Int J Mol Sci. 2025 Apr 5;26(7):3419. doi: 10.3390/ijms26073419.

ABSTRACT

Recent studies show that patients with Alzheimer's disease (AD) harbor specific methylation marks in the brain that, if accessible, could be used as epigenetic biomarkers. Liquid biopsy enables the study of circulating cell-free DNA (cfDNA) fragments originated from dead cells, including neurons affected by neurodegenerative processes. Here, we isolated and epigenetically characterized plasma cfDNA from 35 patients with AD and 35 cognitively healthy controls by using the Infinium® MethylationEPIC BeadChip array. Bioinformatics analysis was performed to identify differential methylation positions (DMPs) and regions (DMRs), including APOE ε4 genotype stratified analysis. Plasma pTau181 (Simoa) and cerebrospinal fluid (CSF) core biomarkers (Fujirebio) were also measured and correlated with differential methylation marks. Validation was performed with bisulfite pyrosequencing and bisulfite cloning sequencing. Epigenome-wide cfDNA analysis identified 102 DMPs associated with AD status. Most DMPs correlated with clinical cognitive and functional tests including 60% for Mini-Mental State Examination (MMSE) and 80% for Global Deterioration Scale (GDS), and with AD blood and CSF biomarkers. In silico functional analysis connected 30 DMPs to neurological processes, identifying key regulators such as SPTBN4 and APOE genes. Several DMRs were annotated to genes previously reported to harbor epigenetic brain changes in AD (HKR1, ZNF154, HOXA5, TRIM40, ATG16L2, ADAMST2) and were linked to APOE ε4 genotypes. Notably, a DMR in the HKR1 gene, previously shown to be hypermethylated in the AD hippocampus, was validated in cfDNA from an orthogonal perspective. These results support the feasibility of studying cfDNA to identify potential epigenetic biomarkers in AD. Thus, liquid biopsy could improve non-invasive AD diagnosis and aid personalized medicine by detecting epigenetic brain markers in blood.

PMID:40244264 | DOI:10.3390/ijms26073419

Int J Mol Sci. 2025 Apr 1;26(7):3283. doi: 10.3390/ijms26073283.

ABSTRACT

Medical treatment of glioblastoma presents a significant challenge. A conventional medication has limited effectiveness, and a single-target therapy is usually effective only in the early stage of the treatment. Recently, there has been increasing focus on multi-target therapies, but the vast range of possible combinations makes clinical experimentation and implementation difficult. From the perspective of systems biology, this study conducted simulations for multi-target glioblastoma therapy based on dynamic analysis of previously established endogenous networks, validated with glioblastoma single-cell RNA sequencing data. Several potentially effective target combinations were identified. The findings also highlight the necessity of multi-target rather than single-target intervention strategies in cancer treatment, as well as the promise in clinical applications and personalized therapies.

PMID:40244148 | DOI:10.3390/ijms26073283

Int J Mol Sci. 2025 Apr 1;26(7):3266. doi: 10.3390/ijms26073266.

ABSTRACT

Metacestodes of Echinococcus multilocularis are the causative agents of alveolar echinococcosis, a neglected, life-threatening, zoonotic disease. To study these metacestodes in vitro, a model system using a culture medium conditioned by rat hepatoma cells is available. A key question is how the parasite interacts with the host and, in particular, which host-derived compounds are taken up. In this study, we focus on the uptake of host-derived proteins. Studies with artificially labeled proteins suggest that this uptake may occur independently of protein size or charge. Closer investigation using proteomics draws, however, a different picture. Of 1170 host (i.e., rat or bovine) proteins as identified by LC-MS/MS-based proteomics present in the culture medium, only 225 are found in metacestode vesicle tissue or fluid. Moreover, their relative abundances differ. Serum albumin, the most abundant culture medium host protein, is only the third most abundant protein in vesicle fluid, where Alpha-2-HS-glycoprotein becomes the most abundant protein. In vesicle fluid obtained ex vivo from experimentally infected mice, the situation is again different, with histone isoforms as the most abundant proteins. This suggests that while maintaining their internal milieu constant, metacestodes may adjust the spectrum of host proteins taken up. Potential uptake mechanisms and functions are discussed.

