CN-121987197-A - Cerebral apoplexy monitoring method and system based on cerebral oxygen metabolism rate imaging

CN121987197ACN 121987197 ACN121987197 ACN 121987197ACN-121987197-A

Abstract

The invention discloses a cerebral apoplexy monitoring method and system based on cerebral oxygen metabolic rate imaging, and relates to the technical field of medical care. The cerebral apoplexy monitoring system based on cerebral oxygen metabolism rate imaging comprises a data acquisition and calculation module, a data preprocessing module, a data fusion and imaging module, a brain state prediction module and a comprehensive evaluation and planning module. According to the cerebral oxygen monitoring method and system based on the cerebral oxygen metabolic rate imaging, the current cerebral oxygen metabolic rate of the brain tissue is obtained through calculation based on the preprocessed blood oxygen parameter data and the relative blood flow index data and through a formula, the oxygen supply and oxygen consumption conditions of the brain tissue can be reflected at the same time, the oxygen supply and demand states of the brain are comprehensively reflected, more comprehensive assessment is provided for the cerebral oxygen metabolic state, and the assessment accuracy of the cerebral apoplexy monitoring method and system based on the cerebral oxygen metabolic rate imaging is improved.

Inventors

YUAN ZHEN
WANG QUAN
Xia Xiaoluan
Fan Rongai
SU JINXIANG

Assignees

珠海澳大科技研究院
珠海中信大有科技有限公司
珠海横琴中信堂健康管理有限公司

Dates

Publication Date: 20260508
Application Date: 20260121

Claims (8)

