Non-small cell lung cancer bone metastasis prediction model and application thereof

Abstract

The invention discloses a non-small cell lung cancer bone metastasis prediction model and application thereof, and belongs to the technical field of cancer metastasis prediction models. The non-small cell lung cancer bone metastasis prediction model based on the autoantibody combination system inflammation index is constructed and verified, and the bone metastasis risk of a non-small cell lung cancer patient is predicted. The non-small cell lung cancer bone metastasis prediction model provides a new method for non-small cell lung cancer bone metastasis screening and auxiliary diagnosis, and has a good application prospect.

Description

Technical Field

[0001] The present invention belongs to the technical field of cancer metastasis prediction models, and particularly relates to a non-small cell lung cancer bone metastasis prediction model and its use. Background Art

[0002] During the cancer progression of 20% - 40% of non-small cell lung cancer (NSCLC) patients, varying degrees of bone metastasis will occur, and the median survival period is only 6 - 10 months. Early detection of bone metastasis can help clinicians implement early intervention, reduce the medical burden, and improve the prognosis of patients. Currently, the clinical diagnosis of NSCLC bone metastasis mainly relies on clinical symptoms and imaging examinations. However, when imaging shows that the tumor has metastasized, the patient often has reached the advanced stage and has a poor prognosis. Finding a NSCLC bone metastasis prediction model with unified standards, ideal effects, and being cost-effective is an urgent problem to be solved.

[0003] Autoantibodies can not only be detected in tumors but may also be related to the occurrence and development of malignant tumors. Autoantibodies in cancer patients may become potential markers for diagnosing and differentiating metastatic and non-metastatic cancers. Currently, there are no reports on the related research of autoantibodies and NSCLC bone metastasis.

[0004] Inflammatory reactions play an important role in the tumor microenvironment and are closely related to tumor occurrence, development, invasion, and metastasis. In recent years, multiple systemic inflammation indicators, due to their advantages of low cost and convenient detection, have been repeatedly reported to be related to the occurrence and development of NSCLC. These indicator changes can not only reflect the systemic immune and inflammatory status of NSCLC patients but may also be associated with the occurrence and progression of bone metastasis.

[0005] Constructing a NSCLC bone metastasis prediction model with unified standards, ideal effects, and being cost-effective to identify high-risk groups of NSCLC bone metastasis has important research value for helping clinicians implement early intervention, reducing the medical burden, and improving the prognosis of patients. Summary of the Invention

[0006] In order to solve the above problems existing in the prior art, the purpose of the present invention is to provide a non-small cell lung cancer bone metastasis prediction model and its use.

[0007] In order to achieve the above purpose, the present invention adopts the following technical solutions:

[0008] The present invention provides a prediction model for predicting the bone metastasis risk of NSCLC patients, and the prediction model includes the following modules:

[0009] I. Data input module

[0010] For inputting the characteristic data of patients, the characteristic data being gender, pathological type, TNM stage, anti-ENAs, LMR, and ELR; the anti-ENAs being anti-extractable nuclear antigen antibodies, the LMR being the ratio of lymphocytes to monocytes, and the ELR being the ratio of eosinophils to lymphocytes; the patients being non-small cell lung cancer patients;

[0011] II. Model construction module

[0012] Construct a nomogram for predicting the risk of bone metastasis in non-small cell lung cancer patients using the characteristic data of the input module;

[0013] III. Prediction module

[0014] Input the characteristic data of the patient to be predicted into the nomogram constructed in step II, and output the prediction result of the probability of the occurrence of bone metastasis risk in non-small cell lung cancer patients.

[0015] Furthermore, the prediction result in step III is the probability of the occurrence of bone metastasis in non-small cell lung cancer corresponding to the total score obtained by adding the scores of the characteristic data of gender, pathological type, TNM stage, anti-ENAs, LMR, and ELR in the nomogram.

