17 results about "Clinical variables" patented technology

The application provides a liver cancer risk stratification and treatment recommendation information generation method and device, the method comprises the following steps: obtaining liver cancer related structured clinical data of a target object; based on a pre-set electronic medical record narrative template, converting the structured clinical data into an electronic medical record style narrative text containing clinical semantics, embedding key clinical variable markers corresponding to the structured clinical data into the electronic medical record style narrative text, and constructing a model input sequence; inputting the model input sequence into a pre-trained large language model for inference operation, outputting a structured text stream, and the large language model is obtained through decision tree constraint based on a liver cancer diagnosis and treatment guideline and multi-objective reinforcement learning strategy training; and extracting comprehensive decision assistance information from the structured text stream by using a pre-set analysis rule, realizing high-credibility clinical assistance decision, and significantly improving the accuracy and logical consistency of the treatment recommendation information.

The application discloses a type 2 diabetes heart failure risk prediction system and method, and belongs to the technical field of medical diagnosis and risk assessment. The prediction system comprises the following modules: a data and feature engineering module, which is responsible for standardization processing of data and screening of key prediction factors, and obtains a core feature set for machine learning; a model construction and selection module, which uses the core feature set and trains multiple machine learning algorithms in parallel; through cross-validation and comprehensive performance evaluation, the best model is selected as a prediction model; and a model deployment and application module, which converts output results of the prediction model into a clinically usable static nomogram or online tool, and performs visual output. The application predicts by integrating clinical variables and adopting a machine learning algorithm, and provides a static nomogram and a dynamic Web application, realizes heart failure risk assessment without relying on NT-proBNP detection, and can improve the prevention and management efficiency of cardiovascular diseases.

The application discloses a kind of multi-disease prediction systems based on binocular fundus image, including fundus feature extraction module, similarity-difference feature fusion module, clinical feature extraction module, multi-modal feature fusion module and multi-disease classification prediction module.System uses unified backbone network to extract left and right fundus image features, by similarity-difference feature fusion module, the features of two-way are added to each element to obtain preliminary fusion features, and the absolute value is obtained by subtracting each element to obtain different features, after normalization and 1 complement operation generates dissimilarity weight and similarity weight, separate out dissimilarity information and similarity information after weighting fusion according to hyperparameter α Fusion fundus feature is obtained;Multi-dimensional clinical variable features are extracted by two fully connected layers, after fundus features and clinical features are spliced, input multi-disease classification prediction module, and output disease prevalence probability.The application makes full use of the correlation and complementarity of binocular fundus image, and realizes high-precision, low-cost multi-disease screening combined with clinical indicators.

This invention discloses a method and system for pontine infarct segmentation and END prediction based on multimodal joint learning. The method includes acquiring and preprocessing multimodal data; constructing a wavelet transform-based feature encoding network to perform multi-scale decomposition and detail preservation of features; constructing a dual-task guided fusion module to align the deep semantics of clinical variables and imaging features and generate task-specific representations; constructing a Mamba-based global feature aggregation module to model sequence dependencies using a state-space model; constructing a multimodal second-order fusion classifier to enhance the clinical-image interaction modeling using second-order statistics; and employing a two-stage joint training strategy for training and prediction, and outputting the prediction results. This invention utilizes the DWT / IWT mechanism to significantly improve the accuracy of capturing small pontine infarct lesions; it achieves explicit interaction between segmentation evidence and prediction signals, significantly improving the segmentation accuracy of small lesions and the reliability of stroke risk assessment.

This invention discloses an early ovarian cancer auxiliary diagnostic system based on multimodal fusion, belonging to the field of early ovarian cancer auxiliary diagnostic technology. It includes a data acquisition module, an image feature extraction module, a non-image feature extraction module, a multimodal fusion module, and a diagnostic output module. The data acquisition module acquires raw ultrasound images, serum biomarker test values, and clinical variable data; the image feature extraction module extracts image depth features; the non-image feature extraction module extracts non-image features; the multimodal fusion module dynamically calculates the attention weights of the two modalities and generates fused features; and the diagnostic output module outputs risk probability values. This invention overcomes the limitations of single-modal diagnosis by dynamically and adaptively fusing multimodal data such as ultrasound images, serum biomarkers, and clinical variables, significantly improving the accuracy of early ovarian cancer diagnosis, while also possessing interpretability and clinical deployment feasibility.

The application discloses a blood sample-based 1 The application discloses a method for constructing a metabolic related fatty liver disease typing model of H-NMR metabolomics, a typing system, a method and application. 1 The application screens 2 clinical variables and 10 H-NMR metabolic variables based on an NMR metabolomics analysis platform and a machine learning method, and establishes a scoring formula of 5 metabolic pathways by using a logistic regression model. Based on the scoring of the 5 metabolic pathways, unsupervised clustering is performed by using K-means clustering analysis, MASLD patients are divided into 3 categories, and a new disease typing model is established, and good verification is obtained in a verification queue. The 3 disease typing methods provided by the application have significant differences in metabolic disease new-onset risk and death risk and fat-reducing treatment response. The body fluid sample required by the typing system or method is convenient to extract and causes less pain to patients, and the application provides a new direction for clinical outcome prediction and clinical treatment of MASLD.

The application discloses an adnexal mass benign and malignant risk assessment method based on exosome OCS and O-RADS combination. The biomarker combination provided by the application comprises ovarian cancer OCS score and imaging grading score, and the ovarian cancer OCS score is obtained based on the concentration of CA125 protein, HE4 protein and C5a protein in exosomes. The application first uses the OCS score in combination with the O-RADS imaging score, and by comprehensively combining the biological information and the imaging characteristics, the accurate diagnosis capability for the O-RADS 3 / 4 gray area patients is significantly improved, and unnecessary excessive medical treatment and missed diagnosis are reduced. The application also includes various clinical variables such as age, menopausal status and ultrasonic characteristics, and constructs a multi-dimensional combined diagnosis model, and further improves the diagnosis performance.