PMID:40244114 | DOI:10.3390/ijms26073266

Int J Mol Sci. 2025 Mar 28;26(7):3116. doi: 10.3390/ijms26073116.

ABSTRACT

The utilization of photosynthetic microbes, such as cyanobacteria and microalgae, offers sustainable solutions to addressing global resource shortages and pollution. While these microorganisms have demonstrated significant potential in biomanufacturing, their industrial application is limited by suboptimal photosynthetic efficiency. Synthetic biology integrates molecular biology, systems biology, and engineering principles to provide a powerful tool for elucidating photosynthetic mechanisms and rationally optimizing photosynthetic platforms. This review summarizes recent advancements in regulating photosynthesis in cyanobacteria and microalgae via synthetic biology, focusing on strategies to enhance light energy absorption, optimize electron transport chains, and improve carbon assimilation. Furthermore, we discuss key challenges in translating these genetic modifications to large-scale bioproduction, highlighting specific bottlenecks in strain stability, metabolic burden, and process scalability. Finally, we propose potential solutions, such as AI-assisted metabolic engineering, synthetic microbial consortia, and next-generation photobioreactor designs, to overcome these limitations. Overall, while synthetic biology holds great promise for enhancing photosynthetic efficiency in cyanobacteria and microalgae, further research is needed to refine genetic strategies and develop scalable production systems.

PMID:40243859 | DOI:10.3390/ijms26073116

Int J Mol Sci. 2025 Mar 25;26(7):2958. doi: 10.3390/ijms26072958.

ABSTRACT

The phosphate starvation response (PHR) transcription factor family play central regulatory roles in nutrient signaling, but its relationship with other abiotic stress remains elusive. In the woody plant Eucalyptus grandis, we characterized 12 EgPHRs, which were phylogenetically divided into three groups, with group I exhibiting conserved structural features (e.g., unique motif composition and exon number). Notably, a protein-protein interaction network analysis revealed that EgPHR had a species-specific protein-protein interaction network: EgPHR6 interacted with SPX proteins of multiple species, while Eucalyptus and poplar PHR uniquely bound to TRARAC-kinesin ATPase, suggesting functional differences between woody and herbaceous plants. A promoter sequence analysis revealed a regulatory network of 59 transcription factors (TFs, e.g., BPC, MYBs, ERFs and WUS), mainly associated with tissue differentiation, abiotic stress, and hormonal responses that regulated EgPHRs' expression. Transcriptomics and RT-qPCR gene expression analyses showed that all EgPHRs dynamically responded to phosphate (Pi) starvation, with the expression of EgPHR2 and EgPHR6 exhibiting sustained induction, and were also regulated by salt, cold, jasmonic acid, and boron deficiency. Strikingly, nitrogen starvation suppressed most EgPHRs, highlighting crosstalk between nutrient signaling pathways. These findings revealed the multifaceted regulatory role of EgPHRs in adaptation to abiotic stresses and provided insights into their unique evolutionary and functional characteristics in woody plants.

PMID:40243569 | DOI:10.3390/ijms26072958

Int J Mol Sci. 2025 Mar 22;26(7):2899. doi: 10.3390/ijms26072899.

ABSTRACT

The inevitable exposure of humans to micro/nanoplastics has become a pressing global environmental issue, with growing concerns regarding their impact on health. While the direct effects of micro/nanoplastics on human health remain largely unknown, increasing attention is being given to their potential role as carriers of environmental pollutants and organic substances. This study investigates the direct toxicity of 500 nm polystyrene nanoplastics (NPs) on human hepatocytes (HepG2) in vitro, both alone and in combination with cadmium (Cd), a hazardous heavy metal and a prevalent environmental pollutant. One-hour exposure to 100 µg/mL of NPs causes a significant increase in ROS production (+25% compared to control) but cell viability remains unaffected even at concentrations much higher than environmental levels. Interestingly, NPs significantly reduce Cd cytotoxicity at LC50 concentrations (cell viability compared to control: 55.4% for 50 µM Cd, 66.9% for 50 µM Cd + 10 µg/mL NPs, 68.4% for 50 µM Cd + 100 µg/mL NPs). Additionally, NPs do not alter the cellular lipid content after short-term exposure (24 h). However, when Cd and fatty acids are added to the medium, NPs appear to sequester fatty acids, reducing their availability and impairing their uptake by cells in a dose-dependent manner. We confirmed by Dynamic Light Scattering and Scanning Electron Microscopy the interaction between NPs, Cd and free fatty acids. Although polystyrene NPs exhibited minimal cytotoxicity in our experimental model, collectively our findings suggest that predicting the effects of cell exposure to NPs is extremely challenging, due to the potential interaction between NPs, environmental pollutants and specific components of the biological matrix.

