Computers in biology and medicine最新文献

{"title":"Inhibitors as a therapeutic frontier in lung cancer: Mechanism, opportunities, and molecular docking studies","authors":"Riya Patel ,&nbsp;Darshan Rathod ,&nbsp;Nehal Shah ,&nbsp;Vipul Vaghela ,&nbsp;Nikunj Patadiya","doi":"10.1016/j.compbiomed.2025.110889","DOIUrl":"10.1016/j.compbiomed.2025.110889","url":null,"abstract":"<div><div>Lung cancer remains one of the leading causes of death worldwide, and novel therapies are urgently needed to enhance patient outcomes. Recent advancements have highlighted the potential of molecular inhibitors that target specific signaling pathways, offering promising treatment alternatives. These inhibitors block key pathways such as EGFR, ALK, and ROS1, which are often dysregulated in lung cancer and play essential roles in cell proliferation, apoptosis, and metastasis. By selectively targeting these pathways, inhibitors may reduce adverse effects, improve therapeutic efficacy, and minimize damage to healthy cells. This paper explores how these inhibitors affect lung cancer cells, examining recent developments and emerging opportunities. It also discusses the role of molecular docking studies, which are crucial in drug development as they reveal the molecular interactions and binding affinities between inhibitors and their target proteins. Molecular docking has significantly advanced the design of inhibitors, enhancing their potency and specificity for clinical use. By analyzing these computational models, this study evaluates molecular inhibitors' real-world applications and future potential in lung cancer treatment. The findings emphasize the importance of targeted approaches, highlighting the challenges and successes in developing molecular inhibitors and paving the way toward precision medicine.</div></div>","PeriodicalId":10578,"journal":{"name":"Computers in biology and medicine","volume":"196 ","pages":"Article 110889"},"PeriodicalIF":6.3,"publicationDate":"2025-08-04","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144766652","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Automated detection of lacunes in brain MR images using SAM with robust prompts using self-distillation and anatomy-informed priors","authors":"Pon Deepika,&nbsp;Gouri Shanker,&nbsp;Ramanujam Narayanan,&nbsp;Vaanathi Sundaresan","doi":"10.1016/j.compbiomed.2025.110806","DOIUrl":"10.1016/j.compbiomed.2025.110806","url":null,"abstract":"<div><h3>Background:</h3><div>Lacunes, which are small fluid-filled cavities in the brain, are signs of cerebral small vessel disease and have been clinically associated with various neurodegenerative and cerebrovascular diseases. Hence, accurate detection of lacunes is crucial and is one of the initial steps for the precise diagnosis of these diseases. However, developing a robust and consistently reliable method for detecting lacunes is challenging because of the heterogeneity in their appearance, contrast, shape, and size.</div></div><div><h3>Method:</h3><div>In this study, we propose a lacune detection method using the Segment Anything Model (SAM), guided by point prompts from a candidate prompt generator. The prompt generator initially detects potential lacunes with a high sensitivity using a composite loss function. The true lacunes are then selected using SAM by discriminating their characteristics from mimics such as the sulcus and enlarged perivascular spaces, imitating the clinicians’ strategy of examining the potential lacunes along all three axes. False positives are further reduced by adaptive thresholds based on the region wise prevalence of lacunes.</div></div><div><h3>Results:</h3><div>We evaluated our method on two diverse, multi-centric MRI datasets, VALDO and ISLES, comprising only FLAIR sequences. Despite diverse imaging conditions and significant variations in slice thickness (0.5–6 mm), our method achieved sensitivities of 84% and 92%, with average false positive rates of 0.05 and 0.06 per slice in ISLES and VALDO datasets respectively.</div></div><div><h3>Conclusions:</h3><div>The proposed method demonstrates robust performance across varied imaging conditions and outperformed the state-of-the-art methods, demonstrating its effectiveness in lacune detection and quantification.</div></div>","PeriodicalId":10578,"journal":{"name":"Computers in biology and medicine","volume":"196 ","pages":"Article 110806"},"PeriodicalIF":6.3,"publicationDate":"2025-08-04","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144766649","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 MMP-7 in idiopathic pulmonary fibrosis: An integrative in vivo and in silico evaluation of the therapeutic potential of Tylophora indica","authors":"Dibyendu Halder,&nbsp;Md Sabbir Khan,&nbsp;Than Sue,&nbsp;Mohammad Ali,&nbsp;Umme Mim Sad Jahan,&nbsp;Md Atiquzzaman,&nbsp;Marina Khatun,&nbsp;Md Khairul Islam,&nbsp;Rabeya Bashree Keya,&nbsp;Arif Hasan,&nbsp;Md Asaduzzaman Sikder","doi":"10.1016/j.compbiomed.2025.110867","DOIUrl":"10.1016/j.compbiomed.2025.110867","url":null,"abstract":"<div><h3>Background</h3><div>Idiopathic pulmonary fibrosis (IPF) is a degenerative pulmonary condition marked by a substantial accumulation of extracellular matrix and chronic inflammation. Matrix metalloproteinase-7 (MMP-7) is integral to fibrosis and likely a curative focus. This study investigates the therapeutic potential of <em>Tylophora indica (T. indica)</em> plant extract for treating IPF, utilizing <em>in vivo</em> and <em>in silico</em> approaches that target MMP-7.