1. A DLBCL-associated biomarker composition comprising at least three of IL-2, IL-4, IL-6, IL-10, TNF- α, IFN- γ, and IL-17.

2. The biomarker composition of claim 1, wherein the biomarker composition comprises at least three of IL-6, IL-10, TNF-a, and IFN- γ.

3. The biomarker composition according to claim 1, wherein the biomarker composition is IL-6, IL-10, TNF-a and IFN- γ.

4. Use of a biomarker composition according to any of claims 1 to 3, for the prediction of the effect of the prognosis of DLBCL.

5. The DLBCL prognostic effect prediction model is characterized in that the prediction model is an SVM prediction model, the type of the SVM prediction model is C-classification, an SVM core is Linear or Polynomial or Radial Basis Function (RBF), and C is 0.500-0.999; the initial data are the amounts of each cytokine in the biomarker compositions according to claim 1.

6. The predictive model of claim 5, wherein the biomarker composition is IL-6, IL-10, TNF-a, and IFN- γ.

7. The prediction model according to any one of claims 5 to 6, wherein the prediction model is constructed by a method comprising the steps of:

s1, determining the content of the biomarker composition in the blood of the DLBCL patient after treatment, and performing statistical analysis;

and S2, applying the data result of each cytokine content obtained in the S1 to the support vector machine, and constructing an SVM prediction model for obtaining the DLBCL prognosis effect.

8. The prediction model of claim 5, wherein the SVM core is an RBF core.

9. The prediction model of claim 8, wherein the optimal C of the SVM prediction model is 0.895.

Background

Diffuse large B-cell lymphoma (DLBCL) is a mature B-cell malignancy, most common in adult non-Hodgkin lymphoma (NHL). In China, the incidence rate of DLBCL accounts for 40.1% of NHL, and the DLBCL is more and more concerned due to high incidence rate and high lethality rate.

The pathogenesis of DLBCL is unclear and several studies have shown that it may be associated with inflammatory-induced immune dysfunction, which is usually mediated by cytokines. As a first-line classical treatment for DLBCL, the current R-CHOP regimen provides effective remission in most patients, but a significant proportion of patients are not susceptible to R-CHOP, or a proportion of patients experience early relapse after treatment. This high heterogeneity in prognosis makes prognostic stratification (i.e., classifying patients first and then evaluating the prognosis separately in each class) a very important consideration in optimizing treatment.

The IPI scoring standard is used as a gold standard for the current prognosis evaluation of DLBCL patients, which fails to reflect the role of immune escape in the occurrence and development of diseases, so that it is difficult to avoid the situation that the prognosis effect prediction is inaccurate when immune escape occurs. Furthermore, the accuracy of the IPI scoring criteria has decreased since standard therapy with rituximab (R) in combination. Foreign studies show that the expression levels of some cytokines have close relationship with the onset, severity and prognosis of DLBCL. At present, related research in China is few, and in view of domestic and foreign environments, patient disease factor difference and high heterogeneity of DLBCL, a new biomarker is searched, a prognosis effect of a DLBCL patient can be predicted more accurately and efficiently is explored, and further, an IPI scoring system is very necessary to be perfected.

Disclosure of Invention

In order to improve the accuracy of the prediction of the DLBCL prognosis effect, the application provides a DLBCL related biomarker composition, application thereof and a DLBCL prognosis effect prediction model.

In a first aspect, the present application provides a DLBCL-related biomarker composition, using the following technical scheme: a DLBCL-associated biomarker composition comprising at least three of IL-2, IL-4, IL-6, IL-10, TNF- α, IFN- γ, and IL-17.

By adopting the technical scheme, the application discovers biomarkers related to DLBCL through a large number of experiments, and the content of the cytokines (IL-2, IL-4, IL-6, IL-10, TNF-alpha, IFN-gamma and IL-17) of the patient suffering from DLBCL is obviously different from that of healthy people (higher or obviously higher than that of the healthy people), so that the occurrence of DLBCL can be qualitatively judged through the content of the biomarker composition.

Preferably, the biomarker composition may be used for the prediction of the effect of the DLBCL prognosis.

The content of cytokines in the above biomarker composition in the patients of the therapeutically effective group has a significantly elevated tendency compared to that of the therapeutically ineffective group. Therefore, the method for predicting the DLBCL prognosis effect by adopting the content of the cell factors in the biomarker composition is feasible and effective, and provides a new idea for the DLBCL prognosis analysis and clinical treatment.

