Liangrui Pan , Xiang Wang , Qingchun Liang , Jiandong Shang , Wenjuan Liu , Liwen Xu , Shaoliang Peng
{"title":"DEDUCE: Multi-head attention decoupled contrastive learning to discover cancer subtypes based on multi-omics data","authors":"Liangrui Pan , Xiang Wang , Qingchun Liang , Jiandong Shang , Wenjuan Liu , Liwen Xu , Shaoliang Peng","doi":"10.1016/j.cmpb.2024.108478","DOIUrl":"10.1016/j.cmpb.2024.108478","url":null,"abstract":"<div><h3>Background and Objective:</h3><div>Given the high heterogeneity and clinical diversity of cancer, substantial variations exist in multi-omics data and clinical features across different cancer subtypes.</div></div><div><h3>Methods:</h3><div>We propose a model, named DEDUCE, based on a symmetric multi-head attention encoders (SMAE), for unsupervised contrastive learning to analyze multi-omics cancer data, with the aim of identifying and characterizing cancer subtypes. This model adopts a unsupervised SMAE that can deeply extract contextual features and long-range dependencies from multi-omics data, thereby mitigating the impact of noise. Importantly, DEDUCE introduces a subtype decoupled contrastive learning method based on a multi-head attention mechanism to simultaneously learn features from multi-omics data and perform clustering for identifying cancer subtypes. Subtypes are clustered by calculating the similarity between samples in both the feature space and sample space of multi-omics data. The fundamental concept involves decoupling various attributes of multi-omics data features and learning them as contrasting terms. A contrastive loss function is constructed to quantify the disparity between positive and negative examples, and the model minimizes this difference, thereby promoting the acquisition of enhanced feature representation.</div></div><div><h3>Results:</h3><div>The DEDUCE model undergoes extensive experiments on simulated multi-omics datasets, single-cell multi-omics datasets, and cancer multi-omics datasets, outperforming 10 deep learning models. The DEDUCE model outperforms state-of-the-art methods, and ablation experiments demonstrate the effectiveness of each module in the DEDUCE model. Finally, we applied the DEDUCE model to identify six cancer subtypes of AML.</div></div><div><h3>Conclusion:</h3><div>In this paper, we proposed DEDUCE model learns features from multi-omics data through SMAE, and the subtype decoupled contrastive learning consistently optimizes the model for clustering and identifying cancer subtypes. The DEDUCE model demonstrates a significant capability in discovering new cancer subtypes. We applied the DEDUCE model to identify six subtypes of AML. Through the analysis of GO function enrichment, subtype-specific biological functions, and GSEA of AML using the DEDUCE model, the interpretability of the DEDUCE model in identifying cancer subtypes is further enhanced.</div></div>","PeriodicalId":10624,"journal":{"name":"Computer methods and programs in biomedicine","volume":"257 ","pages":"Article 108478"},"PeriodicalIF":4.9,"publicationDate":"2024-10-30","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142587373","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}
André Mourato , Rodrigo Valente , José Xavier , Moisés Brito , Stéphane Avril , António C. Tomás , José Fragata
{"title":"Comparative analysis of Zero Pressure Geometry and prestress methods in cardiovascular Fluid-Structure Interaction","authors":"André Mourato , Rodrigo Valente , José Xavier , Moisés Brito , Stéphane Avril , António C. Tomás , José Fragata","doi":"10.1016/j.cmpb.2024.108475","DOIUrl":"10.1016/j.cmpb.2024.108475","url":null,"abstract":"<div><h3>Background and Objective:</h3><div>Modelling patient-specific aortic biomechanics with advanced computational techniques, such as Fluid–Structure Interaction (FSI), can be crucial to provide effective decision-making indices to enhance current clinical practices. To effectively simulate Ascending Thoracic Aortic Aneurysms (ATAA), the stress-free configuration must be defined. The Zero Pressure Geometry (ZPG) and the Prestress Tensor (PT) are two of the main approaches to tackle this issue. However, their impact on the numerical results is yet to be analysed. Computed Tomography Angiography (CTA) and Magnetic Resonance Imaging (MRI) data were used to develop patient-specific 2-way FSI frameworks.</div></div><div><h3>Methods:</h3><div>Three models were developed considering different tissue prestressing approaches to account for the reference configuration and their numerical results were compared. The selected approaches were: (i) ZPG, (ii) PT and (iii) a combination of the PT approach with a regional mapping of material properties (PTCAL).