21 results about "Clinical correlation" patented technology

The invention discloses a security and robustness automatic testing method for a large language model in the medical field, and the method comprises the steps: constructing a field perception harmful problem generator which can automatically learn and map medical knowledge to potential harmful queries; according to the method and the system, the real security boundary of the large language model in the medical field can be efficiently and systematically evaluated, and particularly, the evaluation depth and the evaluation breadth are provided in complicated situations involving professional knowledge misleading or potential harm. A generator is optimized through a customized weighted composite reward function, so that the model is effectively guided to generate a field perception harmful problem which has high harmfulness, is coherent and smooth and conforms to grammar, and the quality and pertinence of a data set are remarkably improved; according to the framework, the dependence on a large amount of time-consuming manual intervention is fundamentally eliminated by integrating automatic scoring and a multi-index similarity filtering mechanism of a hazard detection model.

The application discloses a kind of dabigatran bleeding risk prediction methods based on machine learning, it is related to computer-aided drug risk management technical field, the method is by obtaining patient baseline and follow-up data, multiple imputation method is handled missing value;With HAS-BLED score as the basis, candidate variables are screened from potential risk factors using LassoCV algorithm, and anemia, insulin use and antifungal agent use are determined as new prediction variables through clinical correlation analysis to construct a set of prediction variables;Finally, a random forest or gradient boosting algorithm is used to train the model to output the bleeding risk assessment results of the individual to be tested.The application combines machine learning algorithm, improves the prediction accuracy of dabigatran bleeding risk in Chinese non-valvular atrial fibrillation patients, solves the problem of insufficient prediction accuracy of traditional HAS-BLED score, and provides a more reliable basis for clinical individualized anticoagulant therapy decision.

The invention relates to the technical field of biological detection and toxicological evaluation, and discloses a neural behavior multi-parameter integration-based caenorhabditis elegans exposure evaluation method which comprises the following specific steps: S1, synchronous culture of caenorhabditis elegans; s2, pollutant exposure treatment; s3, multi-parameter behavior detection; s4, data standardization processing; and S5, constructing a comprehensive evaluation model, and performing weighted integration on four core parameters including the movement speed, the chemotactic index, the synaptic spot density and the calcium signal amplitude after standardization to construct a neurotoxicity index NTI. According to the method disclosed by the invention, a multi-layer and multi-dimensional neurotoxicity evaluation system is constructed by integrating four types of neurobehavior parameters including motion behaviors, chemotactic reactions, synaptic morphology and neuron calcium signals, and the defect that a traditional single endpoint index cannot comprehensively reflect overall disturbance of a neural regulation network is overcome; by utilizing the high homology of the nervous system of caenorhabditis elegans and the human nerve gene, the clinical correlation and mechanism analysis capability of environmental pollutant neurotoxicity evaluation are remarkably improved.

A system and method for improving predictive modeling of patient outcomes by transforming ctDNA-linked genomic data into engineered pathway and network features and training machine-learning models that explicitly incorporate ctDNA status and amount while controlling selection bias. By broadening the cohort to include both ctDNA-positive and ctDNA-negative patients, harmonizing ctDNA across assays, and iteratively reducing features using Shapley-based importance, the approach improves predictive accuracy, calibration, and clinical relevance.

The invention relates to a mouse spinal cord injury bionic compression modeling system and a method thereof.The system comprises a mechanical arm, a magneto-rheological-pneumatic loading compression module, a multi-mode optical fiber sensing probe and a control module, the magneto-rheological-pneumatic loading compression module comprises a shell, an air bag, a magneto-rheological elastomer and an air supply part, the shell is installed at the driving end of the mechanical arm, and the air bag is installed at the driving end of the mechanical arm; the air bag is installed in an inner cavity of the shell, the electromagnetic part is fixedly installed at the bottom of the inner cavity of the shell, the air supply part supplies air to the air bag, and the magnetorheological elastomer is fixedly installed at the end, away from the bottom of the shell, of the air bag. The multi-mode optical fiber sensing probe is installed at the end, away from the air bag, of the magnetorheological elastomer and used for detecting the microenvironment of the deep spinal cord tissue of the mouse. The injured spinal cord chronic compression animal model constructed by the scheme is high in fidelity, the clinical correlation of the model can be remarkably improved, and the experimental variability is reduced, so that a more reliable and more comprehensive research platform is provided for mechanism research and treatment strategy evaluation of degenerative spinal cord diseases.