Preferably, the biomarker composition comprises at least three of IL-6, IL-10, TNF- α, and IFN- γ.

The relevant data results in the application show that the content of IL-6 and IL-10 in the DLBCL patients is obviously increased compared with that in the healthy population, the content of IFN-gamma is also obviously increased, and the content of TNF-alpha has no obvious statistical difference between the DLBCL patients and the healthy population, but still has obvious increasing trend. Therefore, selecting any three of IL-6, IL-10, TNF- α, and IFN- γ enables more accurate and efficient analysis and prediction of DLBCL, and enables high-accuracy prediction of the prognostic effect of DLBCL.

Preferably, the biomarker compositions are IL-6, IL-10, TNF- α, and IFN- γ.

The four cytokines of IL-6, IL-10, TNF-alpha and IFN-gamma are jointly used as biomarkers for predicting the DLBCL prognosis treatment effect, and the DLBCL prognosis treatment effect can be efficiently and accurately predicted.

In a second aspect, the present application provides a use of a biomarker composition, using the following technical scheme:

use of a biomarker composition as described above for the prediction of the effect of the prognosis of DLBCL.

The current R-CHOP regimen, as a first-line classical treatment regimen for DLBCL, can provide remission in most patients, but a significant proportion of patients remain insensitive to R-CHOP or relapse early. This high heterogeneity in prognosis, enabling the use of appropriate prognostic stratification, is very important for optimizing treatment. Among them, the IPI score, which is the gold standard for the current prognosis evaluation of DLBCL patients, also fails to reflect the role of immune escape in the development of disease, and the accuracy of IPI has decreased since standard therapy with rituximab (R) in combination. By adopting the technical scheme, when the effect prediction of DLBCL prognosis is carried out, a plurality of appropriate cytokines (namely biomarker compositions) are selected, and the effect prediction of DLBCL prognosis is carried out by combining the content data of the appropriate cytokines after DLBCL prognosis. The method can effectively improve the effect prediction accuracy of DLBCL patient prognosis, and makes up the defect that the IPI scoring accuracy is reduced after the standard treatment is carried out by jointly using rituximab.

In a third aspect, the present application provides a DLBCL prognosis effect prediction model, which adopts the following technical solutions: the prediction model is an SVM prediction model, the type of the SVM prediction model is C-classification, an SVM core is Linear or Polynomial or Radial Basis Function (RBF), and C is 0.500-0.999; the initial data are the amounts of each cytokine in the biomarker compositions according to claim 1.

The method detects the expression condition of each cytokine of the DLBCL patient through flow cytometry, discusses a detection and analysis system of the cytokine expression profile of the DLBCL patient, and provides a new thought for the prognosis analysis and clinical treatment of the DLBCL. The SVM is used for predicting and evaluating the DLBCL prognosis effect, a plurality of appropriate cell factors are obtained through screening, the content values of the cell factors are input into an SVM algorithm as initial data, an SVM prediction model is obtained through establishment and optimization, and then the DLBCL prognosis effect prediction model with high prediction accuracy is finally obtained.

Preferably, the biomarker compositions are IL-6, IL-10, TNF- α, and IFN- γ.

By adopting the technical scheme, the contents of IL-6, IL-10, TNF-alpha and IFN-gamma are taken as initial data, and an SVM prediction model is combined, so that a DLBCL prognosis effect prediction model with higher prediction accuracy can be obtained.

Preferably, the prediction model is constructed by a method comprising the following steps:

s1, determining the content of each cytokine in the biomarker composition in the blood of the treated DLBCL patient, and performing statistical analysis;

and S2, applying the data result of each cytokine content obtained in the S1 to the support vector machine, and constructing an SVM prediction model for obtaining the DLBCL prognosis effect.

Preferably, the SVM core is an RBF core.

Preferably, the optimal C of the SVM predictive model is 0.895.

By adopting the technical scheme, the optimal SVM prediction model is combined with DLBCL prognosis content data of IL-6, IL-10, TNF-alpha and IFN-gamma, the effect of DLBCL prognosis (short-term treatment) can be more accurately predicted, and the prediction accuracy is up to 75.00-82%.

