{"title":"TCKAN: a novel integrated network model for predicting mortality risk in sepsis patients.","authors":"Fanglin Dong, Shibo Li, Weihua Li","doi":"10.1007/s11517-024-03245-2","DOIUrl":"10.1007/s11517-024-03245-2","url":null,"abstract":"<p><p>Sepsis poses a major global health threat, accounting for millions of deaths annually and significant economic costs. Accurately predicting the risk of mortality in sepsis patients enables early identification, promotes the efficient allocation of medical resources, and facilitates timely interventions, thereby improving patient outcomes. Current methods typically utilize only one type of data-either constant, temporal, or ICD codes. This study introduces a novel approach, the Time-Constant Kolmogorov-Arnold Network (TCKAN), which uniquely integrates temporal data, constant data, and ICD codes within a single predictive model. Unlike existing methods that typically rely on one type of data, TCKAN leverages a multi-modal data integration strategy, resulting in superior predictive accuracy and robustness in identifying high-risk sepsis patients. Validated against the MIMIC-III and MIMIC-IV datasets, TCKAN surpasses existing machine learning and deep learning methods in accuracy, sensitivity, and specificity. Notably, TCKAN achieved AUCs of 87.76% and 88.07%, demonstrating superior capability in identifying high-risk patients. Additionally, TCKAN effectively combats the prevalent issue of data imbalance in clinical settings, improving the detection of patients at elevated risk of mortality and facilitating timely interventions. These results confirm the model's effectiveness and its potential to transform patient management and treatment optimization in clinical practice. Although the TCKAN model has already incorporated temporal, constant, and ICD code data, future research could include more diverse medical data types, such as imaging and laboratory test results, to achieve a more comprehensive data integration and further improve predictive accuracy.</p>","PeriodicalId":49840,"journal":{"name":"Medical & Biological Engineering & Computing","volume":" ","pages":"1013-1025"},"PeriodicalIF":2.6,"publicationDate":"2025-04-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142669767","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
{"title":"Artificial blood-hope and the challenges to combat tumor hypoxia for anti-cancer therapy.","authors":"Rishabh Sharma, Manju Kashyap, Hatem Zayed, Lucky Krishnia, Manoj Kumar Kashyap","doi":"10.1007/s11517-024-03233-6","DOIUrl":"10.1007/s11517-024-03233-6","url":null,"abstract":"<p><p>The blood plays a vital role in the human body and serves as an intermediary between various physiological systems and organs. White blood cells, which are a part of the immune system, defend against infections and regulate the body temperature and pH balance. Blood platelets play a crucial role in clotting, the prevention of excessive bleeding, and the promotion of healing. Blood also serves as a courier system that transports hormones to facilitate communication and synchronization between different organs and systems in the body. The circulatory system, comprised of arteries, veins, and capillaries, plays a crucial role in the efficient transportation and connection of vital nutrients and oxygen. Despite the importance of natural blood, there are often supply shortages, compatibility issues, and medical conditions, which make alternatives such as artificial blood necessary. This is particularly relevant in cancer treatment, which was the focus of our study. In this study, we investigated the potential of artificial blood in cancer therapy, specifically to address tumor hypoxia. We also examined the potential of red blood cell substitutes such as hemoglobin-based oxygen carriers and perfluorocarbons. Additionally, we examined the production of hemoglobin using E. coli and the role of hemoglobin in oncogenesis. Furthermore, we explored the potential use of artificial platelets for cancer treatment. Our study emphasizes the significance of artificial blood in improving cancer treatment outcomes.</p>","PeriodicalId":49840,"journal":{"name":"Medical & Biological Engineering & Computing","volume":" ","pages":"933-957"},"PeriodicalIF":2.6,"publicationDate":"2025-04-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142755760","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Everton F Baro, Luiz S Oliveira, Alceu de Souza Britto
