Computer Methods And Programs In Biomedicine[JOURNAL]
Sun
200 papers
Actinic keratosis staging in multimodal image data.
Slian A, Korecka K, Polańska A … +1 more , Czajkowska J
Comput Methods Programs Biomed · 2026 Jul · PMID 41990405 · Publisher ↗
BACKGROUND AND OBJECTIVE: Actinic Keratosis (AK) is a common skin condition, usually appearing on sun-exposed areas, whose progression is associated with characteristic dermatoscopic and structural changes. Early detecti...
BACKGROUND AND OBJECTIVE: Actinic Keratosis (AK) is a common skin condition, usually appearing on sun-exposed areas, whose progression is associated with characteristic dermatoscopic and structural changes. Early detection of AK is crucial, as cancer progression may occur in changed skin. This study aimed to develop a multimodal, machine-learning-based framework combining dermatoscopic and high-frequency ultrasound (HFUS) data to automatically stage AK and identify early lesions. METHODS: A dataset containing 222 pairs of dermatoscopic and HFUS images was clinically evaluated using the 3-point Zalaudek scale. Dermatoscopic images underwent ROI selection, hair removal, and extensive feature extraction (color, erythema, pigmentation, vessels, scales, pixel intensities, GLCM/LBP texture). HFUS images were divided into entry echo, sub-epidermal low-echoic band (SLEB), and dermis using a deep neural network, and then features describing the morphology and structure of the skin for each layer were extracted. A pre-trained EfficientNet network was used for feature extraction. Logistic Regression, k-Nearest Neighbors, Random Forests, Support Vector Machines and Multilayer Perceptrons with Sequential Feature Selection using 5-fold patient-wise cross-validation were used for feature-based classification. Additionally, multimodal TwinCNN was evaluated, with various pre-trained models as feature extractors. RESULTS: Combining dermatoscopic and HFUS features consistently outperformed single-modality models. Depending on the defined task, the models achieved over 80% accuracy (healthy, AK1-AK3), 78% (AK1-AK3), and almost 90% in the case of early AK detection vs. healthy and advanced AK on multimodal features. The TwinCNN model performed worse than classical machine-learning approaches, likely due to the limited size of the dataset and class imbalance. CONCLUSIONS: A multimodal framework integrating dermatoscopic and HFUS imaging enables accurate AK classification, surpassing single-modality approaches. Future work should expand multicenter datasets, improve automation of pre-processing steps, and explore enhanced neural multimodal fusion architectures.
Connection density affects the behavior of functional brain network metrics.
Song X, Zhang S, Du C … +2 more , Chai L, Zhang J
Comput Methods Programs Biomed · 2026 Jul · PMID 41980296 · Publisher ↗
BACKGROUND AND OBJECTIVE: Many neuropsychiatric disorders are associated with alterations in functional brain network (FBN) connectivity. Functional network metrics reveal alterations in FBNs by quantifying changes in ne...
BACKGROUND AND OBJECTIVE: Many neuropsychiatric disorders are associated with alterations in functional brain network (FBN) connectivity. Functional network metrics reveal alterations in FBNs by quantifying changes in network segregation and integration. In general, calculating functional network metrics requires a reasonable thresholding or binarization of the network, which is a challenging task. We observe that the same functional network metric yields conflicting conclusions across different studies: some report higher values in patient groups compared to healthy groups, while others report lower values in patient groups. This paper postulates the hypothesis that arbitrary choices of connection density are responsible for the inconsistency of experimental results. METHODS: We investigate the behavior of 16 functional network metrics using three independent datasets that include patients with Alzheimer's disease (AD), mild cognitive impairment (MCI), and schizophrenia (SZ). Our analysis covers connection densities from 1% to 99% (in 1% intervals) and examines both binary and weighted networks in time and wavelet domains. RESULTS: Our results reveal a "reversal phenomenon" in many functional network metrics, where the difference between patient and healthy groups reverses as connection density increases. This provides a plausible explanation for the inconsistent conclusions reported in the literature. We further find that the metrics showing significant differences vary across analytical modes (domain and network type). Moreover, the significant metrics differ across diseases, reflecting disease-related heterogeneity. CONCLUSION: To avoid the "reversal phenomenon" and maximize the inter-group differences, we establish optimal connection density ranges for FBN analyses across various neuropsychiatric disorders, thereby improving the consistency and comparability of research results. In addition, we identify a set of metrics that demonstrate robustness across multiple datasets, providing a reliable reference for subsequent analyses. Our work sheds new light into the widespread use of functional network metrics and emphasizes that standardized connection density selection is crucial for achieving consistent results.
PRDM: Position-aware robust reconstruction with diffusion model for emotion recognition From EEG.
Xu W, Lin Y, Wang X … +2 more , Liu J, Ren Y
Comput Methods Programs Biomed · 2026 Jul · PMID 41980295 · Publisher ↗
BACKGROUND AND OBJECTIVE: Emotion recognition is significant in domains like smart education, mental health assistance, and brain-computer interface. Electroencephalography-based emotion recognition methods provide an ob...
