[go: up one dir, main page]

CN118072960B - A method for predicting cachexia after radiotherapy for head and neck tumors - Google Patents

A method for predicting cachexia after radiotherapy for head and neck tumors Download PDF

Info

Publication number
CN118072960B
CN118072960B CN202410465916.8A CN202410465916A CN118072960B CN 118072960 B CN118072960 B CN 118072960B CN 202410465916 A CN202410465916 A CN 202410465916A CN 118072960 B CN118072960 B CN 118072960B
Authority
CN
China
Prior art keywords
cachexia
data
radiotherapy
occurrence
probability
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Active
Application number
CN202410465916.8A
Other languages
Chinese (zh)
Other versions
CN118072960A (en
Inventor
贾燕
雷蕾
陈雪峰
陆菲清
傅晓炜
蔡亚南
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Zhejiang Cancer Hospital
Original Assignee
Zhejiang Cancer Hospital
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Zhejiang Cancer Hospital filed Critical Zhejiang Cancer Hospital
Priority to CN202410465916.8A priority Critical patent/CN118072960B/en
Publication of CN118072960A publication Critical patent/CN118072960A/en
Application granted granted Critical
Publication of CN118072960B publication Critical patent/CN118072960B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Classifications

    • G—PHYSICS
    • G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
    • G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
    • G16H50/00—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics
    • G16H50/30—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics for calculating health indices; for individual health risk assessment
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/23—Clustering techniques
    • G06F18/232—Non-hierarchical techniques
    • G06F18/2321—Non-hierarchical techniques using statistics or function optimisation, e.g. modelling of probability density functions
    • G06F18/23213—Non-hierarchical techniques using statistics or function optimisation, e.g. modelling of probability density functions with fixed number of clusters, e.g. K-means clustering
    • G—PHYSICS
    • G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
    • G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
    • G16H10/00—ICT specially adapted for the handling or processing of patient-related medical or healthcare data
    • G16H10/60—ICT specially adapted for the handling or processing of patient-related medical or healthcare data for patient-specific data, e.g. for electronic patient records
    • G—PHYSICS
    • G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
    • G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
    • G16H50/00—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics
    • G16H50/70—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics for mining of medical data, e.g. analysing previous cases of other patients
    • Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
    • Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
    • Y02A—TECHNOLOGIES FOR ADAPTATION TO CLIMATE CHANGE
    • Y02A90/00—Technologies having an indirect contribution to adaptation to climate change
    • Y02A90/10—Information and communication technologies [ICT] supporting adaptation to climate change, e.g. for weather forecasting or climate simulation

Landscapes

  • Engineering & Computer Science (AREA)
  • Data Mining & Analysis (AREA)
  • Medical Informatics (AREA)
  • Public Health (AREA)
  • Health & Medical Sciences (AREA)
  • Primary Health Care (AREA)
  • General Health & Medical Sciences (AREA)
  • Epidemiology (AREA)
  • Pathology (AREA)
  • Databases & Information Systems (AREA)
  • Biomedical Technology (AREA)
  • Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • Bioinformatics & Cheminformatics (AREA)
  • General Physics & Mathematics (AREA)
  • General Engineering & Computer Science (AREA)
  • Evolutionary Computation (AREA)
  • Evolutionary Biology (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Bioinformatics & Computational Biology (AREA)
  • Artificial Intelligence (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Probability & Statistics with Applications (AREA)
  • Radiation-Therapy Devices (AREA)

Abstract

The application provides a method for predicting concurrent malignant fluid after radiotherapy of head and neck tumor, and relates to the technical field of disease prediction. In the process of predicting the concurrent cachexia of the head and neck tumor after radiotherapy, the conditions of physique, living habit, disease state, character characteristics, selected treatment scheme and the like of a patient are considered, and the change condition of relevant sign data of the patient in the period of radiotherapy is considered, so that a cachexia prediction result determination model is trained through big data, and the probability of occurrence of the concurrent cachexia of the head and neck tumor after radiotherapy can be obtained accurately, and the method can be provided for doctors to formulate reasonable treatment schemes.

