Tumor treatment responsiveness evaluation method, equipment and device
By detecting the peripheral blood cell subpopulations of patients with advanced tumors, using single-factor logistic regression and XGBoost model, significant immune cell subpopulations were screened out, and a classification prediction model was established, which solved the problem of difficulty in accurately assessing the response of small-dose rhythm chemotherapy in the existing technology, and achieved dynamic and accurate tumor treatment response evaluation.
Patent Information
- Application Number
- CN202510230529.0
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-02-28
- Publication Date
- 2025-07-29
AI Technical Summary
The prior art is difficult to accurately predict the response of patients with advanced tumors to low-dose rhythmic chemotherapy, and the compliance is poor, and there is a lack of effective models for evaluating the possibility of patients' treatment benefits.
By detecting the peripheral blood cell subpopulations of patients with advanced tumors, a single-factor logistic regression model was used to screen the significant immune cell subpopulations, establish an XGBoost classification prediction model, and dynamically evaluate the patient's treatment response status, including responding and non-responsiveness.
It has achieved a non-invasive, dynamic and accurate assessment of the possibility of chemotherapy benefits in patients with advanced tumors in pan-cancer species, which has improved the reference value of the treatment plan selection, has a wide range of application and a significant prediction effect.
Smart Images
Figure CN120388723A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of tumor treatment evaluation, and particularly relates to a method, device and apparatus for evaluating tumor treatment responsiveness. Background Art
[0002] Currently, the treatment methods for cancer mainly include comprehensive means such as surgical operation combined with radiotherapy, chemotherapy or targeted drug therapy. Among them, chemotherapy plays an extremely important role in the treatment plan for advanced tumors. Conventional chemotherapy uses the maximum tolerated dose (MTD), which can directly kill tumor cells by using cytotoxic drugs. However, chemotherapy based on MTD has a killing effect on all proliferating cells, with obvious adverse reactions and poor patient compliance. Metronomic chemotherapy refers to the administration of relatively low doses (usually 1 / 10 - 1 / 3 of the conventional dose) of cytotoxic drugs (such as cyclophosphamide, vincristine, topotecan, cisplatin, etc.) closely and regularly. The main anti-tumor mechanism of metronomic chemotherapy is to inhibit tumor-related angiogenesis and stimulate the body's immunity, rather than the direct cytotoxic effect on tumor cells. It has been confirmed that metronomic chemotherapy has small adverse reactions in various tumors. Currently, for advanced tumor patients, a comprehensive treatment plan mainly based on metronomic chemotherapy is usually adopted, that is, the "tumor treatment" described in the patent. The prognostic evaluation of tumor treatment is of great significance for improving the survival rate of patients, optimizing the medication plan and reducing unnecessary treatments. The existing technologies mainly focus on the benefit prediction or prognostic evaluation of neoadjuvant chemotherapy or adjuvant chemotherapy for a specific type of cancer, and the modeling data used are mostly transcriptome data of surgical tissue samples, which are relatively difficult to obtain.
[0003] Therefore, in order to more accurately predict whether patients can benefit from metronomic chemotherapy, it is urgent to develop a reliable tumor treatment responsiveness prediction model. Summary of the Invention
[0004] The object of the present invention is to provide a method, device and apparatus for evaluating tumor treatment responsiveness, which can evaluate the peripheral blood immune pattern of patients by detecting peripheral blood cell subsets of advanced tumor patients, and then accurately and dynamically predict the benefit possibility of pan-cancer advanced tumor patients receiving metronomic chemotherapy, providing great reference value for the selection of treatment plans.
[0005] To achieve the above object, on the one hand, the present invention provides a method for evaluating tumor treatment responsiveness, including:
[0006] A single-factor logistic regression model was used to analyze the effects of peripheral blood immune cell subsets on the response to low-dose rhythmic chemotherapy, and immune cell subsets that had a significant effect on patients with treatment response compared to those without treatment response were obtained (judged by the OR value).
[0007] Based on the selected immune cell subsets, a classification prediction model was established. The model was used to predict the treatment status of patients receiving low-dose rhythmic chemotherapy, and the treatment status included response and non-response.
[0008] Furthermore, based on multi-color flow cytometry analysis, peripheral blood samples of patients receiving low-dose rhythmic chemotherapy were analyzed to obtain evaluation data.