</div></div><div><h3>Results:</h3><div>The pressure fields estimated by all models were equivalent. The estimation of Wall Shear Stress (WSS) based metrics revealed good correspondence between all models except the Relative Residence Time (RRT). Regarding ATAA wall mechanics, the proposed extension to the PT approach presented a closer agreement with the ZPG model than its counterpart. Additionally, the PT and PTCAL approaches required around 60% fewer iterations to achieve cycle-to-cycle convergence than the ZPG algorithm.</div></div><div><h3>Conclusion:</h3><div>Using a regional mapping of material properties in combination with the PT method presented a better correspondence with the ZPG approach. The outcomes of this study can pave the way for advancing the accuracy and convergence of ATAA numerical models using the PT methodology.</div></div>","PeriodicalId":10624,"journal":{"name":"Computer methods and programs in biomedicine","volume":"257 ","pages":"Article 108475"},"PeriodicalIF":4.9,"publicationDate":"2024-10-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142581381","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}
Kyle Baylous , Brandon Kovarovic , Rodrigo R. Paz , Salwa Anam , Ryan Helbock , Marc Horner , Marvin Slepian , Danny Bluestein
{"title":"Thrombogenic Risk Assessment of Transcatheter Prosthetic Heart Valves Using a Fluid-Structure Interaction Approach","authors":"Kyle Baylous , Brandon Kovarovic , Rodrigo R. Paz , Salwa Anam , Ryan Helbock , Marc Horner , Marvin Slepian , Danny Bluestein","doi":"10.1016/j.cmpb.2024.108469","DOIUrl":"10.1016/j.cmpb.2024.108469","url":null,"abstract":"<div><h3>Background and Objective</h3><div>Prosthetic heart valve interventions such as TAVR have surged over the past decade, but the associated complication of long-term, life-threatening thrombotic events continues to undermine patient outcomes. Thus, improving thrombogenic risk analysis of TAVR devices is crucial. In vitro studies for thrombogenicity are typically difficult to perform. However, revised ISO testing standards include computational testing for thrombogenic risk assessment of cardiovascular implants. We present a fluid-structure interaction (FSI) approach for assessing thrombogenic risk of transcatheter aortic valves.</div></div><div><h3>Methods</h3><div>An FSI framework was implemented via the incompressible computational fluid dynamics multi-physics solver of the ANSYS LS-DYNA software. The numerical modeling approach for flow analysis was validated by comparing the derived flow rate of the 29 mm CoreValve device from benchtop testing and orifice areas of commercial TAVR valves in the literature to in silico results. Thrombogenic risk was analyzed by computing stress accumulation (SA) on virtual platelets seeded in the flow fields via ANSYS EnSight. The integrated FSI-thrombogenicity methodology was subsequently employed to examine hemodynamics and thrombogenic risk of TAVR devices with two approaches: 1) engineering optimization and 2) clinical assessment.</div></div><div><h3>Results</h3><div>Simulated effective orifice areas for commercial valves were in reported ranges. In silico cardiac output and flow rate during the positive pressure differential period matched experimental results by approximately 93 %. The approach was used to analyze the effect of various TAVR leaflet designs on hemodynamics, where platelets experienced instantaneous stresses reaching around 10 Pa. Post-TAVR deployment hemodynamics in patient-specific bicuspid aortic valve anatomies revealed varying degrees of thrombogenic risk with the highest median SA around 70 dyn·s/cm<sup>2</sup> - nearly double the activation threshold - despite those being clinically classified as “mild” paravalvular leaks.</div></div><div><h3>Conclusions</h3><div>Our methodology can be used to improve the thromboresistance of prosthetic valves from the initial design stage to the clinic. It allows for unparalleled optimization of devices, uncovering key TAVR leaflet design parameters that can be used to mitigate thrombogenic risk, in addition to patient-specific modeling to evaluate device performance. This work demonstrates the utility of advanced in silico analysis of TAVR devices that can be utilized for thrombogenic risk assessment of other blood recirculating devices.</div></div>","PeriodicalId":10624,"journal":{"name":"Computer methods and programs in biomedicine","volume":"257 ","pages":"Article 108469"},"PeriodicalIF":4.9,"publicationDate":"2024-10-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142496563","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}