[0016] Furthermore, the gender is a categorical variable, represented by the value of gender, with female assigned a value of 0 and male assigned a value of 1;

[0017] The pathological type is a categorical variable, represented by the value of the pathological type, with squamous cell carcinoma of the lung assigned a value of 1 and adenocarcinoma of the lung assigned a value of 2;

[0018] The TNM stage is a categorical variable, represented by the value of the TNM stage variable, with stage I assigned a value of 1, stage II assigned a value of 2, stage III assigned a value of 3, and stage IV assigned a value of 4;

[0019] The anti-ENAs is a categorical variable, and the value of anti-ENAs is the value of the anti-extractable nuclear antigen antibody, including immunoglobulin IgG antibodies against 14 different antigens: nRNP, Sm, SS-A, Ro-52, SS-B, Scl-70, Jo-1, CENP B, PCNA, ds-DNA, nucleosome, histone, ribosomal P protein, and AMAM2. If any one of the antibodies is positive, the anti-ENAs test is positive. Represented by the value of anti-ENAs, anti-ENAs negative is assigned a value of 0 and anti-ENAs positive is assigned a value of 1;

[0020] The LMR is a continuous variable, and the value of LMR is the ratio of lymphocyte count to monocyte count. Represented by the value of LMR, ≤2.7 is assigned a value of 0, and >2.7 is assigned a value of 1;

[0021] The ELR is a continuous variable, and the value of ELR is the ratio of eosinophil count to lymphocyte count, which is represented by the value of ELR. A value ≤ 0.106 is assigned 0, and a value > 0.106 is assigned 1.

[0022] Furthermore, the assignment schemes of the LMR and ELR are determined in the following manner:

[0023] Calculate the optimal cut-off values of the patient's LMR and ELR through the ROC curve and convert them into binary variables.

[0024] Furthermore, in the model construction module, the method of constructing a nomogram for predicting the bone metastasis risk of non-small cell lung cancer patients using the feature data of the input module is to perform multivariate Logistic regression analysis on the feature data of the input module to construct and draw the nomogram.

[0025] The present invention also provides the use of the above prediction model in the preparation of a device for predicting the bone metastasis risk of non-small cell lung cancer patients.

[0026] The present invention also provides a computer-readable storage medium, on which the above prediction model is stored.

[0027] The present invention has achieved the following beneficial effects:

[0028] The present invention constructs and validates a prediction model for non-small cell lung cancer bone metastasis based on autoantibodies combined with systemic inflammation indexes to predict the risk of bone metastasis in non-small cell lung cancer patients. The prediction model for non-small cell lung cancer bone metastasis of the present invention provides a new method for the screening and auxiliary diagnosis of bone metastasis in non-small cell lung cancer and has good application prospects.

[0029] Obviously, based on the above content of the present invention, according to the common general knowledge and conventional means in the art, without departing from the above basic technical idea of the present invention, various other forms of modifications, substitutions or changes can be made.

[0030] The following is a further detailed description of the above content of the present invention through specific embodiments in the form of examples. However, this should not be construed as limiting the scope of the above subject matter of the present invention to the following examples. All technologies implemented based on the above content of the present invention fall within the scope of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0031] Figure 1 For screening relevant risk factors for NSCLC bone metastasis.

[0032] Figure 2 For the receiver operating characteristic curve and Calibration calibration curve of the prediction model for non-small cell lung cancer bone metastasis of the present invention.

[0033] Figure 3 These are the results of the DCA and CIC analyses of the non-small cell lung cancer bone metastasis prediction model of the present invention. Detailed implementation manner

[0034] The raw materials and equipment used in the present invention are all known products and are obtained by purchasing commercially available products.

[0035] A retrospective analysis was conducted on 351 NSCLC patients in the Affiliated Hospital of Southwest Medical University from January 2020 to July 2024. General clinical data, autoantibody results, and blood routine indexes of the patients were collected, and the systemic inflammation index was calculated. Among them, 119 cases had bone metastasis. The collected data were randomly divided into a training set (n = 252) and a validation set (n = 99) at a ratio of 7:3. LASSO regression analysis was used to determine risk factors, and further multivariate logistics regression was used to screen independent risk factors to construct a nomogram model. The area under the receiver operating characteristic curve (ROC), calibration curve, decision curve analysis (DCA), and clinical impact curve (CIC) were used to evaluate the discrimination, calibration ability, prediction ability, and clinical utility of the model.