Description

Method for predicting concurrent cachexia after radiotherapy of head and neck tumor
Technical Field
The invention belongs to the technical field of disease prediction, and particularly relates to a method for predicting concurrent malignant fluid quality after radiotherapy of head and neck tumors.
Background
The cachexia is an important factor affecting the survival time of patients with head and neck tumor after radiotherapy, so how to accurately predict the probability of cachexia and perform intervention treatment early is a problem to be solved urgently for prolonging the life of patients.
In the prior art, diagnosis can be generally performed only after the patient has malignant fluid, and the best treatment time is often missed. The research shows that the occurrence of cachexia is affected by various factors including the demographic characteristics, life habit characteristics, character characteristics, disease state characteristics and treatment scheme of patients, and can be reflected by the change condition of various sign indexes in the radiotherapy process of patients.
Disclosure of Invention
The invention aims to provide a method for predicting concurrent cachexia after radiotherapy of head and neck tumors, which aims to solve the technical problem that the concurrent cachexia after radiotherapy of head and neck tumors cannot be accurately predicted in the prior art.
A method for predicting concurrent cachexia after radiation therapy of a head and neck tumor, the method comprising:
s1: generating an input vector according to the personal condition of a user to be predicted, inputting the input vector into a first cachexia prediction result determination model, determining a first clustering center, and determining a first cachexia occurrence probability according to the input vector and the first clustering center;
The step S1 specifically comprises the following substeps:
S11: acquiring demographic data, life habit data, disease state data, character characteristic data and selected treatment scheme data of a user to be predicted, and preprocessing the data;
S12: forming a first input vector according to the preprocessed data, inputting the first input vector into a first cachexia prediction result determining model, and determining a first clustering center;
s13: generating a first cacheline occurrence probability of the user to be predicted according to the Euclidean distance between the first input vector and the first clustering center;
The step S13 specifically comprises the following substeps:
S131: judging whether the first clustering center has the condition of generating cachexia, if so, turning to a step S132, otherwise, ending and outputting the first cachexia generation probability as 0;
S132: generating a first cacheline occurrence probability according to the severity condition of the first clustering center and the Euclidean distance;
wherein the first cachexia occurrence probability is inversely related to the euclidean distance;
S2: collecting first sign data of the user to be predicted at a preset time node within a preset time point before radiotherapy and a preset time point after radiotherapy is finished, and forming a first relative sign data sequence;
S3: inputting the first relative sign data sequence into a second cachexia prediction result determination model to output a second cachexia occurrence probability of the user to be predicted;
S4: and inputting the first cachexia occurrence probability and the second cachexia occurrence probability into a third cachexia prediction result determination model, thereby obtaining the final cachexia prediction probability of the user to be predicted.
Preferably, the demographic data includes gender, age, and work type data of the user to be predicted;
The lifestyle data includes smoking history, drinking history, diet preference data and work-rest habit data;
the character characteristic data comprise an inward type and an outward type;
the disease state data comprise tumor parts, past disease history, stage conditions and metastasis conditions;
The treatment plan data include surgical conditions, radiation patterns, induction of chemotherapy, concurrent chemotherapy, and targeted therapy.
Preferably, the first cachexia prediction result determination model is obtained by:
s121: acquiring sample data for training the first cachexia prediction result determination model;
S122: clustering the sample data by adopting a K-MEANS algorithm, and acquiring a plurality of clustering centers;
S123: and labeling the occurrence of the cachexia condition and the severity condition aiming at the clustering centers.
Preferably, the first sign data specifically comprises body weight, BMI, nutritional score, life self-care score, anorexia assessment, psychological assessment, muscle content, fat content, total protein content, grip strength, circumference of lower leg, CRP, white blood cell count, neutrophil count, lymphocyte count, hemoglobin content, total protein content, albumin content, prealbumin content, retinol binding protein content, IL-6 content, IL-1 content, tumor necrosis factor.
Preferably, the forming the first relative sign data sequence specifically includes:
And calculating the change value of each subsequent sign monitoring data based on the first sign monitoring data of the user to be predicted, and forming the first relative sign data sequence based on the change value.