PMID:40243532 | DOI:10.3390/ijms26072899

Microbiol Immunol. 2025 Apr 17. doi: 10.1111/1348-0421.13219. Online ahead of print.

ABSTRACT

Hepatocellular carcinoma (HCC) is the primary form of liver cancer that poses a significant global health concern due to its increasing incidence rates and diverse etiology. Chronic infection induced by hepatitis B virus (HBV) is a prominent etiological factor influencing the development of HCC. Although recent advances in multi-omics approaches have facilitated extensive exploration of HCC molecular characteristics, translating the characteristics of subtypes into clinical applications has been challenging due to parameters like limited sample size and complex classifiers for early detection. In the present study, we performed transcriptomics profiling of HBV-infected HCC patient tissue data to gather comprehensive insights into the intricate molecular mechanisms underlying HBV-associated HCC, specifically, viral protein interactions that influence the expression of oncogenes. The 1059 differentially expressed genes (DEGs) identified across two GEO data sets revealed upregulation of cell cycle and mitosis-related genes, alongside downregulation of genes involved in fatty acid degradation and cytochrome P450 activity. CDK1 and CDC20 which are part of the top cluster and hub gene from interactome analysis were identified as potential markers for HBV-positive HCC through gene expression pattern and overall survival analysis. Additionally, 19 DEGs showing significance in HCC development were identified as interacting partners with HBV proteins. Among them, the interaction of HBsAg with ALB and SHBG and their downregulation correlates to the lower testosterone levels identified in HBV and HCC patients. Together, the study enhances the understanding of the heterogeneity and molecular pathogenesis of HBV-positive HCC.

PMID:40243270 | DOI:10.1111/1348-0421.13219

Front Public Health. 2025 Apr 2;13:1484286. doi: 10.3389/fpubh.2025.1484286. eCollection 2025.

ABSTRACT

The participation of patient organizations in the construction of medical security systems for rare diseases is an important proposition in the theory of welfare pluralism, which indicates that individualized services and diversified connections of patient organizations can fully meet the multiple medical insurance needs of patients with rare diseases. In actual practice, however, patient organizations continue to experience institutional and structural difficulties, such as improper supervision, insufficient capacity and insufficient coordination. The theory of welfare pluralism and the developmental experience of foreign rare disease organizations have suggested that patient organizations should clarify the bridge positioning and specialized development path in the multi-subject cooperation of organizations providing medical security for rare diseases, thereby improving the efficient implementation of medical security policies for rare diseases.

PMID:40241973 | PMC:PMC11999925 | DOI:10.3389/fpubh.2025.1484286

Exp Ther Med. 2025 Apr 4;29(6):110. doi: 10.3892/etm.2025.12860. eCollection 2025 Jun.

ABSTRACT

TAM (TYRO3, AXL, MERTK) receptor tyrosine kinases (RTKs) have intrinsic roles in tumor cell proliferation, migration, chemoresistance, and suppression of antitumor immunity. The overexpression of TAM RTKs is associated with poor prognosis in various types of cancer. Single-target agents of TAM RTKs have limited efficacy because of an adaptive feedback mechanism resulting from the cooperation of TAM family members. This suggests that multiple targeting of members has the potential for a more potent anticancer effect. The present study used a deep-learning based drug-target interaction (DTI) prediction model called molecule transformer-DTI (MT-DTI) to identify commercially available drugs that may inhibit the three members of TAM RTKs. The results showed that fostamatinib, a spleen tyrosine kinase (Syk) inhibitor, could inhibit the three receptor kinases of the TAM family with an IC50 <1 µM. Notably, no other Syk inhibitors were predicted by the MT-DTI model. To verify this result, this study performed in vitro studies with various types of cancer cell lines. Consistent with the DTI results, this study observed that fostamatinib suppressed cell proliferation by inhibiting TAM RTKs, while other Syk inhibitors showed no inhibitory activity. These results suggest that fostamatinib could exhibit anticancer activity as a pan-TAM inhibitor. Taken together, these findings demonstrated that this artificial intelligence model could be effectively used for drug repurposing and repositioning. Furthermore, by identifying its novel mechanism of action, this study confirmed the potential for fostamatinib to expand its indications as a TAM inhibitor.

