Disease Prediction Model Research Articles

Establishing of a risk prediction model for metabolic dysfunction-associated steatotic liver disease: a retrospective cohort study

ObjectivesOver 30% of people worldwide suffer from metabolic dysfunction-associated steatotic liver disease (MASLD), a significant global health issue. Identifying and preventing high-risk individuals for MASLD early is crucial. The purpose of our study is to investigate the factors related to the development of MASLD and develop a risk prediction model for its occurrence.MethodsThe study included 5107 subjects, divided into training and validation groups in a 7:3 ratio using a random number table method. Collinearity diagnosis and Cox regression were used to identify factors associated with MASLD incidence, and a risk prediction model was created. The model’s accuracy, reliability, and clinical applicability were assessed.ResultsOur study indicated that male, body mass index (BMI), low-density lipoprotein cholesterol (LDL-C), high-density lipoprotein cholesterol (HDL-C), fasting plasma glucose (FPG), serum uric acid to creatinine ratio (SUA/Cr) and white blood cell (WBC) were associated with MASLD incidence. The elements were determined to be crucial for creating a risk prediction model. The model showed strong discriminative potential with a C-index of 0.783 and the time-dependent AUCs of 0.781, 0.789, 0.814 and 0.796 for 1–4 years in the training group, and a C-index of 0.788 and the time-dependent AUCs of 0.798, 0.782, 0.787 and 0.825 for 1–4 years in validation. Calibration curves confirmed the model’s accuracy, and decision curve analysis (DCA) validated its clinical utility.ConclusionsThe model may provide clinical physicians with a reliable method for identifying high-risk populations for MASLD and serve as a guide for developing prediction models for other diseases.

P0336 From inflammation to depression: key biomarkers for Inflammatory Bowel Disease-related Major Depressive Disorder

Abstract Background Inflammatory bowel disease (IBD) is a chronic, inflammatory, and autoimmune disorder[1], and its incidence of comorbid with major depressive disorder (MDD) is significantly higher than the general population[2]. However, many patients lack proper recognition and necessary psychological health treatments. We aimed to identify potential biomarkers and mechanisms involved in the development of IBD comorbid with MDD (IBD-MDD). Methods We utilized IBD and MDD-related datasets from the GEO database for differential gene expression analysis, protein-protein interaction (PPI) and pathway enrichment analysis, random forest algorithm, LASSO regression analysis, and construction of a disease prediction model. We assessed the accuracy of the model using ROC curve, explored potential mechanisms through immune infiltration analysis, and validated candidate biomarkers using peripheral blood samples from patients in our center’s cohort. Results We identified 484 IBD-related secreted proteins and 142 key module genes associated with MDD. PPI analysis revealed two crucial modules primarily involved in inflammation and immune regulation. We identified four diagnostic genes (HGF, SPARC, ADAM12, and MMP8) from the 21 shared genes between IBD-related secreted proteins and MDD key module genes, constructed a nomogram model and confirmed its accuracy using ROC curve from an external independent dataset. Immune infiltration analysis revealed significant associations between the four diagnostic genes, and cellular immune dysregulation in MDD. Finally, we validated the expression patterns of the four diagnostic genes in our cohort.The results showed that the serum levels of HGF and MMP8 were significantly elevated in IBD patients with comorbid MDD, SPARC and ADAM12 showed an increasing trend (Figure 1A). Meanwhile, we developed a predictive model for IBD-related MDD based on our cohort (Figure 1B). The calibration curve suggests that the predictive probability of the nomogram model is nearly the same as that of the ideal model (Figure 1C). ROC curve analysis of the efficacy of the joint predictive model based on the 4 diagnostic markers showed that the AUC value was 0.87, indicating that the nomogram have strong diagnostic value for IBD-related MDD (Figure 1D). Conclusion Our study discovered four candidate biomarkers for IBD-MDD, providing new insights for the diagnosis and therapeutic intervention of serum-based IBD comorbid with MDD.

Prediction-powered Inference for Clinical Trials.