1. The cerebral apoplexy monitoring method based on cerebral oxygen metabolism rate imaging is characterized by comprising the following steps of: Transmitting multi-wavelength near infrared light through a frequency domain near infrared spectrometer, collecting frequency domain signals, obtaining an absorption coefficient and a scattering coefficient of brain tissues, obtaining blood oxygen parameter data of the brain tissues through calculation, transmitting coherent laser through a diffusion correlation spectrometer, detecting speckle light intensity signals, obtaining speckle light intensity autocorrelation functions, and obtaining relative blood flow index data of the brain tissues through calculation; Preprocessing blood oxygen parameter data and relative blood flow index data of brain tissues to obtain preprocessed blood oxygen parameter data and relative blood flow index data; Based on the preprocessed blood oxygen parameter data and the relative blood flow index data, calculating to obtain the current brain oxygen metabolism rate of brain tissues through a formula, processing the calculated current brain oxygen metabolism rate data by using an image reconstruction algorithm, and generating a brain oxygen metabolism distribution image of a brain region; Constructing a brain oxygen metabolism rate prediction model based on an improved LSTM model, and inputting the preprocessed blood oxygen parameter data and the relative blood flow index data into the trained brain oxygen metabolism rate prediction model to obtain a predicted brain oxygen metabolism rate of brain tissue; comprehensively evaluating the brain physiological state of the cerebral apoplexy patients according to the current cerebral oxygen metabolism rate and the predicted cerebral oxygen metabolism rate of the brain tissues to obtain the brain physiological state scores of the cerebral apoplexy patients, and planning corresponding treatment schemes for the cerebral apoplexy patients based on the brain physiological state scores.
2. The cerebral apoplexy monitoring method based on cerebral oxygen metabolism rate imaging according to claim 1, wherein the specific steps of preprocessing blood oxygen parameter data and relative blood flow index data of brain tissue to obtain the preprocessed blood oxygen parameter data and the preprocessed relative blood flow index data are as follows: Time synchronization, namely adding a time stamp into blood oxygen parameter data and relative blood flow index data of brain tissues, calculating time difference between the two types of data, and carrying out integral translation on the relative blood flow index data according to the time difference to realize time alignment with the blood oxygen parameter data; Filtering and denoising, namely removing low-frequency and high-frequency interference of blood oxygen parameter data through a band-pass filter, decomposing the blood oxygen parameter data by adopting a wavelet denoising method, performing soft threshold processing on the decomposed high-frequency wavelet coefficient after decomposition is completed, removing noise, performing smoothing processing on relative blood flow index data through Gaussian filtering, removing high-frequency fluctuation, and dynamically tracking the relative blood flow index data and removing motion artifacts by utilizing a self-adaptive filter.
3. The cerebral apoplexy monitoring method based on cerebral oxygen metabolism rate imaging according to claim 2, wherein the specific steps of obtaining the current cerebral oxygen metabolism rate of the brain tissue based on the preprocessed blood oxygen parameter data and the relative blood flow index data through formula calculation are as follows: Combining the preprocessed blood oxygen parameter data with the relative blood flow index data, and calculating according to a formula to obtain the current brain oxygen metabolism rate of the brain tissue, wherein the calculating formula specifically comprises: Wherein Representing the current brain oxygen metabolism rate of the brain tissue, Represents cerebral blood flow, belongs to the relative blood flow index data after pretreatment, And Respectively representing arterial blood oxygen saturation and venous blood oxygen saturation, which belong to the pretreated blood oxygen parameter data, Indicating the concentration of hemoglobin.
4. The cerebral apoplexy monitoring method based on cerebral oxygen metabolism rate imaging according to claim 3, wherein the specific steps of processing the calculated current cerebral oxygen metabolism rate data by using an image reconstruction algorithm and generating a cerebral oxygen metabolism distribution image of a brain region are as follows: The calculated current cerebral oxygen metabolism rate data are arranged, the spatial positions and the three-dimensional coordinates of all data sampling points are clarified, the current cerebral oxygen metabolism rate data are combined with a standard cerebral template, the prior structure knowledge of the standard cerebral template is utilized to assist in reconstructing a distribution image, and the data sampling points are mapped to the corresponding positions of the standard cerebral template through nearest neighbor interpolation or a distance-based weighting method during the process; and using an interpolation algorithm or a diffusion algorithm to diffuse the current cerebral oxygen metabolism rate data and fill the whole cerebral area, mapping the diffused current cerebral oxygen metabolism rate data onto a standard cerebral template, and generating a cerebral oxygen metabolism distribution image of the cerebral area, wherein the cerebral oxygen metabolism distribution image can show the current cerebral oxygen metabolism rate in a thermodynamic diagram form, and the color depth represents the intensity of the current cerebral oxygen metabolism rate.
5. The brain stroke monitoring method based on brain oxygen metabolism rate imaging according to claim 4, wherein said brain oxygen metabolism rate prediction model based on an improved LSTM model comprises an input layer, an LSTM layer, a dropoff layer, an attention mechanism layer, a fully connected layer and an output layer, wherein the input layer is configured to receive the preprocessed blood oxygen parameter data and the relative blood flow index data, wherein the LSTM layer is configured to receive the output of the input layer and capture the time dependency between the data, wherein the dropoff layer is configured to receive the output of the LSTM layer and randomly discard a portion of the neuron outputs, wherein the attention mechanism layer is configured to receive the output of the dropoff layer and importance weight the hidden state of each time step, wherein the fully connected layer is configured to receive the output of the attention mechanism layer and map the output of the attention mechanism layer to the output space, wherein the output layer is configured to receive the output of the fully connected layer and output the predicted brain oxygen metabolism rate of the brain tissue.