PMID:40242601 | PMC:PMC12001310 | DOI:10.3892/etm.2025.12860

Cancer Res Treat. 2025 Apr 15. doi: 10.4143/crt.2025.033. Online ahead of print.

ABSTRACT

PURPOSE: Triple-negative breast cancer (TNBC) is a particularly aggressive subtype of breast cancer, with approximately 30% of patients eventually developing brain metastases (BM), which result in poor outcomes. An understanding of the tumor microenvironment (TME) at both primary and metastatic sites offers insights into the mechanisms underlying BM and potential therapeutic targets.

MATERIALS AND METHOD: Spatial RNA sequencing (spRNA-seq) was performed on primary TNBC and paired BM tissues from three patients, one of whom had previously received immune checkpoint inhibitors before BM diagnosis. Specimen regions were categorized into tumor, proximal, and distal TME based on their spatial locations. Gene expression differences across these zones were analyzed, and immune cell infiltration was estimated using TIMER. A gene module analysis was conducted to identify key gene clusters associated with BM.

RESULTS: Distinct gene expression profiles were noted in the proximal and distal TMEs. In BM, the proximal TME exhibited neuronal gene expression, suggesting neuron-tumor interactions compared to tumor, and upregulation of epithelial genes compared to the distal TME. Immune cell analysis revealed dynamic changes in CD8+ T cells and macrophages across the tumor and TME zones. Gene module analysis identified five key modules, including one related to glycolysis, which correlated with patient survival. Drug repurposing analysis identified potential therapeutic targets, including VEGFA, RAC1, EGLN3, and CAMK1D.

CONCLUSION: This study provides novel insights into the transcriptional landscapes in TNBC BM using spRNA-seq, emphasizing the role of neuron-tumor interactions and immune dynamics. These findings suggest new therapeutic strategies and underscore the importance of further research.

PMID:40241579 | DOI:10.4143/crt.2025.033

Comput Struct Biotechnol J. 2025 Apr 3;27:1440-1448. doi: 10.1016/j.csbj.2025.04.004. eCollection 2025.

ABSTRACT

The generation of omics data sets has become an important approach in modern pharmacological and toxicological research as it can provide mechanistic and quantitative information on a large scale. Analyses of these data frequently revealed a non-linear dose-response relationship underscoring the importance of the modeling process to infer biological exposure limits. A number of tools have been developed for dose-response modeling and various thresholds have been defined as a quantitative representation of the effect of a substance, such as effective concentrations or benchmark doses (BMD). Here we present DoseRider an easy-to-use web application and a companion R package for linear and non-linear dose-response modeling and assessment of BMD at the level of biological pathways or signatures using generalized mixed effect models. This approach allows to analyze custom or provided multi-omics data such as RNA sequencing or metabolomics data and its annotation of a collection of pathways and gene sets from various species. Moreover, we introduce the concept of the trend change doses (TCDs) as a numerical descriptor of effects derived from complex dose-response curves. The usability of DoseRider was demonstrated by analyses of RNA sequencing data of bisphenol AF (BPAF) treatment of a human breast cancer cell line (MCF-7) at 8 different concentrations using gene sets for chemical and genetic perturbations (MSigDB). The BMD for BPAF and a set of genes upregulated by estrogen in breast cancer was 0.2 µM (95 %-CI 0.1-0.5 µM) and the lowest TCD (TCD1) was 0.003 µM (95 %-CI 0.0006-0.01 µM). The comprehensive presentation of the results underlines the suitability of the system for pharmacogenomics, toxicogenomics, and applications beyond.