The present application relates to the technical field of microorganism and the technical field of medicine, and particularly relates to an oral microorganism marker and application thereof in diagnosis of prognosis of gastric cancer. The oral microorganism marker comprises at least one of Aggregatibacter, Filifactor, Moryella and Leptotrichia. The present application firstly discovers that the level of human oral flora changes in different prognosis outcomes of gastric cancer, and by detecting the change of the flora, the prognosis survival period of the patient with gastric cancer can be predicted. The oral microorganism marker has the advantages of high specificity and strong sensitivity in diagnosis. The present application proves the clinical correlation of the oral microorganism flora in the prognosis of gastric cancer, and provides a non-invasive and highly sensitive tool for identifying the patient at risk of poor outcome before treatment.

The application discloses an image diagnosis report generation method and device and a storage medium. The method comprises the following steps: searching a target template matched with a diagnosis text input by a user based on the diagnosis text, through three different levels of matching degrees, i.e. a spatial matching degree between the diagnosis text and the candidate template at an anatomical site level, a semantic matching degree between the diagnosis text and the candidate template related to pathological information, and a clinical correlation degree between the diagnosis text and the candidate template, and further optimizing the search through a time decay factor, a clinical priority and a conflict degree between regions of interest, improving the accuracy of searching the required image diagnosis template for the user, and thus solving the technical problem that the keyword matching effect of the traditional keyword matching algorithm in the prior art is poor in a medical scene, and the accuracy of searching a correct template is reduced.

Improved methods for determining full-genome associations with phenotype data represented by medical images, ECG traces, spirometry-traces, or other high-dimensional phenotype-representing physiosignals are provided. These methods include training an encoder, as pan of an autoencoder, to project input physiosignals into a phenotypically representative set of lower-dimensional latent variables. In some examples, the latent variables are augmented by clinical correlates of the input physiosignals (e.g., a. force vital capacity-determined from a spirometry trace), The latent variables and / or clinical correlates are then used to determine genetic loci that are associated with each of the latent variables. These associations can then be used to focus drug development and / or to predict polygenic scores tor ram diseases for which sufficient, data, may-not be available for a full genome-wide association study or other genomic data-to-phenotype association.

The invention belongs to the technical field of biological medicine, and particularly relates to a cell medicine and a medicine effect evaluation method and application of the cell medicine in bladder cancer medicine. A sensitization PBMC suspension is added into an organoid suspension for co-culture, cells are collected after co-culture is completed to obtain a cell drug, and a bladder cancer efficacy evaluation model is constructed based on the cell drug: the cell drug and a bladder cancer organoid are co-cultured, so that the defect that a traditional model cannot simulate tumor-immune interaction is overcome, and the bladder cancer efficacy evaluation model is constructed based on the cell drug. The drug effect evaluation result has high clinical correlation, the cell drug targeting killing ability is strong, and the method is suitable for formulation of individualized treatment schemes of bladder cancer and high-throughput screening of anti-bladder cancer drugs, and has significant clinical transformation value.

This invention provides a method and system for retrieving patient digital models based on event graphs. The method includes: designing a timeline-based event extraction and structuring method for unstructured or semi-structured data in electronic medical records, constructing an event graph of the electronic medical records; extracting key information from the medical record text in the event graph using an intelligent model; designing a patient digital model retrieval algorithm; establishing a key information retrieval and matching mechanism; and quickly searching for digital models similar to the current patient's condition from a historical case database, providing data support for subsequent diagnosis and treatment analysis. This invention can extract key information from massive heterogeneous clinical data, construct patient digital models with high clinical relevance, interpretability, and dynamic adaptability. Based on event graph technology and large-scale model applications, it improves the modeling accuracy and clinical applicability of patient data in the diagnosis and treatment of complex diseases, significantly improving the accuracy, completeness, and clinical relevance of patient digital models.

