A system for osteosarcoma prognosis risk prediction

Abstract

The application relates to the technical field of medical data processing, and discloses an osteosarcoma prognosis risk prediction system. The osteosarcoma prognosis risk prediction system is constructed by integrating non-invasive and easily-obtained multi-dimensional clinical data (including demographic and baseline clinical data, tumor load and malignancy data, and serum tumor marker data) and training an optimized machine learning model (such as a random forest). The osteosarcoma prognosis risk prediction model and system can realize early warning, real-time monitoring and accurate grading, can realize early, real-time and non-invasive evaluation and grading early warning of the prognosis risk of patients, solve the problems that existing monitoring means are lagging, invasive and cannot be frequently performed, provide objective and quantitative basis for clinical identification of high-risk patients and development of individualized strategies, and provide strong technical support for improvement of the management and prognosis of osteosarcoma patients.

Description

Technical Field

[0001] This invention relates to the field of medical data processing technology, and in particular to a prognostic risk prediction system for osteosarcoma. Background Technology

[0002] Osteosarcoma (OS) is the most common primary malignant bone tumor in children. It is highly malignant and prone to hematogenous metastasis, leading to systemic progression. Currently, the combined treatment strategy of neoadjuvant chemotherapy and surgery achieves a 5-year survival rate of 60%–80%. Despite the promising results in OS treatment, nearly 60% of children develop tumor progression after neoadjuvant chemotherapy due to chemotherapy resistance, with a survival rate below 30%. Currently, chest CT and whole-body PET-CT scans are important tools for monitoring tumor recurrence and progression during neoadjuvant chemotherapy and follow-up examinations in OS patients. However, due to the harmful effects of radiation and the cost of these examinations, they cannot be performed frequently. Furthermore, these imaging examinations can only detect tumor progression when it is clearly established, resulting in a significant lag in monitoring tumor progression after comprehensive treatment in OS patients, failing to achieve early warning. There are currently many studies predicting disease progression in OS patients after chemotherapy resistance, such as tumor tissue molecular markers, bone transformation indicators, histological type, and tumor necrosis rate. Although these indicators can reflect and monitor the outcome of OS patients after chemotherapy to varying degrees, they still have certain limitations. Obtaining molecular markers, histological types, and tumor necrosis rates from tumor tissue requires histological examination using puncture or open biopsy and postoperative specimens, thus making real-time monitoring of disease progression impossible postoperatively. For patients receiving standardized OS treatment, identifying which groups are at high risk of tumor progression during osteosarcoma treatment and how to provide early warning of prognosis at the time of diagnosis has been a serious challenge for both patients and clinicians. Summary of the Invention

[0003] This invention provides a prognostic risk prediction system for osteosarcoma to overcome the shortcomings of existing technologies.

[0004] The present invention provides a method for constructing a prognostic risk prediction model for osteosarcoma, comprising:

[0005] Medical data of the target population are used as modeling features. The target population includes osteosarcoma patients who have relapsed or not relapsed after comprehensive treatment for osteosarcoma. The medical data used as modeling features include any one or any combination of the following: demographic and baseline clinical data, tumor burden and malignancy data, and serum tumor marker data. Using machine learning algorithms, a model is trained to learn the characteristics of medical data corresponding to osteosarcoma patients who died or survived after comprehensive treatment, as well as those who survived with the disease and those who survived without the disease, thus obtaining an osteosarcoma prognostic risk prediction model.

[0006] According to the present invention, a method for constructing a prognostic risk prediction model for osteosarcoma is provided, wherein demographic and baseline clinical data include any one of the following indicators or any combination thereof: age, body mass index (BMI), and tumor site.

[0007] According to the present invention, a method for constructing a prognostic risk prediction model for osteosarcoma is provided, wherein tumor burden and malignancy data include Enneking stage and / or a combination of stage and age (enneking_age).

[0008] According to the present invention, a method for constructing a prognostic risk prediction model for osteosarcoma is provided, wherein serum tumor marker data includes any one of the following indicators or any combination thereof: initial ALP (ALPinput, referring to serum alkaline phosphatase level before treatment), initial LDH (LDHinput, referring to serum lactate dehydrogenase level before treatment), initial α-HBDH (α-HBDHinput, initial α-hydroxybutyrate dehydrogenase), initial ALP / LDH ratio, last ALP (ALPlast, referring to alkaline phosphatase level at follow-up after treatment), last LDH (LDHlast, referring to lactate dehydrogenase level at follow-up after treatment), and ALP change (ALP_change, referring to the change value or rate of change of ALP level before and after treatment).