PMID:40242291 | PMC:PMC12001094 | DOI:10.1016/j.csbj.2025.04.004

Health Care Sci. 2025 Apr 8;4(2):82-93. doi: 10.1002/hcs2.70008. eCollection 2025 Apr.

ABSTRACT

BACKGROUND: Precision medicine (PM) has taken center stage in healthcare since the completion of the genomic project. Developed countries have gradually integrated PM into mainstream patient management. However, Nigeria still grapples with wide acceptance, key translational research and implementation of PM. This study sought to explore the knowledge and attitude of PM among pharmacists as key stakeholders in the healthcare team.

METHODS: A cross-sectional study was conducted in selected tertiary hospitals across the country. A 21-item semi-structured questionnaire was administered by hybrid online and physical methods and the results analyzed with Statistical Package for the Social Sciences Version 25. Descriptive statistics were used to summarize the data. A chi-square test was employed to determine the association of knowledge of PM and the sociodemographic characteristics of the study population.

RESULTS: A total of 167 hospital pharmacists participated in the study. A high proportion of the participants are familiar with artificial intelligence (91.75%), Pharmacogenomics (84.5%), and precision medicine (61%). Overall, 38.9% of the pharmacists had a good knowledge while 13.2% had a poor knowledge of PM and associated terms. The level of knowledge did not correlate significantly with gender (X 2 = 3.21, p = 0.201), age (X 2 = 5, p = 0.27), marital status (X 2 = 3.21, p = 0.201), and professional level (X 2 = 6.85, p = 0.144). The most important value of precision medicine to hospital pharmacists is the ability to minimize the impact of disease through preventive medicine (49%) while a large portion are pursuing and or actively planning to pursue additional education in precision medicine.

CONCLUSIONS: There is a highly positive attitude toward the prospect of PM among hospital pharmacists in Nigeria. Education modules in this field are highly recommended as most do not have a holistic knowledge of terms used in PM. Also, more research aimed at translating PM knowledge into clinical practice is recommended.

PMID:40241984 | PMC:PMC11997455 | DOI:10.1002/hcs2.70008

Respirol Case Rep. 2025 Apr 15;13(4):e70185. doi: 10.1002/rcr2.70185. eCollection 2025 Apr.

ABSTRACT

The coexistence of trisomy 21 and cystic fibrosis (CF) is extremely rare, with fewer than 10 reported cases, all involving homozygous CFTR mutations. However, the impact of a heterozygous CFTR mutation in a patient with trisomy 21 remains unexplored. We present a male infant with trisomy 21 who experienced recurrent respiratory distress and was later found to carry a heterozygous pathogenic CFTR mutation (p.Phe508del). His respiratory complications were severe, requiring tracheostomy and long-term respiratory support. This case highlights the potential interplay between trisomy 21-associated anatomical features and CFTR-related airway abnormalities, possibly exacerbating respiratory morbidity. Given the high burden of respiratory complications in both conditions, clinicians should consider CFTR-related disorders in patients with trisomy 21 presenting with severe respiratory issues. Further research is warranted to determine the clinical significance of CFTR heterozygosity in trisomy 21 and its implications for disease severity and management.

PMID:40241998 | PMC:PMC12000535 | DOI:10.1002/rcr2.70185

Phys Imaging Radiat Oncol. 2025 Mar 26;34:100749. doi: 10.1016/j.phro.2025.100749. eCollection 2025 Apr.

ABSTRACT

BACKGROUND AND PURPOSE: In order to optimize the radiotherapy treatment and minimize toxicities, organs-at-risk (OARs) and clinical target volume (CTV) must be segmented. Deep Learning (DL) techniques show significant potential for performing this task effectively. The availability of a large single-institute data sample, combined with additional numerous multi-centric data, makes it possible to develop and validate a reliable CTV segmentation model.

MATERIALS AND METHODS: Planning CT data of 1822 patients were available (861 from a single center for training and 961 from 8 centers for validation). A preprocessing step, aimed at standardizing all the images, followed by a 3D-Unet capable of segmenting both right and left CTVs was implemented. The metrics used to evaluate the performance were the Dice similarity coefficient (DSC), the Hausdorff distance (HD), and its 95th percentile variant (HD_95) and the Average Surface Distance (ASD).