BACKGROUND AND OBJECTIVE: Emotion recognition is significant in domains like smart education, mental health assistance, and brain-computer interface. Electroencephalography-based emotion recognition methods provide an objective approach to evaluating emotional states by analyzing the brain's electrophysiological activity. In this research, we aim to address the challenges of EEG data scarcity and noise that vary depending on the electrode position, both of which significantly impact the quality of emotion recognition. METHODS: We propose position-aware robust reconstruction with diffusion model (PRDM) for emotion recognition from EEG. PRDM integrates electrode positional encoding with features in each level and fuses various temporal scale features, reconstructing noise-independent features from the corrupted features to enrich the training sample. Additionally, we propose a spatial-temporal adaptive convolutional network that adaptively adjusts parameters based on EEG features for more precise classification. RESULTS: The developed framework is validated on DEAP and DREAMER datasets, demonstrating excellent performance in both valence and arousal classification tasks, achieving 95.01% and 95.62% accuracy on DEAP, and 84.77% and 87.30% accuracy on DREAMER, respectively. A comparative analysis of EEG topographic maps further confirms the model's effectiveness in reconstructing noise-independent EEG features. Ablation experiments further explored the contribution of each component to model performance, proving that our proposed method effectively enhances the performance of emotion recognition. CONCLUSION: The proposed PRDM mitigates the impact of data scarcity in EEG signals and inconsistent noise distribution across electrode positions, suggesting improvements in emotion recognition accuracy based on EEG signals.
Non-invasive classification of coronary perfusion pressure during CPR using smartphone-based skin video and deep learning.
Kwon S, Shin DA, Kim T … +6 more , Kim KS, Suh GJ, Sim J, Hur S, Park J, Lee JC
Comput Methods Programs Biomed · 2026 Jul · PMID 41967268 · Publisher ↗
BACKGROUND AND OBJECTIVE: Coronary perfusion pressure (CPP) is an important determinant of myocardial blood flow and an indicator during cardiopulmonary resuscitation (CPR). However, conventional CPP monitoring methods a...
BACKGROUND AND OBJECTIVE: Coronary perfusion pressure (CPP) is an important determinant of myocardial blood flow and an indicator during cardiopulmonary resuscitation (CPR). However, conventional CPP monitoring methods are invasive and unsuitable for out-of-hospital settings. This study proposes a non-invasive approach to classify CPP levels using skin video recorded with a smartphone camera and deep learning. METHODS: Video and biosignal data were collected simultaneously from 15 pigs during CPR. An integrated deep learning model that combines backbone model (CNN, EfficientNetV2-B0, ConvNeXt-Nano, FastViT-T8) with gated recurrent unit (GRU) was developed to classify whether CPP exceeded the clinically relevant threshold of 20 mmHg. Gradient-weighted Class Activation Mapping (Grad-CAM) was used to identify regions attended to by the model, and statistical analysis conducted. RESULTS: Among four different backbone models used for training and evaluation, the EfficientNetV2-B0-GRU architecture demonstrated the best performance, achieving accuracy of 84.60 % and F1-score of 76.04 %. Statistical analysis revealed significant differences in YCrCb channel values between correctly classified groups. In addition, active regions of Grad-CAM showed greater variances than inactive regions, indicating that the model focused on regions with higher color variabilities related to perfusion. CONCLUSION: A deep learning-based approach using skin video enables the non-invasive classification of CPP during CPR. Grad-CAM and quantitative YCrCb analyses improve interpretability, while the proposed method provides proof-of-concept for cost-effective and accessible perfusion-related assessment using skin video recorded with a smartphone, with potential future utility for real-time CPR quality assessment and decision-making in emergency or out-of-hospital settings.
Image pre-processing impact on generative model performance for Unsupervised Venous Malformation Segmentation.
Fraissenon A, Kugusheva A, Ladraa S … +4 more , Guibaud L, Canaud G, Clarysse P, Roux E
Comput Methods Programs Biomed · 2026 Jul · PMID 41966797 · Publisher ↗
BACKGROUND AND OBJECTIVE: Venous malformations (VM), commonly observed in PIK3CA-related overgrowth spectrum (PROS), are infiltrative and widely distributed lesions. Repositioning of targeted therapy has recently been pr...
BACKGROUND AND OBJECTIVE: Venous malformations (VM), commonly observed in PIK3CA-related overgrowth spectrum (PROS), are infiltrative and widely distributed lesions. Repositioning of targeted therapy has recently been proposed to treat this condition, requiring an accurate volumetric quantification of these vascular lesions for monitoring treatment efficacy and adjusting dosages accordingly. So far, these malformations have only been coarsely estimated using thresholding techniques on MRI acquisitions. However, such thresholding strategies are poorly reproducible and still require manual removal of all liquid physiological structures that are erroneously over-segmented. METHODS: In this work, we developed and compared several unsupervised approaches based on reconstruction error to generate a pre-segmentation mask of VMs on mice whole-body MRI scans. Investigated deep models were trained on whole-body MRI scans of healthy mice (n=36) and evaluated on a MRI test set of PIK3CA-mutated mice (n=5). The performance of the tested models - autoencoders, generative adversarial networks (GANs), and diffusion models (DDPM) - are compared in terms of F1-score (Dice), Precision and Recall. The impact of three pre-processings onto the results is also investigated. RESULTS: While baseline segmentation obtained with Otsu thresholding can reach high Dice by widely over-segmenting the MRI scans, both metrics variability and over-segmentation were improved using deep models revealing a better generalization ability. The best trade-off performance was obtained with DDPM model when using background removal pre-processing (Dice 0.50 ± 0.03) and the GAN trained on edge maps (Dice 0.47 ± 0.04). CONCLUSION: Our results demonstrate the importance of using edge maps as input to the GAN model, and the superiority of lesion masks obtained from the diffusion model for clinical applications.
