Computer methods and programs in biomedicine最新文献

{"title":"RPF-Net: A multimodal model for the postoperative UISS risk stratification of non-metastatic ccRCC based on CT and whole-slide images","authors":"Jiahui Mao ,&nbsp;Wuchao Li ,&nbsp;Xinhuan Sun ,&nbsp;Bangkang Fu ,&nbsp;Junjie He ,&nbsp;Chongzhe Yan ,&nbsp;Jianguo Zhu ,&nbsp;Zhuxue Zhang ,&nbsp;Jiahui Mao ,&nbsp;Zhangxin Hong ,&nbsp;Qi Tang ,&nbsp;Zhen Liu ,&nbsp;Pinhao Li ,&nbsp;Yan Zhang ,&nbsp;Rongpin Wang","doi":"10.1016/j.cmpb.2025.108836","DOIUrl":"10.1016/j.cmpb.2025.108836","url":null,"abstract":"<div><h3>Background and objectives</h3><div>Postoperative non-metastatic clear cell renal cell carcinoma (nccRCC) patients face the risk of tumor recurrence and metastasis. However, prognosis assessment for nccRCC remains time-consuming and subjective. In the current diagnostic landscape, computed tomography (CT) images provide macro-scale anatomical information, and whole-slide images (WSIs) offer micro-scale details that are inaccessible to CT imaging. To address this gap, the study proposes a multimodal approach that leverages both CT and WSI data to develop an automated model for postoperative risk stratification in nccRCC.</div></div><div><h3>Methods</h3><div>This study proposes a multimodal model named the Radiology-Pathology Fusion Network (RPF-Net), which employs self-attention, graph-attention, and dynamic attention fusion mechanisms to integrate CT images and WSIs for classifying nccRCC patients into low-risk and intermediate-high-risk groups per the University of California, Los Angeles, Integrated Staging System (UISS) criteria. The proposed model is divided into three steps. First, the ResNet-50 and 3D ResNet-50 are used as feature extractors to respectively extract representative feature maps from WSIs and CT images. Second, a dual-branch module is designed to extract global and local features of the WSIs. Finally, a multilayer dynamic attention fusion (MDAF) module is developed to facilitate cross-modal feature interaction and predict the risk stratification results.</div></div><div><h3>Results</h3><div>The area under the curve (AUC), accuracy, precision, and F1 Score of the RPF-Net on the internal validation set are 0.949±0.013, 0.894±0.019, 0.895±0.020, and 0.894±0.019, respectively. Furthermore, the RPF-Net shows robust generalization, achieving an AUC of 0.901 on the external validation set and 0.924 on the public dataset.</div></div><div><h3>Conclusions</h3><div>The RPF-Net models the diagnostic process of multimodal data and shows strong generalization and excellent performance. This model may be a potential tool to facilitate clinical risk stratification and management for postoperative nccRCC patients.</div></div>","PeriodicalId":10624,"journal":{"name":"Computer methods and programs in biomedicine","volume":"268 ","pages":"Article 108836"},"PeriodicalIF":4.9,"publicationDate":"2025-05-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144072207","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Multi-task learning for joint prediction of breast cancer histological indicators in dynamic contrast-enhanced magnetic resonance imaging","authors":"Rong Sun ,&nbsp;Xiujuan Li ,&nbsp;Baosan Han ,&nbsp;Yuanzhong Xie ,&nbsp;Shengdong Nie","doi":"10.1016/j.cmpb.2025.108830","DOIUrl":"10.1016/j.cmpb.2025.108830","url":null,"abstract":"<div><h3>Objectives</h3><div>Achieving efficient analysis of multiple pathological indicators has great significance for breast cancer prognosis and therapeutic decision-making. In this study, we aim to explore a deep multi-task learning (MTL) framework for collaborative prediction of histological grade and proliferation marker (Ki-67) status in breast cancer using multi-phase dynamic contrast-enhanced magnetic resonance imaging (DCE-MRI).</div></div><div><h3>Methods</h3><div>In the novel design of hybrid multi-task architecture (HMT-Net), co-representative features are explicitly distilled using a feature extraction backbone. A customized prediction network is then introduced to perform soft-parameter sharing between two correlated tasks. Specifically, task-common and task-specific knowledge is transmitted into tower layers for informative interactions. Furthermore, low-level feature maps containing tumor edges and texture details are recaptured by a hard-parameter sharing branch, which are then incorporated into the tower layer for each subtask. Finally, the probabilities of two histological indicators, predicted in the multi-phase DCE-MRI, are separately fused using a decision-level fusion strategy.</div></div><div><h3>Results</h3><div>Experimental results demonstrate that the proposed HMT-Net achieves optimal discriminative performance over other recent MTL architectures and deep models based on single image series, with the area under the receiver operating characteristic curve of 0.908 for tumor grade and 0.694 for Ki-67 status.