[0036] The gender, age, smoking status, pathological type, TNM stage, ANA (antinuclear antibody), ANA main fluorescence nuclear type titer, anti-ENAs (anti-extractable nuclear antigen antibody), WBC (white blood cell count), NEU (neutrophil count), LYM (lymphocyte count), MONO (monocyte count), EOS (eosinophil count), BASO (basophil count), PLT (platelet count), LER (ratio of lymphocyte count to monocyte count), NLR (ratio of neutrophil count to lymphocyte count), PLR (ratio of platelet count to lymphocyte count), LWR (ratio of lymphocyte count to white blood cell count), ELR (ratio of eosinophil count to lymphocyte count), BLR (ratio of basophil count to lymphocyte count), SII (neutrophil count × platelet count / lymphocyte count), and SIRI (neutrophil count × monocyte count / lymphocyte count) of the patients were collected. WBC, NEU, LYM, MONO, EOS, BASO, and PLT were converted into categorical variables according to the reference value range. The optimal cut-off value of each systemic inflammation index in the training set was calculated through the ROC curve and converted into a binary categorical variable. Age was converted into a categorical variable according to different age groups. The assignment of each index is shown in Table 1.

[0037] The bone metastasis is a categorical variable, represented by the value of bone metastasis. If non-small cell lung cancer bone metastasis has not occurred, it is assigned a value of 0; if non-small cell lung cancer bone metastasis has occurred, it is assigned a value of 1;

[0038] The gender is a categorical variable, represented by the value of gender. If female, it is assigned a value of 0; if male, it is assigned a value of 1;

[0039] The age is a continuous variable. According to different age groups, the age is converted into a categorical variable, represented by the value of age. If 30 < age ≤ 50, it is assigned a value of 1; if 50 < age ≤ 70, it is assigned a value of 2; if 70 < age ≤ 90, it is assigned a value of 3.

[0040] The smoking history is a categorical variable, represented by the value of smoking history. If there is no smoking history, it is assigned a value of 0; if there is a smoking history, it is assigned a value of 1;

[0041] The pathological type is a categorical variable, represented by the value of pathological type. If it is squamous cell carcinoma of the lung, it is assigned a value of 1; if it is adenocarcinoma of the lung, it is assigned a value of 2;

[0042] The TNM stage is a categorical variable, represented by the value of the TNM stage variable. If it is stage I, it is assigned a value of 1; if it is stage II, it is assigned a value of 2; if it is stage III, it is assigned a value of 3; if it is stage IV, it is assigned a value of 4;

[0043] The ANA is a categorical variable. The value of ANA is the value of antinuclear antibody, including homogeneous type, nuclear granular type, cytoplasmic granular type, centromere type, nuclear dot type, nuclear membrane type, nucleolar type, Golgi body type, etc. An ANA titer ≥ 1:100 is defined as positive. It is represented by the value of ANA. If ANA is negative, it is assigned a value of 0; if ANA is positive, it is assigned a value of 1;

[0044] The titer of the main fluorescence nuclear type of ANA is a categorical variable, represented by the value of the titer of the main fluorescence nuclear type of ANA. If the titer < 1:100, it is assigned a value of 1; if the titer = 1:100, it is assigned a value of 2; if the titer = 1:320, it is assigned a value of 3; if the titer > 1:1000, it is assigned a value of 4; if the titer > 1:3200, it is assigned a value of 5;

[0045] The anti-ENAs is a categorical variable. The value of anti-ENAs is the value of anti-extractable nuclear antigen antibody, including immunoglobulin IgG antibodies against 14 different antigens: nRNP, Sm, SS-A, Ro-52, SS-B, Scl-70, Jo-1, CENP B, PCNA, ds-DNA, nucleosome, histone, ribosomal P protein (Ribosomal Pprotein, RIB-P), and AMAM2. If any one of the antibodies is positive, the anti-ENAs test is positive. It is represented by the value of anti-ENAs. If anti-ENAs is negative, it is assigned a value of 0; if anti-ENAs is positive, it is assigned a value of 1;

[0046] The WBC is a continuous variable, and the value of WBC is the value of white blood cell count, with the unit of 10^9 / L. It is converted into a categorical variable according to the reference value range, represented by the value of WBC. When WBC ≤ 3.5, it is assigned a value of 0; when 3.5 < WBC ≤ 9.5, it is assigned a value of 1; when WBC > 9.5, it is assigned a value of 2.

