Sepsis encephalopathy spectrum diagnosis and treatment platform

Abstract

The present application belongs to the technical field of sepsis encephalopathy, and particularly relates to a sepsis encephalopathy energy spectrum diagnosis and treatment platform, which comprises a multi-dimensional feature set, a dynamic evaluation model and an intelligent multi-modal data analysis module, the multi-dimensional feature set comprises power distribution of electroencephalogram signals and cross-band energy spectrum, the power distribution of electroencephalogram signals is used for reflecting local brain area energy state, and the cross-band energy spectrum is used for revealing global network synergy, and the dynamic evaluation model comprises time-varying characteristics of real-time analysis energy spectrum and inflammation factor level of sepsis-related encephalopathy patients; through the above structure, the power distribution of electroencephalogram signals and the cross-band energy spectrum are combined to construct the multi-dimensional feature set, the sensitivity and specificity for pathological changes of sepsis-related encephalopathy are enhanced, and then the time-varying characteristics of real-time analysis energy spectrum are combined to obtain the inflammation factor level of sepsis-related encephalopathy patients, so that corresponding diagnosis and treatment can be made for the patients in time.

Description

Technical Field

[0001] This invention belongs to the field of septic encephalopathy technology, specifically a spectral diagnostic and therapeutic platform for septic encephalopathy. Background Technology

[0002] Sepsis-associated encephalopathy (SAE) is a brain dysfunction caused by systemic infection without direct central nervous system infection. It typically manifests as impaired consciousness and cognitive function, ranging from mild delirium to deep coma. SAE is a common and serious complication of sepsis, and its pathological mechanism is complex, involving blood-brain barrier disruption, neuroinflammatory responses, and mitochondrial dysfunction.

[0003] However, existing technologies often have the following drawbacks: Currently, there is a lack of specific biomarkers in clinical practice, and diagnosis mainly relies on clinical manifestations and neuroimaging, which has problems such as low sensitivity and high invasiveness. Traditional electroencephalography (EEG) analysis focuses on the energy changes of a single frequency band (such as alpha waves and beta waves), which is difficult to fully reflect the complex brain network damage of SAE. Although classification methods based on functional connectivity can identify abnormalities in specific brain regions, they do not combine the dynamic characteristics of energy distribution and cannot quantify the association between brain metabolic state and inflammatory damage.

[0004] Therefore, this invention provides a spectrum diagnostic and therapeutic platform for sepsis encephalopathy. Summary of the Invention

[0005] In order to overcome the shortcomings of the prior art, at least one technical problem raised in the background art is solved.

[0006] The technical solution adopted by the present invention to solve its technical problem is: the septic encephalopathy spectrum diagnosis and treatment platform of the present invention includes a multidimensional feature set, a dynamic evaluation model and an intelligent multimodal data analysis module; The multidimensional feature set includes the power distribution and cross-band energy spectrum of the EEG signal. The power distribution of the EEG signal is used to reflect the energy state of local brain regions, and the cross-band energy spectrum is used to reveal the global network synergy. The dynamic assessment model includes real-time analysis of the time-varying characteristics of the energy spectrum and the levels of inflammatory factors in patients with sepsis-associated encephalopathy. The real-time analysis of the time-varying characteristics of the energy spectrum is used to analyze low-frequency energy decay and high-frequency energy fluctuations. The intelligent multimodal data analysis module includes a data integration layer, a feature model training layer, and a real-time diagnostic unit. The data integration layer is used to collect data from different sources. The feature model training layer is used to automatically extract key features through deep learning algorithms and to build a prediction model through unsupervised learning algorithms to assess disease risk, disease progression, and treatment response.

[0007] As a preferred embodiment of the present invention, the inflammatory factor levels of patients with sepsis-associated encephalopathy can be analyzed from interleukin-6 and C-reactive protein.

[0008] As a preferred technical solution of the present invention, the key feature data includes, but is not limited to, clinical symptoms, laboratory test results, genomic information, imaging scan results, and patient lifestyle data.

[0009] As a preferred technical solution of the present invention, the establishment of the prediction model includes target variables, data collection and preparation, model selection and evaluation, and model analysis and deployment.