6. The cerebral apoplexy monitoring method based on cerebral oxygen metabolism rate imaging of claim 5, wherein the training of the cerebral oxygen metabolism rate prediction model, the specific steps of obtaining the trained cerebral oxygen metabolism rate prediction model are as follows: Acquiring historical blood oxygen parameter data and historical relative blood flow index data of brain tissues, preprocessing the historical blood oxygen parameter data and the historical relative blood flow index data, and dividing the preprocessed historical blood oxygen parameter data and the preprocessed historical relative blood flow index data into a training set, a verification set and a test set according to the proportion of 70%, 15% and 15%; Inputting sample data in the training set into a brain oxygen metabolism rate prediction model to obtain a brain oxygen metabolism rate prediction result of brain tissue, and calculating the loss between the brain oxygen metabolism rate prediction result of the brain tissue and the actual brain oxygen metabolism rate of the brain tissue by using a mean square error loss function; reversely calculating the gradient of each layer of parameters from the loss through a chain rule, and adjusting the parameters of a brain oxygen metabolic rate prediction model by using a gradient descent algorithm so as to minimize the loss; Dividing the training set into a plurality of small batches, repeating the two steps in a batch-by-batch mode, and traversing the whole training set for a plurality of times until the brain oxygen metabolism rate prediction model converges or reaches the preset iteration times; And (3) periodically evaluating various losses and indexes of the brain oxygen metabolism rate prediction model through sample data in the verification set, preventing overfitting, and evaluating the final performance of the brain oxygen metabolism rate prediction model through sample data in the test set after training is completed to obtain the trained brain oxygen metabolism rate prediction model.
7. The cerebral apoplexy monitoring method based on cerebral oxygen metabolism rate imaging according to claim 6, wherein the comprehensive evaluation of the brain physiological state of the cerebral apoplexy patient according to the current cerebral oxygen metabolism rate and the predicted cerebral oxygen metabolism rate of the brain tissue, the brain physiological state score of the cerebral apoplexy patient is obtained, and the specific steps of planning the corresponding treatment scheme for the cerebral apoplexy patient based on the brain physiological state score are as follows: based on the current brain oxygen metabolism rate of brain tissues and the predicted brain oxygen metabolism rate, the brain physiological state score of the cerebral apoplexy patient is calculated by a formula, wherein the calculation formula is specifically as follows: Wherein A brain physiological state score representing a stroke patient, Representing a predicted brain oxygen metabolism rate of the brain tissue, Represents the standard brain oxygen metabolism rate of brain tissue, And Represents a weight coefficient, an ; And grading the brain physiological states of the cerebral apoplexy patients according to the brain physiological states of the cerebral apoplexy patients, dividing the cerebral apoplexy patients into corresponding brain physiological state risk levels based on preset grading rules, and planning corresponding-level treatment schemes for the cerebral apoplexy patients according to the brain physiological state risk levels.
8. A cerebral apoplexy monitoring system based on cerebral oxygen metabolism rate imaging, which is applied to the cerebral apoplexy monitoring method based on cerebral oxygen metabolism rate imaging as claimed in any one of claims 1 to 7, and is characterized by comprising the following steps: The data acquisition and calculation module is used for transmitting multi-wavelength near infrared light through the frequency domain near infrared spectrometer and collecting frequency domain signals to obtain the absorption coefficient and the scattering coefficient of the brain tissue, calculating to obtain blood oxygen parameter data of the brain tissue, transmitting coherent laser through the diffusion correlation spectrometer and detecting speckle light intensity signals to obtain speckle light intensity autocorrelation functions, and calculating to obtain relative blood flow index data of the brain tissue; the data preprocessing module is used for preprocessing blood oxygen parameter data and relative blood flow index data of brain tissues to obtain preprocessed blood oxygen parameter data and relative blood flow index data; the data fusion and imaging module is used for calculating the current cerebral oxygen metabolism rate of the brain tissue through a formula based on the preprocessed blood oxygen parameter data and the relative blood flow index data, processing the calculated current cerebral oxygen metabolism rate data through an image reconstruction algorithm, and generating a cerebral oxygen metabolism distribution image of a brain region; The brain state prediction module is used for constructing a brain oxygen metabolism rate prediction model based on the improved LSTM model, inputting the preprocessed blood oxygen parameter data and the relative blood flow index data into the trained brain oxygen metabolism rate prediction model to obtain the predicted brain oxygen metabolism rate of brain tissues; And the comprehensive evaluation and planning module is used for comprehensively evaluating the brain physiological state of the cerebral apoplexy patients according to the current cerebral oxygen metabolic rate and the predicted cerebral oxygen metabolic rate of the brain tissues to obtain the brain physiological state scores of the cerebral apoplexy patients, and planning corresponding treatment schemes for the cerebral apoplexy patients based on the brain physiological state scores.

