Nat Commun. 2025 Feb 8;16(1):1469. doi: 10.1038/s41467-025-56749-2.

ABSTRACT

Enhancers serve as pivotal regulators of gene expression throughout various biological processes by interacting with transcription factors (TFs). While transcription factor binding sites (TFBSs) are widely acknowledged as key determinants of TF binding and enhancer activity, the significant role of their surrounding context sequences remains to be quantitatively characterized. Here we propose the concept of transcription factor binding unit (TFBU) to modularly model enhancers by quantifying the impact of context sequences surrounding TFBSs using deep learning models. Based on this concept, we develop DeepTFBU, a comprehensive toolkit for enhancer design. We demonstrate that designing TFBS context sequences can significantly modulate enhancer activities and produce cell type-specific responses. DeepTFBU is also highly efficient in the de novo design of enhancers containing multiple TFBSs. Furthermore, DeepTFBU enables flexible decoupling and optimization of generalized enhancers. We prove that TFBU is a crucial concept, and DeepTFBU is highly effective for rational enhancer design.

PMID:39922842 | DOI:10.1038/s41467-025-56749-2

Sci Rep. 2025 Feb 8;15(1):4695. doi: 10.1038/s41598-024-84367-3.

ABSTRACT

Understanding how the traits of lineages are related to diversification is key for elucidating the origin of variation in species richness. Here, we test whether traits are related to species richness among lineages of trees from all major biogeographical settings of the lowland wet tropics. We explore whether variation in mortality rate, breeding system and maximum diameter are related to species richness, either directly or via associations with range size, among 463 genera that contain wet tropical forest trees. For Amazonian genera, we also explore whether traits are related to species richness via variation among genera in mean species-level range size. Lineages with higher mortality rates-faster life-history strategies-have larger ranges in all biogeographic settings and have higher mean species-level range sizes in Amazonia. These lineages also have smaller maximum diameters and, in the Americas, contain dioecious species. In turn, lineages with greater overall range size have higher species richness. Our results show that fast life-history strategies influence species richness in all biogeographic settings because lineages with these ecological strategies have greater range sizes. These links suggest that dispersal has been a key process in the evolution of the tropical forest flora.

PMID:39922807 | DOI:10.1038/s41598-024-84367-3

Trends Cell Biol. 2025 Feb 7:S0962-8924(25)00003-0. doi: 10.1016/j.tcb.2025.01.003. Online ahead of print.

ABSTRACT

The selective removal of mitochondria by mitophagy proceeds via multiple mechanisms and is essential for human well-being. The PINK1/Parkin and NIX/BNIP3 pathways are strongly linked to mitochondrial dysfunction and hypoxia, respectively. Both are regulated by ubiquitylation and mitochondrial import. Recent studies have elucidated how the ubiquitin kinase PINK1 acts as a sensor of mitochondrial import stress through stable interaction with a mitochondrial import supercomplex. The stability of BNIP3 and NIX is regulated by the SCFFBXL4 ubiquitin ligase complex. Substrate recognition requires an adaptor molecule, PPTC7, whose availability is limited by mitochondrial import. Unravelling the functional implications of each mode of mitophagy remains a critical challenge. We propose that mitochondrial import stress prompts a switch between these two pathways.

PMID:39922712 | DOI:10.1016/j.tcb.2025.01.003

Arch Biochem Biophys. 2025 Feb 6:110330. doi: 10.1016/j.abb.2025.110330. Online ahead of print.

ABSTRACT

Glucose provides substrate for the predominant anaplerotic pathway which involves the activity of pyruvate carboxylase (PC). PC-mediated anaplerosis has been extensively studied as a metabolic regulator in glycolytic cells during tumorigenesis and metastasis. Herein, inaccuracies in established methods to measure relative intracellular flux through PC are highlighted and a compartmentalized condensed metabolic network (CCMN) is used to resolve the total malate pool into relative contributions from PC and other sources by metabolic flux analysis (MFA) with [U-13C6]glucose tracing. Performance of the CCMN method is evaluated in breast cancer cell lines that are exposed to small molecules targeting metabolism. Across conditions and cell lines, the CCMN approach yields results nearest to an accepted gold-standard methodology, using [3-13C]glucose, or even exposes the gold standard's limitations. The CCMN method does not require a separate experiment with a much more costly and generally less informative metabolic tracer, such as [3-13C]glucose, and in some cases, may outperform its application.

