Computers in biology and medicine最新文献

{"title":"Influence of variance reduction techniques on conventional radiotherapy simulations with TOPAS MC","authors":"André L.E. Fidelis ,&nbsp;Felipe M.L. de Souza ,&nbsp;Juliana de M. Nascimento ,&nbsp;Ruy S.R. Neto ,&nbsp;Luiz A.R. da Rosa ,&nbsp;Simone C. Cardoso","doi":"10.1016/j.compbiomed.2025.109797","DOIUrl":"10.1016/j.compbiomed.2025.109797","url":null,"abstract":"<div><h3>Purpose</h3><div>This study aims to apply Geometrical Particle Splitting (GPS) and Importance Sampling (Ant Colony Method) as variance reduction techniques (VRT) to improve computational efficiency in Monte Carlo calculations of the conventional radiotherapy beam quality index, TPR_20,10.</div></div><div><h3>Methods</h3><div>The TOPAS was used to simulate a setup to determine the PDD_20,10 ratio using a water phantom at a source-to-surface distance of 100 cm and a 10 × 10 cm² field. TPR_20,10 values were calculated for each simulation using the 6 MV Elekta Precise linear accelerator IAEA phase space as source. Control simulations without VRTs provided a baseline. The influence of the number of split planes was evaluated for both techniques. The effects of VRT on photons and electrons were evaluated, focusing on computational efficiency compared to simulations without VRT. A Z-test assessed bias by checking compatibility between simulated and experimental TPR_20,10 values.</div></div><div><h3>Results</h3><div>TPR_20,10 results were validated for 3 σ compatibility with experimental data. The GPS technique's best results showed an efficiency gain factor of 4.01. Applying the technique to only photons or both photons and electrons did not yield significant differences. In Importance Sampling, the best results achieved a gain factor of 16.91 for both photons and electrons, but only yielded 0.75 for electrons alone.</div></div><div><h3>Conclusions</h3><div>VRTs significantly improved computational efficiency, though their effectiveness depends on geometry and particle selection. When applied carefully, these techniques enhance precision without greatly increasing processing time.</div></div>","PeriodicalId":10578,"journal":{"name":"Computers in biology and medicine","volume":"187 ","pages":"Article 109797"},"PeriodicalIF":7.0,"publicationDate":"2025-02-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143317325","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":"Harnessing network pharmacology and in silico drug discovery to uncover new targets and therapeutics for Alzheimer's disease","authors":"Haitham Al Madhagi ,&nbsp;Husam Nassar","doi":"10.1016/j.compbiomed.2025.109781","DOIUrl":"10.1016/j.compbiomed.2025.109781","url":null,"abstract":"<div><div>Alzheimer's disease (AD) is the leading cause of progressive neurodegenerative dementia, affecting approximately 50 million individuals globally. Recent studies have highlighted the differential expression of circular RNAs (circRNAs) in AD, which may disrupt the circRNA-miRNA-mRNA regulatory networks in neuronal cells. This work aims to integrate network pharmacology with in silico drug design to identify novel druggable targets for AD and propose promising drug candidates. We analyzed two circRNA datasets from the Gene Expression Omnibus, employing enrichment analysis and constructing a circRNA-miRNA-mRNA network. The RNAenrich platform facilitated the identification of hub genes and potential druggable targets. The identified target was subjected to virtual screening against a chemical drug library comprising over 6000 compounds in clinical trials while ensuring compliance with Lipinski's Rule of Five. Our findings reveal that differentially expressed circRNAs are significantly involved in gland development, apoptosis regulation, hypoxic response, and neuronal death. Notably, CDK-6 emerged as the most promising druggable target, exhibiting strong binding affinity with five selected ligands: DB06963, DB06888, DB07020, DB08683, and DB06976. These ligands demonstrated distinct binding modes and stable interactions over 500 ns of molecular dynamics simulations conducted via Desmond. In conclusion, our study identifies CDK-6 as a viable target for therapeutic intervention in Alzheimer's disease. The top five ligands present a compelling case for further investigation as innovative CDK-6 inhibitors and potential drug candidates for AD treatment.