[0009] According to the method for constructing a prognostic risk prediction model for osteosarcoma provided by the present invention, after the step of acquiring medical data of the target population, the method further includes the following step: Preprocessing of medical data of the target population includes any one or any combination of the following: missing value imputation, outlier handling, variable standardization, derived variable construction, and feature selection.

[0010] According to the present invention, a method for constructing a prognostic risk prediction model for osteosarcoma is provided. The method utilizes machine learning algorithms to train the model to learn the characteristics of medical data corresponding to osteosarcoma patient groups who have died or survived after comprehensive treatment, those with disease-associated survival, and those without disease-associated survival, thereby obtaining the osteosarcoma prognostic risk prediction model, including: The machine learning algorithm used is the random forest algorithm; During training, five-fold cross-validation and early stopping mechanisms are used, along with grid search and Bayesian optimization to determine the optimal hyperparameters for model training. Additionally, a conservative adjustment strategy is employed to conservatively adjust the output probabilities.

[0011] According to the method for constructing a prognostic risk prediction model for osteosarcoma provided by the present invention, the optimal hyperparameters include any one of the following or any combination thereof: the number of decision trees is 80, the maximum depth is 6, the minimum number of split samples is 15, the minimum number of leaf node samples is 10, and the maximum number of features is the square root, that is, n_estimators=80, max_depth=6, min_samples_split=15, min_samples_leaf=10, max_features=sqrt.

[0012] The method for constructing a prognostic risk prediction model for osteosarcoma according to the present invention further includes: Based on the prediction results of the osteosarcoma prognostic risk prediction model during the training process, features that contribute highly to the prediction results are selected from medical data as inference features. The features that contribute highly to the prediction results include any one of the following or any combination thereof: Enneking stage, initial ALP, ALP change value, initial α-HBDH, and a combination of stage and age.

[0013] This invention also provides a prognostic risk prediction system for osteosarcoma, comprising: The data receiving module is used to: receive medical data of a subject as inference features from at least one terminal, wherein the subject is an osteosarcoma patient, and the medical data used as inference features includes at least one or any combination thereof: demographic and baseline clinical data, tumor burden and malignancy data, and serum tumor marker data; The prediction module is used to: input medical data used as inference features into the osteosarcoma prognostic risk prediction model obtained by the construction method of the osteosarcoma prognostic risk prediction model described above, and obtain the osteosarcoma prognostic risk prediction result of the subject. The data output module is used to send the osteosarcoma prognostic risk prediction results of the subject to at least one terminal.

[0014] According to the present invention, an osteosarcoma prognostic risk prediction system includes a prediction module comprising: The prediction submodule is used to: input medical data used as inference features into the osteosarcoma prognostic risk prediction model to obtain the predicted probability of osteosarcoma prognostic risk for the subject. The risk level classification submodule is used to: obtain the osteosarcoma prognostic risk level of the test subject based on the predicted probability of osteosarcoma prognostic risk of the test subject using a preset risk level classification standard, and use it as the predicted result of osteosarcoma prognostic risk of the test subject. The preset risk level classification standard includes: when the predicted probability of osteosarcoma prognostic risk is less than 0.3, it is judged as low risk; when the predicted probability of osteosarcoma prognostic risk is within the interval [0.3, 0.7], it is judged as medium risk; and when the predicted probability of osteosarcoma prognostic risk is greater than 0.7, it is judged as high risk.

[0015] It should be noted that a terminal refers to an input / output device connected to a computer system. Depending on the function, terminals can be divided into various types: smart terminals or intelligent terminals, dumb terminals, interactive terminals or online terminals. Specifically, a terminal can be various mobile communication devices, such as mobile phones and tablets. This article aims to provide users with the function of inputting data and outputting data.

[0016] The present invention also provides an electronic device, including a processor and a memory storing a computer program, wherein the processor executes the computer program to implement the method for constructing the osteosarcoma prognostic risk prediction model described above.

[0017] The present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the method for constructing the osteosarcoma prognostic risk prediction model described above.

[0018] The present invention also provides a computer program product, the computer program product comprising a computer program that can be stored on a non-transitory computer-readable storage medium, and when the computer program is executed by a processor, the computer is able to execute any of the above-described methods for constructing an osteosarcoma prognostic risk prediction model.