BoneMesh: An open-source 3D slicer framework for automated mesh generation and material mapping from CT to finite elements.
Strack D, Chundi G, Rehtanz N … +5 more , Soltani Z, Pieper SD, Pinter C, Subburaj K, Alkalay RN
Comput Methods Programs Biomed · 2026 Jul · PMID 41966796 · Publisher ↗
BACKGROUND AND OBJECTIVE: Finite Element Analysis (FEA) is a powerful computational technique used to assess bone strength and fracture risk based on patient-specific anatomical data. However, the widespread clinical and...
BACKGROUND AND OBJECTIVE: Finite Element Analysis (FEA) is a powerful computational technique used to assess bone strength and fracture risk based on patient-specific anatomical data. However, the widespread clinical and research adoption of this technique is often hindered by the complex and labor-intensive preprocessing steps required to convert clinical imaging data into analysis-ready finite element models, including segmentation, mesh generation, material property mapping, and solver-specific file formatting. METHODS: To address this bottleneck, we developed BoneMesh, an automated meshing and material-mapping software module integrated into the open-source 3D Slicer platform. BoneMesh streamlines the generation of high-quality, patient-specific tetrahedral meshes directly from CT-based bone segmentations and efficiently assigns bone mineral density values derived from the imaging data. RESULTS: We verified the performance of BoneMesh against established commercial and freeware software (Abaqus, Simpleware and Bonemat) using CT-scans and segmentations from frequently analyzed human bones (Femur, Tibia, Lumbar vertebra, and Thoracic vertebra), demonstrating similar or superior performance in achieving targeted mesh quality parameters, including mesh edge length accuracy, element shape factors, and aspect ratios. CONCLUSION: BoneMesh significantly reduces manual intervention, preprocessing time, and the potential for human errors, thus offering an open-source, accessible, reliable, and efficient pipeline for advancing patient-specific finite element simulations in clinical research and potentially in routine clinical assessments.
Performance evaluation of quantum support vector machine for COVID-19 biomarker analysis.
Choi J, Yu C, Jung KL … +5 more , Foo SS, Chen W, Comhair SA, Jehi L, Jung JU
Comput Methods Programs Biomed · 2026 Jul · PMID 41966795 · Publisher ↗
BACKGROUND AND OBJECTIVE: Identifying key biomarkers from multi-omics data is essential for advancing COVID-19 diagnosis and understanding disease mechanisms. Quantum machine learning approaches, particularly the quantum...
BACKGROUND AND OBJECTIVE: Identifying key biomarkers from multi-omics data is essential for advancing COVID-19 diagnosis and understanding disease mechanisms. Quantum machine learning approaches, particularly the quantum support vector machine, offer potential for biomarker analysis. This study aimed to assess the applicability of the quantum support vector machine for biomarker evaluation under a performance-based feature importance framework. METHODS: Proteomic and metabolomic biomarker data from two independent cohorts (Cleveland Clinic and Swedish Medical Center) were analyzed. Biomarkers were ranked using ridge regression and grouped into higher- and lower-importance sets. These groups were used to compare classification performance between the classical support vector machine and the quantum support vector machine. Multiple quantum kernels were examined, including amplitude encoding, angle encoding, the ZZ feature map, and the projected quantum kernel. RESULTS: Across diverse experimental conditions, the quantum support vector machine achieved classification performance comparable to, and in some settings slightly higher than, that of the classical support vector machine. Moreover, the quantum support vector machine performance consistently aligned with biomarker importance rankings derived from ridge regression. CONCLUSIONS: Within a performance-based feature importance framework, these results highlight the quantum support vector machine as a promising approach for multi-omics data analysis in biomedical research. The findings suggest that the quantum support vector machine can support biomarker importance evaluation in complex diseases such as COVID-19 through model performance-driven analysis.
Multiscale insights into thrombus growth and detachment under non-physiological blood flow.
Xu Z, Shi Y, Wu X … +5 more , Wang S, He F, Hao P, Chen Z, Zhang X
Comput Methods Programs Biomed · 2026 Jul · PMID 41962341 · Publisher ↗
BACKGROUND AND OBJECTIVE: Flow dynamics play a fundamental role in modulating thrombus evolution, serving as a primary driver for mass transport, cell-protein interactions, and structural stability. While it is well-esta...