Preferably, the second cachexia prediction result determination model is specifically obtained by training in the following manner:
Firstly, extracting physical sign data of a patient from a preset time point before radiotherapy to a preset time point after the radiotherapy is finished aiming at a plurality of patients which are selected to be subjected to radiotherapy due to head and neck tumors, further forming a plurality of second physical sign data sequences, and labeling each piece of sample data according to the probability of cachexia occurrence according to the final treatment result;
And secondly, taking a plurality of second sign data sequences as input, taking the bad liquid occurrence probability as output, training a convolutional neural network model, and obtaining the second bad liquid prediction result determination model after the training accuracy reaches the preset requirement.
The method for predicting the concurrent cachexia after the radiotherapy of the head and neck tumor provided by the application is used for obtaining more accurate probability of occurrence of the concurrent cachexia after the radiotherapy of the head and neck tumor by training a cachexia prediction result determination model through big data in the process of predicting the concurrent cachexia after the radiotherapy of the head and neck tumor, taking the conditions of physique, living habit, disease state, character characteristics, selected treatment scheme and the like of a patient into consideration, and taking the condition of relevant sign data change of the patient in the period of the radiotherapy into consideration, so that the method can be provided for doctors to formulate reasonable treatment schemes.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the embodiments or the prior art will be briefly described below. It will be apparent to those of ordinary skill in the art that the drawings in the following description are exemplary only and that other implementations can be obtained from the extensions of the drawings provided without inventive effort.
FIG. 1 is a flow chart showing the execution of a method for predicting the concurrent cachexia after radiotherapy of a head and neck tumor according to the present invention.
Detailed Description
The following description of the embodiments of the present invention will be made clearly and completely with reference to the accompanying drawings, in which it is apparent that the embodiments described are only some embodiments of the present invention, but not all embodiments. All other embodiments, which can be made by those skilled in the art based on the embodiments of the invention without making any inventive effort, are intended to be within the scope of the invention.
The present invention will now be described in detail with reference to the drawings and the specific embodiments thereof, wherein the exemplary embodiments and the description are for the purpose of illustrating the invention only and are not to be construed as limiting the invention.
A method for predicting concurrent cachexia after radiotherapy of a head and neck tumor according to the present invention is described in detail below.
A method for predicting concurrent cachexia after radiotherapy of head and neck tumor, the specific flow is shown in figure 1, comprises the following steps:
s is S1 the method comprises the following steps: and generating an input vector according to the personal condition of the user to be predicted, inputting the input vector into a first cachexia prediction result determination model, determining a first clustering center, and determining the occurrence probability of the first cachexia according to the input vector and the first clustering center.
The risk of complications of cachexia following radiotherapy of head and neck tumors is often greatly related to the patient's constitution, lifestyle, disease state, personality characteristics and the treatment regimen selected. Therefore, in this step, it is necessary to extract the personal situation of the user to be predicted, and perform preliminary prediction of the cachexia prediction result according to the personal situation of the user to be predicted, so as to form a first cachexia prediction result, that is, a first cachexia occurrence probability.
The step S1 specifically comprises the following sub-steps:
S11: and acquiring demographic data, life habit data, disease state data, character characteristic data and selected treatment scheme data of the user to be predicted, and preprocessing the data.
Specifically, the demographic data includes gender, age, and work type data of the user to be predicted. The lifestyle data includes smoking history, drinking history, diet preference data, and work-rest habit data. The character characteristic data comprises an inward type and an outward type. The disease state data includes tumor sites, past disease history, stage conditions, and metastasis conditions. The treatment plan data include surgical conditions, radiation patterns, induction of chemotherapy, concurrent chemotherapy, and targeted therapy.
The preprocessing is specifically to label the acquired data of the user to be predicted according to the requirement of the input vector, reject the data which deviate from the normal value obviously, and fill blank items according to the data average value.
S12: and forming a first input vector according to the preprocessed data, inputting the first input vector into a first cachexia prediction result determining model, and determining a first clustering center.