Miguel López-Pérez , Alba Morquecho , Arne Schmidt , Fernando Pérez-Bueno , Aurelio Martín-Castro , Javier Mateos , Rafael Molina
{"title":"The CrowdGleason dataset: Learning the Gleason grade from crowds and experts","authors":"Miguel López-Pérez , Alba Morquecho , Arne Schmidt , Fernando Pérez-Bueno , Aurelio Martín-Castro , Javier Mateos , Rafael Molina","doi":"10.1016/j.cmpb.2024.108472","DOIUrl":"10.1016/j.cmpb.2024.108472","url":null,"abstract":"<div><h3>Background:</h3><div>Currently, prostate cancer (PCa) diagnosis relies on the human analysis of prostate biopsy Whole Slide Images (WSIs) using the Gleason score. Since this process is error-prone and time-consuming, recent advances in machine learning have promoted the use of automated systems to assist pathologists. Unfortunately, labeled datasets for training and validation are scarce due to the need for expert pathologists to provide ground-truth labels.</div></div><div><h3>Methods:</h3><div>This work introduces a new prostate histopathological dataset named CrowdGleason, which consists of 19,077 patches from 1045 WSIs with various Gleason grades. The dataset was annotated using a crowdsourcing protocol involving seven pathologists-in-training to distribute the labeling effort. To provide a baseline analysis, two crowdsourcing methods based on Gaussian Processes (GPs) were evaluated for Gleason grade prediction: SVGPCR, which learns a model from the CrowdGleason dataset, and SVGPMIX, which combines data from the public dataset SICAPv2 and the CrowdGleason dataset. The performance of these methods was compared with other crowdsourcing and expert label-based methods through comprehensive experiments.</div></div><div><h3>Results:</h3><div>The results demonstrate that our GP-based crowdsourcing approach outperforms other methods for aggregating crowdsourced labels (<span><math><mrow><mi>κ</mi><mo>=</mo><mn>0</mn><mo>.</mo><mn>7048</mn><mo>±</mo><mn>0</mn><mo>.</mo><mn>0207</mn></mrow></math></span>) for SVGPCR vs.(<span><math><mrow><mi>κ</mi><mo>=</mo><mn>0</mn><mo>.</mo><mn>6576</mn><mo>±</mo><mn>0</mn><mo>.</mo><mn>0086</mn></mrow></math></span>) for SVGP with majority voting). SVGPCR trained with crowdsourced labels performs better than GP trained with expert labels from SICAPv2 (<span><math><mrow><mi>κ</mi><mo>=</mo><mn>0</mn><mo>.</mo><mn>6583</mn><mo>±</mo><mn>0</mn><mo>.</mo><mn>0220</mn></mrow></math></span>) and outperforms most individual pathologists-in-training (mean <span><math><mrow><mi>κ</mi><mo>=</mo><mn>0</mn><mo>.</mo><mn>5432</mn></mrow></math></span>). Additionally, SVGPMIX trained with a combination of SICAPv2 and CrowdGleason achieves the highest performance on both datasets (<span><math><mrow><mi>κ</mi><mo>=</mo><mn>0</mn><mo>.</mo><mn>7814</mn><mo>±</mo><mn>0</mn><mo>.</mo><mn>0083</mn></mrow></math></span> and <span><math><mrow><mi>κ</mi><mo>=</mo><mn>0</mn><mo>.</mo><mn>7276</mn><mo>±</mo><mn>0</mn><mo>.</mo><mn>0260</mn></mrow></math></span>).</div></div><div><h3>Conclusion:</h3><div>The experiments show that the CrowdGleason dataset can be successfully used for training and validating supervised and crowdsourcing methods. Furthermore, the crowdsourcing methods trained on this dataset obtain competitive results against those using expert labels. Interestingly, the combination of expert and non-expert labels opens the door to a future of massive labeling by incorporating both expert and non-expert pathologist an","PeriodicalId":10624,"journal":{"name":"Computer methods and programs in biomedicine","volume":"257 ","pages":"Article 108472"},"PeriodicalIF":4.9,"publicationDate":"2024-10-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142564128","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}
Zhenya Zang, Quan Wang, Mingliang Pan, Yuanzhe Zhang, Xi Chen, Xingda Li, David Day Uei Li