</div></div><div><h3>Conclusions</h3><div>Benefiting from the innovative HMT-Net, our proposed method elucidates its strong robustness and flexibility in the collaborative prediction task of breast biomarkers. Multi-phase DCE-MRI is expected to contribute valuable dynamic information for breast cancer pathological assessment in a non-invasive manner.</div></div>","PeriodicalId":10624,"journal":{"name":"Computer methods and programs in biomedicine","volume":"267 ","pages":"Article 108830"},"PeriodicalIF":4.9,"publicationDate":"2025-05-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143912345","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Fluid-structure interaction analysis of bioinspired polymeric heart valves with experimental validation","authors":"Xinying Liu ,&nbsp;Aeryne Lee ,&nbsp;Yiqi Wang ,&nbsp;Thanh Phuong Hoang ,&nbsp;Karinna Shay Yee ,&nbsp;Luke Mosse ,&nbsp;Nils Karajan ,&nbsp;David S. Winlaw ,&nbsp;Sina Naficy ,&nbsp;David F. Fletcher","doi":"10.1016/j.cmpb.2025.108839","DOIUrl":"10.1016/j.cmpb.2025.108839","url":null,"abstract":"<div><h3>Background and objectives</h3><div>Valvular heart disease, when not addressed adequately, can result in heart failure, serious heart-related health problems, and in some cases, death. Polymeric heart valves (PHVs) are promising valve replacement technologies that may offer improved durability and better biological performance. Notably, PHVs have the potential to accommodate highly innovative valve designs. Given this feature of PHVs, it is important to shortlist the best performing valve designs prior to committing to extensive <em>in vitro</em> hemodynamic validation prototypes.</div></div><div><h3>Methods</h3><div>This study presents a computational fluid-structure interaction (FSI) workflow, which integrates computational fluid dynamics (CFD) and finite element analysis (FEA), to simulate the hemodynamic performance of PHVs with two different valve designs under physiological conditions.</div></div><div><h3>Results</h3><div>The model accurately predicts cardiac output (CO), effective orifice area (EOA) and regurgitant fraction (RF) and these predictions have been successfully validated using experimental data. Consistent with experimental findings, increasing valve thickness results in a decrease in EOA, with RF trends varying between different valve designs. The fully opened and unfolded valve exhibited the lowest WSS on the leaflet surfaces. Both valve design and thickness significantly influence stress distribution along the leaflets with the thinnest valves showing lower von Mises stresses during opening and higher stresses during closing. Detailed analysis of flow patterns, wall shear stress (WSS), valve opening and closing behaviors, and mechanical stress distribution are presented.</div></div><div><h3>Conclusions</h3><div>This work demonstrates the potential of FSI simulations in predicting the hydrodynamic and mechanical behavior of PHVs, offering valuable insights into valve durability and design optimization for improved patient outcomes. This approach can significantly accelerate valve development by reducing reliance on extensive <em>in vitro</em> and <em>in vivo</em> testing.</div></div>","PeriodicalId":10624,"journal":{"name":"Computer methods and programs in biomedicine","volume":"268 ","pages":"Article 108839"},"PeriodicalIF":4.9,"publicationDate":"2025-05-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143929080","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"A systematic review of AI as a digital twin for prostate cancer care","authors":"Annette John ,&nbsp;Reda Alhajj ,&nbsp;Jon Rokne","doi":"10.1016/j.cmpb.2025.108804","DOIUrl":"10.1016/j.cmpb.2025.108804","url":null,"abstract":"<div><div>Artificial Intelligence (AI) and Digital Twin (DT) technologies are rapidly transforming healthcare, offering the potential for personalized, accurate, and efficient medical care. This systematic review focuses on the intersection of AI-based digital twins and their applications in prostate cancer pathology. A digital twin, when applied to healthcare, creates a dynamic, data-driven virtual model that simulates a patient’s biological systems in real-time. By incorporating AI techniques such as Machine Learning (ML) and Deep Learning (DL), these systems enhance predictive accuracy, enable early diagnosis, and facilitate individualized treatment strategies for prostate cancer.</div><div>This review systematically examines recent advances (2020-2025) in AI-driven digital twins for prostate cancer, highlighting key methodologies, algorithms, and data integration strategies. The literature analysis also reveals substantial progress in image processing, predictive modeling, and clinical decision support systems, which are the basic tools used when implementing digital twins for prostate cancer care.