The first clustering center is the clustering center nearest to the first input vector.
The first cachexia prediction result determination model is obtained by the following steps:
s121: sample data for training the first cachexia prediction result determination model is obtained.
Specifically, for a specified number of head and neck tumor patients, data extraction is performed according to the data requirements of the input vector, and cachexia condition labeling is performed for each sample data. The cachexia includes whether or not cachexia has occurred and the severity after a specified period of time.
S122: and clustering the sample data by adopting a K-MEANS algorithm, and acquiring a plurality of clustering centers.
Because patients with different conditions usually have a certain rule of cachexia risk and severity, after the sample data are clustered, the patients with the same type can be classified into one type, so that a guide can be provided for prediction of cachexia prediction results.
For example, an input vector is formed for one sample data [ male, 42 years old, heavy physical work, no smoking, drinking, preference for heavy oil heavy salt, night-stay habit, inward personality, nasopharyngeal carcinoma, hypertension, diabetes, four phases, distant metastasis, no surgery, TOMO radiotherapy, TP-induced chemotherapy, cisplatin concurrent chemotherapy, thioxin-targeted therapy ], and an input vector is formed for other sample data according to the above format. And then, clustering a plurality of input vectors corresponding to the sample data by adopting a K-MEANS algorithm, thereby forming a plurality of clustering centers.
S123: and labeling the occurrence of the cachexia condition and the severity condition aiming at the clustering centers.
After clustering is performed on the input vectors corresponding to the sample data, the generated cluster centers complete simple images of similar patients. Then, for patients with similar situations, labeling is needed for the situation and severity of whether the cachexia occurs in the corresponding cluster center.
Specifically, for a plurality of sample data corresponding to each cluster center, the condition and severity of cachexia are counted. For example, for the first cluster center, if there are 10 sample data corresponding to the first cluster center, 10 cases of cachexia and 10 cases of severity are obtained respectively, and the cases of cachexia and the cases of severity of the first cluster center are obtained through statistical operations, such as counting operations, weighted summation operations, and the like. The occurrence of cachexia includes the occurrence of cachexia, the severity condition includes a severity classification, the severity condition is divided into 3 classes, 3 classes are the most severe, and the 3 classes of severity condition correspond to the occurrence probabilities of cachexia of 33%, 66% and 99%, respectively.
S13: and generating a first cacheline occurrence probability of the user to be predicted according to the Euclidean distance between the first input vector and the first clustering center.
The step S13 specifically includes:
S131: and judging whether the first clustering center has the condition of generating the cachexia, if so, turning to a step S132, otherwise, ending and outputting the first cachexia generation probability as 0.
S132: and generating a first cachexia occurrence probability according to the cachexia occurrence condition of the first clustering center and the Euclidean distance.
The probability of occurrence of the first cachexia is inversely related to the Euclidean distance, that is, the smaller the Euclidean distance is, the larger the probability of occurrence of the first cachexia is.
For example, if the severity of the first cluster center is 2-level, the probability of occurrence of cachexia is 66%. Next, the Euclidean distance between the user to be predicted and the clustering center is normalized to be between [0,1], and the first cachexia occurrence probability is obtained according to the normalization result, specifically, if the normalization result is 0.5, the first cachexia occurrence probability is 33% + (33% ×0.5), namely 49.5%.
S2: and acquiring first sign data of the user to be predicted at a preset time point from a preset time point before radiotherapy to a preset time point after radiotherapy is finished, and forming a first relative sign data sequence.
The cachexia after radiotherapy of head and neck tumor is divided into three stages of cachexia early stage, cachexia and refractory cachexia, and the rapid decrease of muscle content, fat content and total protein content can be accompanied in the course of cachexia development, so that the cachexia development condition of a patient can be predicted more accurately by monitoring the above physical sign data in real time in a preset time period from before radiotherapy to after radiotherapy is finished.
The first sign data may specifically include body weight, BMI, nutritional score, lifestyle score, anorexia assessment, psychological assessment, muscle content, fat content, total protein content, grip strength, calf circumference, CRP, white blood cell count, neutrophil count, lymphocyte count, hemoglobin, total protein, albumin, prealbumin, retinol binding protein, IL-6, IL-1, tumor necrosis factor, and the like.