[0009] Furthermore, the immune cell subsets that had a significant effect included: IFNγ+PD1-CD8+T cell, PMN-MDSC, WBC, Neutrophils, PD1+CD8+T cell, IFNγ-PD1+CD8+T cell.
[0010] Furthermore, the classification prediction model was trained using the XGBoost model. The data was divided into a training set and a validation set, and the SHAP contribution values (SHapley Additive exPlanations) of each immune cell subset were calculated. The SHAP contribution values of the immune cell subsets from high to low were: Neutrophils, WBC, IFNγ+PD1-CD8+T cell, PMN-MDSC, PD1+CD8+T cell, IFNγ-PD1+CD8+T cell.
[0011] Furthermore, the classification prediction model also included using LASSO, support vector machine, random forest, and neural network as prediction models.
[0012] On the other hand, the present invention also provides a tumor treatment response evaluation system based on machine learning, including:
[0013] A feature screening module that uses a single-factor logistic regression model to analyze the effects of peripheral blood immune cell subsets on the response to low-dose rhythmic chemotherapy, and obtains immune cell subsets that have a significant effect on patients with treatment response compared to those without treatment response;
[0014] A prediction and analysis module that, based on the selected immune cell subsets, establishes a classification prediction model respectively. The model is used to predict the treatment status of patients receiving low-dose rhythmic chemotherapy, and the treatment status includes response and non-response.
[0015] On the other hand, the present invention also provides an electronic device, which includes a processor and a memory. One or more programs are stored in the memory and configured to be executed by the processor to implement the above method.
[0016] On the other hand, the present invention also provides a storage device. The storage medium stores multiple instructions, and the instructions are suitable for being loaded by a processor to execute the steps in the above tumor treatment response evaluation method.
[0017] The present invention has the following technological progress and beneficial effects:
[0018] 1. The prediction data used in the present invention are cell subsets identified by multi-color flow cytometry analysis of peripheral blood samples of advanced patients at different time points (i.e., the time points when the follow-up patients come for chemotherapy each time). It can non-invasively and dynamically obtain samples while evaluating the peripheral blood immune pattern of advanced patients at different time points, and is suitable for clinical transformation and promotion.
[0019] 2. The XGBoost classifier model selected in the present invention belongs to a high-order machine learning model that takes into account both interpretability and non-linear predictability. The AUC value of the training set is as high as 0.949, with significant effects and high precision.
[0020] 3. The present invention can predict the benefit possibility of patients receiving low-dose rhythmic chemotherapy by evaluating the peripheral blood immune pattern of pan-cancer patients, with a wide range of applications and high clinical application value. BRIEF DESCRIPTION OF THE DRAWINGS
[0021] In order to more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings required for the description of the embodiments or the prior art. Obviously, the following drawings are only some embodiments of the present invention. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.
[0022] Figure 1 It is a method flow chart of a tumor treatment response evaluation method according to an embodiment of the present invention.
[0023] Figure 2 It is a comparison chart of OR values for analyzing the influence of each peripheral blood immune cell subset on tumor treatment reactivity by a single-factor logistic regression model in the present invention.
[0024] Figure 3 It is a schematic diagram of the average SHAP contribution value of each immune cell subset in the present invention.
[0025] Figure 4 It is a schematic diagram of the area under the ROC curve of the training set and the validation set of the XGBoost model in the present invention. Detailed implementation manners
[0026] The following will clearly and completely describe the technical solutions in the embodiments of the present invention with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts belong to the scope of protection of the present invention.
[0027] In the following description, "some embodiments" are involved, which describe a subset of all possible embodiments. However, it can be understood that "some embodiments" can be the same subset or different subsets of all possible embodiments, and can be combined with each other without conflict.
[0028] If similar descriptions such as "first / second" appear in the application documents, the following explanation will be added. In the following description, the terms "first, second, third" only distinguish similar objects and do not represent a specific order for the objects. It can be understood that "first, second, third" can be interchanged with a specific order or sequence when allowed, so that the embodiments of the present application described here can be implemented in an order other than that illustrated or described here.
[0029] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by those of ordinary skill in the technical field to which this application belongs. The terms used herein are only for the purpose of describing the embodiments of this application and are not intended to limit this application.
[0030] The present invention provides a method, device and apparatus for evaluating tumor treatment responsiveness, which can evaluate the peripheral blood immune pattern of patients by detecting peripheral blood cell subsets of advanced tumor patients, and then accurately and dynamically predict the benefit possibility of advanced tumor patients with pan-cancer types receiving low-dose rhythmic chemotherapy, providing great reference value for the selection of treatment plans.