</div></div><div><h3>Methods and materials</h3><div><em>T. indica</em> extract was administered to a bleomycin-induced IPF mouse model at 200 and 300 mg/kg/day doses. Efficacy was evaluated through histological analysis and quantitative RT-PCR to measure MMP-7 expression. <em>In silico</em> molecular dynamics simulation and molecular docking identified bioactive compounds from <em>T. indica</em> that could inhibit MMP-7. ADMET profiling was used to evaluate these substances' pharmacological potential and safety.</div></div><div><h3>Results</h3><div><em>T. indica</em> extract at 300 mg/kg/day significantly reduced fibrosis and inflammation, improving histopathological scores and lowering MMP-7 expression. <em>In silico</em> analysis identified pergularinine, tylophorine, quercetin, kaempferol, and tylophorinidine as potent MMP-7 inhibitors with stronger binding affinities than pirfenidone, a standard anti-fibrotic drug. Molecular dynamics simulations confirmed the stability of these interactions, and the compounds showed favorable safety profiles in ADMET assessments.</div></div><div><h3>Conclusion</h3><div><em>T. indica</em> extract demonstrated significant antifibrotic activity by downregulating MMP-7 expression and improving lung histopathology in the IPF mouse model. The identified phytochemicals show strong potential as natural MMP-7 inhibitors, suggesting <em>T. indica</em> as a prospective therapeutic agent for IPF. Additional clinical studies are required to validate these results.</div></div>","PeriodicalId":10578,"journal":{"name":"Computers in biology and medicine","volume":"196 ","pages":"Article 110867"},"PeriodicalIF":6.3,"publicationDate":"2025-08-04","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144766653","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 lightweight YOLOv8-based model for gastric cancer detection","authors":"Seung-Won Jeong ,&nbsp;Shabir Ahmad ,&nbsp;Jae-Seoung Kim ,&nbsp;Taegkeun Whangbo","doi":"10.1016/j.compbiomed.2025.110689","DOIUrl":"10.1016/j.compbiomed.2025.110689","url":null,"abstract":"<div><div>Recent research on deep learning-based gastric cancer detection has demonstrated high performance, with capabilities comparable to or exceeding those of medical professionals. However, the performance of deep learning models depends on the performance of processors such as GPU and CPU, and real-world medical environments encompass a diverse array of computer processor. Recognizing the necessity of research that accommodates varying processors, this study proposes a gastric cancer detection model based on YOLOv8, aiming to achieve real-time performance with reduced sensitivity to performance fluctuations of computer environments. YOLOv8-n was adopted as the baseline, with Ghost conv applied to the backbone for compression. Lightweight channel-wise attention was introduced in the neck and the head via SE blocks to enhance feature representation without sacrificing real-time performance. By comparing the detection precision and speed in CPU and four GPUs with different performances, this study explores the feasibility of applying a deep learning-based gastric cancer detector in processors in actual medical field. Experimental results demonstrate that the proposed model maintains real-time inference speed on GPUs of various performance levels. Moreover, it achieved 77.5 mean Average Precision (mAP) which outperformed the mAP of 76.5 of YOLOv8-m, while outperforming 74.4 mAP of YOLOv8-n (baseline) by 4.16%. The complexity of the proposed model was minimal, with only 2.8M parameters and 7.7 GFLOPs, demonstrating the achievement of high detection precision with a reduced number of parameters.</div></div>","PeriodicalId":10578,"journal":{"name":"Computers in biology and medicine","volume":"196 ","pages":"Article 110689"},"PeriodicalIF":6.3,"publicationDate":"2025-08-04","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144766651","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":"Comparison of machine learning algorithms for predicting length of stay in chronic kidney disease patients","authors":"Birgül Yabana Kiremit ,&nbsp;Durmuş Özkan Şahin","doi":"10.1016/j.compbiomed.2025.110825","DOIUrl":"10.1016/j.compbiomed.2025.110825","url":null,"abstract":"<div><div>The length of stay (LOS) for patients in hospitals is crucial for workforce planning, resource allocation, and bed capacity management, impacting healthcare costs, future needs and financial planning. This study focuses on calculating the LOS for Chronic Kidney Disease (CKD) patients admitted as inpatients and estimating