RESULTS: The segmentation model achieved high performance on the validation set (DSC: 0.90; HD: 20.5 mm; HD_95: 10.0 mm; ASD: 2.1 mm; epoch 298). Furthermore, the model predicted smoother contours than the clinical ones along the cranial-caudal axis in both directions. When applied to internal and external data the same metrics demonstrated an overall agreement and model transferability for all but one (Inst 9) center.

CONCLUSION: . A 3D-Unet for CTV segmentation trained on a large single institute cohort consisting of planning CTs and manual segmentations was built and externally validated, reaching high performance.

PMID:40242807 | PMC:PMC12002654 | DOI:10.1016/j.phro.2025.100749

KDD. 2025 Aug;2025(v1):2779-2790. doi: 10.1145/3690624.3709402. Epub 2025 Jul 20.

ABSTRACT

Sepsis is an organ dysfunction caused by a deregulated immune response to an infection. Early sepsis prediction and identification allow for timely intervention, leading to improved clinical outcomes. Clinical calculators (e.g., the six-organ dysfunction assessment of SOFA in Figure 1) play a vital role in sepsis identification within clinicians' workflow, providing evidence-based risk assessments essential for sepsis diagnosis. However, artificial intelligence (AI) sepsis prediction models typically generate a single sepsis risk score without incorporating clinical calculators for assessing organ dysfunctions, making the models less convincing and transparent to clinicians. To bridge the gap, we propose to mimic clinicians' workflow with a novel framework SepsisCalc to integrate clinical calculators into the predictive model, yielding a clinically transparent and precise model for utilization in clinical settings. Practically, clinical calculators usually combine information from multiple component variables in Electronic Health Records (EHR), and might not be applicable when the variables are (partially) missing. We mitigate this issue by representing EHRs as temporal graphs and integrating a learning module to dynamically add the accurately estimated calculator to the graphs. Experimental results on real-world datasets show that the proposed model outperforms state-of-the-art methods on sepsis prediction tasks. Moreover, we developed a system to identify organ dysfunctions and potential sepsis risks, providing a human-AI interaction tool for deployment, which can help clinicians understand the prediction outputs and prepare timely interventions for the corresponding dysfunctions, paving the way for actionable clinical decision-making support for early intervention.

PMID:40242786 | PMC:PMC11998859 | DOI:10.1145/3690624.3709402

Exp Ther Med. 2025 Apr 4;29(6):110. doi: 10.3892/etm.2025.12860. eCollection 2025 Jun.

ABSTRACT

TAM (TYRO3, AXL, MERTK) receptor tyrosine kinases (RTKs) have intrinsic roles in tumor cell proliferation, migration, chemoresistance, and suppression of antitumor immunity. The overexpression of TAM RTKs is associated with poor prognosis in various types of cancer. Single-target agents of TAM RTKs have limited efficacy because of an adaptive feedback mechanism resulting from the cooperation of TAM family members. This suggests that multiple targeting of members has the potential for a more potent anticancer effect. The present study used a deep-learning based drug-target interaction (DTI) prediction model called molecule transformer-DTI (MT-DTI) to identify commercially available drugs that may inhibit the three members of TAM RTKs. The results showed that fostamatinib, a spleen tyrosine kinase (Syk) inhibitor, could inhibit the three receptor kinases of the TAM family with an IC50 <1 µM. Notably, no other Syk inhibitors were predicted by the MT-DTI model. To verify this result, this study performed in vitro studies with various types of cancer cell lines. Consistent with the DTI results, this study observed that fostamatinib suppressed cell proliferation by inhibiting TAM RTKs, while other Syk inhibitors showed no inhibitory activity. These results suggest that fostamatinib could exhibit anticancer activity as a pan-TAM inhibitor. Taken together, these findings demonstrated that this artificial intelligence model could be effectively used for drug repurposing and repositioning. Furthermore, by identifying its novel mechanism of action, this study confirmed the potential for fostamatinib to expand its indications as a TAM inhibitor.

PMID:40242601 | PMC:PMC12001310 | DOI:10.3892/etm.2025.12860

Fundam Res. 2023 Jun 18;5(2):880-889. doi: 10.1016/j.fmre.2023.04.018. eCollection 2025 Mar.