{"title":"Predicting hospitalization with LLMs from health insurance data.","authors":"Everton F Baro, Luiz S Oliveira, Alceu de Souza Britto","doi":"10.1007/s11517-024-03251-4","DOIUrl":"10.1007/s11517-024-03251-4","url":null,"abstract":"<p><p>Predictions of hospitalizations can help in the development of applications for health insurance, hospitals, and medicine. The data collected by health insurance has potential that is not always explored, and extracting features from it for use in machine learning applications requires demanding processes and specialized knowledge. With the emergence of large language models (LLM) there are possibilities to use this data for a wide range of applications requiring little specialized knowledge. To do this, it is necessary to organize and prepare this data to be used by these models. Therefore, in this work, an approach is presented for using data from health insurance in LLMs with the objective of predict hospitalizations. As a result, pre-trained models were generated in Portuguese and English with health insurance data that can be used in several applications. To prove the effectiveness of the models, tests were carried out to predict hospitalizations in general and due to stroke. For hospitalizations in general, F1-Score = 87.8 and AUC = 0.955 were achieved, and for hospitalizations due to stroke, the best model achieved F1-Score = 88.7 and AUC of 0.964. Considering the potential for use, the models were made available to the scientific community.</p>","PeriodicalId":49840,"journal":{"name":"Medical & Biological Engineering & Computing","volume":" ","pages":"1215-1226"},"PeriodicalIF":2.6,"publicationDate":"2025-04-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142856411","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Kai Zhang, Wei Liang, Peng Cao, Zhaoyang Mao, Jinzhu Yang, Osmar R Zaiane
{"title":"CorLabelNet: a comprehensive framework for multi-label chest X-ray image classification with correlation guided discriminant feature learning and oversampling.","authors":"Kai Zhang, Wei Liang, Peng Cao, Zhaoyang Mao, Jinzhu Yang, Osmar R Zaiane","doi":"10.1007/s11517-024-03247-0","DOIUrl":"10.1007/s11517-024-03247-0","url":null,"abstract":"<p><p>Recent advancements in deep learning techniques have significantly improved multi-label chest X-ray (CXR) image classification for clinical diagnosis. However, most previous studies neither effectively learn label correlations nor take full advantage of them to improve multi-label classification performance. In addition, different labels of CXR images are usually severely imbalanced, resulting in the model exhibiting a bias towards the majority class. To address these challenges, we introduce a framework that not only learns label correlations but also utilizes them to guide the learning of features and the process of oversampling. In this paper, our approach incorporates self-attention to capture high-order label correlations and considers label correlations from both global and local perspectives. Then, we propose a consistency constraint and a multi-label contrastive loss to enhance feature learning. To alleviate the imbalance issue, we further propose an oversampling approach that exploits the learned label correlation to identify crucial seed samples for oversampling. Our approach repeats 5-fold cross-validation process experiments three times and achieves the best performance on both the CheXpert and ChestX-Ray14 datasets. Learning accurate label correlation is significant for multi-label classification and taking full advantage of label correlations is beneficial for discriminative feature learning and oversampling. A comparative analysis with the state-of-the-art approaches highlights the effectiveness of our proposed methods.</p>","PeriodicalId":49840,"journal":{"name":"Medical & Biological Engineering & Computing","volume":" ","pages":"1045-1058"},"PeriodicalIF":2.6,"publicationDate":"2025-04-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142752064","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Shih-Hsien Tseng, Kung-Min Wang, Ting-Yang Su, Kung-Jeng Wang
{"title":"Survivability prognosis of lung cancer patients with comorbidities-a Gaussian Bayesian network model.","authors":"Shih-Hsien Tseng, Kung-Min Wang, Ting-Yang Su, Kung-Jeng Wang","doi":"10.1007/s11517-024-03261-2","DOIUrl":"10.1007/s11517-024-03261-2","url":null,"abstract":"<p><p>Comorbidities are influencing factors that cause lung cancer. An accurate survivability prediction model is required considering these confounding factors (a variety of comorbidities and treatments). The study developed a conditional Gaussian Bayesian network (CGBN) model to predict the related survival time with likelihood under various conditions. The lung cancer patients were collected from the National Health Insurance Research Database in Taiwan. Six major chronic diseases (i.e., pulmonary tuberculosis, COPD, kidney failure, diabetes mellitus, stroke, and liver disease) are investigated. A total of 2875 lung cancer cases with key comorbidities were selected. This study examined three types of lung cancer treatment: surgery, chemotherapy, and targeted therapy. The study outcomes provided the likelihood of survival time occurrences. Survival analysis indicates that diabetes mellitus and liver disease are significantly riskier than the other comorbidities for lung cancer patients. The proposed CGBN model achieved high accuracy as compared to the existing literature. The proposed CGBN model is advantageous for modeling the relationship between numerical and categorical influencing factors and response variables for lung cancer with comorbidities. The proposed model facilitates the flexible and accurate estimation of various lung cancer-related queries.</p>","PeriodicalId":49840,"journal":{"name":"Medical & Biological Engineering & Computing","volume":" ","pages":"1201-1213"},"PeriodicalIF":2.6,"publicationDate":"2025-04-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142856413","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