</div></div>","PeriodicalId":10578,"journal":{"name":"Computers in biology and medicine","volume":"187 ","pages":"Article 109781"},"PeriodicalIF":7.0,"publicationDate":"2025-02-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143180256","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":"Autoregressive exogenous neural structures for synthetic datasets of olive disease control model with fractional Grünwald-Letnikov solver","authors":"Nabeela Anwar ,&nbsp;Muhammad Asif Zahoor Raja ,&nbsp;Adiqa Kausar Kiani ,&nbsp;Iftikhar Ahmad ,&nbsp;Muhammad Shoaib","doi":"10.1016/j.compbiomed.2025.109707","DOIUrl":"10.1016/j.compbiomed.2025.109707","url":null,"abstract":"<div><div>A fundamental element of the Mediterranean diet, olive oil is abundant in heart-healthy monounsaturated fats and antioxidants, lowering the risk of cardiovascular diseases. However, the olive oil industry confronts hurdles arising from olive tree diseases, despite the numerous health advantages associated with its consumption. In pursuit of research goals, this study endeavors to employ cutting-edge intelligent computing paradigms, specifically nonlinear autoregressive exogenous neural networks utilizing the Levenberg-Marquardt scheme (NNLMS), to comprehensively analyze the complex dynamic interactions of the fractional-order olive disease control (FO-ODC) model. In the realm of nonlinear fractional differential modeling, this study explores a system governed by four distinct populations: the branches and leaves of healthy olive trees, olive trees affected by a detrimental fungus, a pathogenic filamentous fungus causing infection and damage to olive leaves, and branches, and the microbial organisms residing in the phyllosphere. The research aims to scrutinize the transmission patterns of olive disease within this complex ecological framework. Employing the fractional Grünwald-Letnikov backward finite difference method, this study undertakes the generation of a synthetic dataset that accurately illustrates variations in several key parameters, including the rate of healthy leaf production, natural mortality rate, growth rate of beneficial fungi, nutrient acquisition rate by pathogens from infected leaves, the scaling factor governing food acquisition in their mutualistic relationship, and the rate at which leaves are adversely affected or degrade due to the influence of harmful fungi. In each iteration of the NNLMS application, the synthetic dataset is arbitrarily segmented into training, testing, and validation samples, facilitating the computation of an approximate solution for the dynamics embedded in the nonlinear FO-ODC model. The viability of the design approach is evaluated/assessed by consistently matching outcomes with reference solutions through numerous variations of the FO-ODC model. The reliability and efficiency of the design approach are measured using various measures, such as regression analysis, absolute errors, mean errors, autocorrelations and error histograms.</div></div>","PeriodicalId":10578,"journal":{"name":"Computers in biology and medicine","volume":"187 ","pages":"Article 109707"},"PeriodicalIF":7.0,"publicationDate":"2025-02-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143317352","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":"Dynamic blinking feature extraction for automated facial nerve paralysis detection","authors":"Akara Supratak ,&nbsp;Watsaporn Pornwatanacharoen ,&nbsp;Varit Rungbanapan ,&nbsp;Skonlawut Tasnaworanun ,&nbsp;Rachata Chopdamrongtham ,&nbsp;Thanapon Noraset ,&nbsp;Manachaya Prukajorn ,&nbsp;Pimkwan Jaru-ampornpan","doi":"10.1016/j.compbiomed.2025.109722","DOIUrl":"10.1016/j.compbiomed.2025.109722","url":null,"abstract":"<div><div>Facial nerve paralysis (FNP) impair eyelid closure and blinking, risking ophthalmic complications and vision loss. Current detection methods primarily rely on static facial asymmetries, overlooking the dynamic eyelid movements during blinking that are important for evaluating treatment outcomes such as blink restoration. In this