BACKGROUND AND OBJECTIVE: Flow dynamics play a fundamental role in modulating thrombus evolution, serving as a primary driver for mass transport, cell-protein interactions, and structural stability. While it is well-established that local flow patterns significantly influence thrombus growth and morphological changes, the precise biomechanical mechanisms linking varying flow conditions to the dynamic processes of accumulation and detachment remain to be fully elucidated. This study focuses on the intricate correlation between flow-mediated forces and thrombus stability, aiming to uncover how fluidic environments regulate the multiscale transition from cellular adhesion to macroscopic thrombus formation. METHOD: Based on dissipative particle dynamics and a coarse-grained cell model, this study establishes a mesoscopic-scale model for simulating platelet activation, adhesion, and fibrin formation. The proposed method enables high-resolution numerical simulation of thrombus growth, achieving multi-scale computations spanning protein-cell-thrombus levels. Ultimately, it allows for analysis and prediction of thrombus growth status, compositional changes, and detachment processes during thrombus development. RESULT: By combining microfluidic experiments and multiscale computational method, we systematically elucidated the dynamics of thrombus formation and detachment under non-physiological shear flow conditions. Our results indicate that flow intensity significantly modulates the cellular-to-fibrin ratio within thrombus. Through combined experimental and computational analyses, we identified two distinct thrombus detachment mechanisms: shear-driven boundary fragmentation detachment and pressure gradient-induced internal layer separation via thrombus fissuring. Diverging from traditional views that predominantly implicate fluid shear stress in thrombus detachment, our quantitative assessments reveal that momentum transfer from blood cell collisions is a pivotal factor in the detachment process. This insight highlights the interplay and competition between hydrodynamic and cellular kinetics in thrombus growth evolution.
A Parameter-free unsupervised framework for fMRI data analysis using batch learning growing neural gas and spatial-temporal false positive control.
Khani TH, Tajarrod AH, Shamsi M … +1 more , Zarei A
Comput Methods Programs Biomed · 2026 Jul · PMID 41955908 · Publisher ↗
BACKGROUND AND OBJECTIVE: Clustering methods are essential for analyzing functional magnetic resonance imaging (fMRI) time-series data to identify active brain regions and elucidate functional neural patterns. Despite re...
BACKGROUND AND OBJECTIVE: Clustering methods are essential for analyzing functional magnetic resonance imaging (fMRI) time-series data to identify active brain regions and elucidate functional neural patterns. Despite recent advancements in enhancing the stability and automation of cluster number selection, many methods still rely on user-defined parameters, which complicates their application. METHODS: To deal with these issues, this study introduces a novel parameter-free hierarchical topological structure learning clustering algorithm, Batch Learning Growing Neural Gas (BL-GNG), which builds on the Growing Neural Gas (GNG) model to improve convergence speed and eliminate the need for manual parameter tuning. Then, a two-stage false positive rate control mechanism, based on randomization inference and three-dimensional neighborhood criteria, further enhances the algorithm's robustness is proffered. RESULTS: The performance of BL-GNG was evaluated on real fMRI data targeting auditory cortex activity, using the General Linear Model in SPM with a Family-Wise Error-corrected threshold of p < 0.05 as the ground truth. Compared to K-means, Fuzzy C-means (FCM), Neural Gas (NG), and GNG. BL-GNG achieved a Jaccard Coefficient of 0.99 and an Area Under the ROC Curve of 0.97, demonstrating superior exactitude and stability across 50 iterative runs. With an average execution time of 26 s, the algorithm also offers significant computational efficiency. CONCLUSIONS: These results highlight BL-GNG's potential as a powerful tool for fMRI analysis, with applications in diagnosing brain disorders, investigating neural subnetworks, and advancing cognitive neuroscience research.
PAC-KENReg: An interpretable hypergraph framework capturing nonlinear and dynamic functional connectivity for neuropsychiatric disease diagnosis.
Yan H, Wang P, Lin X … +6 more , Wang Z, Wu Z, Yu X, Han J, Tong J, Zeng S
Comput Methods Programs Biomed · 2026 Jul · PMID 41955907 · Publisher ↗
BACKGROUND AND OBJECTIVE: Hypergraph-based analysis has become a pivotal method for modeling functional connectivity networks to elucidate neuropathological mechanisms in psychiatric disorders. However, existing approach...
BACKGROUND AND OBJECTIVE: Hypergraph-based analysis has become a pivotal method for modeling functional connectivity networks to elucidate neuropathological mechanisms in psychiatric disorders. However, existing approaches are constrained by two key limitations: their reliance on linear assumptions, which limits the capture of complex nonlinear interactions among brain regions, and their failure to incorporate the neurophysiological meaning of hyperedge weights. To address these challenges, this study introduces an enhanced hypergraph framework that integrates nonlinear modeling with biologically meaningful neural information for both static and dynamic functional network analysis. METHODS: This study proposes a novel framework, "Phase-Amplitude Coupling-weighted Kernelized Elastic Net Regularization" (PAC-KENReg). The method employs a kernelized Elastic Net to capture high-order nonlinear interactions and construct the hypergraph topology. It then innovatively weights hyperedges using phase-amplitude coupling (PAC) strength to incorporate neurophysiological information. Furthermore, we extend the framework to dynamic analysis by introducing temporal stability matrices that capture the time-varying characteristics of brain network organization. RESULTS: Validated on public EEG datasets of Major Depressive Disorder (MDD) and Attention Deficit Hyperactivity Disorder (ADHD), PAC-KENReg achieved a classification accuracy of 78.39% and a balanced accuracy of 78.33% for MDD, significantly outperforming baseline models. The method also demonstrated superior discriminative power for ADHD, effectively identified key pathology-associated brain regions, and revealed abnormal dynamic connectivity profiles in patients. CONCLUSION: The proposed PAC-KENReg framework exhibits excellent generalization capability and robustness, providing a reliable and interpretable tool for characterizing brain connectivity patterns. It offers an effective computational approach for the objective diagnosis and mechanistic investigation of psychiatric disorders.