[0019] The present invention provides a method and system for constructing a prognostic risk prediction model for osteosarcoma, which can bring at least the following beneficial effects: This invention constructs a prognostic risk prediction model for osteosarcoma based on readily available medical data such as demographic and baseline clinical data, tumor burden and malignancy data, and serum tumor marker data. This model enables proactive assessment of the prognosis of osteosarcoma patients without relying on invasive biopsies or expensive imaging examinations. Risk prediction can be performed using readily available indicators during routine or follow-up procedures (such as age, BMI, tumor location, Enneking stage, and serum ALP, LDH, α-HBDH, etc.), significantly reducing monitoring costs and patient burden. This makes dynamic and frequent risk assessments during treatment and follow-up possible, thereby achieving the goal of early warning.

[0020] This invention specifically incorporates serum tumor marker data (such as initial ALP, initial LDH, initial α-HBDH, last ALP, last LDH, ALP change, etc.) that can be conveniently and non-invasively repeated throughout the entire treatment process (initial values ​​before treatment and last values ​​after treatment) and during follow-up. By combining these dynamically changing serum indicators with machine learning models, it is possible to continuously and in real time assess the risk of tumor progression or recurrence in OS patients during long-term follow-up after comprehensive treatment (including postoperatively), thus compensating for the temporal and spatial limitations of histological testing methods.

[0021] This invention employs machine learning algorithms (such as random forests) to train and learn from the data characteristics of osteosarcoma patients with a known prognosis (relapse or no relapse). This allows for the discovery of complex, nonlinear predictive patterns, providing a more comprehensive and systematic assessment of the patient's overall condition and tumor biological behavior. Specific hyperparameters used during model training (such as the number and depth of decision trees) and strategies employed, including five-fold cross-validation, early stopping mechanisms, grid search, and Bayesian optimization, ensure the robustness and generalization ability of the osteosarcoma prognostic risk prediction model. In particular, key predictive features identified through feature contribution analysis (such as Enneking stage, initial ALP, ALP change, initial α-HBDH, and the stage-age composite term) provide high-predictive-value inputs to the model, enabling it to more accurately predict osteosarcoma prognostic risk.

[0022] The osteosarcoma prognostic risk prediction system constructed in this invention can receive patient test data, output quantitative prognostic risk prediction probabilities through a trained osteosarcoma prognostic risk prediction model, and further classify risk levels according to preset criteria (e.g., probability <0.3 for low risk, 0.3-0.7 for intermediate risk, and >0.7 for high risk). This intuitive and graded output can directly assist clinicians in quickly identifying those with a high risk of tumor progression during standardized treatment, thereby providing objective and quantitative decision-making basis for developing individualized enhanced monitoring plans, adjusting treatment strategies, or carrying out early intervention, and is expected to improve the survival prognosis of high-risk patients. Attached Figure Description

[0023] To more clearly illustrate the technical solutions in this invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of this invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0024] Figure 1 This is a flowchart illustrating a method for constructing a prognostic risk prediction model for osteosarcoma provided by the present invention.

[0025] Figure 2 This is a schematic diagram illustrating the performance evaluation of the osteosarcoma prognostic risk prediction model.

[0026] Figure 3 This shows the contribution of each indicator in the medical data to the prediction results of the osteosarcoma prognostic risk prediction model.

[0027] Figure 4 This is a schematic diagram of the structure of an osteosarcoma prognostic risk prediction system provided by the present invention.

[0028] Figure 5 This is a schematic diagram of the structure of the electronic device provided by the present invention. Detailed Implementation

[0029] To make the objectives, technical solutions, and advantages of this invention clearer, the technical solutions of this invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, embodiments of this invention, and should not be construed as limiting the invention. All other embodiments obtained by those skilled in the art based on the embodiments of this invention without creative effort are within the scope of protection of this invention. In the description of this invention, it should be understood that the terminology used is for descriptive purposes only and should not be construed as indicating or implying relative importance.

[0030] Figure 1 This is a flowchart illustrating a method for constructing a prognostic risk prediction model for osteosarcoma provided by the present invention. The execution entity of this method can be any applicable terminal-side device or network-side device, such as an osteosarcoma prognostic risk prediction device.