PMID:39922407 | DOI:10.1016/j.abb.2025.110330

Cell Syst. 2025 Feb 5:101168. doi: 10.1016/j.cels.2025.101168. Online ahead of print.

ABSTRACT

Biomolecular condensates have been proposed to buffer intracellular concentrations and reduce noise. However, concentrations need not be buffered in multicomponent systems, leading to a non-constant saturation concentration (csat) when individual components are varied. Simplified equilibrium considerations suggest that noise reduction might be closely related to concentration buffering and that a fixed saturation concentration is required for noise reduction to be effective. Here, we present a theoretical analysis to demonstrate that these suggestions do not apply to mesoscopic fluctuating systems. We show that concentration buffering and noise reduction are distinct concepts, which cannot be used interchangeably. We further demonstrate that concentration buffering and a constant csat are neither necessary nor sufficient for noise reduction to be effective. Clarity about these concepts is important for studying the role of condensates in controlling cellular noise and for the interpretation of concentration relationships in cells. A record of this paper's transparent peer review process is included in the supplemental information.

PMID:39922189 | DOI:10.1016/j.cels.2025.101168

J Exp Bot. 2025 Feb 8:eraf053. doi: 10.1093/jxb/eraf053. Online ahead of print.

ABSTRACT

Arbuscular mycorrhizal (AM) symbiosis can prime plant defenses, leading to mycorrhiza-induced resistance (MIR) against different attackers, including insect herbivores. Still, our knowledge of the complex molecular regulation leading to MIR is very limited. Here, we show that the AM fungus Funneliformis mosseae protects tomato plants against two different chewing herbivores, Spodoptera exigua and Manduca sexta. We explore the underlying molecular mechanism through genome-wide transcriptional profiling, bioinformatics network analyses, and functional bioassays. Herbivore-triggered JA-regulated defenses were primed in leaves of mycorrhizal plants, while ET biosynthesis and signaling were also higher both before and after herbivory. We hypothesized that fine-tuned ET signaling is required for the primed defensive response leading to MIR. ET is a complex regulator of plant responses to stress and is generally considered a negative regulator of plant defenses against herbivory. However, ET-deficient or insensitive lines did not show AM-primed JA biosynthesis or defense response, and were unable to develop MIR against any of the herbivores. Thus, we demonstrate that hormone crosstalk is central to the priming of plant immunity by beneficial microbes, with ET fine-tuning being essential for the primed JA biosynthesis and boosted defenses leading to MIR in tomato.

PMID:39921876 | DOI:10.1093/jxb/eraf053

STAR Protoc. 2025 Feb 7;6(1):103614. doi: 10.1016/j.xpro.2025.103614. Online ahead of print.

ABSTRACT

Owing to inconsistencies in human B cell classification and the difficulty in distinguishing heterogeneous subpopulations, we present a protocol to construct gene regulatory networks and gene activity landscapes for human B cell developmental stages. We describe steps for acquiring bone marrow data; conducting single-cell downstream analysis; and leveraging the St. Jude Algorithm for the Reconstruction of Accurate Cellular Networks (SJARACNe), Network-based Bayesian Inference of Drivers (NetBID2), and single-cell Mutual Information-based Network Engineering Ranger (scMINER) algorithms for network-based analysis. Our protocol elucidates the biological characteristics of developmental stages in human B cells. For complete details on the use and execution of this protocol, please refer to Huang et al.1.

PMID:39921864 | DOI:10.1016/j.xpro.2025.103614

STAR Protoc. 2025 Feb 6;6(1):103618. doi: 10.1016/j.xpro.2025.103618. Online ahead of print.

ABSTRACT

Network analysis is a powerful tool for investigating complex interactions between different microbial taxa within a community. We present a protocol to study the gut microbial community in a mouse model of breast cancer using a network-based approach. Here, we describe the procedures for tumor cell production and inoculation and 16S rRNA data processing. We then detail steps for constructing co-occurrence networks based on correlations between microbial abundances. For complete details on the use and execution of this protocol, please refer to Wu-Chuang et al.1.

PMID:39921860 | DOI:10.1016/j.xpro.2025.103618

Sci Rep. 2025 Feb 8;15(1):4783. doi: 10.1038/s41598-025-89272-x.

