Computers in biology and medicine最新文献

{"title":"KC-UNIT: Multi-kernel conversion using unpaired image-to-image translation with perceptual guidance in chest computed tomography imaging","authors":"Changyong Choi ,&nbsp;Doa Kim ,&nbsp;Seungjoo Park ,&nbsp;Hyunki Lee ,&nbsp;Hyunwook Kim ,&nbsp;Sang Min Lee ,&nbsp;Namkug Kim","doi":"10.1016/j.compbiomed.2025.110846","DOIUrl":"10.1016/j.compbiomed.2025.110846","url":null,"abstract":"<div><div>Computed tomography (CT) images are reconstructed from raw datasets including sinogram using various convolution kernels through back projection. Kernels are typically chosen depending on the anatomical structure being imaged and the specific purpose of the scan, balancing the trade-off between image sharpness and pixel noise. Generally, a sinogram requires large storage capacity, and storage space is often limited in clinical settings. Thus, CT images are generally reconstructed with only one specific kernel in clinical settings, and the sinogram is typically discarded after a week. Therefore, many researchers have proposed deep learning-based image-to-image translation methods for CT kernel conversion. However, transferring the style of the target kernel while preserving anatomical structure remains challenging, particularly when translating CT images from a source domain to a target domain in an unpaired manner, which is often encountered in real-world settings. Thus, we propose a novel kernel conversion method using unpaired image-to-image translation (KC-UNIT). This approach utilizes discriminator regularization, using feature maps from the generator to improve semantic representation learning. To capture content and style features, cosine similarity content and contrastive style losses were defined between the feature map of generator and semantic label map of discriminator. This can be easily incorporated by modifying the discriminator's architecture without requiring any additional learnable or pre-trained networks. The KC-UNIT demonstrated the ability to preserve fine-grained anatomical structure from the source domain during transfer. Our method outperformed existing generative adversarial network-based methods across most kernel conversion methods in three kernel domains. The code is available at <span><span>https://github.com/cychoi97/KC-UNIT</span><svg><path></path></svg></span>.</div></div>","PeriodicalId":10578,"journal":{"name":"Computers in biology and medicine","volume":"196 ","pages":"Article 110846"},"PeriodicalIF":7.0,"publicationDate":"2025-07-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144704492","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":"Charting a finite element, mechanical atlas of dermatologic wound closure","authors":"Congzhou M. Sha","doi":"10.1016/j.compbiomed.2025.110839","DOIUrl":"10.1016/j.compbiomed.2025.110839","url":null,"abstract":"<div><div>Wound geometry and the mechanical properties of human skin govern the failure modes of partially healed or scarred tissue. Though dermatologists and surgeons develop an intuitive understanding of the mechanical characteristics of skin through clinical practice, finite element models of wounds can aid in formalizing intuition. In this work, we explore the effect of wound geometry and primary intention closure on the propagation of mechanical stresses through skin. We use a two-layer, orthotropic, hyperelastic model of the epidermis, dermis, and subcutis to accurately capture the mechanical and geometric effects at work. We highlight the key assumptions which must be made when modeling closure of wounds by primary intention, clearly delineating promising areas for model improvement. Models are implemented in DOLFINx, an open-source finite element framework, and reference code is provided for reproducible and extensible science. We provide a framework for modeling the mechanical properties of skin wounds, reporting on the results of extensive finite element simulations of the skin. We perform sanity checks on the results and highlight the problems which remain in enabling clinical use of such modeling. We present finite element methods and reference code for modeling skin wounds and skin wound closure by primary intention, with a focus on reproducibility and critical examination of underlying assumptions.