</div><div>Our survey also critically evaluates the strengths and limitations of current approaches, identifying gaps such as the need for real-time data integration, improved explainability in AI models, and more robust clinical validation. It concludes with a discussion of future research directions, emphasizing the importance of integrating multi-modal data with Large Language Models (LLMs) and Vision-Language Models (VLMs), scalability, and ethical considerations in advancing AI-driven digital twins for prostate cancer diagnosis and treatment.</div><div>This paper provides a comprehensive resource for researchers and clinicians, offering insights into how AI-based digital twins can enhance precision medicine and improve patient outcomes in prostate cancer care.</div></div>","PeriodicalId":10624,"journal":{"name":"Computer methods and programs in biomedicine","volume":"268 ","pages":"Article 108804"},"PeriodicalIF":4.9,"publicationDate":"2025-05-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143921593","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Bayesian adaptive enrichment design in multi-arm clinical trials: The BayesAET package for R users","authors":"Denghuang Zhan ,&nbsp;Yongdong Ouyang ,&nbsp;Fidel Vila-Rodriguez ,&nbsp;Mohammad Ehsanul Karim ,&nbsp;Hubert Wong","doi":"10.1016/j.cmpb.2025.108833","DOIUrl":"10.1016/j.cmpb.2025.108833","url":null,"abstract":"<div><h3>Background</h3><div>Randomized controlled trials seldom assess treatment effect heterogeneity across subpopulations, potentially leading to suboptimal treatment recommendations and inefficient use of healthcare resources. Adaptive enrichment designs seek to identify patient subpopulations most likely to benefit from the treatment. This manuscript introduces BayesAET, an R package developed to support Bayesian adaptive enrichment trial designs. The package helps identify optimal treatments for pre-specified subpopulations within a broader patient population, improving the efficiency and relevant inference of clinical trials.</div></div><div><h3>Methods</h3><div>BayesAET integrates Bayesian multi-arm multi-stage designs with adaptive enrichment strategies. It allows for the incorporation of historical data through Bayesian priors, supports adaptive randomization and interim analyses. These features facilitate flexible but robust modifications to trial parameters based on accumulated data, including early stopping, dropping ineffective treatments, and adjusting randomization probabilities. The package supports various outcome types, including continuous, binary, and count outcomes.</div></div><div><h3>Results</h3><div>We showcase BayesAET through a case study of a trial evaluating repetitive transcranial magnetic stimulation for depression and anxiety. The trial involved three treatment protocols and two subpopulations (with and without benzodiazepine use). Simulations demonstrate that BayesAET effectively identifies differential treatment effects, adapts trial parameters based on interim data, and improves precision in treatment effect estimation.</div></div><div><h3>Conclusion</h3><div>BayesAET provides a comprehensive tool for designing and analyzing Bayesian adaptive enrichment trials to identify the optimal treatments with pre-specified subpopulations.</div></div>","PeriodicalId":10624,"journal":{"name":"Computer methods and programs in biomedicine","volume":"268 ","pages":"Article 108833"},"PeriodicalIF":4.9,"publicationDate":"2025-05-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143936218","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Development of patient-specific apparent blood viscosity predictive models for computational fluid dynamics analysis of intracranial aneurysms with machine learning approaches","authors":"Takashi Suzuki ,&nbsp;Hiroyuki Takao ,&nbsp;Tomoaki Suzuki ,&nbsp;Soichiro Fujimura ,&nbsp;Shunsuke Hataoka ,&nbsp;Tomonobu Kodama ,&nbsp;Ken Aoki ,&nbsp;Toshihiro Ishibashi ,&nbsp;Makoto Yamamoto ,&nbsp;Hideki Yamamoto ,&nbsp;Yuichi Murayama","doi":"10.1016/j.cmpb.2025.108831","DOIUrl":"10.1016/j.cmpb.2025.108831","url":null,"abstract":"<div><h3>Background and Objectives</h3><div>A model to predict patient-specific apparent viscosity as a computational condition in computational fluid dynamics (CFD) analysis, which is used in research on intracranial aneurysms, is important. The purpose of this study was to develop a model to predict patient-specific apparent viscosity from clinical blood test results.