ABSTRACT

This study aims to identify and evaluate the most common drugs associated with the risks of autoimmune hepatitis (AIH) using the FDA Adverse Event Reporting System (FAERS) database. Adverse drug events (ADEs) associated with drug-induced AIH (DI-AIH) were retrieved from the FAERS database (January 2004-June 2024). Disproportionality analysis was performed to identify drugs significantly linked to DI-AIH, and time-to-onset (TTO) analyses were conducted to evaluate the timing and risk profiles of DI-AIH adverse reactions. Our study identified 2,511 ADEs linked to autoimmune hepatitis. Disproportionality analysis identified 22 drugs significantly associated with AIH risk, including 4 antibiotics, 3 antivirals, 4 cardiovascular drugs, 5 antitumor agents, 2 immunomodulators, 2 nonsteroidal anti-inflammatory drugs, and 1 drug each from the respiratory and nervous system categories. The highest DI-AIH risks were observed with minocycline (ROR = 53.97), nitrofurantoin (ROR = 57.02), and doxycycline (ROR = 16.12). Antitumor drugs had the shortest median TTO (77.00 days), whereas cardiovascular drugs exhibited the longest (668.30 days). Through a comprehensive analysis of the FAERS database, our study identified drugs strongly associated with AIH. Preventing DI-AIH requires careful drug selection and monitoring. This study provides evidence-based insights into implicated drugs, aiming to optimize clinical management and mitigate risks.

PMID:39922875 | DOI:10.1038/s41598-025-89272-x

Eur J Hosp Pharm. 2025 Feb 8:ejhpharm-2024-004352. doi: 10.1136/ejhpharm-2024-004352. Online ahead of print.

ABSTRACT

OBJECTIVES: This study addressed the gaps in the disclosure of statin-associated adverse drug reactions (ADRs) in China's official database and the inadequacy of cases reported relative to the population size in public ADR databases.

METHODS: To address these limitations, we conducted a retrospective trial-based analysis using data from Chinese journals to comprehensively assess statin-associated ADRs from 1992 to 2023, focusing on liver (2895 studies, n = 163 810) and muscle (2888 studies, n = 161 714) related outcomes.

RESULTS: For large sample size clinical trial analysis (n≥100), our analysis encompassed data from 31 763 participants for muscle-related ADRs (incidence rate: 0.004-0.006, common effect model; 0.002-0.006, random effects model) and 31 281 participants for liver-related ADRs (incidence rate: 0.004-0.006, common effect model; 0.003-0.006, random effects model), covering various statins, including atorvastatin, simvastatin, rosuvastatin, fluvastatin, pitavastatin, pravastatin and lovastatin. Notably, muscle-related ADRs, particularly rhabdomyolysis, were most prevalent with fluvastatin, lovastatin and pravastatin, showing rates of 0.90%, 0.74% and 0.53%, respectively. Pitavastatin and atorvastatin were frequently associated with liver-related ADRs such as abnormal liver function and elevated enzymes, with rates of 5.36% and 1.819%, respectively.

CONCLUSIONS: This study underscores significant variations in ADR incidence among different statins in the Chinese population, providing critical insights for healthcare professionals and policymakers to enhance patient safety and optimise clinical decisions regarding statin therapy.

PMID:39922684 | DOI:10.1136/ejhpharm-2024-004352

Clin Exp Med. 2025 Feb 8;25(1):51. doi: 10.1007/s10238-025-01576-4.

ABSTRACT

Hepatocellular carcinoma (HCC) is the most common primary liver cancer. Sorafenib is the first FDA-approved systemic therapy for advanced HCC. This study investigates the influence of IL-23R (rs7517847) and ATG-10 (rs10514231) genetic polymorphisms on Sorafenib response, survival outcomes, average tolerable dose, and adverse events. This prospective open-label cohort study included 100 HCC patients, assessing IL-23R and ATG-10 genotypes via real-time polymerase chain reaction (RT-PCR). Patient's responses were evaluated using modified RECIST criteria. Statistical analyses evaluated the association of genetic variants with response, progression-free survival (PFS), overall survival (OS), average tolerable Sorafenib dose, and adverse events. IL-23R TT carriers had the highest Sorafenib response rate (80%) compared to GT (13.3%) and GG (6.7%) (P = 0.021), while ATG-10 TT carriers had a 13.9-fold increased response likelihood (P = 0.001). The T allele in ATG-10 significantly predicted longer PFS (P = 0.025) and OS (P = 0.011), suggesting a potential prognostic role. IL-23R GG carriers received significantly higher Sorafenib doses than TT (P = 0.0174) and GT (P = 0.0227), whereas ATG-10 had no effect on dosage. However, its CT genotype was significantly associated with a higher risk of Hand-Foot Syndrome (P = 0.012), and independent of dose (P = 0.0018). IL-23R and ATG-10 polymorphisms influence Sorafenib response, survival, and tolerability in HCC patients. Genetic screening may improve personalized treatment strategies by optimizing Sorafenib efficacy and minimizing toxicity.This trial was registered on clinicaltrials.gov with registration number NCT06030895, registered on "September 11th, 2023," retrospectively.