</div></div>","PeriodicalId":10578,"journal":{"name":"Computers in biology and medicine","volume":"196 ","pages":"Article 110839"},"PeriodicalIF":7.0,"publicationDate":"2025-07-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144702523","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":"Effects of pathological mutations on the CHCHD2 monomer structure: A study by AlphaFold3 linked to the generation of conformational ensembles","authors":"Zeyneb Azra Özmen ,&nbsp;Fatma Nilsu Çaylı ,&nbsp;Vladimir N. Uversky ,&nbsp;Junga Alexa Woo ,&nbsp;David E. Kang ,&nbsp;Orkid Coskuner-Weber","doi":"10.1016/j.compbiomed.2025.110810","DOIUrl":"10.1016/j.compbiomed.2025.110810","url":null,"abstract":"<div><div>CHCHD2 is a mitochondrial protein linked to neurodegenerative diseases such as Parkinson's disease (PD), Alzheimer's disease (AD), and frontotemporal dementia (FTD). To investigate the structural effects of disease-associated mutations, we analyzed 19 pathogenic variants using AlphaFold3 and conformational ensemble modeling with AFflecto. While the radius of gyration and end-to-end distances remained largely unchanged, mutations significantly altered secondary structure elements and contact maps, particularly in local folding. Intrinsic disorder and LLPS analyses revealed that mutations modulate the protein's droplet-forming capacity and interaction flexibility. These changes may impact protein-protein interactions, phase behavior, and mitochondrial function. Our findings indicate that pathogenic CHCHD2 mutations cause subtle but functionally relevant structural perturbations rather than global destabilization. This study underscores the importance of ensemble-based modeling in understanding mutation-induced dysfunction in intrinsically disordered proteins involved in neurodegeneration.</div></div>","PeriodicalId":10578,"journal":{"name":"Computers in biology and medicine","volume":"196 ","pages":"Article 110810"},"PeriodicalIF":7.0,"publicationDate":"2025-07-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144702520","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"A hyperspectral imaging dataset and Grassmann manifold method for intraoperative pixel-wise classification of metastatic colon cancer in the liver","authors":"Ivica Kopriva ,&nbsp;Dario Sitnik ,&nbsp;Laura-Isabelle Dion-Bertrand ,&nbsp;Marija Milković Periša ,&nbsp;Arijana Pačić ,&nbsp;Mirko Hadžija ,&nbsp;Marijana Popović Hadžija","doi":"10.1016/j.compbiomed.2025.110841","DOIUrl":"10.1016/j.compbiomed.2025.110841","url":null,"abstract":"<div><div>Hyperspectral imaging (HSI) holds significant potential for transforming the field of computational pathology. However, the number of HSI-based research studies remains limited, and in many cases, the advantages of HSI over traditional RGB imaging have not been conclusively demonstrated, particularly for specimens collected intraoperatively. To address these challenges we present: (<em>i</em>) a database consisted of 27 HSIs of hematoxylin-eosin stained frozen sections, collected from 14 patients with colon adenocarcinoma metastasized to the liver. It is aimed to validate pixel-wise classification for intraoperative tumor resection; (<em>ii</em>) a novel method which combines Grassmann points with nearest subspace classifier for pixel-wise classification of HSIs. The HSIs were acquired in the spectral range of 450 nm–800 nm, with a resolution of 1 nm, resulting in images of 1384 × 1035 pixels. Pixel-wise annotations were performed by two pathologists and one medical expert. To overcome challenges such as experimental variability and the lack of annotated data, we applied Grassmann manifold (GM) approach in combination with spectral-spatial features extracted by tensor singular spectrum analysis (TSSA) method to non-overlapping patches of 230 × 258 pixels. Using only 1 % of labeled pixels per class, the GM-TSSA method achieved a micro balanced accuracy (BACC) of 0.963 and a micro F₁-score of 0.959 on the HSI dataset. The GM-TSSA approach outperformed six deep learning architectures trained with 63 % of labeled pixels. Data are available at: <span><span>https://data.fulir.irb.hr/islandora/object/irb:538</span><svg><path></path></svg></span>, and code is available at: <span><span>https://github.com/ikopriva/ColonCancerHSI</span><svg><path></path></svg></span>.</div></div>","PeriodicalId":10578,"journal":{"name":"Computers in biology and medicine","volume":"196 ","pages":"Article 110841"},"PeriodicalIF":7.0,"publicationDate":"2025-07-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144702519","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":"Benchmarking of open-source algorithms for heart rate estimation from motion-corrupted photoplethysmography","authors":"Marcello Sicbaldi ,&nbsp;Luca Palmerini ,&nbsp;Serena Moscato ,&nbsp;Paola di Florio ,&nbsp;Alessandro Silvani ,&nbsp;Lorenzo Chiari","doi":"10.1016/j.compbiomed.2025.110808","DOIUrl":"10.1016/j.compbiomed.2025.110808","url":null,"abstract":"<div><div>Photoplethysmography holds promise for continuous, non-intrusive heart rate monitoring through wearable devices. However, motion artifacts can impact the reliability of heart rate estimates. The integration of accelerometer data has been proven helpful in mitigating these artifacts. Although several algorithms that combine photoplethysmography and accelerometer data for heart rate estimation have been proposed, it remains unclear which performs best. We performed a systematic and comprehensive search and evaluation of all relevant published algorithms (N = 126) and benchmarked all available