{"title":"Towards high-performance deep learning architecture and hardware accelerator design for robust analysis in diffuse correlation spectroscopy","authors":"Zhenya Zang, Quan Wang, Mingliang Pan, Yuanzhe Zhang, Xi Chen, Xingda Li, David Day Uei Li","doi":"10.1016/j.cmpb.2024.108471","DOIUrl":"10.1016/j.cmpb.2024.108471","url":null,"abstract":"<div><div>This study proposes a compact deep learning (DL) architecture and a highly parallelized computing hardware platform to reconstruct the blood flow index (BFi) in diffuse correlation spectroscopy (DCS). We leveraged a rigorous analytical model to generate autocorrelation functions (ACFs) to train the DL network. We assessed the accuracy of the proposed DL using simulated and milk phantom data. Compared to convolutional neural networks (CNN), our lightweight DL architecture achieves 66.7% and 18.5% improvement in MSE for BFi and the coherence factor <em>β</em>, using synthetic data evaluation. The accuracy of rBFi over different algorithms was also investigated. We further simplified the DL computing primitives using subtraction for feature extraction, considering further hardware implementation. We extensively explored computing parallelism and fixed-point quantization within the DL architecture. With the DL model's compact size, we employed unrolling and pipelining optimizations for computation-intensive for-loops in the DL model while storing all learned parameters in on-chip BRAMs. We also achieved pixel-wise parallelism, enabling simultaneous, real-time processing of 10 and 15 autocorrelation functions on Zynq-7000 and Zynq-UltraScale+ field programmable gate array (FPGA), respectively. Unlike existing FPGA accelerators that produce BFi and the <em>β</em> from autocorrelation functions on standalone hardware, our approach is an encapsulated, end-to-end on-chip conversion process from intensity photon data to the temporal intensity ACF and subsequently reconstructing BFi and <em>β</em>. This hardware platform achieves an on-chip solution to replace post-processing and miniaturize modern DCS systems that use single-photon cameras. We also comprehensively compared the computational efficiency of our FPGA accelerator to CPU and GPU solutions.</div></div>","PeriodicalId":10624,"journal":{"name":"Computer methods and programs in biomedicine","volume":"258 ","pages":"Article 108471"},"PeriodicalIF":4.9,"publicationDate":"2024-10-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142616209","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}
Emmanuel Eghan-Acquah , Alireza Y Bavil , David Bade , Martina Barzan , Azadeh Nasseri , David J Saxby , Stefanie Feih , Christopher P Carty
{"title":"Enhancing biomechanical outcomes in proximal femoral osteotomy through optimised blade plate sizing: A neuromusculoskeletal-informed finite element analysis","authors":"Emmanuel Eghan-Acquah , Alireza Y Bavil , David Bade , Martina Barzan , Azadeh Nasseri , David J Saxby , Stefanie Feih , Christopher P Carty","doi":"10.1016/j.cmpb.2024.108480","DOIUrl":"10.1016/j.cmpb.2024.108480","url":null,"abstract":"<div><div>Proximal femoral osteotomy (PFO) is a frequently performed surgical procedure to correct hip deformities in the paediatric population. The optimal size of the blade plate implant in PFO is a critical but underexplored factor influencing biomechanical outcomes. This study introduces a novel approach to refine implant selection by integrating personalized neuromusculoskeletal modelling with finite element analysis. Using computed tomography scans and walking gait data from six paediatric patients with various pathologies and deformities, we assessed the impact of four distinct implant width-to-femoral neck diameter (W-D) ratios (30 %, 40 %, 50 %, and 60 %) on surgical outcomes. The results show that the risk of implant yield generally decreases with increasing W-D ratio, except for Patient P2, where the yield risk remained below 100 % across all ratios. The implant factor of safety (FoS) increased with larger W-D ratios, except for Patients P2 and P6, where the highest FoS was 2.60 (P2) and 0.49 (P6) at a 60 % W-D ratio. Bone-implant micromotion consistently remained below 40 µm at higher W-D ratios, with a 50 % W-D ratio striking the optimal balance for mechanical stability in all patients except P6. Although interfragmentary and principal femoral strains did not display consistent trends across all patients, they highlight the need for patient-specific approaches to ensure effective fracture healing. These findings highlight the importance of patient-specific considerations in implant selection, offering surgeons a more informed pathway to enhance patient outcomes and extend implant longevity. Additionally, the insights gained from this study provide valuable guidance for manufacturers in designing next-generation blade plates tailored to improve biomechanical performance in paediatric orthopaedics.