PMID:39921803 | DOI:10.1007/s10238-025-01576-4

Clin Transl Sci. 2025 Feb;18(2):e70154. doi: 10.1111/cts.70154.

ABSTRACT

This trial aimed to identify the effects of providing pharmacogenomic (PGx) results and recommendations for patients with chronic pain treated in primary care practices compared to standard care. An open-label, prospective, largely virtual, type-2 hybrid effectiveness trial randomized participants to PGx or standard care arms. Adults with pain ≥ 3 months who were treated with tramadol, codeine, or hydrocodone enrolled. Alternative analgesics were recommended for CYP2D6 intermediate or poor metabolizers (IM/PMs). Prescribing decisions were at providers' discretion. The trial randomized 253 participants. A modified intent-to-treat primary analysis assessed change in pain intensity over 3 months among IM/PMs (PGx: 49; Standard care: 57). The PGx and standard care arms showed no difference in pain intensity change (-0.10 ± 0.63 vs. -0.21 ± 0.75 standard deviation; p = 0.74) or PGx-aligned care (69% vs. 63%; standardized difference [SD] = 0.13). In IM/PMs, secondary analyses of pain intensity change suggested improvements with PGx-aligned (n = 70; -0.21 ± 0.70) vs. unaligned care (n = 36; -0.06 ± 0.69) (SD = -0.22), with this difference increasing when examining IM/PMs with an analgesic change (aligned: n = 31, -0.28 ± 0.76; unaligned: n = 36, -0.06 ± 0.69; SD = -0.31). This approach to PGx implementation for chronic pain was not associated with different prescribing (i.e., similar proportions of PGx-aligned care) or clinical outcomes. Secondary analyses suggest that prescribing aligned with PGx recommendations showed a small improvement in pain intensity. However, the proportion of patients with a clinically meaningful improvement (≥ 30%) in pain intensity was similar. Future efforts should identify effective implementation methods.

PMID:39921243 | DOI:10.1111/cts.70154

Respir Med. 2025 Feb 5:107982. doi: 10.1016/j.rmed.2025.107982. Online ahead of print.

ABSTRACT

BACKGROUND/RATIONALE: Lung transplant (LTx) candidates with cystic fibrosis (CF) have ventilatory and musculoskeletal limitations and benefit from pulmonary rehabilitation (PR). Their training response has not been well characterized. The study aims to: 1) characterize the effect of outpatient PR and 2) evaluate the clinical characteristics associated with their PR response.

METHODS: Single-center retrospective cohort study of CF LTx candidates (July 2009-June 2019) with available pre-transplant exercise data, who participated in PR 2 to 3 times/week until transplantation. Demographics, CF-related characteristics, aerobic and muscle training volumes, and six-minute walk distance (6MWD) were characterized using descriptive statistics, paired t-tests and Spearman correlations to describe relationships between CF-related characteristics and training volumes.

RESULTS: In 86 CF LTx candidates (32±10 years, 49% males, FEV1: 23±5%; listing 6MWD 421±89 meters), the median PR time was 87 days (24-36 sessions). 78% had at least one exacerbation and 55% required hospitalization. 88% used supplemental oxygen and 37% required home non-invasive ventilation. Treadmill speed (1.7±0.5 mph); biceps (50 IQR [40-70] lbs*reps) and quadriceps (30 IQR [30-40] lbs*reps) training volumes improved with PR (p< 0.05), whereas 6MWD remained unchanged. The presence of ≥ 1 respiratory exacerbation was associated with a lower progression in treadmill speed [-0.36 mph 95%CI (-0.67 to -0.04), p=0.028].

CONCLUSION: CF LTx candidates participating in PR increased treadmill speed and muscle training volumes, with preservation of 6MWD. Respiratory exacerbations were prevalent and important determinants of aerobic training.

PMID:39921065 | DOI:10.1016/j.rmed.2025.107982

Discov Oncol. 2025 Feb 8;16(1):135. doi: 10.1007/s12672-025-01908-6.