study, we present an automated system for objectively extracting dynamic blink features from high-frame-rate videos to address these limitations. We develop algorithms for dynamic blink feature extraction using a facial landmark detection model to capture eyelid movements and derive parameters for each blink. These parameters are processed with an Isolation Forest model to learn the typical distribution of combined parameters from both eyes, generating normality scores for each blink pair to indicate the degree of abnormality in upper eyelid movement while reducing noise from landmark detection and head movements. Our evaluation, which included 103 subjects (86 healthy and 17 with FNP), shows that the machine learning model trained to detect FNP using normality scores outperformed those trained with static parameters (with an increase of 75% in F1-score) and dynamic parameters (with an increase of 35% in F1-score). Notably, the normality score of the closing blink velocity, representing the speed at which the upper eyelid margin moves during the eye-closing phase, was the most distinguishing feature for FNP detection. These findings highlight the potential of the dynamic blink features in FNP detection and suggest further exploration to assess their effectiveness as objective measures for diagnosing FNP in addition to the facial asymmetry features proposed in other studies.</div></div>","PeriodicalId":10578,"journal":{"name":"Computers in biology and medicine","volume":"187 ","pages":"Article 109722"},"PeriodicalIF":7.0,"publicationDate":"2025-02-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143317355","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"REMED-T2D: A robust ensemble learning model for early detection of type 2 diabetes using healthcare dataset","authors":"Le Thi Phan ,&nbsp;Rajan Rakkiyappan ,&nbsp;Balachandran Manavalan","doi":"10.1016/j.compbiomed.2025.109771","DOIUrl":"10.1016/j.compbiomed.2025.109771","url":null,"abstract":"<div><div>Early diagnosis and timely treatment of diabetes are critical for effective disease management and the prevention of complications. Undiagnosed diabetes can lead to an increased risk of several health issues. Although numerous machine learning (ML) models have been designed to detect diabetes, many exhibit unsatisfactory performance, are not publicly available, and lack validation on external datasets. To address these limitations, we have developed REMED-T2D, an advanced ensemble ML approach that enhances predictive accuracy and robustness through the integration of diverse ML algorithms. Our approach involves a rigorous data preprocessing process and systematic evaluation of 20 different algorithms, encompassing both conventional ML and deep learning for diabetes prediction. Firstly, we applied an under-sampling approach to an imbalanced Pima Indian Diabetes dataset and generated five balanced datasets. Using these datasets, we investigated various computational strategies to select the optimal model for accurate diabetes classification. Our results demonstrate that REMED-T2D outperformed state-of-the-art methods on the training dataset, with notable improvements in ACC (1.40–4.60%) and MCC (3.50–9.80%). Extensive external validations revealed that the model trained on a five-feature subset achieved ACC of 92.61 % and 92.26 % on the RTML1 and Pabna datasets, respectively. Moreover, a model based on a seven-feature subset improved ACC by 2.80 % and MCC by 13.27 % on the RTML2 dataset. These results suggest the potential of REMED-T2D to predict diabetes in Asian females. Notably, this is the first study to conduct such a comprehensive analysis using the Pima dataset, incorporating a diverse set of ML algorithms. Furthermore, we have developed a publicly accessible web server (<span><span>https://balalab-skku.org/REMED-T2D/</span><svg><path></path></svg></span>) to facilitate self-monitoring and timely medical interventions. We believe REMED-T2D will assist healthcare professionals in detecting diabetes earlier and implementing preventive measures, ultimately improving health outcomes for those at risk.