Prediction-powered inference (PPI) [1] and its subsequent development called PPI++ [2] provide a novel approach to standard statistical estimation leveraging machine learning systems to enhance unlabeled data with predictions. We use this paradigm in clinical trials. The predictions are provided by disease progression models, providing prognostic scores for all the participants as a function of baseline covariates. The proposed method would empower clinical trials by providing untreated digital twins of the treated patients while remaining statistically valid. The potential implications of this new estimator of the treatment effect in a two-arm randomized clinical trial (RCT) are manifold. First, it leads to an overall reduction of the sample size required to reach the same power as a standard RCT. Secondly, it advocates for an imbalance of controls and treated patients, requiring fewer controls to achieve the same power. Finally, this technique directly transfers any disease prediction model trained on large cohorts to practical and scientifically valid use. In this paper, we demonstrate the theoretical properties of this estimator and illustrate them through simulations. We show that it is asymptotically unbiased for the Average Treatment Effect and derive an explicit formula for its variance. An application to an Alzheimer's disease clinical trial showcases the potential to reduce the sample size.

ML-powered Internet of Medical Things Structure for Heart Disease Prediction

Background Machine Learning-powered Internet of Medical Things (MLIoMT) is a burgeoning framework poised to transform healthcare, particularly in the timely identification of heart disease. Purpose This article proposes an innovative MLIoMT structure aimed at leveraging machine learning (ML) algorithms for heart disease detection. Methods Through the integration of wearable sensors, mobile applications, cloud computing, and advanced ML techniques, MLIoMT enables continuous monitoring of vital signs and cardiac health indicators in real time. By analyzing this data stream, abnormalities indicative of heart disease can be detected early, facilitating timely intervention and personalized healthcare recommendations. The MLIoMT framework employs diverse ML methods such as deep learning and ensemble techniques to enhance the accuracy and reliability of heart disease prediction models. Results The proposed structure holds promise for revolutionizing preventive healthcare, enabling proactive management of cardiac health and ultimately reducing the burden of heart disease. Results in terms of accuracy, precision, recall and F1 score show that the proposed system has better performance and efficiency. Conclusion Overall, MLIoMT represents a significant advancement in healthcare technology, with the potential to improve patient outcomes and enhance overall quality of life.

An Intelligent Ensemble Deep Learning Techniques with Improved Owl Search Algorithm-Aided Optimal Feature Selection for Predicting the Presence of Heart Diseases

Cardiovascular disease evaluation and prediction are critical medical duties to ensure precise classification, which allows heart specialists to offer treatment to patients. Heart illness is one of the deadliest diseases, particularly a sudden cardiac arrest, which strikes an individual so suddenly that there is no opportunity to manage it, and this condition is extremely hard to detect. Numerous medical machine learning, as well as data mining methods, have been employed to extract significant information about coronary artery disease prognosis. However, the precision of the intended outcomes is not adequate. An effective deep learning-assisted heart disease prediction model is presented to solve the above difficulties. At first, the data are accumulated from a standard dataset. After this process, the features from the data are optimally selected using the Improved Owl Search Algorithm (IOSA). These features are further transferred to the prediction phase. Here, the heart disease is predicted by the Ensemble Deep Attention with Convolution Network (EDACNet), and this model is composed by the combination of “Long Short-Term Memory (LSTM)”, “Gated Recurrent Unit (GRU)”, “Deep Markov Random Field (Deep MRF)” and “1Dimensional Convolutional Neural Network (1DCNN)”. The prediction scores obtained from each structure are multiplied by an optimized weight, and it is subjected to a weighted averaging process. Here, the weight is optimized by the same IOSA. The final results after the weighted averaging process provide the final predicted outcomes over heart diseases. At last, the validation process is carried out to measure the usefulness of a designed prediction model. The result of the developed model achieves better performance in terms of accuracy and precision. It shows the value of accuracy to be 96 and also the precision attained the value of 97. Therefore, the entire validation developed a model that provides effective performance over existing techniques.

A Novel Point-of-Care Prediction Model for Steatotic Liver Disease: Expected Role of Mass Screening in the Global Obesity Crisis.

