.

[0173] It is believed that increasing the number of microRNA types input to the logistic regression leads to a wider difference in scores between the two groups to be distinguished, thus improving diagnostic accuracy. Therefore, it can be said that a larger number of microRNAs used in the diagnosis of refractory Kawasaki disease according to this invention is preferable.

[0174] This invention provides an early diagnostic method for refractory Kawasaki disease that utilizes microRNA expression profiles as biomarkers, enabling the early identification of patients with refractory Kawasaki disease and the rapid development of appropriate treatment strategies. This will prevent serious complications, improve the quality of life for patients, and contribute to reducing medical costs. In particular, if the development of Kawasaki disease coronary artery aneurysms can be prevented through early diagnosis and appropriate treatment, the need for expensive treatment costs and long-term drug therapy will be reduced, bringing significant benefits to the healthcare and welfare of the public. Furthermore, the development and dissemination of the novel biomarkers of this invention in related industries will promote innovation in diagnostic technology and the expansion of the medical device market, resulting in economic ripple effects. The widespread use of diagnostic technology utilizing microRNA profile biomarkers is expected to be applicable not only to Kawasaki disease but also to the early diagnosis of other diseases, contributing to the improvement of medical technology and the development of industry. Thus, early diagnosis of refractory Kawasaki disease using novel microRNA profile biomarkers is expected to make a significant contribution to improving the health and welfare of the public and to the development of related industries.

[0175] This invention, utilizing a novel microRNA biomarker for early diagnosis of refractory Kawasaki disease, is expected to be widely used in general hospitals with pediatric departments throughout Japan once it is commercialized. Furthermore, under the guidance of the Japanese Society of Pediatric Cardiology and the Japanese Society for Kawasaki Disease, efforts will be made to enable early initiation of aggressive treatments such as immunosuppressants for patients predicted to develop refractory Kawasaki disease. By aiming for early cure within 10 days of the onset of Kawasaki disease, when coronary artery aneurysms begin to form, it will be possible to prevent the most severe type of refractory Kawasaki disease: giant coronary artery aneurysms. Moreover, if the usefulness and effectiveness of this diagnostic kit are demonstrated, it is highly likely that it will be used for early diagnosis of refractory Kawasaki disease worldwide, and is expected to contribute to the international medical market.

Claims

1. A predictive biomarker for refractory Kawasaki disease, containing at least one microRNA belonging to the hsa-miR-548 family.

2. The biomarker for predicting refractory Kawasaki disease according to claim 1, wherein the microRNA belonging to the hsa-miR-548 family is at least one microRNA selected from the group consisting of hsa-miR-548d-3p, hsa-miR-548b-5p, hsa-miR-548at-3p, hsa-miR-548au-5p, and hsa-miR-548as-5p.

3. The microRNA belonging to the hsa-miR-548 family is at least one microRNA selected from the group consisting of hsa-miR-548d-3p, hsa-miR-548b-5p, and hsa-miR-548at-3p, and is a biomarker for predicting IVIG response group according to claim 1 or 2.

4. Additionally, hsa-miR-3649, hsa-miR-500b-3p, hsa-miR-4700-5p, hsa-miR-455-5p, hsa-miR-4738-5p, hsa-miR-4786-3p, hsa-miR-5091, hsa-m iR-8060, hsa-miR-515-5p, hsa-miR-4650-3p, hsa-miR-553, hsa-miR-675-5p, hsa-miR-6833-5p, hsa-miR-148a-5p, hsa-miR-6782-3p, h A biomarker for predicting IVIG response group according to claim 3, comprising at least one microRNA selected from the group consisting of sa-miR-2114-3p, hsa-miR-6727-3p, hsa-miR-4427, hsa-miR-29c-3p, hsa-miR-4723-5p, hsa-miR-362-3p, hsa-miR-5010-5p, hsa-miR-4799-3p, hsa-miR-1178-3p, hsa-miR-6783-3p, and hsa-miR-1306-5p.

5. A biomarker for predicting IVIG response group according to claim 3, comprising hsa-miR-3649, hsa-miR-500b-3p, hsa-miR-4700-5p, hsa-miR-455-5p, hsa-miR-4738-5p, hsa-miR-4786-3p, hsa-miR-548d-3p, hsa-miR-5091, hsa-miR-515-5p, hsa-miR-553, and hsa-miR-148a-5p.

6. hsa-miR-3649, hsa-miR-500b-3p, hsa-miR-4700-5p, hsa-miR-455-5p, hsa-miR-4738-5p, hsa-miR-4786-3p, hsa-miR-548d-3p, hsa-miR-50 91, hsa-miR-8060, hsa-miR-515-5p, hsa-miR-4650-3p, hsa-miR-553, hsa-miR-675-5p, hsa-miR-6833-5p, hsa-miR-148a-5p, hsa-miR-548b- A biomarker for predicting IVIG response group according to claim 3, comprising 5p, hsa-miR-6782-3p, hsa-miR-2114-3p, hsa-miR-6727-3p, hsa-miR-4427, hsa-miR-29c-3p, hsa-miR-4723-5p, hsa-miR-362-3p, hsa-miR-5010-5p, hsa-miR-4799-3p, hsa-miR-1178-3p, hsa-miR-548at-3p, hsa-miR-6783-3p, and hsa-miR-1306-5p.