open-source methods (N = 11) using the same real-world dataset. A robust methodological framework was employed for assessing these algorithms, featuring a comprehensive set of performance metrics. Out of 126 retrieved articles, 11 provided open-source implementations and were included in the benchmarking. We found that deep learning algorithms consistently outperformed model-based algorithms and algorithms that did not correct for accelerometer data, particularly in dynamic conditions with substantial motion artifacts. The BeliefPPG algorithm performed best across all metrics, with an estimation bias of 0.7 ± 0.8 bpm, an estimation variability of 4.4 ± 2.0 bpm, and a Spearman's correlation of 0.73 ± 0.14 bpm with the heart rate ground truth. These findings underscore the potential of deep learning techniques to enhance the reliability of photoplethysmography-based heart rate monitoring through integration with accelerometer data in real-world conditions. This work provides valuable insights into the performance of these algorithms and highlights the importance of developing broader, more diverse datasets to enhance generalizability in future research.</div></div>","PeriodicalId":10578,"journal":{"name":"Computers in biology and medicine","volume":"196 ","pages":"Article 110808"},"PeriodicalIF":7.0,"publicationDate":"2025-07-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144702521","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":"Targeting the dengue virus NS5-Methyltransferase SAM binding site with limonoids: Molecular docking, dynamics simulation, DFT and ADMET analysis","authors":"Purna Chandra Pal ,&nbsp;Bhanuranjan Das","doi":"10.1016/j.compbiomed.2025.110784","DOIUrl":"10.1016/j.compbiomed.2025.110784","url":null,"abstract":"<div><div>Dengue virus (DENV) infects over 100 million people annually, yet no approved antiviral therapies are available. The DENV genome is a positive-sense single-stranded RNA (+) ssRNA) encoding ten proteins: three structural (capsid, membrane, and envelope) and seven non-structural (NS1–NS5). Among these, the NS5 methyltransferase (NS5-MTase) is essential for viral replication and is a promising drug target due to the absence of approved inhibitors. NS5-MTase has two binding sites: one for S-adenosyl-L-methionine (SAM) and another for RNA. The RNA-binding site is shallow and solvent-exposed, making the SAM-binding site a more suitable target for small-molecule inhibitors. Phytocompounds, particularly limonoids—a class of tetraterpenoids with known pharmacological activities are promising candidates in antiviral drug discovery. In this study, 500 limonoids were screened through molecular docking against the SAM-binding site of DENV NS5-MTase. This is the first large-scale in silico virtual screening of limonoids targeting this site. Three top compounds were identified: 7-deacetyl-21-hydroxyneotrichilenonelide (DHC), demethyl-3-detigloyl-iso-swietenine (DDIS), and demethyl-iso-swietenolide (DIS). The docking scores of the three compounds were −9.4, −9.5, and −9.9 kcal/mol, respectively. These three compounds were further evaluated using molecular dynamics (MD) simulations to assess the stability and interaction profiles of the protein-ligand complexes. DIS exhibited the most stable binding conformation (forming 3-4 H-bond/ns and maintaining a RMSD below 0.15 nm) and a more favorable interaction profile compared to the other compounds during MD simulations. Additionally, binding free energy (MM-PBSA), in silico ADMET analysis, and DFT calculations, also indicate that DIS is the most promising candidate against DENV.</div></div>","PeriodicalId":10578,"journal":{"name":"Computers in biology and medicine","volume":"196 ","pages":"Article 110784"},"PeriodicalIF":7.0,"publicationDate":"2025-07-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144702522","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":"How photoplethysmography can be used to detect major depressive disorder among patients with obstructive sleep apnea during sleep","authors":"Vikash Shaw ,&nbsp;Quoc Cuong Ngo ,&nbsp;Nemuel Daniel Pah ,&nbsp;Ahsan Habib Khandoker ,&nbsp;Prasant Kumar Mahapatra ,&nbsp;Dinesh Pankaj ,&nbsp;Dinesh K. Kumar","doi":"10.1016/j.compbiomed.2025.110814","DOIUrl":"10.1016/j.compbiomed.2025.110814","url":null,"abstract":"<div><div>Major Depressive Disorder (MDD) frequently coexists with Obstructive Sleep Apnea (OSA), yet it remains underdiagnosed in OSA populations due to overlapping symptoms and limited access to psychiatric sleep evaluations. Earlier studies have explored photoplethysmography (PPG) for screening either OSA or MDD individually but have not investigated the use of PPG to analyze comorbid MDD in patients with OSA. Additionally, beat-level entropy and complexity features, which quantify subtle nonlinear variations in pulse morphology and may reflect autonomic nervous system dysregulation associated with this comorbidity, have