</div></div>","PeriodicalId":10624,"journal":{"name":"Computer methods and programs in biomedicine","volume":"257 ","pages":"Article 108480"},"PeriodicalIF":4.9,"publicationDate":"2024-10-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142567736","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":"Discovering explainable biomarkers for breast cancer anti-PD1 response via network Shapley value analysis","authors":"Chenxi Sun, Zhi-Ping Liu","doi":"10.1016/j.cmpb.2024.108481","DOIUrl":"10.1016/j.cmpb.2024.108481","url":null,"abstract":"<div><h3>Background and objective</h3><div>Immunotherapy holds promise in enhancing pathological complete response rates in breast cancer, albeit confined to a select cohort of patients. Consequently, pinpointing factors predictive of treatment responsiveness is of paramount importance. Gene expression and regulation, inherently operating within intricate networks, constitute fundamental molecular machinery for cellular processes and often serve as robust biomarkers. Nevertheless, contemporary feature selection approaches grapple with two key challenges: opacity in modeling and scarcity in accounting for gene-gene interactions</div></div><div><h3>Methods</h3><div>To address these limitations, we devise a novel feature selection methodology grounded in cooperative game theory, harmoniously integrating with sophisticated machine learning models. This approach identifies interconnected gene regulatory network biomarker modules with priori genetic linkage architecture. Specifically, we leverage Shapley values on network to quantify feature importance, while strategically constraining their integration based on network expansion principles and nodal adjacency, thereby fostering enhanced interpretability in feature selection. We apply our methods to a publicly available single-cell RNA sequencing dataset of breast cancer immunotherapy responses, using the identified feature gene set as biomarkers. Functional enrichment analysis with independent validations further illustrates their effective predictive performance</div></div><div><h3>Results</h3><div>We demonstrate the sophistication and excellence of the proposed method in data with network structure. It unveiled a cohesive biomarker module encompassing 27 genes for immunotherapy response. Notably, this module proves adept at precisely predicting anti-PD1 therapeutic outcomes in breast cancer patients with classification accuracy of 0.905 and AUC value of 0.971, underscoring its unique capacity to illuminate gene functionalities</div></div><div><h3>Conclusion</h3><div>The proposed method is effective for identifying network module biomarkers, and the detected anti-PD1 response biomarkers can enrich our understanding of the underlying physiological mechanisms of immunotherapy, which have a promising application for realizing precision medicine.</div></div>","PeriodicalId":10624,"journal":{"name":"Computer methods and programs in biomedicine","volume":"257 ","pages":"Article 108481"},"PeriodicalIF":4.9,"publicationDate":"2024-10-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142564115","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}
Junqi Wang , Hailong Li , Kim M Cecil , Mekibib Altaye , Nehal A Parikh , Lili He
{"title":"DFC-Igloo: A dynamic functional connectome learning framework for identifying neurodevelopmental biomarkers in very preterm infants","authors":"Junqi Wang , Hailong Li , Kim M Cecil , Mekibib Altaye , Nehal A Parikh , Lili He","doi":"10.1016/j.cmpb.2024.108479","DOIUrl":"10.1016/j.cmpb.2024.108479","url":null,"abstract":"<div><h3>Background and Objective</h3><div>Very preterm infants are susceptible to neurodevelopmental impairments, necessitating early detection of prognostic biomarkers for timely intervention. The study aims to explore possible functional biomarkers for very preterm infants at born that relate to their future cognitive and motor development using resting-state fMRI. Prior studies are limited by the sample size and suffer from efficient functional connectome (FC) construction algorithms that can handle the noisy data contained in neonatal time series, leading to equivocal findings. Therefore, we first propose an enhanced functional connectome construction algorithm as a prerequisite step. We then apply the new FC construction algorithm to our large prospective very preterm cohort to explore multi-level neurodevelopmental biomarkers.</div></div><div><h3>Methods</h3><div>There exists an intrinsic relationship between the structural connectome (SC) and FC, with a notable coupling between the two. This observation implies a putative property of graph signal smoothness on the SC as well. Yet, this property has not been fully exploited for constructing intrinsic dFC. In this study, we proposed an advanced dynamic FC (dFC) learning model, dFC-Igloo, which leveraged SC information to iteratively refine dFC estimations by applying graph signal smoothness to both FC and SC. The model was evaluated on artificial small-world graphs and simulated graph signals.