[0010] As a preferred technical solution of the present invention, the data integration layer includes basic personal information and medical history, physiological parameter monitoring, laboratory test results, imaging examination results, and diagnosis and treatment process records. Physiological parameter monitoring is used to detect blood pressure and heart rate. Laboratory test results are represented by white blood cell count and C-reactive protein level. Imaging examination results are represented by CT and MRI images. Diagnosis and treatment process records are represented by drug use.

[0011] As a preferred technical solution of the present invention, the real-time diagnostic unit generates the best treatment strategy for an individual based on the patient's specific biomarkers, genetic information and clinical manifestations, combined with treatment response measures.

[0012] As a preferred technical solution of the present invention, the target variable of the prediction model is selected from at least one of the following medical evaluation parameters; a) Binary variables, representing Boolean logic judgments on the occurrence status of sepsis-related encephalopathy; b) Multilevel categorical variables, representing the grading assessment results of the severity of sepsis-related encephalopathy.

[0013] As a preferred technical solution of the present invention, the data collection and preparation collects data from multiple sources, including electronic health records, laboratory information systems, image archiving and communication systems, real-time physiological monitoring devices, mobile applications, and wearable devices.

[0014] The beneficial effects of this invention are as follows: 1. The sepsis encephalopathy energy spectrum diagnosis and treatment platform of the present invention constructs a multi-dimensional feature set by combining the power distribution of electroencephalogram signals with cross-band energy spectrum, thereby enhancing the sensitivity and specificity of pathological changes in sepsis-related encephalopathy. In addition, by combining real-time analysis of the time-varying characteristics of energy spectrum, the level of inflammatory factors in patients with sepsis-related encephalopathy can be obtained, and corresponding diagnosis and treatment can be given to patients in a timely manner. 2. The septic encephalopathy spectrum diagnosis and treatment platform of the present invention can combine data from multiple different sources for analysis through an intelligent multimodal data analysis module, and use machine learning algorithms to perform in-depth analysis of these data to quickly and intelligently assess the patient's disease risk, disease progression, and treatment response. Based on the combination of the patient's multi-source data, it analyzes whether the patient has sepsis-related encephalopathy, and then assesses the severity of the patient's sepsis-related encephalopathy to determine the grading results, so as to make the best diagnosis and treatment strategy for the patient and enhance the accuracy of the diagnosis and treatment plan. Attached Figure Description

[0015] The invention will now be further described with reference to the accompanying drawings.

[0016] Figure 1 This is a schematic diagram of the steps in this invention; Figure 2 This is a flowchart illustrating the image acquisition module in this invention; Figure 3 The diagrams show the waveforms, power distribution, energy spectrum, and power spectral density (PSD) of the rat at different frequency bands, in sequence. Figure 4 This is a comparison chart of the power distribution results and power spectral density (PSD) results 48 hours after modeling in this invention. Detailed Implementation

[0017] To make the technical means, creative features, objectives and effects of this invention easier to understand, the invention will be further described below in conjunction with specific embodiments.

[0018] Reference Figure 1 - Figure 4 This invention provides two technical solutions: Example 1: The spectral diagnostic and treatment platform for sepsis encephalopathy includes a multidimensional feature set, a dynamic assessment model, and an intelligent multimodal data analysis module; The multidimensional feature set includes the power distribution and cross-band energy spectrum of EEG signals. The power distribution of EEG signals is used to reflect the energy state of local brain regions, while the cross-band energy spectrum is used to reveal the global network synergy. The dynamic assessment model includes real-time analysis of the time-varying characteristics of the energy spectrum and the levels of inflammatory factors in patients with sepsis-associated encephalopathy. The real-time analysis of the time-varying characteristics of the energy spectrum is used to analyze low-frequency energy decay and high-frequency energy fluctuations. The intelligent multimodal data analysis module includes a data integration layer, a feature model training layer, and a real-time diagnostic unit. The data integration layer is used to collect data from different sources. The feature model training layer is used to automatically extract key features through deep learning algorithms and to build a predictive model through unsupervised learning algorithms to assess disease risk, disease progression, and treatment response.