The incidence of steatotic liver disease (SLD) is increasing across all age groups as the incidence of obesity increases worldwide. The existing noninvasive prediction models for SLD require laboratory tests or imaging and perform poorly in the early diagnosis of infrequently screened populations such as young adults and individuals with healthcare disparities. We developed a machine learning-based point-of-care prediction model for SLD that is readily available to the broader population with the aim of facilitating early detection and timely intervention and ultimately reducing the burden of SLD. We retrospectively analyzed the clinical data of 28,506 adults who had routine health check-ups in South Korea from January to December 2022. A total of 229,162 individuals were included in the external validation study. Data were analyzed and predictions were made using a logistic regression model with machine learning algorithms. A total of 20,094 individuals were categorized into SLD and non-SLD groups on the basis of the presence of fatty liver disease. We developed three prediction models: SLD model 1, which included age and body mass index (BMI); SLD model 2, which included BMI and body fat per muscle mass; and SLD model 3, which included BMI and visceral fat per muscle mass. In the derivation cohort, the area under the receiver operating characteristic curve (AUROC) was 0.817 for model 1, 0.821 for model 2, and 0.820 for model 3. In the internal validation cohort, 86.9% of individuals were correctly classified by the SLD models. The external validation study revealed an AUROC above 0.84 for all the models. As our three novel SLD prediction models are cost-effective, noninvasive, and accessible, they could serve as validated clinical tools for mass screening of SLD.

A transformer-based model for next disease prediction using electronic health records

A transformer-based model for next disease prediction using electronic health records

Advances in the clinical application of machine learning in acute pancreatitis: a review.

Traditional disease prediction models and scoring systems for acute pancreatitis (AP) are often inadequate in providing concise, reliable, and effective predictions regarding disease progression and prognosis. As a novel interdisciplinary field within artificial intelligence (AI), machine learning (ML) is increasingly being applied to various aspects of AP, including severity assessment, complications, recurrence rates, organ dysfunction, and the timing of surgical intervention. This review focuses on recent advancements in the application of ML models in the context of AP.

Construction and evaluation of the prediction model for advanced disease in well-differentiated colorectal neuroendocrine neoplasms less than 2cm in diameter.

Advanced lesions are often ignored in well-differentiated colorectal neuroendocrine neoplasms (NENs) smaller than 2cm, and we aimed to develop an effective nomogram for these lesions. We extracted data from the Surveillance, Epidemiology, and End Results (SEER) database and used a logistic regression model to identify independent risk factors for advanced disease. All these identified factors were included to construct the prediction model, and the receiver operating characteristic (ROC) curve, calibration plot and DCA curve were utilized to assess the predictive value. The data obtained from the National Cancer Center were utilized for external validation. In total, 3223 patients were enrolled in the training set, including 2947 (91.4%) with early disease and 276 (8.6%) with advanced disease. The logistic analysis showed that age (odds ratio (OR)=1.486, 95% confidence interval (CI): 1.102-2.003, P=0.009), tumor size (OR=11.071, 95% CI: 8.229-14.893, P<0.001), tumor location (OR=7.882, 95% CI: 5.784-10.743, P<0.001) and tumor grade (OR=1.768, 95% CI: 1.206-2.593, P=0.004) were independent variables for advanced disease. All of them were included in the final prediction model. The area under the ROC curve (AUC) was 0.838 (95% CI: 0.807-0.868). The calibration plot and Hosmer‒Lemeshow test (P=0.108) indicated favorable consistency between the predicted probabilities and actual probabilities of advanced disease. The Brier score was 0.108, indicating acceptable overall performance. The DCA curve presented a significant clinical net benefit. In the validation set, both the ROC curve and calibration plot exhibited an acceptable discrimination ability (AUC=0.807 (95% CI 0.702-0.913) and calibration (Hosmer Lemeshow P=0.997), respectively. The prediction model had good value for identifying advanced disease from well-differentiated colorectal NENs smaller than 2cm.

A Novel Blockchain-Enabled Fuzzy CLSTM Model for Secure and Scalable Heart Disease Prediction in Healthcare