been largely overlooked. This study investigates whether beat-to-beat PPG-derived features can distinguish among healthy controls (CO), individuals with OSA only, referred to as OSA-, and those with OSA and comorbid MDD, referred to as OSA+. PPG recordings from 60 participants (CO: 25, OSA-: 20, OSA+: 15) were preprocessed to extract artifact-free segments. For each segment, skewness (Sk) and kurtosis (Ku) were computed on a beat-to-beat basis, followed by the extraction of Approximate Entropy (ApEn), Hjorth Activity (HA), Hjorth Mobility (HM), and Hjorth Complexity (HC) parameters to quantify signal variability. Feature selection was conducted within a 5-fold nested cross-validation framework using Spearman correlation, and only features that satisfied both a correlation threshold and statistical significance (p &lt; 0.05) were retained for SVM classification. For the CO vs. OSA task, ApEn(Sk) and ApEn(Ku) emerged as the most discriminative features, achieving an accuracy of 84 % and an AUC of 0.91. For the OSA− vs. OSA + task, three features—ApEn(Sk), ApEn(Ku), and HM(Sk)—were selected, yielding an accuracy of 76 % and an AUC of 0.88. This study demonstrates that beat-to-beat variability in PPG morphology can effectively identify MDD within the OSA population. Unlike prior work, which did not investigate comorbid classification or entropy-based features, our approach addresses this gap and supports the feasibility of sleep-based PPG for mental health screening.</div></div>","PeriodicalId":10578,"journal":{"name":"Computers in biology and medicine","volume":"196 ","pages":"Article 110814"},"PeriodicalIF":7.0,"publicationDate":"2025-07-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144696600","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":"Biomedical simulations of electroosmotic non-Newtonian hybrid nanofluid (blood) with hematocrit viscosity through a porous overlapping irregular stenosed artery","authors":"Pramod Kumar Yadav,&nbsp;Aditya Singh","doi":"10.1016/j.compbiomed.2025.110575","DOIUrl":"10.1016/j.compbiomed.2025.110575","url":null,"abstract":"<div><div>Inspired by the therapeutic possibilities of drug delivery in cardiovascular disease management, the authors’ objective is to investigate the hemodynamic phenomenon during the electroosmotic flow of non-Newtonian hybrid nanofluid through an artery featuring two types of stenosis <span><math><mrow><mo>(</mo><mi>i</mi><mo>)</mo></mrow></math></span> irregular and <span><math><mrow><mo>(</mo><mi>i</mi><mi>i</mi><mo>)</mo></mrow></math></span> overlapping. In the present work, the authors have incorporated <span><math><mrow><mi>A</mi><mi>g</mi><mo>−</mo><mi>T</mi><mi>i</mi><msub><mrow><mi>O</mi></mrow><mrow><mn>2</mn></mrow></msub></mrow></math></span> as the hybrid nanoparticles in the blood and the Sutterby fluid is utilized to represent the non-Newtonian nature of the blood. Here, the authors’ objective is to make the model more realistic for real life situation by considering a hematocrit-dependent blood viscosity which is due to the presence of RBCs in the blood. In the proposed model, the authors have incorporated the magnetic field and electric field in the radial and axial direction of the flow, respectively. Additionally, the authors have taken into account the effect of viscous dissipation, chemical reaction, Joule heating, thermal radiation, and body acceleration on the blood flow characteristics. The present model is governed by the non-linear PDEs’ constituting continuity, momentum, energy, and concentration equations with suitable boundary and initial conditions. The authors have utilized the ’Explicit Finite Difference Method’ with the help of MATLAB ‘R2023a’ software to get the graphical results for the wall shear stress, flow resistance, Nusselt number, Sherwood number, concentration, temperature, and velocity. Further, in the present work, the streamlines are plotted for various hemodynamic parameters. It is noticed from the results of the proposed work that <span><math><mrow><mi>A</mi><mi>g</mi><mo>−</mo><mi>T</mi><mi>i</mi><msub><mrow><mi>O</mi></mrow><mrow><mn>2</mn></mrow></msub></mrow></math></span>/blood flow velocity dominates the velocity of <span><math><mrow><mi>A</mi><mi>g</mi></mrow></math></span>/blood. It is also observed that while increasing the nanoparticle volume fraction of 1%, the blood flow velocity rises 41.08% for the irregular stenosed artery and 37.93% for the overlapping stenosed artery compared to pure blood(base fluid). Moreover, the wall shear stress has higher magnitude for irregular stenosis than overlapping stenosis. Additionally, the authors have also validated their findings with the previous literature. The present simulations have a broad range of applications in biomedical engineering, diagnosis of tumors and brain aneurysms, extraction of blood clots, and magnetic targeted drug delivery.