The application comprises a RAS gene cluster-based pediatric AML prognosis model construction method, which relates to the cross field of bioinformatics and medical technology. It comprises: S1: data acquisition and integration, standardization and integration processing of expression data; S2: sample typing, obtaining at least two subtypes with different prognoses; S3: prognosis model construction; S4: clinical correlation verification, constructing a visual prognosis evaluation system integrating risk stratification and clinical characteristics; S5: immune infiltration correlation verification, establishing the correlation between risk stratification and immune infiltration characteristics; S6: drug sensitivity correlation verification, determining the correlation between risk stratification and drug sensitivity; S7: key gene verification. The application realizes the precise risk stratification of pediatric AML patients by integrating RAS signal pathway gene characteristics and clinical data, and the prognosis model is verified by training set and verification set, which provides a standardized scientific research model framework and reliable data support, and provides an efficient auxiliary tool for pediatric AML related scientific research.

The application belongs to the technical field of biological medicine, and specifically discloses an ischemia modified albumin detection kit and method. The kit comprises reagent 1 and reagent 2; the reagent 1 comprises a water-soluble salt containing divalent cobalt ions and a surfactant, and has a pH of 7.1-8.5; the reagent 2 comprises dithiothreitol and a reducing protective agent, and has a pH of 4.0-7.5; the reducing protective agent is selected from at least one of tartaric acid, 2-hydroxy succinic acid, tris(2-carboxyethyl) phosphine, thiourea, mannitol and sodium sulfite. The kit is used to detect ischemia modified albumin by an albumin-cobalt combination test method, has the advantages of strong stability, high precision, good clinical correlation with mainstream manufacturers, wide linear range and the like, and can well meet the needs of clinical examination.

This invention discloses a method for constructing a risk prediction model for blinding diabetic retinopathy (STDR), belonging to the field of diabetic lesion detection. It addresses the problem that STDR screening at the grassroots level relies on specialized resources and that existing models have poor adaptability. The method includes: screening eligible type 2 diabetic patients and organizing clinical data; detecting 15 core laboratory indicators and calculating derived indicators as candidate variables; dividing patients into non-STDR and STDR groups according to DR classification and DME diagnosis results; selecting variables using a dual-dimensional strategy of "statistical significance + clinical relevance" through binary logistic regression, multicollinearity test, and incorporating six indicators including age, duration of diabetes, and MLR to construct a mathematical prediction model; and then building a nomogram visualization prediction model based on the model formula. All indicators included in this model are routinely available in clinical practice. The model has been validated with an AUC of 0.825, sensitivity of 87.2%, specificity of 67.8%, good fit, and good clinical applicability. It can quickly assess the risk of STDR and is suitable for grassroots medical scenarios.

The application discloses a kind of quality control classification methods of intelligent DR image, belong to medical image processing and artificial intelligence technical field.The application solves the problems that existing intelligent DR quality control technology is based on image appearance correlation judgment, cannot trace back the root cause of quality problem, poor cross-device generalization, lack of clinical relevance and insufficient explainability.The application is based on X-ray imaging physical process to construct causal model including device parameter node, imaging process node, image quality node and clinical influence node, determines the type, severity and root cause of image quality problem through causal inference algorithm, and generates quality control classification results and targeted improvement suggestions in combination with clinical correlation evaluation.The application can be used stably without retraining for different brands and models of DR equipment, effectively reduces unnecessary re-shooting, reduces patient radiation dose, and significantly improves the clinical trust and acceptance of AI quality control results.

A system and method are disclosed for generating a History of Present Illness (HPI) summary by extracting and synthesizing clinical data from unstructured or semi-structured medical documents. The invention utilizes natural language processing (NLP) and optical character recognition (OCR), where applicable, to identify relevant patient information such as symptom onset, duration, location, and progression. A disease-specific workflow determines the clinical relevance of extracted elements, which are then compiled into a structured or narrative HPI summary. The system further includes confidence scoring, physician validation interfaces, and dynamic regeneration of the HPI based on user corrections. Additional components include disease-aware document retrieval, timeline-based visualization of patient events, and AI-generated sections of a consultation note beyond the HPI. This invention enhances clinical efficiency, accuracy, and interoperability by automating and contextualizing the HPI generation process for use in electronic health record (EHR) systems.

The invention discloses an antibody combined with an E2-E2 antibody compound, and relates to the field of antibodies. The antibody combined with the E2-E2 antibody compound comprises a heavy chain complementarity determining region and a light chain complementarity determining region, the antibody is strongly combined with the E2 and the compound formed by the E2 antibody, is not combined with the free E2 antibody or is not combined with the free E2, the antibody provides an important raw material source for the detection of the E2, and the detection cost is reduced. And the antibody has good sensitivity and activity, good detection linear relationship and relatively high clinical correlation.