</div></div>","PeriodicalId":10578,"journal":{"name":"Computers in biology and medicine","volume":"187 ","pages":"Article 109771"},"PeriodicalIF":7.0,"publicationDate":"2025-02-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143317353","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":"Two-stage CNN-based framework for leukocytes classification","authors":"Siraj Khan ,&nbsp;Muhammad Sajjad ,&nbsp;José Escorcia-Gutierrez ,&nbsp;Sami Dhahbi ,&nbsp;Mohammad Hijji ,&nbsp;Khan Muhammad","doi":"10.1016/j.compbiomed.2024.109616","DOIUrl":"10.1016/j.compbiomed.2024.109616","url":null,"abstract":"<div><div>Leukocytes are pivotal markers in health, crucial for diagnosing diseases like malaria and viral infections. Peripheral blood smear tests provide pathologists with vital insights into various medical conditions. Manual leukocyte counting is challenging and error-prone due to their complex structure. Accurate segmentation and classification of leukocytes remain challenging, impacting both accuracy and efficiency in blood microscopic image analysis. To overcome these limitations, we propose a robust two-stage CNN framework that integrates YOLOv8 for precise segmentation and MobileNetV3 for effective classification. Initially, WBCs are segmented using YOLOv8m-seg, extracting ROIs for subsequent analysis. Then, features from segmented ROIs are used to train MobileNetV3, classifying WBCs into lymphocytes, monocytes, basophils, eosinophils, and neutrophils. This framework significantly advances leukocyte categorization, enhancing diagnostic performance and patient outcomes. The proposed technique achieved impressive accuracy rates of 99.56 %, 99.19 % and 98.89 % during segmentation and 99.28 %, 99.63 % and 98.49 % during classification on Raabin-WBC, PBC and LISC datasets, respectively, outperforming state-of-the-art methods.</div></div>","PeriodicalId":10578,"journal":{"name":"Computers in biology and medicine","volume":"187 ","pages":"Article 109616"},"PeriodicalIF":7.0,"publicationDate":"2025-02-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143180255","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":"You get the best of both worlds? Integrating deep learning and traditional machine learning for breast cancer risk prediction","authors":"João Mendes ,&nbsp;Bernardo Oliveira ,&nbsp;Carolina Araújo ,&nbsp;Joana Galrão ,&nbsp;Nuno C. Garcia ,&nbsp;Nuno Matela","doi":"10.1016/j.compbiomed.2025.109733","DOIUrl":"10.1016/j.compbiomed.2025.109733","url":null,"abstract":"<div><div>Breast Cancer is the most commonly diagnosed cancer worldwide. While screening mammography diminishes the burden of this disease, it has some flaws related to the presence of false negatives. Adapting screening to each woman’s needs could help overcome these challenges. While traditional risk models are valuable tools, we propose an image-based approach. Since artificial intelligence has proven effective in aiding the diagnosis of breast cancer, we aim to translate this technology to risk prediction.</div><div>A 3-year risk prediction model, with a case-control age-matched approach, was developed based on the analysis of “prior” healthy mammograms. Two classes were defined – “risk” and “control” – based on the assessment done on the most recent examination: if the case was diagnosed with cancer, the prior mammogram was assigned to the “risk” class; otherwise, the prior mammogram was allocated to the “normal” class. In total, we found 3720 available controls and 1471 risk cases. Every mammogram used in this study was taken 3 years before the assessment used for class definition.</div><div>Risk prediction was aimed through three methodologies: traditional machine learning, deep learning, and a combination of both. The AUCs obtained on the test set were 0.68 for the traditional machine learning, and 0.76 for the other two. No statistically significant differences were found among methods.</div><div>Our findings suggest that the use of image-based deep learning methods holds promise on the field of Breast Cancer risk prediction, with further validation being needed to confirm their clinical applicability.