[0047] The NEU is a continuous variable, and the value of NEU is the value of neutrophil count, with the unit of 10^9 / L. It is converted into a categorical variable according to the reference value range, represented by the value of NEU. When NEU ≤ 1.8, it is assigned a value of 0; when 1.8 < NEU ≤ 6.3, it is assigned a value of 1; when NEU > 6.3, it is assigned a value of 2.

[0048] The LYM is a continuous variable, and the value of LYM is the value of lymphocyte count, with the unit of 10^9 / L. It is converted into a categorical variable according to the reference value range, represented by the value of LYM. When LYM ≤ 1.1, it is assigned a value of 0; when 1.1 < LYM ≤ 3.2, it is assigned a value of 1; when LYM > 3.2, it is assigned a value of 2.

[0049] The MONO is a continuous variable, and the value of MONO is the value of monocyte count, with the unit of 10^9 / L. It is converted into a categorical variable according to the reference value range, represented by the value of MONO. When MONO ≤ 0.1, it is assigned a value of 0; when 0.1 < MONO ≤ 0.6, it is assigned a value of 1; when MONO > 0.6, it is assigned a value of 2.

[0050] The EOS is a continuous variable, and the value of EOS is the value of eosinophil count, with the unit of 10^9 / L. It is converted into a categorical variable according to the reference value range, represented by the value of EOS. When EOS ≤ 0.2, it is assigned a value of 0; when 0.2 < EOS ≤ 0.52, it is assigned a value of 1; when EOS > 0.52, it is assigned a value of 2.

[0051] The BASO is a continuous variable, and the value of BASO is the value of basophil, with the unit of 10^9 / L. It is converted into a categorical variable according to the reference value range, represented by the value of BASO. When BASO ≤ 0.06, it is assigned a value of 0; when BASO > 0.06, it is assigned a value of 1.

[0052] The PLT is a continuous variable, and the value of PLT is the value of platelet count, with the unit of 10^9 / L. It is converted into a categorical variable according to the reference value range, represented by the value of PLT. When PLT ≤ 125, it is assigned a value of 0; when 125 < PLT ≤ 350, it is assigned a value of 1; when PLT > 350, it is assigned a value of 2.

[0053] The LMR is a continuous variable, and the value of LMR is the ratio of lymphocyte count to monocyte count. The optimal cut-off value of LMR in the training set is calculated through the ROC curve and converted into a binary categorical variable, represented by the value of LMR. When LMR ≤ 2.7, it is assigned a value of 0; when LMR > 2.7, it is assigned a value of 1.

[0054] The NLR is a continuous variable, and the value of NLR is the ratio of neutrophil count to lymphocyte count. The optimal cut-off value of NLR in the training set is calculated through the ROC curve and converted into a binary variable, represented by the value of NLR. A value ≤ 4.35 is assigned 0, and a value > 4.35 is assigned 1;

[0055] The PLR is a continuous variable, and the value of PLR is the ratio of platelet count to lymphocyte count. The optimal cut-off value of PLR in the training set is calculated through the ROC curve and converted into a binary variable, represented by the value of PLR. A value ≤ 176 is assigned 0, and a value > 176 is assigned 1;

[0056] The LWR is a continuous variable, and the value of LWR is the ratio of lymphocyte count to white blood cell count. The optimal cut-off value of LWR in the training set is calculated through the ROC curve and converted into a binary variable, represented by the value of LWR. A value ≤ 176 is assigned 0, and a value > 176 is assigned 1;