[0019] The intelligent multimodal data analysis module can combine data from various sources for analysis and use machine learning algorithms to perform in-depth analysis of this data. It can quickly and intelligently assess the patient's disease risk, disease progression, and treatment response. Based on the combination of the patient's multi-source data, it can analyze whether the patient has sepsis-associated encephalopathy and then assess the severity of the patient's sepsis-associated encephalopathy to determine the best treatment strategy and improve the accuracy of the treatment plan.

[0020] Example 2: In patients with sepsis-associated encephalopathy, inflammatory factor levels can be analyzed using interleukin-6 (IL-6) and C-reactive protein (CRP). IIL-6 is a crucial mediator of systemic inflammatory responses, inducing the release of other pro-inflammatory factors and exacerbating the inflammatory cascade. Due to its central role in the inflammatory response, IIL-6 levels are often used as a biomarker to assess the severity of sepsis. High levels of IIL-6 typically predict a more severe inflammatory state and poor prognosis. Therefore, monitoring IIL-6 levels in patients with sepsis-associated encephalopathy helps to: identify the risk of disease progression earlier; guide treatment decisions, such as whether stronger anti-inflammatory therapy or other supportive measures are needed; assess treatment efficacy; and track patient recovery.

[0021] Key characteristic data include, but are not limited to, clinical symptoms, laboratory test results, genomic information, imaging scan results, and patient lifestyle data. Clinical symptoms of sepsis-associated encephalopathy (SAE) can be viewed from the following states: patients exhibit varying degrees of confusion; cognitive impairment; symptoms of anxiety and agitation, and may experience myoclonus or seizures; to improve the specific analysis of symptom manifestations and assessment tests, a control experiment was conducted on rats.

[0022] Establishing a predictive model includes target variables, data collection and preparation, model selection and evaluation, and model analysis and deployment. Model selection and evaluation can be combined with the analysis of IL-6 levels and changes in body temperature to obtain the result of predicting whether a patient has the disease.

[0023] The data integration layer includes basic personal information and medical history, physiological parameter monitoring, laboratory test results, imaging examination results, and treatment process records. Physiological parameter monitoring is used to detect blood pressure and heart rate. Laboratory test results are represented by white blood cell count and C-reactive protein level. Imaging examination results are represented by CT and MRI images. Treatment process records are represented by medication use.

[0024] The real-time diagnostic unit generates the best treatment strategy for each individual based on the patient's specific biomarkers, genetic information, and clinical manifestations, combined with diagnostic and treatment approaches. The real-time diagnostic unit facilitates the formulation of the best treatment plan for the patient, thereby improving the accuracy of disease diagnosis and treatment.

[0025] The target variable of the prediction model is selected from at least one of the following medical evaluation parameters; a) Binary variables, representing Boolean logic judgments on the occurrence status of sepsis-related encephalopathy; b) Multilevel categorical variables, representing the grading assessment results of the severity of sepsis-related encephalopathy.

[0026] Binary variables can be used to determine whether a patient has sepsis-related encephalopathy based on data analysis, while multi-level variables can be divided into four levels: asymptomatic, mild, moderate, and severe.

[0027] Data collection and preparation involves gathering data from multiple sources, including electronic health records, laboratory information systems, image archives and communication systems, real-time physiological monitoring devices, mobile applications, and wearable devices. This data collection ensures a sufficiently long time span to cover changes in the condition at different stages, facilitating the determination of whether a patient is ill and at what level.