[0031] See Figure 1 The present invention provides a method for constructing a prognostic risk prediction model for osteosarcoma, which may include: S110. Obtain medical data of the target population as modeling features. The target population includes osteosarcoma patients who have died or survived after comprehensive treatment, those who are alive with the disease, and those who are alive without the disease. The medical data used as modeling features include: demographic and baseline clinical data, tumor burden and malignancy data, and serum tumor marker data.

[0032] In one embodiment, demographic and baseline clinical data include any one of the following indicators or any combination thereof: age, body mass index (BMI), and tumor site.

[0033] In one embodiment, tumor burden and malignancy data include Enneking stage and / or a combination of stage and age (enneking_age).

[0034] In one embodiment, serum tumor marker data includes any one of the following indicators or any combination thereof: initial ALP (ALPinput, referring to serum alkaline phosphatase level before treatment), initial LDH (LDHinput, referring to serum lactate dehydrogenase level before treatment), initial α-HBDH (α-HBDHinput, initial α-hydroxybutyrate dehydrogenase), initial ALP / LDH ratio, last ALP (ALPlast, referring to alkaline phosphatase level at follow-up after treatment), last LDH (LDHlast, referring to lactate dehydrogenase level at follow-up after treatment), and ALP change (ALP_change, referring to the change in ALP level before and after treatment).

[0035] In one embodiment, after obtaining the medical data of the target population, the medical data of the target population can be preprocessed, wherein the preprocessing includes any one or any combination of the following: missing value imputation, outlier handling, variable standardization, derived variable construction, and feature selection.

[0036] S120. Using machine learning algorithms, the model is trained to learn the characteristics of medical data corresponding to osteosarcoma patients who have relapsed or not relapsed after comprehensive treatment, and an osteosarcoma prognostic risk prediction model is obtained.

[0037] In one embodiment, S120 may include: The machine learning algorithm used is the random forest algorithm; During training, a five-fold cross-validation and early stopping mechanism were used, combined with grid search and Bayesian optimization to determine the optimal hyperparameters for model training. The optimal hyperparameters included: 80 decision trees, a maximum depth of 6, a minimum number of split samples of 15, a minimum number of leaf samples of 10, and a maximum number of features equal to the square root, i.e., n_estimators=80, max_depth=6, min_samples_split=15, min_samples_leaf=10, max_features=sqrt; A conservative adjustment strategy is adopted to conservatively adjust the output probability, specifically by nonlinearly compressing the high-risk probability to improve clinical safety.

[0038] Figure 2This diagram illustrates the performance evaluation of the osteosarcoma prognostic risk prediction model trained using medical data from 847 osteosarcoma patients in the Peking University People's Hospital database. The model includes the ROC curve (AUC=0.725), probability calibration curve, risk score distribution plot, and predicted vs. actual risk distribution plot. During model training, 5-fold cross-validation was used, with a 7:3 split between the training and test sets. The osteosarcoma prognostic risk prediction model trained in this embodiment achieved an accuracy of 0.796 and an AUC of 0.900 on the training set, and an AUC of 0.722 and an accuracy of 0.757 on the test set. The cross-validation accuracy was 0.741 (+ / - 0.041). Overfitting was diagnosed with an AUC difference of 0.178 and an accuracy difference of 0.039. The osteosarcoma prognostic risk prediction model trained in this embodiment supports risk stratification (low / medium / high) and visualization analysis, solving the problems of lagging traditional imaging monitoring, high cost, and invasiveness. It is suitable for clinical decision support and personalized treatment formulation.

[0039] In one embodiment, features that contribute significantly to the prediction results can be selected from medical data as key predictive features based on the prediction results of the osteosarcoma prognostic risk prediction model during training. See [link to relevant documentation] in this embodiment. Figure 3 The top five features that contribute significantly to the prediction results are: Enneking stage (23.8%), initial ALP (10.6%), ALP change (8.5%), initial α-HBDH (8.4%), and the combination of stage and age (8.3%).