</div></div><div><h3>Methods</h3><div>The data were from 15 patients with intracranial aneurysms in whom blood viscosity and density were measured and blood tests were performed on the same day. The dataset was divided into two, a training dataset and a test dataset at a ratio of 4:1. The training dataset was used in constructing regression models with shear rate and 12 blood test items (the flexible model) or hematocrit (the simple model) as input, and the measured apparent viscosity as output. CFD analysis was implemented with and without coil geometries, and the viscosity models were evaluated.</div></div><div><h3>Results</h3><div>The root mean squared error (RMSE) of viscosity predicted with the flexible model and the simple model was 0.136 mPa·s and 0.226 mPa·s, respectively. The RMSE of time-averaged and space-averaged velocity and time-averaged and space-averaged wall shear stress computed in CFD analysis were &lt;0.01 m/s and &lt;0.21 Pa, respectively.</div></div><div><h3>Conclusions</h3><div>Regression models to predict patient-specific apparent blood viscosity from shear rate and blood test items were constructed with machine learning. There is a possibility that, using this predictive model, patient-specific blood apparent viscosity can be predicted with high accuracy from the blood test results of individual patients.</div></div>","PeriodicalId":10624,"journal":{"name":"Computer methods and programs in biomedicine","volume":"268 ","pages":"Article 108831"},"PeriodicalIF":4.9,"publicationDate":"2025-05-04","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143936219","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Automated cohort database system for cardiopulmonary physiotherapy: A comprehensive tool supporting research on cardiac surgery patients—framework design, development and validation","authors":"Xiang Liu ,&nbsp;Yuqiang Wang ,&nbsp;Zeruxin Luo ,&nbsp;Tingting Xi ,&nbsp;Wei Huang ,&nbsp;Xiu Zhang ,&nbsp;Tingqian Cao ,&nbsp;Pengming Yu ,&nbsp;Yingqiang Guo","doi":"10.1016/j.cmpb.2025.108825","DOIUrl":"10.1016/j.cmpb.2025.108825","url":null,"abstract":"<div><h3>Background</h3><div>The shared goal of clinical physicians and cardiopulmonary physiotherapists is to tailor optimal comprehensive rehabilitation strategies for each patient undergoing cardiac surgery to improve outcomes. The sustained and large-scale acquisition of patient course and rehabilitation treatment-related data faces numerous challenges. This necessitates research and analysis based on large sample size data from cardiac surgery patients.</div></div><div><h3>Objective</h3><div>The Cardiopulmonary Physiotherapists Database System (CPPTherapists-DBS) was developed to enhance the management and analysis of data for researching risk factors associated with postoperative pulmonary complications (PPCs) in cardiac surgery patients. This system aims to establish comprehensive system design standards, frameworks, and validation procedures to support research in both cardiac surgery and cardiopulmonary physiotherapy.</div></div><div><h3>Methods</h3><div>The development of the CPPTherapists-DBS involved: (1) establishing system design standards and frameworks through a detailed software engineering requirements analysis, where clinical researchers defined data collection standards, business scope, and identification rules based on international guidelines and previous research; (2) designing and developing the system to integrate advanced functionalities for data management and analysis within the established frameworks; (3) validating the system by constructing a retrospective cohort for PPCs and developing and evaluating a predictive model based on the collected data.</div></div><div><h3>Results</h3><div>The CPPTherapists-DBS successfully established design standards and frameworks for system development. It has collected clinical data from 27,027 cardiac surgery patients across multiple medical centers from 2010 to 2021. Since January 2022, it has also included physical rehabilitation treatment records for 5,335 patients. The system’s CCVPRA tool provides advanced visualization capabilities, enabling rapid data modeling for 6,608 patients and development of a predictive model for PPCs. The model demonstrated strong performance with an AUC of 0.78 in the training set and 0.76 in the testing set.</div></div><div><h3>Conclusions</h3><div>The CPPTherapists-DBS effectively automates the collection and management of clinical and rehabilitation data, adhering to established system design standards and frameworks. It offers powerful tools for data visualization and modeling, representing a significant advancement in cohort database systems and providing a replicable model for supporting research on cardiac surgery patients and cardiopulmonary physiotherapy.