{"title":"Understanding the implication of task conditions on asymmetry in gait of post-stroke individuals using an Integrated Wearable System.","authors":"Shashi Ranjan, Priya Darji, Shraddha J Diwan, Uttama Lahiri","doi":"10.1007/s11517-024-03249-y","DOIUrl":"10.1007/s11517-024-03249-y","url":null,"abstract":"<p><p>Hemiplegic individuals often demonstrate gait abnormality causing asymmetry in lower-limb muscle activation-related (implicit) and gait-related (explicit) measures (offering complementary information on one's gait) while walking. Added to hemiplegia, such asymmetry can be aggravated while walking under varying task conditions, namely, walking without speaking (single task), walking while counting backwards (dual task), and walking while holding an object and counting backwards (multiple task). This emphasizes the need to quantify the extent of aggravated implication of multiple-task and dual-task on gait asymmetry compared to single task. Here, we used Integrated Wearable System and carried out a study with a group of age-matched hemiplegic (Grp_S) and healthy (Grp_H) individuals to understand the potential of our system in quantifying asymmetry in explicit and implicit measures of gait, implication of hemiplegic condition and varying task conditions on these asymmetry measures along with their clinical relevance. Results showed the potential of our system in quantifying asymmetry in both explicit and implicit measures of gait, and these measures were statistically higher (p-value < 0.05) in Grp_S than Grp_H irrespective of the task conditions. Also, for Grp_S, these asymmetry measures became more pronounced as task demand increased, and again, these measures have shown a correlation with their risk of fall specifically during more attention-demanding tasks that could be clinically relevant.</p>","PeriodicalId":49840,"journal":{"name":"Medical & Biological Engineering & Computing","volume":" ","pages":"1227-1248"},"PeriodicalIF":2.6,"publicationDate":"2025-04-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142856423","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
{"title":"Automated detection of retinal artery occlusion in fundus photography via self-supervised deep learning and multimodal interpretability using a multimodal AI chatbot.","authors":"Sun Young Ryu, Joon Yul Choi, Tae Keun Yoo","doi":"10.1007/s11517-025-03353-7","DOIUrl":"https://doi.org/10.1007/s11517-025-03353-7","url":null,"abstract":"<p><p>Retinal artery occlusion (RAO) is a sight-threatening condition that requires prompt diagnosis to prevent irreversible vision loss. This study presents an innovative AI-driven approach for RAO detection from fundus images, marking the first application of deep learning for this purpose. Using a self-supervised learning (SSL) framework with SimCLR, our model addresses the challenge of limited labeled RAO data. The ResNet50 model pretrained with SimCLR demonstrated high diagnostic accuracy, achieving areas under the receiver operating characteristic curve (AUC) of 0.924 and 0.988 on two external validation datasets, highlighting its robustness and generalizability in RAO detection. To enhance transparency in clinical AI, we incorporated a multimodal interpretability approach using a ChatGPT-4-based AI chatbot. This chatbot, combined with Grad-CAM visualizations, provides detailed clinical explanations of the model's predictions, emphasizing key RAO features such as retinal whitening and cherry-red spots. This multimodal interpretability framework improves clinicians' understanding of the model's decision-making process, facilitating clinical adoption and trust. By automating RAO detection, this AI model serves as a valuable tool for the early identification of ocular and systemic vascular risks, enabling timely intervention. These findings highlight the potential of fundus imaging for RAO detection and broader cardiovascular risk assessment, advancing AI's role in predictive healthcare.</p>","PeriodicalId":49840,"journal":{"name":"Medical & Biological Engineering & Computing","volume":" ","pages":""},"PeriodicalIF":2.6,"publicationDate":"2025-03-31","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143755557","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Juanyi Wang, Yi Zhang, Yang Yang, Xiao Wen, Xiaodong Xing, Zhiyong Zhang, Chen Dong