[0031] In a specific embodiment, for the convenience of medical staff to use during work, the present invention uses a PC or a mobile phone as an electronic device for running the method of the present invention. The electronic device includes: a processor and a memory for storing executable instructions of the processor. Among them, the executable instructions stored in the memory are configured to run a program for a method of evaluating tumor treatment responsiveness of the present invention, and the program can communicate with other terminals, servers or other forms of devices to complete background program calculations.
[0032] The following describes a method, device and apparatus for evaluating tumor treatment responsiveness according to an embodiment of the present invention with reference to the accompanying drawings. First, a method for evaluating tumor treatment responsiveness according to an embodiment of the present invention will be described with reference to the accompanying drawings.
[0033] Figure 1 It is a flowchart of a method for evaluating tumor treatment responsiveness according to an embodiment of the present invention.
[0034] As Figure 1 shown, the evaluation method includes the following steps:
[0035] S1. Analyze the peripheral blood samples of patients receiving low-dose metronomic chemotherapy based on multi-color flow cytometry analysis to obtain data for evaluation.
[0036] Specifically, in this embodiment, the patient data uses the data of 83 patients with pan-cancer (colon cancer, gastric cancer) receiving low-dose metronomic chemotherapy. First, peripheral blood samples of patients are drawn at different time points (i.e., the time points when the patients come for chemotherapy during follow-up), and then the peripheral blood samples of the patients are analyzed by multi-color flow cytometry to obtain, including 185 dynamic immune pattern monitors of peripheral blood as evaluation data.
[0037] S2. Analyze the influence of each immune cell subset in peripheral blood on the response to low-dose metronomic chemotherapy using a univariate logistic regression model to obtain the immune cell subsets that play a significant role in patients with treatment response compared to patients without treatment response.
[0038] Specifically, as Figure 2 shown, by constructing a univariate logistic regression model to preliminarily analyze the influence of each immune cell subset in peripheral blood on the responsiveness to low-dose metronomic chemotherapy, it is found that for patients with treatment response (radiological evaluation is PR / CR) compared to patients without treatment response (radiological evaluation is PD / SD), IFNγ+PD1-CD8+T cells have a positive impact on the occurrence of PR / CR (OR>1), while PMN-MDSC, WBC, Neutrophils, PD1+CD8+T cells, and IFNγ-PD1+CD8+T cells have a negative impact (OR<1).
[0039] S3. Based on the selected 6 groups of immune cell subsets, establish a classification prediction model, which is used to predict the treatment status of patients receiving low-dose metronomic chemotherapy, and the treatment status includes response and non-response.
[0040] Specifically, in this embodiment, the classification prediction model is trained using the XGBoost model, and the data is divided into a training set (80% of the aforementioned data set) and a validation set (20% of the aforementioned data set).
[0041] In a specific embodiment, during the data preparation and model training process:
[0042] First, set a random seed to ensure the reproducibility of the results. Then, separate the feature and target variables in the training set and test set and convert them into matrix format to be compatible with XGBoost. Next, use the Bayesian optimization method (BayesianOptimization) to search for the best hyperparameters (such as eta, max_depth, min_child_weight, subsample, etc.) of the XGBoost model and evaluate the model performance through 5-fold cross-validation. Train the XGBoost model with the optimized hyperparameters.
[0043] Result prediction and model evaluation:
[0044] Use the trained model to predict the training set and test set, and the prediction type is probability value. Then, draw the ROC curve through the plot.roc function in the pROC package and calculate the AUC value. The plot.roc function can accept the true label and prediction probability as inputs, automatically calculate the AUC value and display it in the graph, and thus can evaluate the classification performance of the model at different thresholds. Then perform SHAP (SHapley Additive exPlanations) analysis, that is, calculate the SHAP contribution value to show the impact of each feature on the model output to explain the decision-making process of the model.
[0045] As Figure 3 shown, the average SHAP contribution values of each immune cell subset are shown in the figure, from high to low in turn: Neutrophils, WBC, IFNγ+PD1-CD8+T cell, PMN-MDSC, PD1+CD8+Tcell, IFNγ-PD1+CD8+Tcell.