Frequency-Domain Dynamic Light Scattering (FEDSA) for in vivo screening of breast tissue abnormalities: A proof-of-concept study.
Fernández-Pinto J, Gómez-Torrado Á, Miranda DA
Comput Methods Programs Biomed · 2026 Jul · PMID 41950616 · Publisher ↗
BACKGROUND AND OBJECTIVE: Dynamic light scattering (DLS) provides valuable information on nanoscale and microscale dynamics, but its systematic application to in vivo tissue evaluation remains limited. This study introdu...
BACKGROUND AND OBJECTIVE: Dynamic light scattering (DLS) provides valuable information on nanoscale and microscale dynamics, but its systematic application to in vivo tissue evaluation remains limited. This study introduces field effect detection by spectral analysis (FEDSA), a frequency-domain approach designed to analyze backscattered light signals and identify tissue abnormalities associated with the field cancerization effect in breast tissue. The objective was to establish a proof of concept showing that FEDSA can differentiate normal from abnormal tissue. METHODS: A two-stage proof-of-concept study was conducted. First, FEDSA was validated using suspensions of alumina particles (60-300 nm and 100-400 nm) and polystyrene particles (315 nm) to test its performance as a dynamic light scattering technique. Second, in vivo measurements were obtained from 26 women (19 with normal tissue and 7 with abnormal tissue confirmed by imaging or clinical diagnosis). Power spectra were decomposed into frequency bands, transformed through principal component analysis, and analyzed by logistic regression. RESULTS: FEDSA reproduced the expected behavior of a dynamic light scattering-type technique when applied to suspensions of particles. In breast tissue experiments, statistically significant differences were observed between normal and abnormal groups, particularly in the 150-160 kHz frequency band. A PCA-logistic regression model showed discriminatory potential. The ROC analysis yielded an AUC of 0.83; however, cross-validation grouped with patients provided a more conservative performance estimate (AUC ≈ 0.68-0.74), supporting the feasibility of the approach while suggesting uncertainty due to the limited cohort size. CONCLUSIONS: This proof-of-concept study demonstrates the feasibility of FEDSA as a non-invasive, low-cost, and non-ionizing frequency-domain technique inspired by DLS principles to differentiate normal from abnormal breast tissue. Although further validation with larger and more diverse cohorts is required, these findings suggest the potential of FEDSA as a complementary tool for early breast cancer risk assessment.
Skeleton-guided sparse anchors for rotated instance segmentation in cell microscopy.
Wang J, Zhou C, Ming Z … +5 more , Wei L, Chen S, Jiang X, Xu C, Qian D
Comput Methods Programs Biomed · 2026 Jul · PMID 41950615 · Publisher ↗
BACKGROUND AND OBJECTIVE: Accurate instance segmentation of clustered cells in microscopy images remains a major bottleneck, as traditional methods often break down when objects of varying sizes and shapes touch or overl...
BACKGROUND AND OBJECTIVE: Accurate instance segmentation of clustered cells in microscopy images remains a major bottleneck, as traditional methods often break down when objects of varying sizes and shapes touch or overlap. We introduce A2B-IS, a novel one-stage framework that represents each cell with a pixel-level mask and a rotated bounding box, specifically designed to improve segmentation quality in densely packed regions. METHODS: A2B-IS decouples mask and box prediction into parallel branches to simplify the pipeline and reduce error propagation. We incorporate a Gaussian skeleton map that (1) guides anchor placement to focus computations on likely cell centers and suppress background noise, and (2) corrects box predictions near instance boundaries to prevent merged or fragmented detections. To enrich feature representations, we embed an Atrous Attention Block that captures fine-grained, multiscale details at high resolution. Finally, a semi-supervised learning strategy leverages unlabeled images alongside annotated data to further boost model robustness and generalization. The code and dataset are available at https://github.com/wangjuncongyu/A2B-Net. RESULTS: On two large-scale cell microscopy datasets, A2B-IS consistently outperformed leading one- and two-stage segmentation approaches. Compared to baseline models, it achieved higher average precision and recall, with particularly strong gains in densely clustered regions and for small or irregularly shaped cells. CONCLUSIONS: By combining pixel-level masks, rotated boxes, skeleton-guided anchors, attention-based feature extraction, and semi-supervised training, A2B-IS delivers substantial improvements in challenging microscopy segmentation tasks. This advance paves the way for more reliable automated analysis of cell populations without extensive per-image calibration.
An algorithm-enhanced stool DNA system improves the differential diagnosis of colorectal cancer versus Crohn's disease in high-risk symptomatic patients.
Gao L, Guo Z, Wang Z … +1 more , Wang M
Comput Methods Programs Biomed · 2026 Jul · PMID 41950614 · Publisher ↗
BACKGROUND AND OBJECTIVE: Crohn's disease (CD) and colorectal cancer (CRC) share many clinical symptoms, making non-invasive differential diagnosis difficult. FIT-sDNA is sensitive for CRC screening in average-risk popul...