[0057] The ELR is a continuous variable, and the value of ELR is the ratio of eosinophil count to lymphocyte count. The optimal cut-off value of ELR in the training set is calculated through the ROC curve and converted into a binary variable, represented by the value of ELR. A value ≤ 0.106 is assigned 0, and a value > 0.106 is assigned 1;

[0058] The BLR is a continuous variable, and the value of BLR is the ratio of basophil count to lymphocyte count. The optimal cut-off value of BLR in the training set is calculated through the ROC curve and converted into a binary variable, represented by the value of BLR. A value ≤ 0.029 is assigned 0, and a value > 0.029 is assigned 1;

[0059] The SII is a continuous variable, and the value of SII is neutrophil count × platelet count / lymphocyte count. The optimal cut-off value of SII in the training set is calculated through the ROC curve and converted into a binary variable, represented by the value of SII. A value ≤ 940 is assigned 0, and a value > 940 is assigned 1;

[0060] The SIRI is a continuous variable, and the value of SIRI is neutrophil count × monocyte count / lymphocyte count. The optimal cut-off value of SIRI in the training set is calculated through the ROC curve and converted into a binary variable, represented by the value of SIRI. A value ≤ 1.52 is assigned 0, and a value > 1.52 is assigned 1;

[0061] In the training set, using the occurrence of bone metastasis in non-small cell lung cancer as the dependent variable and each indicator as the independent variable, the "glmnet" package is used to screen the predictive variables by LASSO regression and 10-fold cross-validation. The results are asFigure 1 A and Figure 1 as shown in B. To avoid model bloat and overfitting, and for the convenience of clinical application, this study will use the predictor variables corresponding to 1 times the standard error of lgλ for the next step of analysis, and 8 predictor variables are determined: gender, pathological type, TNM stage, ANA, anti-ENAs, BASO, LMR, NLR, ELR, SIRI. ROC analysis was performed on the above variables ( Figure 1 C) and the AUC value was greater than 0.5.

[0062] Table 1 Assignment of each index

[0063]

[0064] Taking whether bone metastasis occurred as the dependent variable and the predictor variables screened by LASSO regression as the independent variables, further included in the multivariate Logistic regression analysis, and the backward stepwise regression analysis method was used. The results are shown in Table 2. Finally, the following 6 indexes were determined: gender, pathological type, TNM stage, anti-ENAs (anti-extractable nuclear antigen antibody), LMR (ratio of lymphocyte count to monocyte count), ELR (ratio of eosinophil count to lymphocyte count).

[0065] Table 2 Results of multivariate Logistic regression analysis

[0066]

[0067] Example 1: Construction method of a prediction model for bone metastasis in non-small cell lung cancer based on autoantibodies combined with systemic inflammation index

[0068] I. Input module

[0069] Collect the indexes of the patient's gender, pathological type, TNM stage, anti-ENAs, LMR and ELR, and enter these indexes into the input module. The assignment of each index is shown in Table 1.

[0070] II. Establishment of a prediction model for bone metastasis in non-small cell lung cancer

[0071] In the training set, use the 6 indexes of the patient's gender, pathological type, TNM stage, anti-ENAs, LMR and ELR to construct and draw a nomogram to predict the risk of bone metastasis in patients with non-small cell lung cancer.

[0072] III. Using the prediction model for bone metastasis in non-small cell lung cancer to predict the risk of bone metastasis in patients

[0073] Collect the gender, pathological type, TNM stage, anti-ENAs, LMR, and ELR of the patients in the validation set, and use the nomogram model constructed in Step 2 to predict the risk of bone metastasis in patients.

[0074] The beneficial effects of the present invention are demonstrated by the following examples.

[0075] Example 1: Model Validation

[0076] Input the data of the aforementioned collected training set (n = 252) and validation set (n = 99) into the non-small cell lung cancer bone metastasis prediction model of the present invention to evaluate the prediction efficacy of the present invention.