</div></div>","PeriodicalId":10578,"journal":{"name":"Computers in biology and medicine","volume":"187 ","pages":"Article 109733"},"PeriodicalIF":7.0,"publicationDate":"2025-02-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143317351","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Sam2Rad: A segmentation model for medical images with learnable prompts","authors":"Assefa Seyoum Wahd ,&nbsp;Banafshe Felfeliyan ,&nbsp;Yuyue Zhou ,&nbsp;Shrimanti Ghosh ,&nbsp;Adam McArthur ,&nbsp;Jiechen Zhang ,&nbsp;Jacob L. Jaremko ,&nbsp;Abhilash Hareendranathan","doi":"10.1016/j.compbiomed.2025.109725","DOIUrl":"10.1016/j.compbiomed.2025.109725","url":null,"abstract":"<div><div>The Segment Anything Model and its variants, such as MedSAM, have demonstrated potential for medical image segmentation. However, they heavily rely on high-quality manual prompts, which are both time-consuming and require domain expertise. Even when using manual prompts, including sparse prompts like boxes, points, or text, and dense prompts such as masks, SAM and its variants like MedSAM (fine-tuned on medical images) fail to segment bones in ultrasound images due to significant domain shift.</div><div>To address these limitations, we propose <strong>Sam2</strong> for <strong>Rad</strong>iology (Sam2Rad), a framework that extends SAM and its recent iteration SAM2 to segment bony regions in ultrasound images without requiring manual prompts. At the core of our approach is a Prompt Predictor Network (PPN) that uses a lightweight cross-attention mechanism to generate bounding box coordinates, mask prompts, and high-dimensional embeddings to be used as prompts. Specifically, PPN leverages hierarchical feature maps extracted from SAM’s image encoder as keys and values, and learnable object embeddings as queries. The output of the cross-attention is then used to predict the bounding box coordinates, mask prompts, and high-dimensional prompts. These predicted prompts are subsequently fed to SAM’s mask decoder to generate the final segmentation mask. By aligning the learned prompts with SAM’s original training scheme, PPN enhances compatibility with SAM’s architecture, requiring no additional standalone encoders.</div><div>To preserve SAM’s extensive world knowledge, we keep all SAM modules frozen and train PPN only. This approach enables efficient parameter utilization while retaining SAM’s generalization capabilities. Additionally, Sam2Rad can operate in three modes: fully autonomous without human supervision, semi-autonomous with <em>human-in-the-loop</em> for iterative refinement, and fully manual for tasks like data labeling.</div><div>We tested the proposed model – Sam2Rad on 3 musculoskeletal US datasets – wrist (3822 images), shoulder rotator cuff (1605 images), and hip (4849 images). Without Sam2Rad, all SAM2 variants failed to segment shoulder US in zero-shot generalization with bounding box prompts. Our model, Sam2Rad, improved the performance of all SAM base networks in all datasets, without requiring manual prompts. The improvement in dice score ranged from a 2.2%–5.8% for hip, 19.6%–32.8% for wrist wrist, and up to 51.3% improvement in Dice score (from 25.2% to 76.5% Sam2 large) on shoulder data. Notably, Sam2Rad could be trained with as few as 10 labeled images and it is compatible with any SAM architecture.</div><div>The code is available at <span><span>https://github.com/aswahd/SamRadiology</span><svg><path></path></svg></span>.</div></div>","PeriodicalId":10578,"journal":{"name":"Computers in biology and medicine","volume":"187 ","pages":"Article 109725"},"PeriodicalIF":7.0,"publicationDate":"2025-02-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143317354","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Unraveling the molecular landscape of non-small cell lung cancer: Integrating bioinformatics and statistical approaches to identify biomarkers and drug repurposing","authors":"Adiba Sultana ,&nbsp;Md Shahin Alam ,&nbsp;Alima Khanam ,&nbsp;Huiying Liang","doi":"10.1016/j.compbiomed.2025.109744","DOIUrl":"10.1016/j.compbiomed.2025.109744","url":null,"abstract":"<div><div>Non-small-cell lung cancer (NSCLC) is one of the most malignant tumors and the leading cause of death from cancer worldwide and is increasing at a massive rate every year. Most NSCLC patients are diagnosed at advanced stages, which is associated with