PLoS One. 2025 Jan 27;20(1):e0316536. doi: 10.1371/journal.pone.0316536. eCollection 2025.

ABSTRACT

Modernizing power systems into smart grids has introduced numerous benefits, including enhanced efficiency, reliability, and integration of renewable energy sources. However, this advancement has also increased vulnerability to cyber threats, particularly False Data Injection Attacks (FDIAs). Traditional Intrusion Detection Systems (IDS) often fall short in identifying sophisticated FDIAs due to their reliance on predefined rules and signatures. This paper addresses this gap by proposing a novel IDS that utilizes hybrid feature selection and deep learning classifiers to detect FDIAs in smart grids. The main objective is to enhance the accuracy and robustness of IDS in smart grids. The proposed methodology combines Particle Swarm Optimization (PSO) and Grey Wolf Optimization (GWO) for hybrid feature selection, ensuring the selection of the most relevant features for detecting FDIAs. Additionally, the IDS employs a hybrid deep learning classifier that integrates Convolutional Neural Networks (CNN) and Long Short-Term Memory (LSTM) networks to capture the smart grid data's spatial and temporal features. The dataset used for evaluation, the Industrial Control System (ICS) Cyber Attack Dataset (Power System Dataset) consists of various FDIA scenarios simulated in a smart grid environment. Experimental results demonstrate that the proposed IDS framework significantly outperforms traditional methods. The hybrid feature selection effectively reduces the dimensionality of the dataset, improving computational efficiency and detection performance. The hybrid deep learning classifier performs better in key metrics, including accuracy, recall, precision, and F-measure. Precisely, the proposed approach attains higher accuracy by accurately identifying true positives and minimizing false negatives, ensuring the reliable operation of smart grids. Recall is enhanced by capturing critical features relevant to all attack types, while precision is improved by reducing false positives, leading to fewer unnecessary interventions. The F-measure balances recall and precision, indicating a robust and reliable detection system. This study presents a practical dual-hybrid IDS framework for detecting FDIAs in smart grids, addressing the limitations of existing IDS techniques. Future research should focus on integrating real-world smart grid data for validation, developing adaptive learning mechanisms, exploring other bio-inspired optimization algorithms, and addressing real-time processing and scalability challenges in large-scale deployments.

PMID:39869576 | DOI:10.1371/journal.pone.0316536

PLoS One. 2025 Jan 27;20(1):e0317999. doi: 10.1371/journal.pone.0317999. eCollection 2025.

ABSTRACT

As people's material living standards continue to improve, the types and quantities of household garbage they generate rapidly increase. Therefore, it is urgent to develop a reasonable and effective method for garbage classification. This is important for resource recycling and environmental improvement and contributes to the sustainable development of production and the economy. However, existing deep learning-based garbage image classification models generally suffer from low classification accuracy, insufficient robustness, and slow detection speed due to the large number of model parameters. To this end, a new garbage image classification model is proposed, with the ResNet-50 network as the core architecture. Specifically, first, a redundancy-weighted feature fusion module is proposed, enabling the model to fully leverage valuable feature information, thereby improving its performance. At the same time, the module filters out redundant information from multi-scale features, reducing the number of model parameters. Second, the standard 3×3 convolutions in ResNet-50 are replaced with depth-separable convolutions, significantly improving the model's computational efficiency while preserving the feature extraction capability of the original convolutional structure. Finally, to address the issue of class imbalance, a weighting factor is added to the Focal Loss, aiming to mitigate the negative impact of class imbalance on model performance and enhance the model's robustness. Experimental results on the TrashNet dataset show that the proposed model effectively reduces the number of parameters, improves detection speed, and achieves an accuracy of 94.13%, surpassing the vast majority of existing deep learning-based waste image classification models, demonstrating its solid practical value.

PMID:39869568 | DOI:10.1371/journal.pone.0317999

PLoS One. 2025 Jan 27;20(1):e0315893. doi: 10.1371/journal.pone.0315893. eCollection 2025.