[0040] After training and obtaining the osteosarcoma prognostic risk prediction model, it can be used to predict the prognostic risk of osteosarcoma, forming a method for predicting the prognostic risk of osteosarcoma, including the following steps: Medical data of the subject is obtained and used as inference features. The subject is an osteosarcoma patient. The medical data used as inference features include at least one or any combination thereof: demographic and baseline clinical data, tumor burden and malignancy data, and serum tumor marker data. In this embodiment, the inference features include at least one or any combination thereof: Enneking stage, initial ALP, ALP change value, initial α-HBDH, and a combination of stage and age. The medical data used as inference features is input into the osteosarcoma prognostic risk prediction model constructed by any of the above-mentioned methods to obtain the osteosarcoma prognostic risk prediction result for the subject. Specifically, the osteosarcoma prognostic risk prediction model outputs the osteosarcoma prognostic risk prediction probability, which can be classified into risk levels according to preset risk level classification criteria. In this embodiment, the preset risk level classification criteria include: when the osteosarcoma prognostic risk prediction probability is less than 0.3, it is judged as low risk; when the osteosarcoma prognostic risk prediction probability is in the interval [0.3, 0.7], it is judged as medium risk; and when the osteosarcoma prognostic risk prediction probability is greater than 0.7, it is judged as high risk.

[0041] The osteosarcoma prognostic risk prediction system provided by the present invention is described below. The osteosarcoma prognostic risk prediction system described below can be referred to in correspondence with the osteosarcoma prognostic risk prediction method described above.

[0042] See Figure 4 The present invention provides an osteosarcoma prognostic risk prediction system, which may include: The data receiving module is used to: receive medical data of a subject as inference features from at least one terminal, wherein the subject is an osteosarcoma patient, and the medical data used as inference features includes at least one of the following or any combination thereof: demographic and baseline clinical data, tumor burden and malignancy data, and serum tumor marker data; The prediction module is used to: input medical data used as inference features into the osteosarcoma prognostic risk prediction model obtained by the construction method of the osteosarcoma prognostic risk prediction model described above, and obtain the osteosarcoma prognostic risk prediction result of the subject. The data output module is used to send the osteosarcoma prognostic risk prediction results of the subject to at least one terminal.

[0043] In one embodiment, the prediction module may include: The prediction submodule is used to: input medical data used as inference features into the osteosarcoma prognostic risk prediction model to obtain the predicted probability of osteosarcoma prognostic risk for the subject. The risk level classification submodule is used to: obtain the osteosarcoma prognostic risk level of the test subject based on the predicted probability of osteosarcoma prognostic risk of the test subject using a preset risk level classification standard, and use it as the predicted result of osteosarcoma prognostic risk of the test subject. The preset risk level classification standard includes: when the predicted probability of osteosarcoma prognostic risk is less than 0.3, it is judged as low risk; when the predicted probability of osteosarcoma prognostic risk is within the interval [0.3, 0.7], it is judged as medium risk; and when the predicted probability of osteosarcoma prognostic risk is greater than 0.7, it is judged as high risk.

[0044] Figure 5An example is a schematic diagram of the physical structure of an electronic device, such as... Figure 5 As shown, the electronic device may include a processor 810, a communications interface 820, a memory 830, and a communication bus 840, wherein the processor 810, the communications interface 820, and the memory 830 communicate with each other via the communication bus 840. The processor 810 can call logical instructions in the memory 830 to execute the steps of the method for constructing the osteosarcoma prognostic risk prediction model and / or the osteosarcoma prognostic risk prediction method described above.

[0045] Furthermore, the logical instructions in the aforementioned memory 830 can be implemented as software functional units and, when sold or used as independent products, can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0046] On the other hand, the present invention also provides a computer program product, the computer program product including a computer program, the computer program being stored on a non-transitory computer-readable storage medium, and when the computer program is executed by a processor, the computer is able to perform the steps of the method for constructing the osteosarcoma prognostic risk prediction model and / or the method for predicting the prognostic risk of osteosarcoma described above.

[0047] In another aspect, the present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the method for constructing and / or predicting the prognostic risk of osteosarcoma as described above.

[0048] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. Those skilled in the art can understand and implement this without any creative effort.

[0049] Through the above description of the embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus necessary general-purpose hardware platforms, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solutions, in essence or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods described in the various embodiments or some parts of the embodiments.

[0050] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A prognostic risk prediction system for osteosarcoma, characterized in that, include: The data receiving module is used to: receive medical data of a subject as inference features from at least one terminal, wherein the subject is an osteosarcoma patient, and the medical data used as inference features includes at least one of the following or any combination thereof: demographic and baseline clinical data, tumor burden and malignancy data, and serum tumor marker data; The prediction module is used to: input medical data used as inference features into a preset osteosarcoma prognostic risk prediction model to obtain the osteosarcoma prognostic risk prediction result for the test subject; The data output module is used to send the osteosarcoma prognostic risk prediction results of the subject to at least one terminal.

