CN-122025115-A - Cancer occurrence risk prediction method based on multi-component fatty acid joint detection

CN122025115ACN 122025115 ACN122025115 ACN 122025115ACN-122025115-A

Abstract

The invention relates to the technical field of cancer prediction, in particular to a cancer occurrence risk prediction method based on multi-component fatty acid joint detection, which comprises the steps of data acquisition, mobility abnormality judgment, metabolic function index determination, inflammatory carcinogenesis index determination, preset model prediction, prediction difference abnormality judgment, threshold period adjustment and risk report generation. According to the invention, multicomponent fatty acid data are converted into higher-order functional marking parameters such as fluidity index, inflammation balance index and the like, nutrition intervention responsiveness and long-term metabolism stability parameters are integrated, a preset prediction model is adopted to perform secondary prediction, calculate prediction difference degree, and a threshold self-adaptive adjustment mechanism is established, so that data-driven decision support is provided for accurate cancer prevention and personalized intervention schemes, and the problems of early warning deficiency and personalized intervention deficiency caused by dependency on invasive examination, single biomarkers and static prediction in the prior art are effectively solved.

Inventors

LIU YINGHUA
ZHAO FENG
LIU XIANGRONG
LIU LU
FENG XIJIA

Assignees

中国人民解放军总医院第一医学中心

Dates

Publication Date: 20260512
Application Date: 20251212

Claims (10)

1. A method for predicting risk of developing cancer based on a combination of multicomponent fatty acids, comprising: Acquiring a fluidity index, an inflammation balance index, an activity index of fatty acid desaturase, a nutrition intervention responsiveness of a cell membrane acquired by multi-component fatty acid in a blood sample based on a target, a detection text of blood sample detection data and metabolic stability of the fatty acid in historical health data in a past preset first historical period; Judging whether the cell membrane fluidity is abnormal or not according to the fluidity index and a preset fluidity threshold value so as to obtain a fluidity abnormal result; determining a metabolic function index from the activity index, the metabolic stability, and the nutritional intervention responsiveness based on the flow anomaly result; Determining an inflammatory oncogenic index from the fluidity index, the inflammatory balance index, the activity index based on the flow anomaly result; Inputting the metabolic function index, the inflammatory carcinogenic index and the detection text into a preset prediction model to obtain a prediction difference degree and a risk level prediction result; judging and outputting an abnormal judgment result or the risk level prediction result according to the prediction difference degree and a preset difference threshold value; adjusting the preset mobility threshold or the preset first history period according to the distribution characteristics of the abnormality judgment result and the risk level prediction result in a preset second history period; and generating a cancer risk prediction report according to the risk level prediction result regenerated after the preset mobility threshold value or the preset first history period, the metabolic function index and the inflammatory cancerogenic index.
2. The method for predicting risk of developing cancer based on the joint detection of multicomponent fatty acids according to claim 1, wherein the abnormal flow result is determined when the abnormality of the cell membrane is determined based on the fluidity index being smaller than the preset fluidity threshold.
3. The method for predicting risk of developing cancer based on the joint detection of multicomponent fatty acids according to claim 2, wherein the process of determining metabolic function index from the activity index, the metabolic stability and the nutritional intervention responsiveness comprises: Performing Z-score standardization treatment on the activity index, the metabolic stability and the nutrition intervention responsiveness to obtain an activity index standard value, a stability standard value and a responsiveness standard value; Constructing a feature vector according to the activity index standard value, the stability standard value and the responsiveness standard value to obtain a metabolic function vector; calculating the mahalanobis distance from the metabolic function vector to a preset metabolic normal cluster center to obtain a reference deviation degree; calculating the mahalanobis distance from the metabolic function vector to a preset metabolic disorder cluster center to obtain a disorder deviation degree; Determining the metabolic function index from the reference deviation and the obstacle deviation.
4. The method for predicting risk of developing cancer based on the joint detection of multicomponent fatty acids as set forth in claim 3, wherein the process of determining the metabolic function index from the reference deviation and the obstacle deviation comprises: calculating Markov probability of the transition of the target to the metabolic dysfunction state based on a preset state transition probability matrix; And calculating the logarithm of the product of the Markov probability and the ratio of the reference deviation degree to the obstacle deviation degree to obtain the metabolic function index.
5. The method for predicting risk of developing cancer based on the joint detection of multicomponent fatty acids according to claim 4, wherein determining an inflammatory carcinogenic index from the fluidity index, the inflammatory equilibrium index, and the activity index comprises: Normalizing the fluidity index, the inflammation balance index and the activity index to obtain a fluidity index normalization value, an inflammation balance normalization value and an activity index normalization value, and constructing a feature vector according to the fluidity index normalization value, the inflammation balance normalization value and the activity index normalization value to obtain an inflammation feature vector; Calculating an included angle cosine value of the inflammation characteristic vector and a preset high-risk vector to obtain risk trend similarity; Calculating the modular length of the inflammation characteristic vector to obtain the system disorder degree; Calculating to obtain a system imbalance degree according to the fluidity index normalization value, the inflammation balance normalization value and the activity index normalization value; calculating the ratio of the product of the risk potential and the system disturbance to obtain the inflammatory carcinogenicity index.
6. The method of claim 5, wherein calculating a systematic imbalance from the fluidity index normalized value, the inflammation balance normalized value, and the activity index normalized value comprises: Constructing a metabolic function triangle by taking the fluidity index normalization value, the inflammation balance normalization value and the activity index normalization value as three side lengths of the triangle, and calculating the area of the metabolic function triangle to obtain the metabolic function area; and calculating the ratio of the metabolic function area to a preset area threshold value to obtain the unbalance degree of the system.
7. The method for predicting risk of developing cancer based on the joint detection of multicomponent fatty acids according to claim 6, wherein the process of determining to output an abnormality determination result or the risk level prediction result according to the prediction difference degree and a preset difference threshold value comprises: when the prediction difference degree is larger than the preset difference threshold value, judging that prediction abnormality exists, and outputting the abnormality judgment result; And when the prediction difference degree is smaller than or equal to the preset difference threshold value, judging that the prediction is normal, and outputting the risk level prediction result.
8. The method of claim 7, wherein adjusting the preset fluidity threshold or the preset first history period according to the distribution characteristics of the abnormality determination result and the risk level prediction result in a preset second history period comprises: Calculating the standard deviation of the time distance between the abnormality judgment result and the adjacent risk level prediction result in the preset second history period to obtain a time fluctuation value; counting the occurrence frequency of the abnormality judgment result in the preset second history period to obtain a history abnormality frequency; and adjusting the preset fluidity threshold value or the preset first history period according to the time fluctuation value and the history abnormal frequency.
9. The method of claim 8, wherein adjusting the preset fluidity threshold or the preset first history period according to the time fluctuation value and the history abnormality frequency comprises: When the time fluctuation value is larger than a preset fluctuation threshold value and the historical abnormal frequency is larger than a preset frequency threshold value, reducing the preset mobility threshold value; And when the time fluctuation value is smaller than a preset fluctuation threshold value and the historical abnormal frequency is smaller than a preset frequency threshold value, increasing the preset first historical period.
10. The method of claim 9, wherein generating a cancer risk prediction report based on the risk level prediction result regenerated after adjusting the preset fluidity threshold or the preset first history period, the metabolic function index, and the inflammatory carcinogenesis index comprises: comprehensively adjusting the preset mobility threshold value or the risk level prediction result regenerated after the preset first history period, the metabolic function index and the inflammatory carcinogenic index to generate a cancer comprehensive prediction result; And generating the cancer risk prediction report output by the comprehensive cancer prediction result.