their hospital bills based on services rendered during their stay. Utilizing data from 5,583 CKD patients and 11 input variables, various machine learning (ML) algorithms were applied to develop regression, and classification models. To optimize the model performance and address potential overfitting issues, feature selection techniques were also employed. The Random Forest (RF) algorithm achieved the highest performance for bill amount estimation, with a Correlation Coefficient (CC) of 0.736. The algorithms predicting LOS showed even more promising results, with all performing above 0.848 on the CC metric. The best performances were obtained from Support Vector Machine (SVM), M5P trees and RF with Mean Absolute Error (MAE) and CC results of 2.580 day-0.875, 2.587 day-0.880 and 2.611 day-0.880, respectively. LOS was categorized as short or long using ML algorithms, with Logistic Regression (LogR) achieving the best classification results: 0.944 on the AUC-ROC (Area Under the ROC Curve) metric and 0.872 on the F-Measure metric. The RF algorithm also excelled in classification based on patient units, producing results of 0.788 on the AUC-ROC and 0.863 for accuracy. Additionally with feature selection revealed that reducing input variables maintained prediction accuracy for bill amount and LOS, but it generally negatively affected classification performance. Feature selection was identified as a critical challenge, particularly in balancing the trade-off between dimensionality reduction and predictive accuracy. While dimensionality reduction can improve computational efficiency, careful selection of input variables is essential to maintain robust classification performance. Given the lengthy treatment processes for CKD patients, accurate predictions of LOS, billing amounts, and admission units will assist health managers in planning for future resource needs, such as medical supplies and workforce. Ultimately, this study provides insights that can enhance the financial sustainability and management of healthcare services.</div></div>","PeriodicalId":10578,"journal":{"name":"Computers in biology and medicine","volume":"196 ","pages":"Article 110825"},"PeriodicalIF":6.3,"publicationDate":"2025-08-04","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144766650","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":"Allele-specific expression analysis: pipelines, applications, challenges, and unmet needs","authors":"Stefania Mattevi ,&nbsp;Francesco Mazzarotto ,&nbsp;Paolo Martini","doi":"10.1016/j.compbiomed.2025.110890","DOIUrl":"10.1016/j.compbiomed.2025.110890","url":null,"abstract":"<div><div>In diploid organisms, genes typically exhibit balanced expression of maternal and paternal alleles. However, exceptions exist, such as autosomal genes with allele-specific expression, where genetic and epigenetic variations can lead to the exclusive or preferential expression of a particular allele. In this context, allele-specific expression analysis serves as a powerful tool for understanding gene regulation, with significant functional and clinical implications.</div><div>Despite their increasing importance, current analysis pipelines face notable limitations including a lack of end-to-end solutions, restricted options for multi-omics integration, and insufficient support for single-cell sequencing technologies.</div><div>This review critically assesses 26 cutting-edge pipelines for allele-specific expression analysis, focusing on their input requirements, capabilities, and applications in the field. Pipelines are categorized based on their ability to handle various data types, support haplotype phasing, employ statistical approaches, and provide graphical outputs. Most pipelines fail to automate preprocessing, integrate multi-omic data, and support high-throughput single-cell sequencing. Future advancements should prioritize the development of automated multi-omic workflows, implementing visualization options, and enhancing compatibility with single-cell technologies. By addressing these gaps, next-generation allele-specific expression pipelines will offer insights into the mechanisms of allele-specific expression regulation, thereby advancing our understanding of its biological and clinical significance.</div></div>","PeriodicalId":10578,"journal":{"name":"Computers in biology and medicine","volume":"196 ","pages":"Article 110890"},"PeriodicalIF":6.3,"publicationDate":"2025-08-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144763943","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":"An effective multi-modality analysis for stress classification: A signal-to-image conversion using local pattern techniques","authors":"L. Susmitha ,&nbsp;A. Shamila Ebenezer ,&nbsp;S. Jeba Priya ,&nbsp;M.S.P. Subathra ,&nbsp;S. Thomas George ,&nbsp;Geno Peter ,&nbsp;Albert Alexander Stonier","doi":"10.1016/j.compbiomed.2025.110847","DOIUrl":"10.1016/j.compbiomed.2025.110847","url":null,"abstract":"<div><div>Stress is an intensified reaction that occurs when humans experience