BACKGROUND AND OBJECTIVE: Crohn's disease (CD) and colorectal cancer (CRC) share many clinical symptoms, making non-invasive differential diagnosis difficult. FIT-sDNA is sensitive for CRC screening in average-risk populations but often gives false positives in CD patients due to inflammation-induced mucosal turnover. This study aimed to develop and validate an algorithm-enhanced system (FIT-sDNA-CA) to improve the specificity of CRC triage using current DNA tests. METHODS: The study enrolled 312 subjects, comprising a training cohort of 234 confirmed patients and a prospective validation cohort of 78 potential patients initially diagnosed by clinicians with either CD or CRC. Machine learning algorithms integrated gender, age, fecal KRAS mutation, BMP3/NDRG4/SDC2 methylation, fecal calprotectin (FC), and fecal immunochemical test (FIT) results. After comparing eight algorithms, polynomial regression (PR) was determined to be the optimal model. RESULTS: The PR model demonstrated superior clinical applicability compared to long short-term memory (LSTM) networks (validation set AUC 0.906 vs 0.794). In CRC differential diagnosis, the FIT-sDNA-CA system achieved a positive predictive value of 69.65 % (95 % CI, 66.73-71.58), significantly higher than FIT (45.93 %) and FC (22.92 %). CONCLUSION: By integrating genetic, epigenetic, and inflammatory biomarkers, the FIT-sDNA-CA system effectively filters out confounding signals from intestinal inflammation, overcoming the low specificity limitation of traditional fecal DNA testing. As a highly accurate non-invasive triage tool, this system facilitates early risk stratification for patients with high-risk colorectal cancer symptoms and significantly reduces unnecessary endoscopic referrals.
Semi-automatic generation of selected cerebral vessels for the objective evaluation of vessel segmentation and their geometric parameters in computed tomography angiography images.
Gumulski J, Żyłkowski J, Szatkowski M … +1 more , Spinczyk D
Comput Methods Programs Biomed · 2026 Jul · PMID 41945973 · Publisher ↗
BACKGROUND AND OBJECTIVE: Segmentation of brain vascular structures is a current challenge in radiology for the diagnosis of human vascular pathologies. Owing to the nature of cerebral vessels and to advances in supervis...
BACKGROUND AND OBJECTIVE: Segmentation of brain vascular structures is a current challenge in radiology for the diagnosis of human vascular pathologies. Owing to the nature of cerebral vessels and to advances in supervised segmentation methods, the process of collecting a case set for segmentation with expert masks was very laborious and ambiguous. Objective analysis of automatic segmentation results remains a current, yet unaddressed challenge. To overcome these difficulties, a method for semi-automatic generating synthetic vessels within original computed tomography angiography images using expert masks was developed. METHODS: Generating synthetic vessels that reflect real vessels enables an objective evaluation of segmentation methods and the geometric parameters of the vessels determined based on their segmentation. In this article, the results of four segmentation methods were examined based on generated vessels embedded in original images: UNETR, V-NET, nnUNET, and the classic Frangi method, which remains the baseline reference method. In addition, an analysis of selected geometric parameters of the segmented vessels was performed, including centerline distances, length, diameter, curvature and tortuosity. RESULTS: This study investigated differences between the results of automatic segmentation of the selected arteries with reference to synthetic data. The obtained results indicate significant correlation between vessel geometric parameters and segmentation quality. CONCLUSIONS: Even nnUNET, commonly considered the most effective vessel segmentation method, exhibits significant statistical differences in the determined vessel parameters. An objective analysis of the segmentation results and their geometric parameters, made possible by the developed vessel generation method, indicates a clear need for further development of vessel segmentation methods.
Assessing apparent cell stiffness on fibrous substrates: A comparison of numerical-analytical and in silico models with a novel thermo-contraction approach.
Prosperi G, Paredes J, Aldazabal J
Comput Methods Programs Biomed · 2026 Jul · PMID 41941851 · Publisher ↗
BACKGROUND AND OBJECTIVE: Understanding how cells sense mechanical properties of fibrous substrates is crucial for designing biomaterials that regulate cellular behaviour. Understanding how cells sense the mechanical pro...
BACKGROUND AND OBJECTIVE: Understanding how cells sense mechanical properties of fibrous substrates is crucial for designing biomaterials that regulate cellular behaviour. Understanding how cells sense the mechanical properties of fibrous substrates is crucial for designing biomaterials that regulate cellular behaviour. This study aims to present an accessible initial model for calculating apparent cell stiffness (k) while visualising the cell on a fibrous substrate. By comparing analytical-numerical methods and in silico simulations, we provide a comprehensive approach to exploring how these techniques can be employed to assess cellular mechanics in fibrous environments. METHODS: The model was initially validated on idealised fibre networks using three independent methods: full in silico contraction, analytical beam theory, and applied in silico force-displacement analysis. These approaches yielded consistent k values. Subsequently, the methodology was applied to a real 3D scaffold geometry reconstructed from confocal laser scanning microscopy images of electrospun structures. RESULTS: The evaluation of apparent stiffness shows that the closer the attachment point is to the intersection of the filaments, the higher the apparent stiffness. Additionally, on the same substrate, the greater the number of attachment points, the slight increase in apparent stiffness we observe. The analytical-numerical beam method represents the most cost-effective and efficient method for evaluating apparent stiffness. The equation incorporates a rotational correction factor (α = 0.85) to account for the semi-rigid behaviour of fibre intersections. CONCLUSIONS: This method offers a predictive and scalable tool for evaluating the impact of fibrous architecture and the influence of the geometry of the cell-substrate mechanics. It can support the rational design of fibrous biomaterials for applications in tissue engineering, disease modelling, and regenerative medicine.