The emerging field of healthcare has taken severe measures to safeguard sensitive patient health-related information especially the information taken from the predictive model. In this study, a novel blockchain-based solution is proposed in correlation with the Fuzzy-enhanced CLSTM model (FCLSTM) for storing and transmitting the data securely for heart disease prediction systems by ensuring data integrity, confidentiality, and access control. The proposed model uses a blockchain-based network which is implemented to prevent the tampering or unauthorized access to patients’ health-related data. The process begins with techniques that incorporate the predicted heart disease information from the patient’s data and is encrypted by using the hashing algorithm. A secure hybrid blockchain-based data management framework (SHB-DMF) is designed for exchanging the patient’s health data which enhances scalability and accessibility to the healthcare environment. The system incorporates a SHAES-256 hybrid model for enhancing the data confidentiality and integrity before transmitting to the neural network (FCLSTM). The proposed model uses a smart contract for regulating data access by ensuring the entry of the authorized entities by providing a suitable decrypting mechanism and interacting with the patient’s data. The smart contracts can automate the data retrieval workflows by integrating the blockchain seamlessly with the prediction model. The security process is a three-phase process that includes defining the nodes, selecting of consensus mechanism, and establishing the governance structure for facilitating secure operations. The security and load testing ensure resilience to potential cyber threats and the scalability required for handling high transaction volumes of medical data. Deploying the proposed system provides a robust infrastructure that is tamper-resistant thus advancing the reliability of the cardiovascular prediction system.

Construction of a prediction model for heart disease: a study based on random forest, KNN and logistic regression

Construction of a prediction model for heart disease: a study based on random forest, KNN and logistic regression

Supervised Learning based Chest Disease Prediction with IOT from Xray Images

New WHO research indicates that there are an increasing number of chest-related illnesses. This results in the deaths of 17.9 million people annually. It gets harder to identify problems and start therapy at an early age as the population grows. However, new developments in technology, such as deep learning and machine learning methods, have sped up research in the medical profession. The creation of a machine learning and deep learning model for heart disease prediction based on pertinent features is the goal of this study. The chest's X-ray images are kept in the cloud for public access in the suggested method. The images retrieved from the cloud via the internet are analyzed for prediction using machine learning methods such as K Nearest Neighbor and Random Forest and deep learning algorithms such as Convolutional Neural Network and ResNet. Results are further stored in the cloud. Doctors, users, and patients can access the saved findings on cloud servers for diagnosis and treatment. We used a benchmark dataset from UCI Chest Disease Prediction for this research study. The proposed method classified X-ray images into normal and chest disease as cardiomegaly, aortic enlargement, and enlarged cardioment. Results are updated in the cloud for doctor diagnosis purposes. From experimental analysis, ResNet produced better results compared to other methods.

Multiangle Correlation Feature Extraction and Disease Prediction Model Construction for Patients With Post-Stroke Dysarthria

Multiangle Correlation Feature Extraction and Disease Prediction Model Construction for Patients With Post-Stroke Dysarthria

Advancing Early Cardiovascular Disease Prediction Model using Improved Beluga Whale Optimization with Ensemble Learning via ECG Signal Analytics

Cardiovascular Disease (CVD) mainly affects the blood vessels and heart such as coronary artery disease, stroke, and heart failure. Early recognition is vital for on-time intervention and enhanced patient results. CVD is a major issue in society nowadays. When compared to the non-invasive model, the electrocardiogram (ECG) is the most effective approach for identifying cardiac defects. However, ECG analysis needs an experienced person with high knowledge and basically, it is a time-consuming task. Emerging a new technique to identify the disease at an early stage increases the quality and efficacy of medicinal care. A state-of-the-art technologies like machine learning (ML) and artificial intelligence (AI) have been gradually being used to increase the efficacy and accuracy of CVD recognition, permitting for faster and more exact analysis, and finally contributing to superior management and prevention tactics for CV health. This research paper designs an Early Cardiovascular Disease Prediction using an Improved Beluga Whale Optimizer with Ensemble Learning (ECVDP-IBWOEL) approach via ECG Signal Analytics. The main intention of the ECVDP-IBWOEL system is to forecast the presence of CVD at the early stage using EEG signals. In the ECVDP-IBWOEL method, the primary phase of data preprocessing is initially implemented to convert the input data into a well-suited layout. Also, the ECVDP-IBWOEL technique follows an ensemble learning (EL) process for CVD detection comprising three models namely long short-term memory (LSTM), deep belief networks (DBNs), and stacked autoencoder (SAE). Finally, the IBWO algorithm-based hyperparameter tuning process takes place which can boost the classifier results of the ensemble models. To certify the enhanced results of the ECVDP-IBWOEL system, an extensive experimental study is made. The experimentation outcomes stated that the ECVDP-IBWOEL system underlines promising performance in the CVD prediction process

Integrating EPSOSA-BP neural network algorithm for enhanced accuracy and robustness in optimizing coronary artery disease prediction