a poor prognosis and a very low 5-year survival rate. Therefore, the purpose of this study is to investigate molecular markers for early diagnosis, prognosis and therapy of NSCLC through the integration of bioinformatics and statistical methods. A total of 93 overlapping differentially expressed genes (oDEGs) were identified between NSCLC and normal samples through Linear Models for Microarray (LIMMA) and Significance Analysis of Microarrays (SAM) methods. Six top-degree oDEGs (CCNA2, CDC6, AURKA, CCNB1, MKI67, and PRC1) were identified as key genes (KGs) through the protein-protein interaction (PPI) network. The predictive accuracy analysis of the identified KGs revealed an accuracy of 96.92 %, with a sensitivity of 91.73 % and a specificity of 98.15 %. KGs associated with 3 molecular functions (MFs), 5 cellular components (CCs), 3 biological processes (BPs), and 4 pathways were identified through FunRich software. Analysis of expression levels using the UALCAN database revealed that KGs are significantly associated with potential early diagnostic biomarkers. Survival analysis using the GEPIA database demonstrated that the KGs possessed strong prognostic power for NSCLC. Finally, seven repurposed candidate drugs ENTRECTINIB, SORAFENIB, CHEMBL1765740, TOZASERTIB, NERVIANO, AZD-1152-HQPA, and SELICICLIB were proposed through molecular docking analysis. In conclusion, the findings of this study have the potential to significantly impact the early diagnosis, prognosis, and treatment of NSCLC.</div></div>","PeriodicalId":10578,"journal":{"name":"Computers in biology and medicine","volume":"187 ","pages":"Article 109744"},"PeriodicalIF":7.0,"publicationDate":"2025-02-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143317356","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Finite element analysis of anatomically-modelled prosthetic incus for ossicular chain reconstruction","authors":"Masoud Mohseni-Dargah ,&nbsp;Christopher Pastras ,&nbsp;Payal Mukherjee ,&nbsp;Kai Cheng ,&nbsp;Khosro Khajeh ,&nbsp;Mohsen Asadnia","doi":"10.1016/j.compbiomed.2025.109770","DOIUrl":"10.1016/j.compbiomed.2025.109770","url":null,"abstract":"<div><div>Ossicular chain reconstruction (OCR) is the gold standard for repairing conductive hearing loss (CHL) using prosthetic devices. However, few techniques can validate prosthesis performance before translational implementation. This study not only validates a Finite Element (FE) model for examining (reconstructed) middle ear biomechanics, but also evaluates the potential of personalised prosthetic devices for OCR. Additionally, this research examines tricalcium phosphate (TCP), as a potential biomaterial, besides titanium and hydroxyapatite (HA), as commonly used biomaterials for use in (personalised) OCR. Here, the designed FE model tested the hypothesis that 3D anatomically modelled prosthetic devices (prosthetic incus) can reliably restore sound transmission using appropriate biomaterials, including titanium, HA, or TCP. Our FE modelling examined middle ear biomechanics before and after prosthesis replacement for each biomaterial assignment. Results demonstrated the FE model is in agreement with experimental vibrometry recordings of ossicular motion and earlier numerical simulations. Additionally, the anatomically modelled prosthetic incus closely mimicked normal middle ear biomechanics, revealing its potential for OCR. FE analysis revealed no significant differences between titanium, HA, and TCP prostheses functions, serving as first-order evidence for their support in OCR. This research establishes a FE-based framework for personalised OCR following imaging, which is valuable for future personalised treatments for patients with CHL due to ossicular dysfunction. FE simulations can evaluate the biomechanics and function of prostheses, helping the surgeon make well-informed decisions regarding OCR for translational outcomes.</div></div>","PeriodicalId":10578,"journal":{"name":"Computers in biology and medicine","volume":"187 ","pages":"Article 109770"},"PeriodicalIF":7.0,"publicationDate":"2025-02-04","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143181296","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}