In summary, the present application has the following beneficial effects:

1. the application firstly provides a DLBCL related biomarker composition which comprises at least three of IL-2, IL-4, IL-6, IL-10, TNF-alpha, IFN-gamma and IL-17; the content of the cell factor reflects the related information of the DLBCL, and the effect prediction (or evaluation) of the DLBCL prognosis can be carried out by the content of the cell factor.

2. Under the premise of the biomarker composition, the biomarker composition related to the DLBCL is further optimized to be IL-6, IL-10, TNF-alpha and IFN-gamma, the four cytokines are more related to the diagnosis of the DLBCL and the effect score of the DLBCL prognosis, and the effect of the DLBCL prognosis can be more accurately and efficiently reflected.

3. According to the application, the content data of IL-6, IL-10, TNF-alpha and IFN-gamma are combined with the SVM algorithm, so that a DLBCL prognosis effect prediction model is obtained, and the accuracy of the prediction model is up to 75.00% -82%.

Drawings

FIG. 1 shows the results of differential expression of IL-6, IL-10, TNF- α and IFN- γ among the groups DLBCL, CLL and healthy controls in the examples of the present application: wherein, the 1-A picture shows the differential expression result of IL-6 among DLBCL, CLL group and healthy control group, the 1-B picture shows the differential expression result of IL-10 among DLBCL, CLL group and healthy control group, the 1-C picture shows the differential expression result of TNF-alpha among DLBCL, CLL group and healthy control group, and the 1-D picture shows the differential expression result of IFN-gamma among DLBCL, CLL group and healthy control group;

FIG. 2 is a graph of the expression levels of IL-6, IL-10, TNF- α and IFN- γ cytokines in DLBCL patients with different IPI scores in the examples of the present application: wherein, the 2-A graph shows the expression level of IL-6 cytokine of DLBCL patients with different IPI scores, the 2-B graph shows the expression level of IL-10 cytokine of DLBCL patients with different IPI scores, the 2-C graph shows the expression level of TNF-alpha cytokine of DLBCL patients with different IPI scores, and the 2-D graph shows the expression level of IFN-gamma cytokine of DLBCL patients with different IPI scores; the 2-E plot shows the population distribution for different IPI scores: there were 10 people with an IPI score of 2, 16 people with an IPI score of 3, 17 people with an IPI score of 4, and 6 people with an IPI score of 5;

FIG. 3 is a graph showing the difference in cytokine expression of IL-6, IL-10, TNF- α and IFN- γ between the short-term treatment effective and ineffective groups in the examples of the present application: wherein, the 3-A graph shows the expression levels of IL-6 cytokines in the treatment effective group and the ineffective group, the 3-B graph shows the expression levels of IL-10 cytokines in the treatment effective group and the ineffective group, the 3-C graph shows the expression levels of TNF-alpha cytokines in the treatment effective group and the ineffective group, the 3-D graph shows the expression levels of IFN-gamma cytokines in the treatment effective group and the ineffective group, and the 3-E graph shows the distribution of the number of people in the treatment effective group and the ineffective group: there were 21 in the treatment-effective group and 28 in the treatment-ineffective group.

Detailed Description

The present application will be described in further detail with reference to the following drawings and examples.

Examples of biomarker compositions for DLBCL

The screening process of the DLBCL biomarker composition specifically comprises the following steps:

(1) 77 patients with B-NHL (B cell-non-Hodgkin lymphoma) confirmed by pathological biopsy in hematology and oncology department of people hospital in Zhejiang province from 1 month 2017 to 1 month 2020 are selected, wherein 11 patients with Chronic lymphocytic leukemia/Small lymphocytic lymphoma (Chronic lymphocytic leukemia/Small lymphocytic lymphoma, CLL/SLL) (6 men, 5 women, 55-85 years old, 72 years old at median age), and 52 patients with DLBCL (33 men, 19 women, 36-91 years old, 65 years old at median age). All patients enrolled in B-NHL were free of autoimmune disease, severe infection, secondary tumors and other hematological disorders. 56 physical healthy subjects were used as a control group (34 men and 22 women, the ages of 24-78 years, and the median age of 55 years). The sex and age differences of the patients in the 3 groups have no statistical significance (P is 0.461 and 0.114), and the patients have comparability. The last follow-up time is 2020, 7 months, and the shortest follow-up time is 6 months.