Machine learning classification of normal and malignant cells on the basis of their viscoelastic properties.
Thomas-Chemin O, Séverac C, Abidine Y … +2 more , Trevisiol E, Dague E
Comput Methods Programs Biomed · 2026 Jul · PMID 41936192 · Publisher ↗
BACKGROUND AND OBJECTIVE: Cell mechanics, elasticity and viscoelasticity, are key markers of biological states like cancer. Atomic force microscopy (AFM) is ideal for such studies, but its low throughput limits large-sca...
BACKGROUND AND OBJECTIVE: Cell mechanics, elasticity and viscoelasticity, are key markers of biological states like cancer. Atomic force microscopy (AFM) is ideal for such studies, but its low throughput limits large-scale use. Two solutions exist: automation for higher throughput, or high-density measurements for richer data. The latter enables machine learning (ML)-based classification, with viscoelastic parameters offering unique insights beyond static measures like Young's modulus. METHODS: This study used dynamic mechanical analysis (DMA) to classify cells, focusing on viscoelastic descriptors (storage/loss moduli) across frequencies. Normal (RWPE-1) and grade IV cancerous (PC3-GFP) prostate cells were probed at 1-200Hz, generating 304 features per cell. The fuzzy logic-based LAMDA algorithm, trained on 19 selected features, classified cells using 40 samples per line. RESULTS: PC3-GFP cells showed higher deformability and heterogeneity, behaving more like viscous fluids at low frequencies. The model achieved 79% classification accuracy. Adding features improved performance, suggesting fewer training samples may suffice with rich datasets. A sensitivity-optimized threshold reduced false negatives in cancer detection. CONCLUSIONS: Combining viscoelastic analysis with ML effectively discriminates normal and malignant cells. Future work could refine training and integrate new features, though acquisition time remains a challenge. This approach offers a promising framework for mechanome-based diagnostics, with applications in cancer and stem cell research.
In silico approaches to tackle coronary artery disease: where we are, where we are going.
De Nisco G, Lodi Rizzini M, Veneziani A … +1 more , Marsden AL
Comput Methods Programs Biomed · 2026 Jun · PMID 41933520 · Publisher ↗
Abstract loading — click title to view on PubMed.
Dual-branch attention-enhanced network integrating CT images and clinicoradiographic features for preoperative ternary classification of IASLC grading in lung adenocarcinoma: A multicenter study.
Zuo Z, Deng J, Zeng Y … +3 more , Qi W, Liu W, Zhang J
Comput Methods Programs Biomed · 2026 Jun · PMID 41933519 · Publisher ↗
BACKGROUND AND OBJECTIVE: Precise preoperative prediction of the International Association for the Study of Lung Cancer (IASLC) grade in lung adenocarcinoma (LUAD) is essential for risk stratification and surgical decisi...
BACKGROUND AND OBJECTIVE: Precise preoperative prediction of the International Association for the Study of Lung Cancer (IASLC) grade in lung adenocarcinoma (LUAD) is essential for risk stratification and surgical decision-making. However, existing approaches remain limited in reliably distinguishing among the three IASLC grades. We developed and validated a dual-branch attention-enhanced network that integrates computed tomography (CT) imaging and clinicoradiographic (CR) information for preoperative ternary IASLC grading. METHODS: This retrospective multicenter study included 1477 thin-slice chest CT examinations from three independent institutions. The proposed framework, DB-CRN, jointly models a preprocessed 3D nodule subvolume and patient-specific CR features through a two-stage dual-branch architecture equipped with a convolutional block attention module and channel attention blocks. Three comparative models were constructed: DB-NoCRN (image-only ablation), ML-CR (CR-only XGBoost model), and a conventional radiomics pipeline. Model performance was evaluated using the macro-averaged area under the receiver operating characteristic curve (macro AUC) and the Obuchowski index. RESULTS: DB-CRN achieved the best overall performance, with a macro AUC of 0.858 and an Obuchowski index of 0.844, outperforming DB-NoCRN (macro AUC 0.837; Obuchowski index 0.826), ML-CR (macro AUC 0.787; Obuchowski index 0.729), and the radiomics model (macro AUC 0.811; Obuchowski index 0.810). CONCLUSION: By fusing multiscale CT representations with CR priors under an attention-enhanced dual-branch design, DB-CRN enables robust preoperative ternary classification of IASLC grades in LUAD. This framework offers a noninvasive and reproducible tool to support risk-adapted surgical planning and personalized management, warranting prospective validation.
Unified stability conditions for explicit finite-difference Biot-type poroelastodynamics: Non-dimensional design maps for time step selection in brain fluid transport.
Chou D
Comput Methods Programs Biomed · 2026 Jun · PMID 41916093 · Publisher ↗
BACKGROUND AND OBJECTIVE: Biot-type poroelastodynamics describes coupled solid deformation and pore-fluid transport, and is increasingly germane to biomedical simulation of tissue mechanics and interstitial fluid motion....