challenges(stressors) due to complex and nonlinear responses. The study proposes a pattern-driven framework that combines signal and image-based modalities, incorporating Local Binary Pattern (LBP), Local Normal Derivative Pattern (LNDP), Local Derivative Pattern (LDP), and Local Tetra Pattern (LTrP) using Spectrogram. Derived from the analyzed patterns, features spanning in the Time Domain, Non-Linear chaos theory, Fractal Dimensions, and Histogram descriptors are extracted. This study employs binary classification techniques to identify stress by distinguishing between baseline and stress states in multi-modality sensor data, such as heart rate (HR), and respiration rate (RR). In signal analysis, statistical and non-linear features related to the predictability of stress attained the high classification accuracy using a Support Vector Machine (SVM) based on its linearity. Logistic Regression (LG) with its data complexity has obtained a good accuracy for overall features. For image analysis, the LBP technique exhibits strong overall classification performance for statistical at 98.7 %, entropy and histogram at 100 %, and fractals at 90 % compared with methods. Specifically, for LNDP and LTrP, Logistic Regression and Ensemble outperform SVM, achieving an impressive accuracy of 100 %, and superior performance metrics such as Precision, Matthews Correlation Coefficient (MCC), and Kappa Score based on the distribution of pixel intensities and directionality of pattern images as 0.4 to 1. The AlexNet gives good classification accuracy for Image Analysis of stress detection. Based on the intricate patterns at different scales, image analysis of spectrograms and pattern techniques yields better classification accuracy compared with signal analysis for stress prediction.</div></div>","PeriodicalId":10578,"journal":{"name":"Computers in biology and medicine","volume":"196 ","pages":"Article 110847"},"PeriodicalIF":6.3,"publicationDate":"2025-08-02","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144757845","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":"Implications of virtual screening for South African natural compounds against Plesiomonas shigelloides, a pathogen with zoonotic potential","authors":"Calvin R. Wei ,&nbsp;Zarrin Basharat ,&nbsp;Prajit Adhikari","doi":"10.1016/j.compbiomed.2025.110882","DOIUrl":"10.1016/j.compbiomed.2025.110882","url":null,"abstract":"<div><div><em>Plesiomonas shigelloides</em> is an emerging pathogen associated with gastroenteritis and poses a growing public health concern, especially in regions with limited access to advanced medical treatments. The purpose of this study was to explore the therapeutic potential of South African natural product compounds against <em>P. shigelloides</em> by targeting the essential enzyme Pyridoxine 5′-phosphate synthase or PPS (encoded by <em>PdxJ</em>). <em>P. shigelloides</em> proteomes (n = 26) were processed using the Bacterial Pan Genome Analysis (BPGA) pipeline to identify conserved targets. Targeting conserved protein ensures the potential for broad-spectrum efficacy. PPS was chosen as drug target and its structure was predicted using AlphaFold, enabling high-confidence modeling. Subsequently, docking was performed using AutoDock Vina, focusing on a library of South African compounds (n &gt; 1000). The three inhibitors demonstrating strong binding affinities to the PPS were Scutiaquinone A, Mesquitol-(4α→5)-3,3′,4′,7,8-pentahydroxyflavonone, and Riccardin C. To further validate the stability and efficacy of these interactions, molecular dynamics (MD) simulations were carried out for 100 ns. The simulations revealed stable interactions between the inhibitors and PPS, suggesting potential inhibition of the PPS enzyme. Mesquitol derivative was found to be the safest and recommended for further experimental validation. This study highlights the promising potential of South African natural compounds in combating <em>P. shigelloides</em> infections, paving the way for the development of novel therapeutic strategies.</div></div>","PeriodicalId":10578,"journal":{"name":"Computers in biology and medicine","volume":"196 ","pages":"Article 110882"},"PeriodicalIF":6.3,"publicationDate":"2025-08-02","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144757969","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":"Interpretable bioinformatics approaches for pheochromocytoma bioactivity and protein interaction analysis","authors":"İlhan Uysal","doi":"10.1016/j.compbiomed.2025.110862","DOIUrl":"10.1016/j.compbiomed.2025.110862","url":null,"abstract":"<div><div>Pheochromocytoma (PCC) is a rare neuroendocrine tumor driven by complex molecular mechanisms, notably involving the oncogenic c-Myc/Max and c-Myc/c-Max protein complexes. Despite their pivotal role in tumor progression, the molecular interactions and bioactive compounds specifically targeting these complexes remain inadequately characterized. This study presents an integrative computational pipeline combining interpretable bioinformatics, network biology, and machine