ABSTRACT

Mammography is the mainstream imaging modality used for breast cancer screening. Identification of microcalcifications associated with malignancy may result in early diagnosis of breast cancer and aid in reducing the morbidity and mortality associated with the disease. Computer-aided diagnosis (CAD) is a promising technique due to its efficiency and accuracy. Here, we demonstrated that an automated deep-learning pipeline for microcalcification detection and classification on mammography can facilitate early diagnosis of breast cancer. This technique can not only provide the classification results of mammography, but also annotate specific calcification regions. A large mammography dataset was collected, including 4,810 mammograms with 6,663 microcalcification lesions based on biopsy results, of which 3,301 were malignant and 3,362 were benign. The system was developed and tested using images from multiple centers. The overall classification accuracy values for discriminating between benign and malignant breasts were 0.8124 for the training set and 0.7237 for the test set. The sensitivity values of malignant breast cancer prediction were 0.8891 for the training set and 0.7778 for the test set. In addition, we collected information regarding pathological sub-type (pathotype) and estrogen receptor (ER) status, and we subsequently explored the effectiveness of deep learning-based pathotype and ER classification. Automated artificial intelligence (AI) systems may assist clinicians in making judgments and improve their efficiency in breast cancer screening, diagnosis, and treatment.

PMID:40242534 | PMC:PMC11997558 | DOI:10.1016/j.fmre.2023.04.018

Inf Fusion. 2025 Aug;120:103079. doi: 10.1016/j.inffus.2025.103079. Epub 2025 Mar 20.

ABSTRACT

Multivariate time series classification (TSC) is critical for various applications in fields such as healthcare and finance. While various approaches for TSC have been explored, important properties of time series, such as shift equivariance and inversion invariance, are largely underexplored by existing works. To fill this gap, we propose a novel multi-view approach to capture patterns with properties like shift equivariance. Our method integrates diverse features, including spectral, temporal, local, and global features, to obtain rich, complementary contexts for TSC. We use continuous wavelet transform to capture time-frequency features that remain consistent even when the input is shifted in time. These features are fused with temporal convolutional or multilayer perceptron features to provide complex local and global contextual information. We utilize the Mamba state space model for efficient and scalable sequence modeling and capturing long-range dependencies in time series. Moreover, we introduce a new scanning scheme for Mamba, called tango scanning, to effectively model sequence relationships and leverage inversion invariance, thereby enhancing our model's generalization and robustness. Experiments on two sets of benchmark datasets (10+20 datasets) demonstrate our approach's effectiveness, achieving average accuracy improvements of 4.01-6.45% and 7.93% respectively, over leading TSC models such as TimesNet and TSLANet. The code is available at: https://drive.google.com/file/d/1fScmALgreb_sE9_P2kIsQCmt9SNxp7GP/view?usp=sharing.

PMID:40242510 | PMC:PMC11997873 | DOI:10.1016/j.inffus.2025.103079

Rep Pract Oncol Radiother. 2025 Mar 21;30(1):122-125. doi: 10.5603/rpor.104144. eCollection 2025.

NO ABSTRACT

PMID:40242416 | PMC:PMC11999009 | DOI:10.5603/rpor.104144

ACS ES T Water. 2025 Mar 24;5(4):1814-1820. doi: 10.1021/acsestwater.4c01186. eCollection 2025 Apr 11.

ABSTRACT

We present a novel approach for imaging diatoms using lensless imaging and deep learning. We used a laser beam to scatter off samples of diatomaceous earth (diatoms) and then recorded and transformed the scattered light into microscopy images of the diatoms. The predicted microscopy images gave an average SSIM of 0.98 and an average RMSE of 3.26 as compared to the experimental data. We also demonstrate the capability of determining the velocity and angle of movement of the diatoms from their scattering patterns as they were translated through the laser beam. This work shows the potential for imaging and identifying the movement of diatoms and other microsized organisms in situ within the marine environment. Implementing such a method for real-time image acquisition and analysis could enhance environmental management, including improving the early detection of harmful algal blooms.

PMID:40242343 | PMC:PMC11997998 | DOI:10.1021/acsestwater.4c01186