[0046] Figure 4 Schematic diagram of the area under the ROC curve for the training set and validation set of the XGBoost model. As Figure 4 shown, the XGBoost model can well predict whether the patient's treatment status belongs to treatment response (radiological evaluation is PR / CR) or treatment non-response (radiological evaluation is PD / SD). The AUC value of the training set is 0.949, and the AUC value of the validation set is 0.789.
[0047] In other embodiments of the present invention, the classification prediction model further includes using LASSO, support vector machine, random forest, neural network as the prediction model.
[0048] On the other hand, the present invention also provides a tumor treatment responsiveness evaluation system based on machine learning, including:
[0049] A feature screening module that uses a single-factor logistic regression model to analyze the effects of various immune cell subsets in peripheral blood on the response to low-dose rhythmic chemotherapy, and obtains the immune cell subsets that play a significant role in patients with treatment response compared to those without treatment response.
[0050] A prediction and analysis module that, based on the screened immune cell subsets, respectively establishes a classification prediction model for predicting the treatment status of patients receiving low-dose rhythmic chemotherapy, where the treatment status includes response and non-response.
[0051] On the other hand, the present invention also provides an electronic device, which includes a processor and a memory, and one or more programs are stored in the memory and configured to be executed by the processor to perform the above method.
[0052] On the other hand, the present invention also provides a storage device, characterized in that the storage medium stores multiple instructions, and the instructions are suitable for being loaded by a processor to execute the steps in a method for evaluating the responsiveness of low-dose rhythmic chemotherapy as described in the above claims.
[0053] Although the embodiments of the present invention have been shown and described above, it can be understood that the above embodiments are exemplary and should not be construed as limiting the present invention. Those of ordinary skill in the art can make changes, modifications, substitutions, and variations to the above embodiments within the scope of the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principle of the present invention shall be included within the protection scope of the present invention.
Claims
1. A method for evaluating the responsiveness of tumor treatment based on machine learning, characterized in that, Comprising: Using a univariate logistic regression model to analyze the effects of peripheral blood immune cell subsets on the response to low-dose rhythmic chemotherapy, and obtaining the immune cell subsets that play a significant role in patients with treatment response compared to patients without treatment response; Based on the screened immune cell subsets, a classification prediction model is established respectively, and the model is used to predict the treatment status of patients receiving low-dose rhythmic chemotherapy, and the treatment status includes response and non-response.
2. The tumor treatment responsiveness evaluation method according to claim 1, characterized in that Based on the analysis of multi-color flow cytometry, the peripheral blood samples of patients receiving low-dose rhythmic chemotherapy are analyzed to obtain evaluation data.
3. The method for evaluating tumor treatment responsiveness according to claim 1, characterized in that The immune cell subsets that play a significant role include: IFNγ+PD1-CD8+T cells, PMN-MDSC, WBC, Neutrophils, PD1+CD8+T cells, IFNγ-PD1+CD8+T cells.
4. The method for evaluating tumor treatment responsiveness according to claim 3, wherein, The classification prediction model is trained using the XGBoost model, the data is divided into a training set and a validation set, and the SHAP contribution values of each immune cell subset are calculated respectively. The SHAP contribution values of the immune cell subsets from high to low are: Neutrophils, WBC, IFNγ+PD1-CD8+T cells, PMN-MDSC, PD1+CD8+T cells, IFNγ-PD1+CD8+T cells.
5. The method for evaluating tumor treatment responsiveness according to claim 4, wherein The classification prediction model also includes using LASSO, support vector machine, random forest, neural network as prediction models.
6. A tumor treatment responsiveness evaluation system based on machine learning, characterized in that, Comprising: A feature screening module that uses a univariate logistic regression model to analyze the effects of peripheral blood immune cell subsets on the response to low-dose rhythmic chemotherapy, and obtains the immune cell subsets that play a significant role in patients with treatment response compared to patients without treatment response; A prediction and analysis module that, based on the screened immune cell subsets, establishes a classification prediction model respectively, and the model is used to predict the treatment status of patients receiving low-dose rhythmic chemotherapy, and the treatment status includes response and non-response.
7. An electronic device, characterized in that, Comprising a processor and a memory, one or more programs are stored in the memory and are configured to be executed by the processor to implement the method according to any one of claims 1-5.
8. A storage device, characterized in that, The storage medium stores multiple instructions, and the instructions are suitable for being loaded by the processor to execute the steps in a method for evaluating tumor treatment responsiveness according to any one of claims 1 to 5.