learning to elucidate key molecular mechanisms and bioactive motifs associated with PCC. A curated dataset of 5000 bioactive molecules was obtained from ChEMBL, and structural motifs associated with bioactivity were identified using a genetic programming-based approach. Random Forest, Support Vector Machines, and Gradient Boosting classifiers were trained and cross-validated using 10-fold cross-validation to predict pIC50 values, achieving high performance (mean accuracy: 0.98, AUC &gt;0.97). Feature importance analysis consistently identified pIC50, molecular weight (MW), lipophilicity (LogP), and hydrogen-bonding properties as primary determinants of bioactivity. PPI networks were built using STRING's experimentally validated interactions and refined using BioGRID and literature cross-validation. Network centrality analysis and community detection using the Girvan–Newman algorithm revealed MYC, MAX, and EP300 as central hubs, with associated protein modules significantly enriched for biological processes including transcriptional regulation, cell cycle control, ubiquitination, and apoptosis. To enhance model interpretability, explainable artificial intelligence (XAI) methods, including SHAP and DALEX, were employed to elucidate the contribution of individual molecular descriptors, mechanistically elucidating compound–target interactions. Despite its robustness, this computational framework lacks experimental validation and independent external datasets. Additionally, STRING's uniform confidence scores limited edge-weight precision in network visualizations during network analyses. Nevertheless, this study demonstrates the potential of a multi-layered computational approach to deepen the understanding of MYC-driven oncogenesis in PCC. By integrating motif discovery, network biology, and interpretable machine learning, the work identifies actionable molecular signatures and critical protein targets, providing a foundation for future experimental validation and the development of targeted therapies in pheochromocytoma as well as other rare cancers.</div></div>","PeriodicalId":10578,"journal":{"name":"Computers in biology and medicine","volume":"196 ","pages":"Article 110862"},"PeriodicalIF":6.3,"publicationDate":"2025-08-02","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144757843","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":"High-efficiency spatially guided learning network for lymphoblastic leukemia detection in bone marrow microscopy images","authors":"Liye Mei ,&nbsp;Chentao Lian ,&nbsp;Suyang Han ,&nbsp;Zhaoyi Ye ,&nbsp;Yuyang Hua ,&nbsp;Meixing Sun ,&nbsp;Jing He ,&nbsp;Zhiwei Ye ,&nbsp;Mengqing Mei ,&nbsp;Yaxiaer Yalikun ,&nbsp;Hui Shen ,&nbsp;Cheng Lei ,&nbsp;Bei Xiong","doi":"10.1016/j.compbiomed.2025.110860","DOIUrl":"10.1016/j.compbiomed.2025.110860","url":null,"abstract":"<div><div>Leukemia is a hematologic tumor that proliferates in bone marrow and seriously affects the survival of patients. Early and accurate diagnosis is crucial for effective leukemia treatment. Traditional diagnostic methods rely on experts’ subjective analysis of bone marrow smears microscopic images. This approach is time-consuming and complex. Despite recent advances in deep learning, automated leukemia detection remains limited due to the scarcity of high-quality datasets, the prevailing focus on single-cell image classification rather than precise cell-level detection in whole slide images, along with challenges such as morphological heterogeneity, uneven staining, scale variation, and occluded cell boundary in bone marrow smears. To address these challenges, we construct a novel dataset comprising 1794 high-quality microscopic images, establishing a new benchmark for lymphocytic leukemia detection. Additionally, we develop a fully automated diagnostic method based on spatially-guided learning (SGLNet), enabling rapid whole slide analysis of leukemia. Specifically, we introduce several innovative enhancements to the baseline algorithm, including the spatially-guided learning framework, scale-aware fusion module, small object-enhancing mechanisms, and efficient intersection over union loss function. These improvements effectively address the impact of morphological similarity and complex backgrounds in leukemia detection, significantly enhancing detection accuracy. Finally, the results show that SGLNet achieves mean average precision scores of 95.9 % and 98.6 % in detecting acute lymphoblastic leukemia and chronic lymphocytic leukemia, respectively. These results demonstrate the efficiency and accuracy of our method in identifying lymphoblastic leukemia cells, significantly enhancing large-scale clinical diagnosis, and supporting clinicians in developing personalized treatment plans.</div></div>","PeriodicalId":10578,"journal":{"name":"Computers in biology and medicine","volume":"196 ","pages":"Article 110860"},"PeriodicalIF":6.3,"publicationDate":"2025-08-02","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144763938","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}