7. Biomarkers for predicting IVIG response group, including hsa-miR-3649, hsa-miR-500b-3p, hsa-miR-4700-5p, hsa-miR-455-5p, hsa-miR-4786-3p, hsa-miR-5091, hsa-miR-515-5p, hsa-miR-553, and hsa-miR-148a-5p.

8. The microRNA belonging to the hsa-miR-548 family is at least one microRNA selected from the group consisting of hsa-miR-548au-5p, hsa-miR-548as-5p, and hsa-miR-548b-5p, and is a biomarker for predicting IVIG refractory group according to claim 1 or 2.

9. In addition, hsa-miR-1292-3p, hsa-miR-4744, hsa-miR-550b-2-5p, hsa-miR-553, hsa-miR-99a-3p, hsa-miR- 301a-5p, hsa-miR-371a-5p, hsa-miR-2682-5p, hsa-miR-6892-3p, hsa-miR-3591-3p, hsa-miR-4660, A biomarker for predicting IVIG refractory group according to claim 8, comprising at least one microRNA selected from the group consisting of hsa-miR-425-3p, hsa-miR-320a, hsa-miR-593-5p, hsa-miR-346, hsa-miR-6877-5p, hsa-miR-8060, hsa-miR-18b-5p, and hsa-miR-626.

10. A biomarker for predicting IVIG refractory group according to claim 8, comprising hsa-miR-548au-5p, hsa-miR-548as-5p, hsa-miR-1292-3p, hsa-miR-550b-2-5p, hsa-miR-553, hsa-miR-99a-3p, hsa-miR-301a-5p, hsa-miR-371a-5p, hsa-miR-3591-3p, and hsa-miR-4660.

11. hsa-miR-548au-5p, hsa-miR-548as-5p, hsa-miR-1292-3p, hsa-miR-4744, hsa-miR-550b-2-5p, hsa-miR- 553, hsa-miR-99a-3p, hsa-miR-301a-5p, hsa-miR-371a-5p, hsa-miR-2682-5p, hsa-miR-6892-3p, hsa-miR A biomarker for predicting IVIG refractory group according to claim 8, comprising -3591-3p, hsa-miR-4660, hsa-miR-425-3p, hsa-miR-320a, hsa-miR-593-5p, hsa-miR-346, hsa-miR-6877-5p, hsa-miR-8060, hsa-miR-18b-5p, hsa-miR-626, and hsa-miR-548b-5p.

12. The microRNA belonging to the hsa-miR-548 family is hsa-miR-548b-5p, and the biomarker for predicting IVIG refractory group according to claim 1 or 2.

13. Additionally hsa-miR-550b-2-5p, hsa-miR-626, hsa-miR-553, hsa-miR-3145-5p, hsa-miR-3649, hsa-miR -3678-3p, hsa-miR-1185-1-3p, hsa-miR-6895-3p, hsa-miR-8060, hsa-miR-371a-5p, hsa-miR-68 A biomarker for predicting IVIG refractory group according to claim 12, comprising at least one microRNA selected from the group consisting of 53-3p, hsa-miR-663b, hsa-miR-4518, hsa-miR-494-3p, hsa-miR-346, hsa-miR-1915-5p, hsa-miR-4325, and hsa-miR-6772-3p.

14. A biomarker for predicting IVIG refractory group according to claim 12, comprising hsa-miR-550b-2-5p, hsa-miR-626, hsa-miR-553, hsa-miR-3145-5p, hsa-miR-3649, hsa-miR-3678-3p, hsa-miR-1185-1-3p, hsa-miR-6895-3p, hsa-miR-8060, hsa-miR-548b-5p, and hsa-miR-663b.

15. hsa-miR-550b-2-5p, hsa-miR-626, hsa-miR-553, hsa-miR-3145-5p, hsa-miR-3649, hsa-miR -3678-3p, hsa-miR-1185-1-3p, hsa-miR-6895-3p, hsa-miR-8060, hsa-miR-371a-5p, hsa-mi A biomarker for predicting IVIG refractory group according to claim 12, comprising R-548b-5p, hsa-miR-6853-3p, hsa-miR-663b, hsa-miR-4518, hsa-miR-494-3p, hsa-miR-346, hsa-miR-1915-5p, hsa-miR-4325, and hsa-miR-6772-3p.

16. A method for predicting refractory Kawasaki disease based on the expression level of at least one microRNA belonging to the hsa-miR-548 family in a biological sample collected from a subject.

17. The method according to claim 16, comprising the following steps: (1) Performing a logistic regression analysis in which a score with the expression level of the microRNA as the explanatory variable is set as a linear predictor according to the following formula: and [Formula 1] Score = Σ((coefficient) × microRNA expression level) + intercept (2) Predicting that if the score is lower than the cutoff value, the subject is in the IVIG unresponsive group, or if the score is higher than the cutoff value, the subject is in the IVIG responsive group.

18. The method according to claim 17, wherein the cutoff value is a cutoff value at which the sensitivity and specificity are 100%.

19. The method according to claim 16 or 17, wherein the subject is a patient prior to IVIG treatment.

20. A kit for predicting refractory Kawasaki disease, comprising means for detecting the expression level of the microRNA described in claim 1.

Images