[0028] To verify the abnormal changes observed in the sepsis-associated encephalopathy group, this embodiment uses rats for the experiment: 1) Validation of the animal model of sepsis-associated encephalopathy (SAE): Endotoxin (LPS) solution was injected intraperitoneally at a dose of 2 mg / kg. 24 hours after LPS injection, cecal ligation and perforation were performed. The animals were anesthetized and fixed on the operating table in a supine position (tail facing the operator). A vertical incision of about 1 cm was made along the linea alba 1-1.5 cm below the xiphoid process. The mesentery and cecum were freed under aseptic conditions. After separating the cecum, it was ligated with 5-0 sutures in the middle of the cecum. The cecum was punctured with a 23G needle at the ligated site, and a small amount of feces was squeezed out. The cecum was then put back in place, and the skin and muscles were sutured with 3-0 sutures in sequence. After suturing, the animals were injected subcutaneously with 1 ml of 0.9% sodium chloride injection for fluid replacement. The sham surgery group was the same as the sepsis group except that it was not injected with sepsis (LPS), and the cecum was not ligated or perforated. The animals were injected with 0.5 ml of 0.9% sodium chloride injection in the morning and afternoon every day after surgery, and were fed some jelly strips. 2) Method: ① Electroencephalography (EEG) electrode implantation and local field potential (LFP) recording On days 2 and 3 after modeling, all rats underwent electroencephalography (EEG) testing. The rats were anesthetized and rendered into a coma. After immobilization, the top of the rat's head was prepared, and the eyes were covered with saline-moistened gauze, which was then cut open. The anterior fontanelle was marked with blue ink and leveled using a positioning device. Electrodes were marked 2.4 mm to the right and 4.3 mm to the back of the anterior fontanelle with blue ink, aligning with the midline of the skull. After locating the remaining three positioning points, holes were drilled using a cranial drill, and the silver electrode wires were fixed with cranial nails. The electrodes were slowly lowered to be level with the skull, and after a slow descent of 2.6 mm, the local electroencephalography (EEG) and electromyography (EMG) recording system of the rat and mouse was used to record the local EEG data. Loal field potentials (LFPs) were analyzed using professional neurophysiological data analysis software (NeuroExplorer). The signals were filtered according to the Delta (0.5–4 Hz), Theta (4–8 Hz), Alpha (8–13 Hz), Beta (13–30 Hz), and Gamma (30–80 Hz) bands. Power distribution and power spectral density (PSD) were then analyzed and statistically analyzed, and the energy spectra of each group of rats were compared. ② Data Analysis Experimental data were analyzed using GraphPad Prism 8.0 software and expressed as mean ± standard error (SEM). All data were analyzed using a two-way ANOVA followed by a Tukey test. A p-value < 0.05 was considered statistically significant. 3) Results Abnormal changes in local field potential (LFP) in the CA1 region of the hippocampus of rats with sepsis-associated encephalopathy As attached Figure 3 As shown in the figure, EGG2 represents the electroencephalogram (EEG). Figure 2The terms "Deita" and "Theta" refer to brain waves during deep sleep or unconsciousness, "Theta" to light sleep, meditation, or memory processing, "Alpha" to relaxed state (e.g., with eyes closed), "Beta" to high-frequency brain waves during focused or tense state, and "Gamma" to ultra-high-frequency brain waves related to higher cognitive activities (e.g., perception and learning). Electroencephalography (EEG) is one of the most sensitive techniques for identifying sepsis-associated encephalopathy (SAE). In SAE patients, the incidence of EEG abnormalities is 77.46%, primarily manifested as an increase in delta waves. Researchers believe that an increased percentage of delta waves in EEG can serve as an important indicator of early sepsis-associated encephalopathy. This study used changes in the percentage of delta waves in EEG as an important indicator for diagnosing sepsis-associated encephalopathy, thus dividing septic rats into encephalopathy and non-encephalopathy groups. We measured the changes at 48 and 72 hours after modeling. Changes in local field potentials (LFPs) in the CA1 region of the hippocampus were detected in the sham-operated group and the model group, respectively. Power distribution results showed that after 72 hours, 6 septic rats showed an increased proportion of delta band frequency (p > 0.05 compared with the sham-operated group). These septic rats were classified into the sepsis-associated encephalopathy group, while septic rats with no significant difference in delta band frequency were classified into the sepsis-nonencephalopathy group (Figure B). At the same time, we compared the energy spectra of rats in each group. The results showed that the energy of the 100s local field potential (LFP) in the CA1 region of the hippocampus of the SAE group was higher than that of the sham-operated group, and the color was more vivid (Figure C). However, the power spectral density (PSD) results showed no significant changes in any frequency band in the sham-operated group, the SAE group, and the sepsis-nonencephalopathy group (Figure D).