Of the 52 DLBCL patients, 3 patients were automatically discharged without treatment, 49 patients were treated with R-CHOP regimen for 21 days as a treatment course, and all patients were treated for 3 treatment courses and then predicted and evaluated for recent therapeutic effects according to the criteria of "blood disease diagnosis and therapeutic effect criteria". Of these, the complete remission and partial remission groups are the therapeutically effective groups, and the disease progression and non-remission groups are the therapeutically ineffective groups.

(2) Cytokine assay measures serum cytokine concentrations in untreated patients and healthy controls, as well as sustained-release cytokine concentrations in remission. The concentration of each cytokine is measured by adopting a flow cytometer, using a TH1/Th2/Th17 CBA cytokine kit to measure the concentration of 7 cytokines (comprising IL2, IL4, IL6, IL10, IL17, TNF-alpha and IFN-gamma) according to an operation instruction, collecting fasting elbow venous blood of all examinees, standing, centrifugally separating serum, storing the serum for later use at 4 ℃, and finishing the measurement of the related cytokines within 24 hours. Specific results are shown in table 1.

TABLE 17 expression of cytokines in DLBCL, CLL and normal control groups (ng/L, median (25%, 75%))

Note:^athe marked data are Z values and P values compared with a DLBCL group and a healthy control group, and the marked data are Z values and P values compared with a CLL group and a healthy control group; wherein the table shows the expression of a certain cytokineThe amount (ng/L), the expression level (ng/L) of the cytokine at the median (25%) and the expression level (ng/L) of the cytokine at the median (75%).

As seen from the data results in Table 1, IL-6 and IL-10 in the DLBCL group patients were significantly increased compared with the healthy control group, and IFN-. gamma.was also significantly increased.

Differential expression of IL-6, IL-10, TNF- α and IFN- γ among the DLBCL, CLL and healthy control groups was then counted, and the results are shown in FIG. 1. As seen in FIG. 1, IL-6 and IL-10 in the DLBCL group patients were significantly increased compared to the healthy control group, and IFN-. gamma.was also significantly increased; TNF-alpha was not statistically significantly different between the two groups, but still showed a significant increase.

Therefore, combining the data results of Table 1 and FIG. 1, it was determined that the biomarker composition of DLBCL includes IL-6, IL-10, TNF- α, and IFN- γ.

Prediction model of prognosis of DLBCL patients example (I) correlation of cytokine expression levels of DLBCL patients with IPI scores

The IPI score distribution for 49 DLBCL patients is shown in fig. 2: in DLBCL patients, the serum IL-6 and IL-10 of each group of patients with different IPI scores have significant difference; TNF- α was not statistically different between groups, but still had a significantly elevated trend in IPI at group 2 compared to group 3 or higher, as shown in FIG. 2. As seen from table 2: the patients with IPI of 2 points and the groups of patients with IPI of 3-5 points are obviously reduced in serum IL-6 and IL-10, and the IPI points are positively correlated with the levels of the serum IL-6 and IL-10. Wherein, the IPI score is 0 or 1, the patient is a low-risk type patient, and the survival rate of the patient after 5 years of treatment is generally considered to be 73%; the patient with an IPI score of 2 is a low-risk patient, and the survival rate of the patient after 5 years of treatment is generally considered to be 51%; a 3 IPI score for patients at mid-high risk, generally considered 43% survival after 5 years of treatment; patients with an IPI score of 4 or 5 are at high risk and generally are considered to have a survival rate of 26% after 5 years of treatment.

TABLE 2 cytokine variability between DLBCL patients at different IPI scores

Note: the meaning of "2 vs 3" in the table is "comparison of IPI score value 2 with IPI score value 3", "2 vs 4", "2 vs 5" and "2 vs 3" are similar in meaning, i.e. "comparison of IPI score value 2 with IPI score value 4" and "comparison of IPI score value 2 with IPI score value 5", respectively.

Therefore, as can be seen from the data in FIG. 2 and Table 2, the four cytokines IL-6, IL-10, TNF- α and IFN- γ all have correlation with the IPI score of the existing prognostic evaluation system, and it is reasonable to predict the prognostic effect of DLBCL by the four cytokines IL-6, IL-10, TNF- α and IFN- γ.

Model for predicting effect of prognosis of DLBCL (DLBCL)

In example 1, of the 52 patients who have been screened for DLBCL, 3 patients were automatically discharged without treatment; the remaining 49 DLBCL patients are treated by chemotherapy according to the R-CHOP scheme, 21 days are a treatment course, all the patients are treated for 3 treatment courses, and then the recent treatment effect is evaluated and predicted according to the standard of 'blood disease diagnosis and treatment effect standard', so that the IPI prognosis evaluation is carried out. Of these, the group with complete remission and partial remission was the therapeutically effective group (28 cases in total), and the group with progression of the disease and no remission was the therapeutically ineffective group (21 cases in total).