{"title":"Digital twin-enabled cardiovascular stent optimization: a virtual reality-driven approach to mitigate angioplasty-related deformations.","authors":"Juanyi Wang, Yi Zhang, Yang Yang, Xiao Wen, Xiaodong Xing, Zhiyong Zhang, Chen Dong","doi":"10.1007/s11517-025-03352-8","DOIUrl":"https://doi.org/10.1007/s11517-025-03352-8","url":null,"abstract":"<p><p>Percutaneous transluminal angioplasty with stenting is extensively applied for treatment of atherosclerosis. However, the effects of dogboning (d<sub>b</sub>), foreshortening (f<sub>s</sub>), longitudinal recoil (l<sub>r</sub>) and radial recoil (R<sub>r</sub>) usually occur to inflict damage to the artery and make the positioning difficult during the cardiovascular stent (CS) expansion to the maximum and after the inflated balloon removing. In the article, the design and manufacture of a CS were carried out based on digital twin (DT) technology rather than traditional expertise- and experience-based methods. The highly kinetic model of a CS was firstly derived from its upfront proposed geometric configuration, governing equations of solid mechanics and boundary conditions to construct its DT through virtual reality (VR). Then global sensitivity analysis (GSA) and dynamic response optimization (DRO) was implemented to optimize the material and processing parameters including Young's modulus (E), isotropic tangent modulus (E<sub>t</sub>), Poisson's ratio (ν), density (ρ) and initial yield stress (σ), in order to obtain a satisfied behavior requirements for effects of d<sub>b</sub>, f<sub>s</sub>, l<sub>r</sub> and R<sub>r</sub>. The prototype experiment result showed that the CS made of shape memory Nitinol with optimal material and processing parameters (ρ = 7050 kg m<sup>-3</sup>, ν = 0.27, E = 205 GPa, E<sub>t</sub> = 675.13 MPa and σ = 198.49 MPa) obtained from its digital twin through VR simulation could have desired behavior performance characteristics, such as weak effect of d<sub>b</sub> and f<sub>s</sub> during the CS expansion to the maximum, and l<sub>r</sub> (-0.9%), distal R<sub>r</sub> (0.4%) and central R<sub>r</sub> (0.7%) after the inflated balloon removing.</p>","PeriodicalId":49840,"journal":{"name":"Medical & Biological Engineering & Computing","volume":" ","pages":""},"PeriodicalIF":2.6,"publicationDate":"2025-03-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143712009","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Fuyuan Liao, Liwan Huang, Pu-Chun Mo, Mansoureh Samadi, Nicolas Kelhofer, Tony Tu, Yih-Kuen Jan
{"title":"Effects of handgrip contraction modes on intermuscular coordination quantified by wavelet-based EMG-EMG coherence between 2 and 300 Hz.","authors":"Fuyuan Liao, Liwan Huang, Pu-Chun Mo, Mansoureh Samadi, Nicolas Kelhofer, Tony Tu, Yih-Kuen Jan","doi":"10.1007/s11517-025-03350-w","DOIUrl":"https://doi.org/10.1007/s11517-025-03350-w","url":null,"abstract":"<p><p>Handgrip exercise is a common rehabilitation intervention but the effects of contraction modes (isometric and dynamic contractions) on modulating intermuscular coordination have not been investigated. Furthermore, coherence has been assessed at the <math><mi>α</mi></math> , <math><mi>β</mi></math> , <math><mi>γ</mi></math> bands of surface electromyography (EMG) and ignores the frequency over 60 Hz. The objective of this study was to assess the effects of muscle contraction modes on intermuscular coordination using wavelet-based EMG-EMG coherence. Sixteen healthy adults performed isometric handgrip (IHG) at 30% of maximal voluntary contraction (MVC) for 90 s and dynamic handgrip (DHG) at 30% MVC at a cadence of 1 Hz for 180 s in a random order with a 20-min rest. Two exercise modes have a similar exercise volume. EMG signals were recorded from the flexor digitorum superficialis (FDS), extensor carpi radialis (ECR), flexor carpi ulnaris (FCU), and extensor carpi ulnaris (ECU) and wavelet coherence of muscle pairs of 4 muscles was computed in seven frequency bands, including 2-5, 8-12, 15-35, 35-60, 60-100, 100-200, and 200-300 Hz. The results showed that IHG and DHG evoked different changes in muscle activation and intermuscular coordination. IHG evoked lower muscle activation compared to DHG. DHG resulted in a significant muscle activation and coherence in 2-5 and 8-12 Hz. IHG resulted in a significant muscle activation and coherence in 15-35, 35-60, and 60-100 Hz compared to DHG. The findings of this study indicate that the mode of handgrip exercise affects the muscle activation and intermuscular coordination between 2 and 300 Hz of EMG signals.</p>","PeriodicalId":49840,"journal":{"name":"Medical & Biological Engineering & Computing","volume":" ","pages":""},"PeriodicalIF":2.6,"publicationDate":"2025-03-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143712010","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Ravinder Kumar, Vishal Kumar, Collin Rich, David Lemmerhirt, Balendra, J Brian Fowlkes, Ashish Kumar Sahani
{"title":"Correction to: Machine learning models based on FEM simulation of hoop mode vibrations to enable ultrasonic cuffless measurement of blood pressure.","authors":"Ravinder Kumar, Vishal Kumar, Collin Rich, David Lemmerhirt, Balendra, J Brian Fowlkes, Ashish Kumar Sahani","doi":"10.1007/s11517-025-03351-9","DOIUrl":"https://doi.org/10.1007/s11517-025-03351-9","url":null,"abstract":"","PeriodicalId":49840,"journal":{"name":"Medical & Biological Engineering & Computing","volume":" ","pages":""},"PeriodicalIF":2.6,"publicationDate":"2025-03-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143712005","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}