</div></div>","PeriodicalId":10624,"journal":{"name":"Computer methods and programs in biomedicine","volume":"268 ","pages":"Article 108825"},"PeriodicalIF":4.9,"publicationDate":"2025-05-02","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143929081","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Multi-objective optimization framework to plan laser ablation procedure for prostate tumors through a genetic algorithm","authors":"Gabriele Adabbo ,&nbsp;Assunta Andreozzi ,&nbsp;Marcello Iasiello ,&nbsp;Giovanni Napoli ,&nbsp;Giuseppe Peter Vanoli","doi":"10.1016/j.cmpb.2025.108827","DOIUrl":"10.1016/j.cmpb.2025.108827","url":null,"abstract":"&lt;div&gt;&lt;h3&gt;Background and objectives&lt;/h3&gt;&lt;div&gt;Prostate cancer is the most common form of cancer in the male population. While the survival rate is high, many patients undergo surgical procedures for prostate cancer that might never progress to clinical significance. As a result, minimally invasive therapies are increasingly preferred over chemotherapy, radiotherapy, or surgical interventions. Laser-induced hyperthermia is emerging as a promising minimally invasive technique that targets tumoral tissue without damaging the surrounding healthy prostate. However, the lack of a standardized protocol makes the procedure highly dependent on the surgeon's expertise. Indeed, besides the cancerous tissue, also the healthy one could be heated and undergo a necrosis. Consequently, two contrasting objectives have to be considered during the treatment design: to treat cancer without damaging healthy tissue. Therefore, in this work, a thorough multi-objective optimization is carried out with reference to the laser-induced thermal ablation framework for prostate tumors. This is achieved by coupling finite element simulations with a genetic algorithm-based optimization to identify the best settings for the procedure.&lt;/div&gt;&lt;/div&gt;&lt;div&gt;&lt;h3&gt;Methods&lt;/h3&gt;&lt;div&gt;A multi-objective optimization was conducted to determine the optimal settings for decision variables to achieve the best outcomes. The procedure was executed by the direct coupling between the genetic algorithm which continuously updated the decision variables, for the optimization, and a finite element commercial code to predict variables. The decision variables employed as input for the model were: procedure time, number of laser probes, their position, dimensions, delivered power, and the number of ON/OFF cycles. Pennes’ bioheat equation was employed to obtain the desired objective functions, say thermal damage in the tumor tissue and healthy prostate, to be maximized and minimized, respectively. Additionally, linear regression and Bayesian artificial neural networks were developed to correlate the design variables with the objective functions, providing a tool for optimizing treatment planning.&lt;/div&gt;&lt;/div&gt;&lt;div&gt;&lt;h3&gt;Results&lt;/h3&gt;&lt;div&gt;Results demonstrate that the multi-objective genetic algorithm is a powerful tool for selecting the optimal settings for treatment. By applying the utopian criterion, the best trade off is achieved, since the optimal solution is the one allowing for a complete tumor necrosis with an acceptable damage rate to the healthy prostate (188 &lt;em&gt;mm&lt;/em&gt;&lt;sup&gt;3&lt;/sup&gt;). Linear regressions proved ineffective for predicting the objective functions, while artificial neural networks yielded better results.&lt;/div&gt;&lt;/div&gt;&lt;div&gt;&lt;h3&gt;Conclusions&lt;/h3&gt;&lt;div&gt;This study introduces an effective methodology for optimizing laser-induced thermal ablation for prostate tumors. By coupling genetic algorithms with finite element simulations, a set of optimal protocols (in terms of time and laser settings) can be select","PeriodicalId":10624,"journal":{"name":"Computer methods and programs in biomedicine","volume":"267 ","pages":"Article 108827"},"PeriodicalIF":4.9,"publicationDate":"2025-05-02","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143924162","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Developing a multi-feature fusion model for exacerbation classification in asthma and COPD","authors":"Yuyang Yan ,&nbsp;Loes van Bemmel ,&nbsp;Frits M.E. Franssen ,&nbsp;Sami O. Simons ,&nbsp;Visara Urovi","doi":"10.1016/j.cmpb.2025.108796","DOIUrl":"10.1016/j.cmpb.2025.108796","url":null,"abstract":"<div><h3>Background and Objective:</h3><div>Deteriorations in respiratory health, also known as exacerbations, are important events in the progression of chronic respiratory diseases such as Chronic Obstructive Pulmonary Disease (COPD) and asthma. Changes in vocal characteristics during episodes of respiratory distress suggest that voice analysis could be a valuable tool for monitoring exacerbations. This study aims to develop a remote monitoring method for automatically detecting exacerbations in COPD and asthma patients using only speech data.