Coronary artery disease represents a formidable health threat to middle-aged and elderly populations worldwide. This research introduces an advanced BP neural network algorithm, EPSOSA-BP, which integrates particle swarm optimization, simulated annealing, and a particle elimination mechanism to elevate the precision of heart disease prediction models. To address prior limitations in feature selection, the study employs single-hot encoding and Principal Component Analysis, thereby enhancing the model’s feature learning capability. The proposed method achieved remarkable accuracy rates of 93.22% and 95.20% on the UCI and Kaggle datasets, respectively, underscoring its exceptional performance even with small sample sizes. Ablation experiments further validated the efficacy of the data preprocessing and feature selection techniques employed. Notably, the EPSOSA algorithm surpassed classical optimization algorithms in terms of convergence speed, while also demonstrating improved sensitivity and specificity. This model holds significant potential for facilitating early identification of high-risk patients, which could ultimately save lives and optimize the utilization of medical resources. Despite implementation challenges, including technical integration and data standardization, the algorithm shows promise for use in emergency settings and community health services for regular cardiac risk monitoring.

Development and validation of a Prediction Model for Chronic Thromboembolic Pulmonary Disease

BackgroundAcute pulmonary embolism (APE) is a critical disease with a high mortality rate, some of the surviving patients may develop chronic thromboembolic pulmonary disease (CTEPD), which affects the patient’s prognosis. However, the research on the early diagnosis of CTEPD is limited. This study aimed to establish a prediction model for earlier identification of CTEPD.MethodsThis prospective study included 464 consecutive patients with APE confirmed between January 2020 and September 2023, at 7 centers from China. After follow-up for at least 3 months, the patients were divided into the CTEPD and non-CTEPD groups based on symptoms and computed tomography pulmonary angiography (CTPA) or pulmonary ventilation perfusion (V/Q) scans showing residual thrombosis. The independent risk factors for CTEPD were identified via univariate and multivariate logistic regression analyses. Next, a nomogram of predictive model was established, and validation was completed via decision curve analysis (DCA) and receiver operating characteristic curve analysis.ResultIn total, 130 (28%) patients presented with CTEPD, 17% (22/130) of CTEPD patients developed chronic thromboembolic pulmonary hypertension (CTEPH). Based on the multivariate analysis, a time interval from symptoms onset to diagnosis (time-to-diagnosis) ≥ 15 days (95% confidence interval [CI]: 3.392–14.972, p < 0.001), recurrent pulmonary embolism (RPE) (95%CI: 1.560–17.300, p = 0.007), right ventricular dysfunction (RVD) (95%CI: 1.042–6.437, p = 0.040), central embolus (95%CI: 1.776–7.383, p < 0.001) and residual pulmonary vascular obstruction (RPVO) > 10% (95%CI: 4.884–21.449, p < 0.001) were identified as the independent predictors of CTEPD. Then, A prediction model with a C-index of 0.895 (95% CI 0.863–0.927) was established for high-risk patients. The nomogram had an excellent predictive performance for earlier identification of CTEPD, with an area under the curve of 0.908 (95%CI: 0.875–0.941) in the training cohort and 0.875 (95%CI: 0.803–0.947) in the validation cohort.ConclusionThe current study established and validated a reliable nomogram for predicting CTEPD, which would assist clinicians identify the high-risk patients for CTEPD earlier.

A Streamlined Vgg-19 Classification Model for Multi-Crop Leaf Disease Prediction

Every country economy is depending heavily on agriculture because it produces crops. Identifying leaf diseases is one of the most crucial parts of keeping a country that is agriculturally advanced. Artificial intelligence (AI) in agriculture has emerged as the most significant application in recent years. An early prediction of the type of disease affecting the plant leaves is an essential objective. To classify the current diseases type and distinguish between healthy and infected leaves, the proposed study mainly employs the Visual Geometry Group (VGG 19) classification model. The features extraction of images using a wavelet transform for extracting the features of the leaves. The model is trained and evaluated using the Apple Leaf Disease Symptoms dataset, and the adaptive wiener filter is used to enhance the images and remove noise from the leaf images, while the k means clustering algorithm divides the regions in the leaf images. Accuracy, sensitivity, specificity, recall, and F1 score were among the performance measure parameters that were computed and tracked. This classifier is implemented using the python software. The findings show that it offers greater accuracy and the results are then contrasted with those of earlier classification methods. By identifying plant illnesses early and implementing the right treatment, this approach aims to assist farmers in protecting resources and avoiding financial loss.