[0077] The results are as Figure 2 shown. The AUCs in the training set and the validation set are 0.875 and 0.817 respectively, indicating that the model has good prediction efficacy. The calibration curve shows good consistency between the predicted results and the actual results in the training set and the validation set.

[0078] The results of DCA and CIC analysis are as Figure 3 shown, indicating that both the training set and the validation set have good clinical benefits in the prediction model of the present invention.

[0079] In summary, the present invention constructs and validates a non-small cell lung cancer bone metastasis prediction model based on autoantibodies combined with systemic inflammation indexes to predict the risk of bone metastasis in non-small cell lung cancer patients. The non-small cell lung cancer bone metastasis prediction model of the present invention provides a new method for bone metastasis screening and auxiliary diagnosis of non-small cell lung cancer and has good application prospects.

Claims

1. A prediction model for predicting the risk of bone metastasis in patients with non-small cell lung cancer, characterized in that, The prediction model includes the following modules:

1. Data input module It is used to input the characteristic data of patients, and the characteristic data are gender, pathological type, TNM stage, anti-ENAs, LMR, and ELR; The anti-ENAs are anti-extractable nuclear antigen antibodies, the LMR is the lymphocyte-to-monocyte ratio, and the ELR is the eosinophil-to-lymphocyte ratio; The patients are non-small cell lung cancer patients; 2. Model construction module Using the characteristic data of the input module to construct a nomogram for predicting the bone metastasis risk of non-small cell lung cancer patients; 3. Prediction module Input the characteristic data of the patient to be predicted into the nomogram constructed in step 2, and output the prediction result of the occurrence probability of the bone metastasis risk of non-small cell lung cancer patients.

2. The prediction model according to claim 1, wherein: The prediction result in step 3 is the probability of bone metastasis in non-small cell lung cancer corresponding to the total score obtained by adding the scores of the characteristic data of gender, pathological type, TNM stage, anti-ENAs, LMR, and ELR in the nomogram.

3. The prediction model according to claim 1, wherein: The gender is a categorical variable, represented by the value of gender, with female assigned 0 and male assigned 1; The pathological type is a categorical variable, represented by the value of the pathological type, with squamous cell carcinoma of the lung assigned 1 and adenocarcinoma of the lung assigned 2; The TNM stage is a categorical variable, represented by the value of the TNM stage variable, with stage I assigned 1, stage II assigned 2, stage III assigned 3, and stage IV assigned 4; The anti-ENAs are categorical variables, and the value of anti-ENAs is the value of the anti-extractable nuclear antigen antibody, including immunoglobulin IgG antibodies against 14 different antigens: nRNP, Sm, SS-A, Ro-52, SS-B, Scl-70, Jo-1, CENP B, PCNA, ds-DNA, nucleosome, histone, ribosomal P protein, and AMAM2. If any one of the antibodies is positive, the anti-ENAs test is positive; represented by the value of anti-ENAs, anti-ENAs negative is assigned 0, and anti-ENAs positive is assigned 1; The LMR is a continuous variable, and the value of LMR is the ratio of lymphocyte count to monocyte count. Represented by the value of LMR, ≤2.7 is assigned 0, and >2.7 is assigned 1; The ELR is a continuous variable, and the value of ELR is the ratio of eosinophil count to lymphocyte count. Represented by the value of ELR, ≤0.106 is assigned 0, and >0.106 is assigned 1.

4. The prediction model according to claim 3, wherein: The assignment schemes of the LMR and ELR are determined by the following method: Calculate the optimal cut-off values of the patient's LMR and ELR through the ROC curve and convert them into binary classification variables.

5. The prediction model according to any one of claims 1-4, characterized in that, In the model construction module, the method of using the characteristic data of the input module to construct a nomogram for predicting the bone metastasis risk of non-small cell lung cancer patients is to perform multivariate Logistic regression analysis on the characteristic data of the input module to construct and draw the nomogram.

6. Use of the prediction model according to any one of claims 1-5 in the preparation of a device for predicting the risk of bone metastasis in patients with non-small cell lung cancer.

7. A computer-readable storage medium, on which the prediction model according to any one of claims 1-5 is stored.

Images