The concentration of each cytokine in the serum of the patients in the treatment effective group and the treatment ineffective group is determined, the concentration of each cytokine is determined by adopting a flow cytometer, using a TH1/Th2/Th17 CBA cytokine kit to determine the concentration of 7 cytokines (comprising IL2, IL4, IL6, IL10, IL17, TNF-alpha and IFN-gamma) according to the operation instruction, all the subjects collect fasting elbow venous blood, stand, centrifugally separate serum, store the standby serum at 4 ℃ and complete the determination of the related cytokine within 24 h. The specific results are shown in Table 3.

TABLE 3 serum cytokines (ng/L, median (25%, 75%)) for DLBCL patients with short-term therapeutic efficacy and therapeutic ineffectiveness

Group of	Treatment failure group	Therapeutically effective group	Z value	P value
					IL-2	1.22(0.67，2.40)	0.98(0.76，2.02)	1.384	0.166
IL4	1.39(0.90，2.44)	1.29(0.76，2.02)	0.778	0.437
					IL-6	87.43(33.53，232.49)	10.47(5.09，25.32)	4.303	0
IL-10	59.31(18.79，230.51)	8.73(2.94，25.04)	3.374	0.001
					TNF-α	2.84(0.88，4.68)	1.62(0.72，2.29)	2.293	0.022
IFN-γ	3.6(2.36，4.68)	2.45(1.44，3.04)	1.283	0.2
					IL-17	2.09(0.31，3.34)	1.35(0.75，3.34)	0.667	0.505

Note: shown in the table are the expression level (ng/L) of a certain cytokine, the expression level (ng/L) of the cytokine at the median (25%) and the expression level (ng/L) of the cytokine at the median (75%).

The results in Table 3 show that patients with no treatment expressed higher serum IL-6 and IL-10(P <0.01) and TNF-. alpha.was also significantly elevated compared to patients with effective treatment, see Table 3.

Differential expression of IL-6, IL-10, TNF- α and IFN- γ between the treatment-effective and treatment-ineffective groups was then counted, and the results are shown in FIG. 3. As seen in FIG. 3, IFN-. gamma.was not statistically different between the two groups, but there was still a marked increase in the treatment-ineffective group compared to the treatment-effective group.

Then, dividing the cytokine data into two groups, wherein 80% of the cytokine data are divided into one group which is used as a training set for training; 20% are in one group, test set, for testing. The preferred SVM model is as follows: the type is C-classification, the SVM core is a Radial Basis Function (RBF) core, and C is 0.895.

Application example of prediction model for DLBCL patient prognosis

The following optimal SVM model was used: the type is C-classification, the SVM core is RBF, and the optimal C is 0.895.

Then, the contents of IL-2, IL-4, IL-6, IL-10, TNF-alpha, IFN-gamma and IL-17 of 10 groups of patients are respectively detected, then the detected data results are applied to the optimized SVM model, 10 cases of the test group are predicted by the optimal SVM model, the predicted grouping condition is obtained, and the accuracy of the predicted testing group is obtained by comparing with the actual grouping condition and is 80.00 percent. Specific results are shown in table 4.

TABLE 4

Note: "1" in the case of the group indicates a group which was detected to show positive tumor residual lesions and thus was considered to have a poor prognostic effect, and "0" in the group indicates a group which was detected to show negative tumor residual lesions and thus was considered to have a good prognostic effect.

Predicting the prognosis effect by IPI scoring in the prior art, wherein the final test accuracy is 62.37-71.43%; in this example, the test accuracy was only 70%. Specific results are shown in table 5.

TABLE 5

Note: "1" in the case of the group indicates a group which was detected to show positive tumor residual lesions and thus was considered to have a poor prognostic effect, and "0" in the group indicates a group which was detected to show negative tumor residual lesions and thus was considered to have a good prognostic effect.

The present embodiment is only for explaining the present application, and it is not limited to the present application, and those skilled in the art can make modifications of the present embodiment without inventive contribution as needed after reading the present specification, but all of them are protected by patent law within the scope of the claims of the present application.