The preset time point can be set according to specific conditions, and preferably, 1 month before radiotherapy and 1 month after radiotherapy is finished can be set as the preset time point.
The preset time node can be set according to specific conditions, and preferably, the preset time node can be set to be 7 days.
The forming the first relative sign data sequence specifically includes:
And calculating the change value of each subsequent sign monitoring data based on the first sign monitoring data of the user to be predicted, and forming the first relative sign data sequence based on the change value.
For example, the radiotherapy period is 1 half month, the preset time points are respectively 1 month before radiotherapy and 1 month after radiotherapy, and the sign monitoring period is 3 half months, and 1 sign monitoring is performed every 7 days. Taking the muscle content as an example of the first sign data, the monitoring values of a plurality of monitoring time points are respectively 50, 49, 48, 46 and 42 … …, and the first relative sign data sequence is [50, -1, -2, -4 … … ], namely the decay rate of the muscle content is reflected in the first sign data sequence.
S3: and inputting the first relative sign data sequence into a second cachexia prediction result determination model to output the second cachexia occurrence probability of the user to be predicted.
The second cachexia prediction result determination model is used for carrying out the probability of cachexia occurrence according to the sign attenuation sequence of the patient.
The second cachexia prediction result determination model is specifically obtained through training in the following way:
First, sample data is acquired. For a plurality of patients selected to be subjected to radiotherapy due to head and neck tumors, extracting physical sign data of the patients from a preset time point before radiotherapy to a preset time point after the radiotherapy is finished by using the preset time point, so as to form a plurality of second physical sign data sequences. Wherein the sign data comprises body weight, BMI, nutritional score, life self-ability score, anorexia assessment, psychological assessment, muscle content, fat content, total protein content, grip strength, calf circumference, CRP, white blood cell count, neutrophil count, lymphocyte count, hemoglobin, total protein, albumin, prealbumin, retinol binding protein, IL-6, IL-1, tumor necrosis factor, etc. In addition, for each piece of sample data, the probability of cachexia occurrence is marked according to the final treatment result.
And secondly, taking a plurality of second sign data sequences as input, taking the bad liquid occurrence probability as output, training a convolutional neural network model, and obtaining the second bad liquid prediction result determination model after the training accuracy reaches the preset requirement.
S4: and inputting the first cachexia occurrence probability and the second cachexia occurrence probability into a third cachexia prediction result determination model, thereby obtaining the final cachexia prediction probability of the user to be predicted.
And the step is to integrate the personal condition of the user to be predicted and the physical sign state change condition within the duration of radiotherapy so as to obtain the final cachexia prediction probability of the user to be predicted.
The third cachexia prediction result determination model is obtained through training in the following way:
first, sample data is acquired. Aiming at a plurality of patients suffering from head and neck tumors, the first cachexia occurrence probability and the second cachexia occurrence probability are respectively obtained according to the individual condition and the sign state change condition within the duration of radiotherapy, and finally cachexia occurrence is marked according to the actual treatment result condition.
And secondly, taking the first and second bad liquid occurrence probabilities of a plurality of sample data as inputs, taking the final bad liquid condition as output, and training a convolutional neural network model to obtain the third bad liquid prediction result determination model.
Preferably, the first cachexia occurrence probability, the second cachexia occurrence probability and the final cachexia prediction probability may be provided to the attending physician for the physician to set up a treatment regimen, respectively.
The method for predicting the concurrent cachexia after the radiotherapy of the head and neck tumor provided by the application is used for obtaining more accurate probability of occurrence of the concurrent cachexia after the radiotherapy of the head and neck tumor by training a cachexia prediction result determination model through big data in the process of predicting the concurrent cachexia after the radiotherapy of the head and neck tumor, taking the conditions of physique, living habit, disease state, character characteristics, selected treatment scheme and the like of a patient into consideration, and taking the condition of relevant sign data change of the patient in the period of the radiotherapy into consideration, so that the method can be provided for doctors to formulate reasonable treatment schemes.
The foregoing description is only of the preferred embodiments of the invention, and all changes and modifications that come within the meaning and range of equivalency of the structures, features and principles of the invention are therefore intended to be embraced therein.