2. The osteosarcoma prognostic risk prediction system according to claim 1, characterized in that, The inference features include at least one of the following or any combination thereof: Enneking staging, initial ALP, ALP change value, initial α-HBDH, and a combination of staging and age.

3. The osteosarcoma prognostic risk prediction system according to claim 1, characterized in that, The method for constructing a pre-defined prognostic risk prediction model for osteosarcoma includes: Medical data of the target population are used as modeling features. The target population includes osteosarcoma patients who have died or survived after comprehensive treatment, those who are alive with the disease, and those who are alive without the disease. The medical data used as modeling features include any one or any combination of the following: demographic and baseline clinical data, tumor burden and malignancy data, and serum tumor marker data. Using machine learning algorithms, a model is trained to learn the characteristics of medical data corresponding to osteosarcoma patients who died or survived after comprehensive treatment, as well as those who survived with the disease and those who survived without the disease, thus obtaining an osteosarcoma prognostic risk prediction model.

4. The osteosarcoma prognostic risk prediction system according to claim 1, characterized in that, Demographic and baseline clinical data include any one of the following indicators or any combination thereof: age, body mass index, tumor location; tumor burden and malignancy data include Enneking stage and / or stage and age composite; serum tumor marker data include any one of the following indicators or any combination thereof: initial ALP, initial LDH, initial α-HBDH, initial ALP / LDH ratio, last ALP, last LDH, ALP change.

5. The method for constructing the osteosarcoma prognostic risk prediction model according to claim 2, characterized in that, The method utilizes machine learning algorithms to train a model that learns the characteristics of medical data corresponding to osteosarcoma patient groups who died and survived, survived with the disease, and survived without the disease after comprehensive treatment for osteosarcoma, thereby obtaining an osteosarcoma prognostic risk prediction model, including: The machine learning algorithm used is the random forest algorithm; During training, five-fold cross-validation and early stopping mechanism are used, combined with grid search and Bayesian optimization to determine the optimal hyperparameters for model training. During training, a conservative adjustment strategy is used to conservatively adjust the output probability.

6. The osteosarcoma prognostic risk prediction system according to claim 5, characterized in that, The optimal hyperparameters include any one of the following or any combination thereof: 80 decision trees, 6 maximum depth, 15 minimum number of split samples, 10 minimum number of leaf nodes, and the square root of the maximum number of features.

7. The osteosarcoma prognostic risk prediction system according to claim 6, characterized in that, Also includes: Based on the prediction results of the osteosarcoma prognostic risk prediction model during the training process, features that contribute highly to the prediction results are selected from medical data as inference features. The features that contribute highly to the prediction results include any one of the following or any combination thereof: Enneking stage, initial ALP, ALP change value, initial α-HBDH, and a combination of stage and age.

8. The osteosarcoma prognostic risk prediction system according to any one of claims 1-7, characterized in that, The prediction module includes: The prediction submodule is used to: input the test data of the test subject into the osteosarcoma prognostic risk prediction model to obtain the predicted probability of the osteosarcoma prognostic risk of the test subject; The risk level classification submodule is used to: based on the predicted probability of osteosarcoma prognosis risk for the subject, and using a preset risk level classification standard, obtain the subject's osteosarcoma prognosis risk level, which serves as the subject's osteosarcoma prognosis risk prediction result.

9. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the program, it performs the following steps: Medical data of the subject is received from at least one terminal as inference features, wherein the subject is an osteosarcoma patient and the medical data includes at least one of the following or any combination thereof: demographic and baseline clinical data, tumor burden and malignancy data, and serum tumor marker data; The medical data used as inference features are input into the preset osteosarcoma prognostic risk prediction model to obtain the osteosarcoma prognostic risk prediction results for the test subjects. The results of the osteosarcoma prognostic risk prediction for the test subject are sent to at least one terminal.

10. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that, When the processor executes the program, it performs the following steps: Medical data of the subject is received from at least one terminal as inference features, wherein the subject is an osteosarcoma patient and the medical data includes at least one of the following or any combination thereof: demographic and baseline clinical data, tumor burden and malignancy data, and serum tumor marker data; The medical data used as inference features are input into the preset osteosarcoma prognostic risk prediction model to obtain the osteosarcoma prognostic risk prediction results for the test subjects. The results of the osteosarcoma prognostic risk prediction for the test subject are sent to at least one terminal.

Images