[0029] The terms "front," "back," "left," "right," "top," and "bottom" all refer to the figures in the accompanying drawings. Figure 1 Based on the perspective of the observer, the side of the device facing the observer is defined as the front, the left side of the observer is defined as the left, and so on.

[0030] In the description of this invention, it should be understood that the terms "center", "longitudinal", "lateral", "front", "rear", "left", "right", "vertical", "horizontal", "top", "bottom", "inner", "outer", etc., indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings. They are only for the convenience of describing this invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, they should not be construed as limiting the scope of protection of this invention.

[0031] The foregoing has shown and described the basic principles, main features, and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The embodiments and descriptions in the specification are merely illustrative of the principles of the invention. Various changes and modifications can be made to the invention without departing from its spirit and scope, and all such changes and modifications fall within the scope of the present invention as claimed. The scope of protection of the present invention is defined by the appended claims and their equivalents.

Claims

1. A spectrum-based diagnostic and therapeutic platform for septic encephalopathy, characterized in that: It includes a multidimensional feature set, a dynamic evaluation model, and an intelligent multimodal data analysis module; The multidimensional feature set includes the power distribution and cross-band energy spectrum of the EEG signal. The power distribution of the EEG signal is used to reflect the energy state of local brain regions, and the cross-band energy spectrum is used to reveal the global network synergy. The dynamic assessment model includes real-time analysis of the time-varying characteristics of the energy spectrum and the levels of inflammatory factors in patients with sepsis-associated encephalopathy. The real-time analysis of the time-varying characteristics of the energy spectrum is used to analyze low-frequency energy decay and high-frequency energy fluctuations. The intelligent multimodal data analysis module includes a data integration layer, a feature model training layer, and a real-time diagnostic unit. The data integration layer is used to collect data from different sources. The feature model training layer is used to automatically extract key features through deep learning algorithms and to build a prediction model through unsupervised learning algorithms to assess disease risk, disease progression, and treatment response.

2. The septic encephalopathy spectrum diagnostic and therapeutic platform according to claim 1, characterized in that: The levels of inflammatory factors in patients with sepsis-associated encephalopathy can be analyzed from interleukin-6 and C-reactive protein.

3. The septic encephalopathy spectrum diagnostic and therapeutic platform according to claim 1, characterized in that: The key characteristic data includes clinical symptoms, laboratory test results, genomic information, imaging scan results, and patient lifestyle data.

4. The septic encephalopathy spectrum diagnostic and therapeutic platform according to claim 1, characterized in that: The establishment of the predictive model includes target variables, data collection and preparation, model selection and evaluation, and model analysis and deployment.

5. The septic encephalopathy spectrum diagnostic and therapeutic platform according to claim 1, characterized in that: The data integration layer includes basic personal information and medical history, physiological parameter monitoring, laboratory test results, imaging examination results, and diagnosis and treatment process records. Physiological parameter monitoring is used to detect blood pressure and heart rate. Laboratory test results are represented by white blood cell count and C-reactive protein level. Imaging examination results are represented by CT and MRI images. Diagnosis and treatment process records are represented by medication use.

6. The septic encephalopathy spectrum diagnostic and therapeutic platform according to claim 1, characterized in that: The real-time diagnostic unit generates the best treatment strategy for each individual based on the patient's specific biomarkers, genetic information, and clinical manifestations, combined with treatment response measures.

7. The septic encephalopathy spectrum diagnostic and therapeutic platform according to claim 4, characterized in that: The target variable of the prediction model is selected from at least one of the following medical evaluation parameters; a) Binary variables, representing Boolean logic judgments on the occurrence status of sepsis-related encephalopathy; b) Multilevel categorical variables, representing the grading assessment results of the severity of sepsis-related encephalopathy.

8. The septic encephalopathy spectrum diagnostic and therapeutic platform according to claim 4, characterized in that: The data collection and preparation process gathers data from multiple sources, including electronic health records, laboratory information systems, image archiving and communication systems, real-time physiological monitoring devices, mobile applications, and wearable devices.

Images