ABSTRACT

BACKGROUND: Ras-GTPase-activating protein (GAP)-binding protein 1 (G3BP1) emerges as a pivotal oncogenic gene across various malignancies, notably including nasopharyngeal carcinoma (NPC). The use of automated image analysis tools for immunohistochemical (IHC) staining of particular proteins is highly beneficial, as it could reduce the burden on pathologists. Interestingly, there have been no prior studies that have examined G3BP1 IHC staining using digital pathology.

METHODS: Whole-slide images (WSIs) were meticulously collected and annotated by experienced pathologists. A model was intricately designed and rigorously tested to yield the quantitative data regarding staining intensity and extent. The collective output data was subjected multiplicative analysis, exploring its correlation with the prognosis.

RESULTS: The G3BP1 molecular marker scoring model was successfully established utilizing deep learning methodologies, with a calculated threshold staining scores of 1.5. Notably, patients with NPC exhibiting higher expression levels of G3BP1 proteins displayed significantly lower for overall survival rates (OS). Multivariate analysis further validated that positive expression of G3BP1 stood as an independent poorer prognostic factors, indicating a poorer prognosis for NPC patients.

CONCLUSION: Computational pathology emerges as a transformative tool capable of substantially reducing the burden on pathologists while concurrently enhancing and diagnostic sensitivity and specificity. The positive expression of G3BP1 protein serves as valuable, independent biomarker, offering predictive insights into a poor prognosis for patients with NPC.

PMID:39869565 | DOI:10.1371/journal.pone.0315893

PLoS One. 2025 Jan 27;20(1):e0317450. doi: 10.1371/journal.pone.0317450. eCollection 2025.

ABSTRACT

In 2019, the novel coronavirus swept the world, exposing the monitoring and early warning problems of the medical system. Computer-aided diagnosis models based on deep learning have good universality and can well alleviate these problems. However, traditional image processing methods may lead to high false positive rates, which is unacceptable in disease monitoring and early warning. This paper proposes a low false positive rate disease detection method based on COVID-19 lung images and establishes a two-stage optimization model. In the first stage, the model is trained using classical gradient descent, and relevant features are extracted; in the second stage, an objective function that minimizes the false positive rate is constructed to obtain a network model with high accuracy and low false positive rate. Therefore, the proposed method has the potential to effectively classify medical images. The proposed model was verified using a public COVID-19 radiology dataset and a public COVID-19 lung CT scan dataset. The results show that the model has made significant progress, with the false positive rate reduced to 11.3% and 7.5%, and the area under the ROC curve increased to 92.8% and 97.01%.

PMID:39869555 | DOI:10.1371/journal.pone.0317450

PLoS One. 2025 Jan 27;20(1):e0317662. doi: 10.1371/journal.pone.0317662. eCollection 2025.

ABSTRACT

Patients with type 1 diabetes and their physicians have long desired a fully closed-loop artificial pancreas (AP) system that can alleviate the burden of blood glucose regulation. Although deep reinforcement learning (DRL) methods theoretically enable adaptive insulin dosing control, they face numerous challenges, including safety and training efficiency, which have hindered their clinical application. This paper proposes a safe and efficient adaptive insulin delivery controller based on DRL. It employed ten tricks to enhance the proximal policy optimization (PPO) algorithm, improving training efficiency. Additionally, a dual safety mechanism of 'proactive guidance + reactive correction' was introduced to reduce the risks of hyperglycemia and hypoglycemia and to prevent emergencies. Performance evaluations in the Simglucose simulator demonstrate that the proposed controller achieved an 87.45% time in range (TIR) median, superior to baseline methods, with a lower incidence of hypoglycemia, notably eliminating severe hypoglycemia and treatment failures. These encouraging results indicate that the DRL-based fully closed-loop AP controller has taken an essential step toward clinical implementation.

PMID:39869550 | DOI:10.1371/journal.pone.0317662

Acta Bioeng Biomech. 2025 Jan 27;26(3):123-134. doi: 10.37190/abb-02474-2024-02. Print 2024 Sep 1.