Claims (5)

1.一种用于预测头颈肿瘤放疗后并发恶液质的方法，其特征在于，该方法包括：1. A method for predicting cachexia after radiotherapy for head and neck tumors, characterized in that the method comprises: S1：根据待预测用户的个人情况生成输入向量，输入第一恶液质预测结果确定模型中，确定第一聚类中心，并根据所述输入向量和所述第一聚类中心确定第一恶液质发生概率；S1: generating an input vector according to the personal situation of the user to be predicted, inputting the input vector into a first cachexia prediction result determination model, determining a first cluster center, and determining the probability of occurrence of the first cachexia according to the input vector and the first cluster center; 所述S1具体包括如下子步骤：The S1 specifically includes the following sub-steps: S11：获取待预测用户的人口学数据、生活习惯数据、疾病状态数据、性格特点数据和所选治疗方案数据，对上述数据进行预处理；S11: Obtaining demographic data, living habit data, disease status data, personality characteristics data and selected treatment plan data of the user to be predicted, and preprocessing the above data; S12：根据预处理后的数据形成第一输入向量，并输入至第一恶液质预测结果确定模型中，确定第一聚类中心；S12: forming a first input vector according to the preprocessed data, and inputting the first input vector into a first cachexia prediction result determination model to determine a first cluster center; S13：根据所述第一输入向量与所述第一聚类中心的欧氏距离，生成所述待预测用户的第一恶液质发生概率；S13: generating a first cachexia occurrence probability of the user to be predicted according to the Euclidean distance between the first input vector and the first cluster center; 所述S13具体包括如下子步骤：The S13 specifically includes the following sub-steps: S131：判断所述第一聚类中心是否存在发生恶液质的情况，若为是则转至步骤S132，否则结束并将所述第一恶液质发生概率输出为0；S131: Determine whether cachexia occurs in the first cluster center. If yes, go to step S132; otherwise, terminate and output the first cachexia occurrence probability as 0; S132：根据所述第一聚类中心的严重程度情况和所述欧氏距离，生成第一恶液质发生概率；S132: generating a first cachexia occurrence probability according to the severity of the first cluster center and the Euclidean distance; 其中，所述第一恶液质发生概率与所述欧氏距离成负相关；Wherein, the probability of occurrence of the first cachexia is negatively correlated with the Euclidean distance; S2：在放疗前预设时间点至放疗结束后预设时间点内，以预设时间节点采集所述待预测用户的第一体征数据，并形成第一相对体征数据序列；所述形成第一相对体征数据序列具体包括：以所述待预测用户的第一次体征监测数据为基础，计算后续每一次体征监测数据的变化值，并以此形成所述第一相对体征数据序列；S2: within a preset time point before radiotherapy to a preset time point after radiotherapy, collecting the first vital sign data of the user to be predicted at preset time nodes, and forming a first relative vital sign data sequence; the forming of the first relative vital sign data sequence specifically includes: based on the first vital sign monitoring data of the user to be predicted, calculating the change value of each subsequent vital sign monitoring data, and forming the first relative vital sign data sequence based on this; S3：将所述第一相对体征数据序列输入至第二恶液质预测结果确定模型中，以输出所述待预测用户的第二恶液质发生概率；S3: inputting the first relative vital sign data sequence into a second cachexia prediction result determination model to output the probability of occurrence of the second cachexia of the user to be predicted; S4：将所述第一恶液质发生概率和所述第二恶液质发生概率输入第三恶液质预测结果确定模型中，从而获得所述待预测用户的最终恶液质预测概率。S4: Inputting the first cachexia occurrence probability and the second cachexia occurrence probability into a third cachexia prediction result determination model, thereby obtaining a final cachexia prediction probability of the user to be predicted. 2.