Description

Cancer occurrence risk prediction method based on multi-component fatty acid joint detection Technical Field The invention relates to the technical field of cancer prediction, in particular to a cancer occurrence risk prediction method based on multi-component fatty acid combined detection. Background Along with the increase of the demands of accurate medicine and large-scale health management, the traditional cancer screening relying on imaging and tissue biopsy is difficult to meet the demands of large-scale and dynamic early warning due to strong invasiveness, high cost and insufficient early sensitivity. The existing method based on single biomarker or static prediction is easily affected by individual difference and short-term fluctuation, and cannot realize comprehensive judgment and continuous self-adaptive adjustment of metabolic and inflammatory microenvironment. Under the background, how to construct a noninvasive and dynamic risk prediction system based on multi-component fatty acid and combine an adaptive threshold value and a comprehensive prediction model to realize early warning and personalized intervention becomes a key technical problem to be solved urgently. The Chinese patent application publication No. CN117912694A discloses a multi-mode cancer survival risk prediction method based on deep learning, which comprises the steps of collecting a full slide image of a cancer patient and corresponding clinical data and gene data, processing the full slide image by adopting a Macehko color normalization method, processing the clinical data by adopting a word embedding method, carrying out feature screening on the gene data by adopting a Cox single factor analysis method, constructing a prediction dataset based on the processed full slide image, the clinical data and the gene data, constructing a survival prediction model by adopting a multi-instance learning algorithm based on attention, acquiring a survival prediction result of the cancer patient based on the survival prediction model and the prediction dataset, and training and predicting the survival prediction model. Therefore, the multi-mode cancer survival risk prediction method based on deep learning has the following problems that the method utilizes gene characteristics to screen, complex interaction between genes is easy to ignore to influence a prediction result, input data of the method are data after a disease is in a state, reflection of metabolism or inflammation mechanism before occurrence of cancer is lacking, a model in the method depends on static data, and dynamic prediction capability of change of an individual health state along with time is lacking. Disclosure of Invention Therefore, the invention provides a cancer occurrence risk prediction method based on multi-component fatty acid joint detection, which is used for overcoming the problems of insufficient early warning and personalized intervention deficiency caused by dependency on invasive examination, single biomarker and static prediction in the prior art by integrating a multi-component fatty acid index system and a dynamic threshold adjustment mechanism and combining an adaptive AI prediction model. To achieve the above object, the present invention provides a method for predicting risk of cancer occurrence based on a multi-component fatty acid combination test, comprising: Acquiring a fluidity index, an inflammation balance index, an activity index of fatty acid desaturase, a nutrition intervention responsiveness of a cell membrane acquired by multi-component fatty acid in a blood sample based on a target, a detection text of blood sample detection data and metabolic stability of the fatty acid in historical health data in a past preset first historical period; Judging whether the cell membrane fluidity is abnormal or not according to the fluidity index and a preset fluidity threshold value so as to obtain a fluidity abnormal result; determining a metabolic function index from the activity index, the metabolic stability, and the nutritional intervention responsiveness based on the flow anomaly result; Determining an inflammatory oncogenic index from the fluidity index, the inflammatory balance index, the activity index based on the flow anomaly result; Inputting the metabolic function index, the inflammatory carcinogenic index and the detection text into a preset prediction model to obtain a prediction difference degree and a risk level prediction result; judging and outputting an abnormal judgment result or the risk level prediction result according to the prediction difference degree and a preset difference threshold value; adjusting the preset mobility threshold or the preset first history period according to the distribution characteristics of the abnormality judgment result and the risk level prediction result in a preset second history period; and generating a cancer risk prediction report according to the risk level prediction result regenerated after the