</div></div><div><h3>Results</h3><div>The proposed model achieved the best and most robust recovery of the ground truth graph across different noise levels and simulated SC pairs from the simulation. The model was further applied to a cohort of very preterm infants from five Neonatal Intensive Care Units, where an enhanced dFC was obtained for each infant. Based on the improved dFC, we identified neurodevelopmental biomarkers for neonates across connectome-wide, regional, and subnetwork scales.</div></div><div><h3>Conclusion</h3><div>The identified markers correlate with cognitive and motor developmental outcomes, offering insights into early brain development and potential neurodevelopmental challenges.</div></div>","PeriodicalId":10624,"journal":{"name":"Computer methods and programs in biomedicine","volume":"257 ","pages":"Article 108479"},"PeriodicalIF":4.9,"publicationDate":"2024-10-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142567734","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}
Jianye Shi , Kiran Manjunatha , Felix Vogt , Stefanie Reese
{"title":"Data-driven reduced order surrogate modeling for coronary in-stent restenosis","authors":"Jianye Shi , Kiran Manjunatha , Felix Vogt , Stefanie Reese","doi":"10.1016/j.cmpb.2024.108466","DOIUrl":"10.1016/j.cmpb.2024.108466","url":null,"abstract":"<div><h3>Background:</h3><div>The intricate process of coronary in-stent restenosis (ISR) involves the interplay between different mediators, including platelet-derived growth factor, transforming growth factor-<span><math><mi>β</mi></math></span>, extracellular matrix, smooth muscle cells, endothelial cells, and drug elution from the stent. Modeling such complex multiphysics phenomena demands extensive computational resources and time.</div></div><div><h3>Methods:</h3><div>This paper proposes a novel non-intrusive data-driven reduced order modeling approach for the underlying multiphysics time-dependent parametrized problem. In the offline phase, a 3D convolutional autoencoder, comprising an encoder and decoder, is trained to achieve dimensionality reduction. The encoder condenses the full-order solution into a lower-dimensional latent space, while the decoder facilitates the reconstruction of the full solution from the latent space. To deal with the 5D input datasets (3D geometry + time series + multiple output channels), two ingredients are explored. The first approach incorporates time as an additional parameter and applies 3D convolution on individual time steps, encoding a distinct latent variable for each parameter instance within each time step. The second approach reshapes the 3D geometry into a 2D plane along a less interactive axis and stacks all time steps in the third direction for each parameter instance. This rearrangement generates a larger and complete dataset for one parameter instance, resulting in a singular latent variable across the entire discrete time-series. In both approaches, the multiple outputs are considered automatically in the convolutions. Moreover, Gaussian process regression is applied to establish correlations between the latent variable and the input parameter.</div></div><div><h3>Results:</h3><div>The constitutive model reveals a significant acceleration in neointimal growth between <span><math><mrow><mn>30</mn><mo>−</mo><mn>60</mn></mrow></math></span> days post percutaneous coronary intervention (PCI). The surrogate models applying both approaches exhibit high accuracy in pointwise error, with the first approach showcasing smaller errors across the entire evaluation period for all outputs. The parameter study on drug dosage against ISR rates provides noteworthy insights of neointimal growth, where the nonlinear dependence of ISR rates on the peak drug flux exhibits intriguing periodic patterns. Applying the trained model, the rate of ISR is effectively evaluated, and the optimal parameter range for drug dosage is identified.</div></div><div><h3>Conclusion:</h3><div>The demonstrated non-intrusive reduced order surrogate model proves to be a powerful tool for predicting ISR outcomes. Moreover, the proposed method lays the foundation for real-time simulations and optimization of PCI parameters.</div></div>","PeriodicalId":10624,"journal":{"name":"Computer methods and programs in biomedicine","volume":"257 ","pages":"Article 108466"},"PeriodicalIF":4.9,"publicationDate":"2024-10-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142564112","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}