Embodiments include systems and methods for determining cardiovascular information for a patient. A method includes receiving patient-specific data regarding a geometry of the patient's vasculature; creating an anatomic model representing at least a portion of the patient's vasculature based on the patient-specific data; and creating a computational model of a blood flow characteristic based on the anatomic model. The method also includes identifying one or more of an uncertain parameter, an uncertain clinical variable, and an uncertain geometry; modifying a probability model based on one or more of the identified uncertain parameter, uncertain clinical variable, or uncertain geometry; determining a blood flow characteristic within the patient's vasculature based on the anatomic model and the computational model of the blood flow characteristic of the patient's vasculature; and calculating, based on the probability model and the determined blood flow characteristic, a sensitivity of the determined fractional flow reserve to one or more of the identified uncertain parameter, uncertain clinical variable, or uncertain geometry.

This invention discloses a risk stratification assessment model, device, and construction method for RUNX1::RUNX1T1-positive children with acute myeloid leukemia. The construction method includes: acquiring sample clinical data; assessing the correlation between clinical predictors and overall survival (OS) and event-free survival (ORS); for continuous variables, determining the optimal risk cutoff value for predicting poor prognosis; converting the continuous variables into binary variables; performing univariate Cox proportional hazards regression analysis to screen for factors significantly associated with OS and ORS; and incorporating these factors into a multivariate Cox proportional hazards regression model to confirm independent prognostic factors. This invention systematically integrates the clinical variable MRD1 and the percentage of peripheral blood blasts at diagnosis as core predictors, constructing a model capable of accurately assessing ORS. + A prognostic model for pAML risk was developed and rigorously validated. This model demonstrated superior predictive performance, effectively and accurately identifying patients with a high actual risk of relapse and death from the traditionally low-risk patient population.

The present application relates to a kind of nasopharyngeal carcinoma distant metastasis risk prediction method and system based on cross-modal Transformer and depth consistency loss.Its method includes: collecting the multi-modal MRI image data and structured clinical data of patient;Using pre-training model to obtain tumor region mask, according to which T1, T1C and T2 three modal MRI image is cut, and the image of region of interest containing peritumoral microenvironment is constructed;Image is input three-dimensional image encoder, and deep image feature is extracted, while clinical variable is embedded coding;Through cross attention mechanism, image and clinical feature are fused to model, and the comprehensive feature representation of patient level is obtained;Based on the fusion feature, continuous type distant metastasis risk score is output, and combined with auxiliary classification branch and depth consistency loss, joint optimization is carried out, and the prediction result is obtained.The present application effectively focuses on tumor and peritumoral microenvironment information, enhances the deep layer interaction of multi-modal feature and early high-risk identification ability, and improves the accuracy of nasopharyngeal carcinoma distant metastasis risk prediction.

This invention relates to the field of medical artificial intelligence technology, and discloses a method, system, device, storage medium, and product for predicting the risk of nocturnal hypertension. It employs single-factor analysis combined with clinical relevance to screen several significant research factors. These significant research factors are then input into a nocturnal hypertension risk prediction model for prediction, which reduces fitting risk and noise, improves prediction accuracy, and achieves an optimal balance between predictive performance and clinical operability, facilitating rapid application in busy clinical environments. The nocturnal hypertension risk prediction model of this invention uses a table diffusion model, which can effectively capture the nonlinear relationships between clinical variables, and its predictive performance is significantly better than that of traditional logistic regression models. This invention also uses survival analysis for validation, which not only examines the model's generalization ability and shelf life over time but also effectively corrects for survivor bias caused by time camouflage.

This invention discloses a model for predicting the risk of severe SA-AKI in ICU sepsis patients based on clinical variables. This study is based on a recent prospective, multicenter, large-sample cohort study conducted in China. Using the latest ADQI definition as the diagnostic criteria for severe SA-AKI, 1715 ICU sepsis patients were included. Clinical variables were collected within 24 hours after diagnosis of ICU sepsis patients, and six feature variables with the strongest association with severe SA-AKI were screened. Six machine learning algorithms were used to construct an early prediction model, among which SVM and Logistic Regression algorithms showed better predictive performance. The SHAP method was introduced to generate feature importance maps and individual prediction interpretation maps, which can intuitively show the contribution of each clinical variable to the risk prediction of a specific patient and explain the reasons for the occurrence of severe SA-AKI, so as to help doctors take targeted preventive measures.

This invention discloses a method for generating interpretable analysis reports based on a multimodal diagnostic model. The method comprises the following steps: S1: Loading a multimodal deep learning model as the object to be interpreted and the test sample set; S2: For multimodal MRI, using the 3D Grad-CAM method to generate individualized activation heatmaps and performing group-level statistical comparisons; S3: For dFC dynamic functional connectivity, combining the SAGE algorithm and temporal attention mechanism to identify key time frames and abnormal functional connectivity patterns; S4: For clinical tabular data, applying the SAGE algorithm to calculate the mean absolute Shapley value of each clinical variable and generating a feature importance ranking; S5: Integrating the interpretation results of multimodal MRI, dFC dynamic functional connectivity, and clinical tabular data to automatically generate a structured interpretable report containing key brain regions, key time frames, and important clinical indicators. By employing submodal attribution technology, the method quantitatively analyzes the contributions of MRI anatomical features, dFC temporal patterns, and clinical tabular data to the final diagnostic decision and generates a visual interpretive report to improve model transparency and clinical credibility.