ABSTRACT

Breast cancer is a leading cause of mortality among women, with recurrence prediction remaining a significant challenge. In this context, artificial intelligence application and its resources can serve as a powerful tool in analyzing large amounts of data and predicting cancer recurrence, potentially enabling personalized medical treatment and improving the patient's quality of life. Thus, the systematic review examines the role of AI in predicting breast cancer recurrence using clinical data, imaging data, and combined datasets. Support Vector Machine (SVM) and Neural Networks, especially when applied to combined data, demonstrate strong potential in improving prediction accuracy. SVMs are effective with high-dimensional clinical data, while Neural Networks in genetic and molecular analysis. Despite these advancements, limitations such as dataset diversity, sample size, and evaluation standardization persist, emphasizing the need for further research. AI integration in recurrence prediction offers promising prospects for personalized care but requires rigorous validation for safe clinical application.

PMID:39921795 | DOI:10.1007/s12672-025-01908-6

Discov Oncol. 2025 Feb 8;16(1):137. doi: 10.1007/s12672-025-01877-w.

ABSTRACT

Cancer is a major global health challenge, with approximately 19.3 million new cases and 10 million deaths estimated by 2020. Laboratory advancements in cancer detection have transformed diagnostic capabilities, particularly through the use of biomarkers that play crucial roles in risk assessment, therapy selection, and disease monitoring. Tumor histology, single-cell technology, flow cytometry, molecular imaging, liquid biopsy, immunoassays, and molecular diagnostics have emerged as pivotal tools for cancer detection. The integration of artificial intelligence, particularly deep learning and convolutional neural networks, has enhanced the diagnostic accuracy and data analysis capabilities. However, developing countries face significant challenges including financial constraints, inadequate healthcare infrastructure, and limited access to advanced diagnostic technologies. The impact of COVID-19 has further complicated cancer management in resource-limited settings. Future research should focus on precision medicine and early cancer diagnosis through sophisticated laboratory techniques to improve prognosis and health outcomes. This review examines the evolving landscape of cancer detection, focusing on laboratory research breakthroughs and limitations in developing countries, while providing recommendations for advancing tumor diagnostics in resource-constrained environments.

PMID:39921787 | DOI:10.1007/s12672-025-01877-w

Magn Reson Med. 2025 Feb 8. doi: 10.1002/mrm.30437. Online ahead of print.

ABSTRACT

PURPOSE: Commonly used MR image quality (IQ) metrics have poor concordance with radiologist-perceived diagnostic IQ. Here, we develop and explore deep feature distances (DFDs)-distances computed in a lower-dimensional feature space encoded by a convolutional neural network (CNN)-as improved perceptual IQ metrics for MR image reconstruction. We further explore the impact of distribution shifts between images in the DFD CNN encoder training data and the IQ metric evaluation.

METHODS: We compare commonly used IQ metrics (PSNR and SSIM) to two "out-of-domain" DFDs with encoders trained on natural images, an "in-domain" DFD trained on MR images alone, and two domain-adjacent DFDs trained on large medical imaging datasets. We additionally compare these with several state-of-the-art but less commonly reported IQ metrics, visual information fidelity (VIF), noise quality metric (NQM), and the high-frequency error norm (HFEN). IQ metric performance is assessed via correlations with five expert radiologist reader scores of perceived diagnostic IQ of various accelerated MR image reconstructions. We characterize the behavior of these IQ metrics under common distortions expected during image acquisition, including their sensitivity to acquisition noise.

RESULTS: All DFDs and HFEN correlate more strongly with radiologist-perceived diagnostic IQ than SSIM, PSNR, and other state-of-the-art metrics, with correlations being comparable to radiologist inter-reader variability. Surprisingly, out-of-domain DFDs perform comparably to in-domain and domain-adjacent DFDs.

CONCLUSION: A suite of IQ metrics, including DFDs and HFEN, should be used alongside commonly-reported IQ metrics for a more holistic evaluation of MR image reconstruction perceptual quality. We also observe that general vision encoders are capable of assessing visual IQ even for MR images.

PMID:39921580 | DOI:10.1002/mrm.30437

Adv Sci (Weinh). 2025 Feb 8:e2412488. doi: 10.1002/advs.202412488. Online ahead of print.