The invention discloses a GEO optimization method and system based on a medical knowledge map and user intention driving, and relates to the technical field of medical data acquisition, and the method comprises the following steps: S1, user input perception: receiving and analyzing a medical health request of a user, and extracting a symptom entity and a preliminary intention in the medical health request according to the medical health request; and S2, deep analysis: carrying out interactive reasoning on the symptom entity and the preliminary intention with a medical knowledge graph. Deep semantic analysis and intention expansion are carried out on user query through a medical knowledge graph, accurate alignment between requirements and geographic medical resources on the knowledge level is achieved, fuzzy and surface appeals of users are converted into accurate and structured medical requirements, the medical requirements are matched with deep capabilities of the resources, and the requirements of the users are matched with the deep capabilities of the resources. The clinical relevance and the user satisfaction degree of the recommendation result are greatly improved, and through a multi-objective optimization framework, the nearest distance is considered, and a personalized, dynamic and optimal medical action path is generated.

The invention discloses an intelligent auscultation system based on multi-mode biological acoustic analysis and a closed-loop diagnosis method, and relates to the technical field of medical diagnosis equipment. The method comprises the following steps: collecting broadband biological acoustic signals covering heart sound and breath sound; extracting an endogenous physiological clock signal from the broadband biological acoustic signal; separating cardiac cycle acoustic data and respiratory cycle acoustic data based on the endogenous physiological clock signal; generating an initial pathological acoustic probability map; acquiring clinical anchor point data associated with the target patient from an external medical information system; and based on the clinical anchor point data, performing constraint calibration on the initial pathological acoustic probability map to generate a final diagnosis result. Closed-loop calibration is carried out on the pure acoustic analysis model by introducing clinical anchor point data, and the accuracy, reliability and clinical correlation of automatic auscultation diagnosis are improved.

This invention provides an intelligent medical document generation method based on retrieval enhancement and multimodal features, belonging to the field of intelligent retrieval technology. The method includes constructing a cross-modal semantic association graph to achieve semantic fusion and path reasoning of multimodal medical data; retrieving relevant document fragments from a knowledge base based on graph representation; generating graph-guided sequences; transforming the document skeleton according to causal logic paths; dynamically fusing candidate fragments while maintaining semantic consistency. This invention can improve the accuracy, logicality, and clinical relevance of medical document generation.

The application discloses a schizophrenia cross-frequency coupling intervention target detection method based on synchronous EEG-fMRI, and belongs to the technical field of medical image processing and brain function analysis.The application can comprehensively depict the abnormal mode of cross-frequency coupling of schizophrenia patients by integrating multi-center resting-state synchronous EEG-fMRI data, extracting MI-based multi-band combined PAC features, performing EEG-fMRI fusion analysis and performing clinical symptom evaluation, and breaks through the limitation of traditional single mode or single frequency band analysis.The specific PAC abnormal brain region and frequency band combination significantly related to clinical symptoms are determined, and the correlation analysis with PANSS scores verifies that the detected features have good clinical correlation, which provides an operable target selection method for precise neural regulation of schizophrenia.The synchronous EEG-fMRI fusion analysis system based on cross-frequency coupling constructed by the application provides innovative technical support for mechanism exploration, intervention target detection and individualized treatment of schizophrenia, and has important scientific research value and clinical transformation potential.

The present application relates to a kind of clinical prediction method and device for the fertility potential of uterine fibroid patient, wherein the method comprises: encoding anatomical integrated dataset to generate interpretable standardized feature matrix, and cross-validation is carried out by Elastic Net elastic network model, the minimum feature subset with significant and clinical relevance to composite reproductive outcome endpoint is screened out, and interpretable baseline prediction model is trained according to the minimum feature subset;The original clinical data of the individual to be predicted is input into interpretable baseline prediction model, and the preliminary individualized fertility potential probability value is output;Individualized fertility potential probability value is corrected, and time-varying calibrated uterine fibroid related fertility potential index is generated and mapped to clinical decision threshold interval to generate suggestion report.The present application carries out longitudinal tracking and individualized correction to fertility potential by dynamic calibration coefficient vector and time-varying UFP-DI index, significantly improves the accuracy and individualized level of uterine fibroid patient fertility potential evaluation.