</div></div><div><h3>Methods:</h3><div>This study proposes a speech-based approach for remote monitoring of asthma and COPD exacerbations, leveraging optimized Mel-Frequency Cepstral Coefficients (MFCC) alongside multi-domain acoustic features. We demonstrate that the optimized MFCC outperforms state-of-the-art feature extraction techniques, while integrating complementary features from the time, frequency, energy, and spectral domains further enhances predictive accuracy. To ensure model transparency and facilitate clinical adoption, we employ SHapley Additive exPlanations (SHAP) to identify key speech biomarkers contributing to exacerbation detection.</div></div><div><h3>Results:</h3><div>Compared with the state-of-the-art methods, our method exhibits excellent classification performance with an accuracy of 0.892 and an AUC of 0.955 on the TACTICAS dataset. Moreover, the most salient features ranked by SHAP values are MFCC-related features and energy features, which explains the reason behind the improvement observed with feature fusion.</div></div><div><h3>Conclusions:</h3><div>Comprehensive experiments and comparisons with existing algorithms highlight the potential of speech-based monitoring for respiratory conditions in real-world settings. The proposed method outperforms state-of-the-art approaches, offering a promising avenue for exacerbation diagnosis and monitoring while potentially reducing the burden on both patients and healthcare providers.</div></div>","PeriodicalId":10624,"journal":{"name":"Computer methods and programs in biomedicine","volume":"268 ","pages":"Article 108796"},"PeriodicalIF":4.9,"publicationDate":"2025-04-30","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143921594","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Tracking of respiratory mechanics at multiple oscillation frequencies","authors":"Gergely Makan ,&nbsp;Joyce K.Y. Wu ,&nbsp;Chung-Wai Chow ,&nbsp;Yushu Zou ,&nbsp;Ronald J. Dandurand ,&nbsp;Dorottya Czövek ,&nbsp;Zoltán Gingl ,&nbsp;Zoltán Hantos","doi":"10.1016/j.cmpb.2025.108818","DOIUrl":"10.1016/j.cmpb.2025.108818","url":null,"abstract":"<div><h3>Background and Objective</h3><div>Intra-breath oscillometry (IBOsc) is an emerging approach to characterize dynamic changes in respiratory mechanical impedance (Zrs). IBOsc utilizes a small-amplitude sinusoidal signal superimposed on quiet breathing to track Zrs with sufficient temporal resolution to find specific time points, such as end-expiration (eE) and end-inspiration (eI). IBOsc has demonstrated superiority to conventional multifrequency oscillometry in detecting abnormal respiratory function and predicting future impairment in several clinical settings. The aim of the present study was to construct intra-breath Zrs spectra from multifrequency recordings to demonstrate how the Zrs spectrum and its measures change during breathing.</div></div><div><h3>Methods</h3><div>Conventional oscillometric recordings from groups of healthy subjects and patients with interstitial lung disease, asthma and chronic obstructive pulmonary disease (N=40 each group) were analyzed. Zrs was computed at each component of the multifrequency (5-37-Hz) signal to establish the Zrs spectra at eE and eI. This multi-frequency tracking method was validated on simulated Zrs data generated by a non-linear model of respiratory mechanics. The 2-way median test and Wilcoxon signed rank test were used to compare Zrs values and derived measures between groups and respiratory phases, respectively.</div></div><div><h3>Results</h3><div>Large intra-breath changes in Zrs were found in all subject groups. Most pairwise comparisons of Zrs measures (such as resistance, resonance frequency, reactance area and effective compliance) revealed significant (<em>P</em>&lt;0.05) or highly significant (<em>P</em>&lt;0.001) differences between groups at eE, which became more uniform at eI. Similarly, the changes between eE and eI were significant in most Zrs measures and subject groups, indicating the tidal improvement of lung mechanics in the obstructive patients.</div></div><div><h3>Conclusions</h3><div>Our results demonstrate that re-processing of archived datasets is feasible and can provide useful additional data to further characterize respiratory mechanical phenotypes. In particular, the estimation of Zrs spectra at zero respiratory flow minimizes the contribution of upper airway nonlinearities and thus improves the assessment of intrapulmonary dynamics. However, as this study points out, most current multifrequency signals are suboptimal for exploiting the potential of IBOsc due to low signal-to-noise ratio and interaction between adjacent frequency components.</div></div>","PeriodicalId":10624,"journal":{"name":"Computer methods and programs in biomedicine","volume":"267 ","pages":"Article 108818"},"PeriodicalIF":4.9,"publicationDate":"2025-04-30","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143907602","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}