根据权利要求1所述的一种用于预测头颈肿瘤放疗后并发恶液质的方法，其特征在于，所述人口学数据包括所述待预测用户的性别、年龄、从事工作类型数据；2. A method for predicting cachexia after radiotherapy for head and neck tumors according to claim 1, characterized in that the demographic data includes the gender, age, and type of work of the user to be predicted; 所述生活习惯数据包括吸烟史、饮酒史、饮食偏好数据和作息习惯数据；The lifestyle data include smoking history, drinking history, dietary preference data and work and rest habits data; 所述性格特点数据包括内向型、外向型；The personality trait data include introversion and extroversion; 所述疾病状态数据包括肿瘤部位、既往疾病史、分期情况、转移情况；The disease status data include tumor location, previous disease history, stage, and metastasis; 所述治疗方案数据包括手术情况、放疗方式、诱导化疗、同步化疗、靶向治疗。The treatment plan data include surgical conditions, radiotherapy methods, induction chemotherapy, concurrent chemotherapy, and targeted therapy. 3.根据权利要求2所述的一种用于预测头颈肿瘤放疗后并发恶液质的方法，其特征在于，所述第一恶液质预测结果确定模型通过如下方式获得：3. A method for predicting cachexia after radiotherapy for head and neck tumors according to claim 2, characterized in that the first cachexia prediction result determination model is obtained by the following method: S121：获取用于训练所述第一恶液质预测结果确定模型的样本数据；S121: Acquire sample data for training the first cachexia prediction result determination model; S122：针对所述样本数据，采用K-MEANS算法进行聚类，并获取若干聚类中心；S122: clustering the sample data using the K-MEANS algorithm and obtaining a number of cluster centers; S123：针对所述若干聚类中心，标注发生恶液质情况和严重程度情况。S123: For the plurality of cluster centers, the occurrence and severity of cachexia are marked. 4.根据权利要求3所述的一种用于预测头颈肿瘤放疗后并发恶液质的方法，其特征在于，所述第一体征数据具体包括体重、BMI、营养评分、生活自理能力评分、厌食评估、心理评估、肌肉含量、脂肪含量、总蛋白含量、握力、小腿周长、CRP、白细胞计数、中性粒细胞计数、淋巴细胞计数、血红蛋白含量、总蛋白含量、白蛋白含量、前白蛋白含量、视黄醇结合蛋白含量、IL-6含量、IL-1含量、肿瘤坏死因子。4. A method for predicting cachexia after radiotherapy for head and neck tumors according to claim 3, characterized in that the first physical sign data specifically includes weight, BMI, nutritional score, self-care ability score, anorexia assessment, psychological assessment, muscle content, fat content, total protein content, grip strength, calf circumference, CRP, white blood cell count, neutrophil count, lymphocyte count, hemoglobin content, total protein content, albumin content, prealbumin content, retinol binding protein content, IL-6 content, IL-1 content, and tumor necrosis factor. 5.根据权利要求1所述的一种用于预测头颈肿瘤放疗后并发恶液质的方法，其特征在于，所述第二恶液质预测结果确定模型具体通过如下方式训练获得：5. The method for predicting cachexia after radiotherapy for head and neck tumors according to claim 1, wherein the second cachexia prediction result determination model is obtained by training in the following manner: 首先，针对若干因头颈肿瘤而选择进行放射治疗的患者，在其放疗前预设时间点至放疗结束后预设时间点，以预设时间节点提取所述患者的体征数据，进而形成若干第二体征数据序列，针对每一条样本数据，根据其最后的治疗结果为其进行恶液质发生概率标注；First, for a number of patients who choose to undergo radiotherapy for head and neck tumors, the patient's vital sign data is extracted at a preset time point before the radiotherapy to a preset time point after the radiotherapy, thereby forming a number of second vital sign data sequences. For each sample data, the probability of cachexia occurrence is marked according to the final treatment result; 其次，将若干所述第二体征数据序列作为输入，将所述恶液质发生概率作为输出，训练卷积神经网络模型，当训练精度达到预设要求后，获得所述第二恶液质预测结果确定模型。Secondly, a number of the second vital sign data sequences are taken as input, the probability of cachexia occurrence is taken as output, and a convolutional neural network model is trained. When the training accuracy reaches a preset requirement, a second cachexia prediction result determination model is obtained.
CN202410465916.8A 2024-04-18 2024-04-18 A method for predicting cachexia after radiotherapy for head and neck tumors Active CN118072960B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202410465916.8A CN118072960B (en) 2024-04-18 2024-04-18 A method for predicting cachexia after radiotherapy for head and neck tumors