Enhancing Sensitivity of Point-of-Care Thyroid Diagnosis via Computational Analysis of Lateral Flow Assay Images Using Novel Textural Features and Hybrid-AI Models.

Lateral flow assays are widely used in point-of-care diagnostics but face challenges in sensitivity and accuracy when detecting low analyte concentrations, such as thyroid-stimulating hormone biomarkers. This study aims to enhance assay performance by leveraging textural features and hybrid artificial intelligence models. A modified Gray-Level Co-occurrence Matrix, termed the Averaged Horizontal Multiple Offsets Gray-Level Co-occurrence Matrix, was utilised to compute the textural features of the biosensor assay images. Significant textural features were selected for further analysis. A deep learning Convolutional Neural Network model was employed to extract features from these textural features. Both traditional machine learning models and hybrid artificial intelligence models, which combine Convolutional Neural Network features with traditional algorithms, were used to categorise these textural features based on the thyroid-stimulating hormone concentration levels. The proposed method achieved accuracy levels exceeding 95%. This pioneering study highlights the utility of textural aspects of assay images for accurate predictive disease modelling, offering promising advancements in diagnostics and management within biomedical research.

A Meta-Analysis of the Human Gut Mycobiome Using Internal Transcribed Spacer Data.

The intestinal mycobiome is closely related to human health. There have been several reports investigating the association between the gut fungi and disease, but there is still a lack of overall assessment of the human gut mycobiome. Here, we performed a meta-analysis based on 2372 ITS (Internal Transcribed Spacer) data collected publicly online. We found that the mycobiome diversity of human gut fungi varies significantly across diseases by using EasyAmplicon, and these fungi are mainly composed of three genera, Saccharomyces, Candida, and Aspergillus. In addition, we performed the construction of disease prediction models based on ITS data by using the random forest model and verified the generalization ability of the models. We hope that our results will provide strong support for subsequent studies of the intestinal mycobiome.

Development and Evaluation of a Comprehensive Prediction Model for Incident Coronary Heart Disease Using Genetic, Social, and Lifestyle-Psychological Factors: A Prospective Analysis of the UK Biobank.

Clinical risk calculators for coronary heart disease (CHD) do not include genetic, social, and lifestyle-psychological risk factors. To improve CHD risk prediction by developing and evaluating a prediction model that incorporated a polygenic risk score (PRS) and a polysocial score (PSS), the latter including social determinants of health and lifestyle-psychological factors. Cohort study. United Kingdom. UK Biobank participants recruited between 2006 and 2010. Incident CHD (myocardial infarction and/or coronary revascularization); 10-year clinical risk based on pooled cohort equations (PCE), Predicting Risk of cardiovascular disease EVENTs (PREVENT), and QRISK3; PRS (Polygenic Score Catalog identification: PGS000018) for CHD (PRSCHD); and PSSCHD from 100 related covariates. Machine-learning and time-to-event analyses and model performance indices. In 388 224 participants (age, 55.5 [SD, 8.1] years; 42.5% men; 94.9% White), the hazard ratio for 1 SD increase in PSSCHD for incident CHD was 1.43 (95% CI, 1.38 to 1.49; P < 0.001) and for 1 SD increase in PRSCHD was 1.59 (CI, 1.53 to 1.66, P < 0.001). Non-White persons had higher PSSCHD than White persons. The effects of PSSCHD and PRSCHD on CHD were independent and additive. At a 10-year CHD risk threshold of 7.5%, adding PSSCHD and PRSCHD to PCE reclassified 12% of participants, with 1.86 times higher CHD risk in the up- versus down-reclassified persons and showed superior performance compared with PCE as reflected by improved net benefit while maintaining good calibration relative to the clinical risk calculators. Similar results were seen when incorporating PSSCHD and PRSCHD into PREVENT and QRISK3. A predominantly White cohort; possible healthy participant effect and ecological fallacy. A PSSCHD was associated with incident CHD and its joint modeling with PRSCHD improved the performance of clinical risk calculators. National Human Genome Research Institute.