Description

Cerebral apoplexy monitoring method and system based on cerebral oxygen metabolism rate imaging Technical Field The invention relates to the technical field of medical care, in particular to a cerebral apoplexy monitoring method and system based on cerebral oxygen metabolic rate imaging. Background Cerebral apoplexy is a first disease which is fatal and disabled for adults in China, has the characteristics of high morbidity, high disability rate, high mortality rate and high recurrence rate, and seriously threatens public health. The incidence and recurrence rate of cerebral stroke have been statistically rising year by year, bringing a heavy burden to medical systems and society. Current clinical monitoring of stroke patients relies mainly on observation of vital signs (e.g. consciousness, pupillary response, muscle strength, etc.), and routine monitoring of blood oxygen saturation (SO 2). However, the monitoring of SO 2 can only reflect the oxygenation level of the whole body, and the local metabolic activity of brain tissues cannot be directly evaluated, SO that hysteresis exists in the evaluation of therapeutic effects such as thrombolysis and the like, and the clinical decision lacks real-time and quantitative basis. International guidelines for stroke in 2019 suggest that SO 2 monitoring should be enhanced. However, the existing devices such as pulse oximetry and the like can only provide systemic data, lack specific monitoring capability on brain tissues, and cannot meet the requirement of personalized diagnosis and treatment on cerebral apoplexy patients. Near infrared spectroscopy (NIRS) has been applied to brain function monitoring. The principle is that the bleeding oxygen parameters (HbO 2, hbR and SO 2) are calculated by utilizing multi-wavelength near infrared light to penetrate tissues and measuring absorption difference. However, the traditional NIRS technology can only provide relative blood oxygen change, is difficult to quantitatively obtain the absolute oxygenation level of tissues, has limited reflection on blood flow dynamics, and cannot fully characterize brain metabolic processes. Diffusion Correlation Spectroscopy (DCS) is a non-invasive monitoring method based on speckle correlation functions that can reflect changes in brain tissue microcirculation blood flow. The method is sensitive to the flow velocity in small blood vessels, and is suitable for continuous monitoring of cerebral blood flow. However, DCS technology lacks the ability to detect blood oxygen concentration and cannot independently calculate brain oxygen metabolism rate (CMRO 2). Therefore, the existing single NIRS or DCS technology has the limitation that the NIRS lacks quantitative capability on blood flow dynamics, the DCS lacks sensitivity on blood oxygen parameters, and complete assessment of brain tissue metabolic level cannot be independently realized. Based on the above situation, the invention provides a cerebral apoplexy monitoring method and a cerebral oxygen metabolism rate imaging-based cerebral apoplexy monitoring system with accurate evaluation, which are used for realizing the fusion of FD-fNIRS and DCS and for imaging and dynamic monitoring of cerebral oxygen metabolism rate (CMRO 2). Disclosure of Invention In order to overcome the defects of the prior art, the invention provides a cerebral apoplexy monitoring method and a cerebral apoplexy monitoring system based on cerebral oxygen metabolism rate imaging, which can realize synchronous monitoring of cerebral oxygenation level and microcirculation blood flow of a cerebral apoplexy patient, and acquire cerebral oxygen metabolism rate parameters through data fusion, thereby providing a real-time, quantitative and visual monitoring means for clinic. A cerebral apoplexy monitoring method based on cerebral oxygen metabolism rate imaging comprises the following steps: Transmitting multi-wavelength near infrared light through a frequency domain near infrared spectrometer, collecting frequency domain signals, obtaining an absorption coefficient and a scattering coefficient of brain tissues, obtaining blood oxygen parameter data of the brain tissues through calculation, transmitting coherent laser through a diffusion correlation spectrometer, detecting speckle light intensity signals, obtaining speckle light intensity autocorrelation functions, and obtaining relative blood flow index data of the brain tissues through calculation; Preprocessing blood oxygen parameter data and relative blood flow index data of brain tissues to obtain preprocessed blood oxygen parameter data and relative blood flow index data; Based on the preprocessed blood oxygen parameter data and the relative blood flow index data, calculating to obtain the current brain oxygen metabolism rate of brain tissues through a formula, processing the calculated current brain oxygen metabolism rate data by using an image reconstruction algorithm, and generating a brain oxygen metabolism distribution