BACKGROUND AND OBJECTIVE: Biot-type poroelastodynamics describes coupled solid deformation and pore-fluid transport, and is increasingly germane to biomedical simulation of tissue mechanics and interstitial fluid motion. Yet explicit finite-difference schemes, despite their economy in memory and execution time, remain hampered by the absence of a unified stability condition for systems in which propagative and dissipative processes coexist. This study derives a framework for explicit time-step selection in the u→-p form of Biot's equations. METHODS: A one-dimensional reduction of the linear Biot poroelastodynamic model was discretised with centred spatial differences, central differencing for the second-order displacement time derivative, and forward differencing for first-order pore-pressure evolution. Von Neumann analysis was then applied to the coupled discrete system, yielding a cubic characteristic polynomial in the amplification factor. By separating diffusion-dominated and wave-dominated asymptotic limits and invoking Schur stability theory, explicit admissibility bounds were obtained in terms of material coefficients, mesh size, time step, and wavenumber. A reformulation exposed the governing scales and enabled the construction of a Δx-Δt stability map. The derived conditions were then examined through root-locus analysis and healthy-brain simulations. RESULTS: The analysis yields two complementary stability restrictions: a diffusion-like condition, Δt/Δx≤1/(2D), and a wave-like condition, Δt/Δx≤1/C, showing that the explicit scheme cannot be characterised adequately by a single universal condition. Their intersection defines a critical mesh size and critical time step, with representative brain parameters giving Δx=2.2×10 m and Δt=6.879×10 s. This crossover identifies the scale at which stability control passes from wave transmission across one grid interval to diffusive pore-pressure relaxation over the same interval. Spectral analysis further reveals distinct failure modes: a real-root-dominated, non-oscillatory breakdown in the diffusion-dominated regime and a conjugate-pair-driven oscillatory instability in the wave-dominated regime. Numerical experiments corroborate the theoretical boundaries and demonstrate bounded physiological responses only when both mesh and time step lie within the admissible region. CONCLUSION: The resulting framework furnishes parameter-aware design rules for explicit poroelastodynamic computation and offers a basis for brain fluid-transport simulation, with broader relevance to other Biot-type porous media systems cast in the same coupled displacement-pressure formulation.
In silico modelling of changes in spinal cord blood flow after endovascular aortic aneurysm repair.
Rasiah MG, Konings TJAJ, Nio A … +10 more , Moriconi S, Patel AS, Smith A, Gkoutzios P, Abdelhalim MA, Delhaas T, Cardoso MJ, Lamata P, Mees BME, Modarai B
Comput Methods Programs Biomed · 2026 Jun · PMID 41911607 · Publisher ↗
AIMS: To develop an in-silico model of the aorta and its spinal cord-supplying branches, using it to characterise haemodynamic changes following aortic aneurysm (AA) repair. The work is motivated by the risk of spinal co...
AIMS: To develop an in-silico model of the aorta and its spinal cord-supplying branches, using it to characterise haemodynamic changes following aortic aneurysm (AA) repair. The work is motivated by the risk of spinal cord ischaemia (SCI) and paraplegia, serious complications that can arise from disruption of spinal cord perfusion during AA surgery. An objective, patient-specific tool capable of predicting changes in spinal cord blood flow pre-intervention would address a critical unmet clinical need. METHODS: SimVascular was used to retrospectively model a 76-year-old female patient's aorta pre- and post-uncomplicated endovascular thoraco-abdominal AA repair. The full extent of the aorta and its branches, including spinal cord-supplying vessels, was segmented. Pulsatile flow simulations were conducted under the assumption of rigid walls, with patient-specific inlet and three-element Windkessel models for the outlet boundary conditions on the SimVascular Gateway Cluster. Haemodynamic changes following (staged) stent graft implantation were evaluated, alongside key surface-based metrics: time-averaged wall shear stress (TAWSS), oscillatory shear index (OSI), relative residence time (RRT) and endothelial cell activation potential (ECAP) were assessed primarily focussing on spinal cord-supplying vessels. RESULTS: Postoperatively, segmental artery flow to the spinal cord decreased by 51.86% following exclusion of lumbar and posterior intercostal arteries by the stent. Spinal cord-supplying arteries showed increased TAWSS (+5.2%) and reduced RRT and ECAP, with minimal change in OSI. Modest flow increases were observed in non-spinal vascular beds, including the legs (+6.09%), reno-visceral vessels (+5.89%), and supra-aortic branches (+5.97%). Across vascular territories, visceral arteries had the highest TAWSS and lowest RRT/ECAP, while leg arteries had the lowest TAWSS and highest RRT/ECAP; supra-aortic vessels exhibited the highest OSI. Simulating a hypothetical first-stage thoracic stent deployment demonstrated an 18.2% reduction in spinal cord-directed flow, compared with the 51.9% reduction after complete repair, illustrating the pipeline's capacity to compare surgical strategies. CONCLUSION: This study lays a foundation for computational prediction of SCI risk. It leverages in-silico modelling, using open-source software and routine medical imaging, to assess spinal cord blood flow alterations after aortic surgery. Scaling to more patients and enriching physiological detail of models may forge a path towards a clinical decision-making tool.
About
- Frequency
- Sun
- Papers found
- 200
- RSS feed
- Subscribe