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202410465916.8A CN118072960B (en) 2024-04-18 2024-04-18 A method for predicting cachexia after radiotherapy for head and neck tumors

Publications (2)

Publication Number Publication Date
CN118072960A CN118072960A (en) 2024-05-24
CN118072960B true CN118072960B (en) 2024-08-02

Family

ID=91106162

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202410465916.8A Active CN118072960B (en) 2024-04-18 2024-04-18 A method for predicting cachexia after radiotherapy for head and neck tumors

Country Status (1)

Country Link
CN (1) CN118072960B (en)

Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN107403072A (en) * 2017-08-07 2017-11-28 北京工业大学 A kind of diabetes B prediction and warning method based on machine learning
CN115287356A (en) * 2021-06-29 2022-11-04 中国人民解放军空军军医大学 Early diagnosis kit for gastric cancer cachexia based on the expression level of exosomal miRNA-432-5p

Family Cites Families (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
RU2352948C1 (en) * 2007-06-15 2009-04-20 Государственное образовательное учреждение высшего профессионального образования "КРАСНОЯРСКАЯ ГОСУДАРСТВЕННАЯ МЕДИЦИНСКАЯ АКАДЕМИЯ ФЕДЕРАЛЬНОГО АГЕНТСТВА ПО ЗДРАВООХРАНЕНИЮ И СОЦИАЛЬНОМУ РАЗВИТИЮ" Method of outcome forecast of disease of patients with urgent surgical pathology of organs of abdominal cavity
CN116798646B (en) * 2023-08-17 2023-11-24 四川互慧软件有限公司 Snake injury prognosis prediction method and device based on clustering algorithm and electronic equipment
CN117238489A (en) * 2023-11-16 2023-12-15 四川省肿瘤医院 Prediction method and system for distant metastasis of head and neck tumors in multimodal nuclear medicine imaging

Patent Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN107403072A (en) * 2017-08-07 2017-11-28 北京工业大学 A kind of diabetes B prediction and warning method based on machine learning
CN115287356A (en) * 2021-06-29 2022-11-04 中国人民解放军空军军医大学 Early diagnosis kit for gastric cancer cachexia based on the expression level of exosomal miRNA-432-5p

Also Published As

Publication number Publication date
CN118072960A (en) 2024-05-24

Similar Documents

Publication Publication Date Title
US10846622B2 (en) Methods and systems for an artificial intelligence support network for behavior modification
Nuankaew et al. Average weighted objective distance-based method for type 2 diabetes prediction
CN110246577B (en) Method for assisting gestational diabetes genetic risk prediction based on artificial intelligence
CN112002419B (en) Disease auxiliary diagnosis system, equipment and storage medium based on clustering
EP1399868A2 (en) Information processing method for disease stratification and assessment of disease progressing
CN113539414B (en) A method and system for predicting the rationality of antibiotic use
US11915833B2 (en) Integrated system and method for personalized stratification and prediction of neurodegenerative disease
Vairale et al. Classification of hypothyroid disorder using optimized SVM method
Singh Prediction of Thyroid Disease using Deep Learning Techniques
Wrobel Diagnosing Parkinson’s disease by means of ensemble classification of patients’ voice samples
CN118072960B (en) A method for predicting cachexia after radiotherapy for head and neck tumors
CN113436745A (en) Artificial intelligence auxiliary diagnosis method based on database analysis
CN116631558B (en) A construction method for Internet-based medical testing projects
KR20160000538A (en) System and method for providing health management service
CN114649071A (en) Real world data-based peptic ulcer treatment scheme prediction system
CN112086200B (en) System, equipment and storage medium for predicting diseases based on bitter taste
CN112259231A (en) High-risk gastrointestinal stromal tumor patient postoperative recurrence risk assessment method and system
Abbas et al. Overlapped Symptoms Detection for Cardiovascular Disease Based on Deep Learning Model
Ramadhan et al. Prediction of hypertension in the upcoming year: feature correlation analysis and handling imbalanced based on random forest
CN119153078B (en) Traditional Chinese medicine rehabilitation scheme recommendation method and system
CN118016263B (en) Digital medical assistant system based on voice recognition
Shaker et al. IID3 Classifier To Diagnosis Of High Blood Glucose Levels During Pregnancy.
CN118445419B (en) Method, device, equipment and medium for constructing prompt automatic labeling medical text based on large model
CN109872815A (en) A kind of Disease Warning Mechanism analysis system based on big data
Faris et al. The detection of PCOS using machine learning algorithms and feature selection by K-means clustering

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
GR01 Patent grant
GR01 Patent grant