ABSTRACT

Novel antifungal drugs that evade resistance are urgently needed for Candida infections. Antifungal peptides (AFPs) are potential candidates due to their specific mechanism of action, which makes them less prone to developing drug resistance. An AFP de novo design method, Deep Learning-Quantitative Structure‒Activity Relationship Empirical Screening (DL-QSARES), is developed by integrating deep learning and quantitative structure‒activity relationship empirical screening. After generating candidate AFPs (c_AFPs) through the recombination of dominant amino acids and dipeptide compositions, natural language processing models are utilized and quantitative structure‒activity relationship (QSAR) approaches based on physicochemical properties to screen for promising c_AFPs. Forty-nine promising c_AFPs are screened, and their minimum inhibitory concentrations (MICs) against C. albicans are determined to be 3.9-125 µg mL-1, of which four leading c_AFPs (AFP-8, -10, -11, and -13) has MICs of <10 µg mL-1 against the four tested pathogenic fungi, and AFP-13 has excellent therapeutic efficacy in the animal model.

PMID:39921483 | DOI:10.1002/advs.202412488

Adv Sci (Weinh). 2025 Feb 8:e2412987. doi: 10.1002/advs.202412987. Online ahead of print.

ABSTRACT

Studying the molecular properties of drugs and their interactions with human targets aids in better understanding the clinical performance of drugs and guides drug development. In computer-aided drug discovery, it is crucial to utilize effective molecular feature representations for predicting molecular properties and designing ligands with high binding affinity to targets. However, designing an effective multi-task and self-supervised strategy remains a significant challenge for the pretraining framework. In this study, a multi-task self-supervised deep learning framework is proposed, MTSSMol, which utilizes ≈10 million unlabeled drug-like molecules for pretraining to identify potential inhibitors of fibroblast growth factor receptor 1 (FGFR1). During the pretraining of MTSSMol, molecular representations are learned through a graph neural networks (GNNs) encoder. A multi-task self-supervised pretraining strategy is proposed to fully capture the structural and chemical knowledge of molecules. Extensive computational tests on 27 datasets demonstrate that MTSSMol exhibits exceptional performance in predicting molecular properties across different domains. Moreover, MTSSMol's capability is validated to identify potential inhibitors of FGFR1 through molecular docking using RoseTTAFold All-Atom (RFAA) and molecular dynamics simulations. Overall, MTSSMol provides an effective algorithmic framework for enhancing molecular representation learning and identifying potential drug candidates, offering a valuable tool to accelerate drug discovery processes. All of the codes are freely available online at https:// github.com/zhaoqi106/MTSSMol.

PMID:39921455 | DOI:10.1002/advs.202412987

Adv Sci (Weinh). 2025 Feb 7:e2413805. doi: 10.1002/advs.202413805. Online ahead of print.

ABSTRACT

Artificial intelligence (AI) in science is a key area of modern research. However, many current machine learning methods lack interpretability, making it difficult to grasp the physical mechanisms behind various phenomena, which hampers progress in related fields. This study focuses on the Poisson's ratio of a hexagonal lattice elastic network as it varies with structural deformation. By employing the Kolmogorov-Arnold Network (KAN), the transition of the network's Poisson's ratio from positive to negative as the hexagonal structural element shifts from a convex polygon to a concave polygon was accurately predicted. The KAN provides a clear mathematical framework that describes this transition, revealing the connection between the Poisson's ratio and the geometric properties of the hexagonal element, and accurately identifying the geometric parameters at which the Poisson's ratio equals zero. This work demonstrates the significant potential of the KAN network to clarify the mathematical relationships that underpin physical responses and structural behaviors.

PMID:39921316 | DOI:10.1002/advs.202413805

Int J Med Robot. 2025 Feb;21(1):e70043. doi: 10.1002/rcs.70043.

ABSTRACT

BACKGROUND: This research aims to use deep learning to create automated systems for better breast cancer detection and categorisation in mammogram images, helping medical professionals overcome challenges such as time consumption, feature extraction issues and limited training models.

METHODS: This research introduced a Lightweight Multihead attention Gannet Convolutional Neural Network (LMGCNN) to classify mammogram images effectively. It used wiener filtering, unsharp masking, and adaptive histogram equalisation to enhance images and remove noise, followed by Grey-Level Co-occurrence Matrix (GLCM) for feature extraction. Ideal feature selection is done by a self-adaptive quantum equilibrium optimiser with artificial bee colony.

RESULTS: The research assessed on two datasets, CBIS-DDSM and MIAS, achieving impressive accuracy rates of 98.2% and 99.9%, respectively, which highlight the superior performance of the LMGCNN model while accurately detecting breast cancer compared to previous models.

CONCLUSION: This method illustrates potential in aiding initial and accurate breast cancer detection, possibly leading to improved patient outcomes.

PMID:39921233 | DOI:10.1002/rcs.70043