</div></div>","PeriodicalId":10578,"journal":{"name":"Computers in biology and medicine","volume":"196 ","pages":"Article 110575"},"PeriodicalIF":7.0,"publicationDate":"2025-07-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144696523","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":"Network and modeling analysis of MAPK signaling cascade uncovers EGR1 regulation through ERK2 protein in breast cancer","authors":"Honey Pavithran ,&nbsp;Preetam Ghosh ,&nbsp;Ranjith Kumavath","doi":"10.1016/j.compbiomed.2025.110830","DOIUrl":"10.1016/j.compbiomed.2025.110830","url":null,"abstract":"<div><div>The study explores the therapeutic relevance of Cardiac glycosides (CGs), including lanatoside C (LC), peruvoside (PS), and strophanthidin (STR) in treating breast cancer, using network pharmacology studies and bioinformatics approaches. Building on our prior <em>in vitro</em> studies and transcriptome profiling, we aimed to explore protein expression alterations influenced by the selected compounds in the present study. The methodology was structured and directed to delineate the active protein targets and their molecular mechanism of action in controlling cancer progression. Initially, we predicted the protein targets of individual compounds using SWISSTargetPrediction, and the results were compared with the differentially expressed genes from the transcriptome data acquired in the preliminary studies. The identified protein targets were further studied for their network relatedness and cross-verified by comparing their expression in cancer and normal patient data from TCGA using the UALCAN algorithm. Additionally, we aimed to identify the candidate biomarkers that potentially served as predictive or prognostic indicators in malignant breast cancer by conducting survival analysis of the crucial proteins using the GEPIA2 database. Overall, the analysis allowed us to understand the co-dependence expression between MAPK1 and EGR1 proteins, further emphasizing their clinical significance in cancer diagnosis and probable therapeutic outcomes. MD simulation studies further verified the most significant protein targets with their interaction scores and structural stability of the compounds, which showed higher structural stability around 300-ns trajectories for MAPK1 and EGR1 proteins. Finally, the pathway simulation studies, modeling on the MAPK/ERK signaling cascade, showed significant alteration in the biochemical parameters and stability of the system depending on the concentration of crucial proteins ERK2, also known as MAPK1 and EGR1 proteins in the pathway. These findings underscore the therapeutic potential of CGs and further highlight the significant role of identified proteins in targeting breast cancer.</div></div>","PeriodicalId":10578,"journal":{"name":"Computers in biology and medicine","volume":"196 ","pages":"Article 110830"},"PeriodicalIF":7.0,"publicationDate":"2025-07-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144696601","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":"DEEP Q-NAS: A new algorithm based on neural architecture search and reinforcement learning for brain tumor identification from MRI","authors":"Md Sabid Hasan ,&nbsp;Md Mostafizur Rahman Komol ,&nbsp;Faiyaz Fahim ,&nbsp;Jahirul Islam ,&nbsp;Tanjina Pervin ,&nbsp;Md Mahbub Hasan","doi":"10.1016/j.compbiomed.2025.110767","DOIUrl":"10.1016/j.compbiomed.2025.110767","url":null,"abstract":"<div><div>A significant obstacle in brain tumor treatment planning is determining the tumor’s actual size. Magnetic resonance imaging (MRI) is one of the first-line brain tumor diagnosis. It takes a lot of effort and mostly depends on the operator’s experience to manually separate the size of a brain tumor from 3D MRI volumes. Machine learning has been vastly enhanced by deep learning and computer-aided tumor detection methods. This study proposes to investigate the architecture of object detectors, specifically focusing on search efficiency. In order to provide more specificity, our goal is to effectively explore the Feature Pyramid Network (FPN) and prediction head of a straightforward anchor-free object detector called DEEP Q-NAS. The study utilized the BraTS 2021 dataset which includes multi-parametric magnetic resonance imaging (mpMRI) scans. The architecture we found outperforms the latest object detection models (like Fast R-CNN, YOLOv7, and YOLOv8) by 2.2 to 7 points with average precision (AP) on the MS COCO 2017 dataset. It has a similar level of complexity and less memory usage, which shows how effective our proposed NAS is for object detection. The DEEP Q-NAS with ResNeXt-152 model demonstrates the highest level of detection accuracy, achieving a rate of 99%.</div></div>","PeriodicalId":10578,"journal":{"name":"Computers in biology and medicine","volume":"196 ","pages":"Article 110767"},"PeriodicalIF":7.0,"publicationDate":"2025-07-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144696521","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}