ABSTRACT

Purpose: Monitoring and assessing the level of lower limb motor skills using the Biodex System plays an important role in the training of football players and in post-traumatic rehabilitation. The aim of this study was to build and test an artificial intelligence-based model to assess the peak torque of the lower limb extensors and flexors. The model was based on real-world results in three groups: hearing (n = 19) and deaf football players (n = 28) and non-training deaf pupils (n = 46). Methods: The research used a 4-layer forward CNN neural network with two hidden layers with typical normalization for small data sets and Multilayer Perceptron (MLP) based on MatlabR2023a software with Neural Networks and Deep Learning toolkits and semiautomated learning algorithm selection using ML.NET. Results: The 70-90% accuracy shown in the article is sufficient here. AI provides a highly accurate, objective and efficient means of assessing neuromuscular performance, which can improve injury prevention and rehabilitation strategies. Conclusions: The high accuracy shows that AI-based models can help with this, but their wider practical implementation requires further cross-disciplinary research. AI, and in particular MLP and CNN can support both training methods and various gaming aspects. The contribution of the research is to use an innovative approach to derive computational rules/guidelines from an explicitly given dataset and then identify the relevant physiological torque of the lower limb extensors and flexors in the knee joint. The model complements existing methodologies for describing physiology of peak torque of lower limbs with using fuzzy logic, with a so-called dynamic norm built into the model.

PMID:39869478 | DOI:10.37190/abb-02474-2024-02

Int J Surg. 2025 Jan 24. doi: 10.1097/JS9.0000000000002232. Online ahead of print.

ABSTRACT

BACKGROUND: The impact of aortic arch (AA) morphology on the management of the procedural details and the clinical outcomes of the transfemoral artery (TF)-transcatheter aortic valve replacement (TAVR) has not been evaluated. The goal of this study was to evaluate the AA morphology of patients who had TF-TAVR using an artificial intelligence algorithm and then to evaluate its predictive value for clinical outcomes.

MATERIALS AND METHODS: A total of 1480 consecutive patients undergoing TF-TAVR using a new-generation transcatheter heart valve at 12 institutes were included in this retrospective study. The AA measurements were evaluated by deep learning, and then the approach index (IA) was determined. The machine learning algorithm was used to construct the predictive model and was validated externally.

RESULTS: The area under the curve of the IA model using random forest and logistic regression was 0.675 [95% confidence interval (CI): 0.586-0.764] and 0.757 (95% CI: 0.665-0.849), respectively. The IA model was validated externally, and consistent distinctions were obtained. After we used a generalized propensity score matching method for continuous exposure, the IA was the strongest correlation factor for major procedural events (odds ratio: 3.87; 95% CI: 2.13-7.59, P < 0.001). When leaflet morphology or transcatheter heart valve type was an interactive item with IA, neither of them was statistically significant in terms of clinical outcomes.

CONCLUSION: IA may be used to identify the impact of AA morphology on procedural and clinical outcomes in patients having TF-TAVR and to help to predict the procedural complications.

PMID:39869394 | DOI:10.1097/JS9.0000000000002232

Int J Surg. 2025 Jan 24. doi: 10.1097/JS9.0000000000002221. Online ahead of print.

ABSTRACT

BACKGROUND: Detecting kidney trauma on CT scans can be challenging and is sometimes overlooked. While deep learning (DL) has shown promise in medical imaging, its application to kidney injuries remains underexplored. This study aims to develop and validate a DL algorithm for detecting kidney trauma, using institutional trauma data and the Radiological Society of North America (RSNA) dataset for external validation.

METHODS: We developed RenoTrNet, a DL model trained on institutional data. We evaluated the model's performance through external validation on randomly selected cases from the RSNA dataset. Performance metrics included the accuracy, sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV). Heatmap visualizations were used to aid interpretability.

RESULTS: In the internal testing dataset, the model achieved an accuracy of 0.88 (95% CI: 0.82-0.92), with a sensitivity of 0.75 (95% CI: 0.62-0.85) and a specificity of 0.95 (95% CI: 0.89-0.98). PPV and NPV were 0.89 (95% CI: 0.76-0.95) and 0.88 (95% CI: 0.81-0.93), respectively. In external RSNA validation, the algorithm c demonstrated robust performance with an accuracy of 0.93 (0.91-0.95), a sensitivity of 0.73 (0.60-0.83), a specificity of 0.94 (0.93-0.96), a PPV of 0.45 (0.35-0.56), and an NPV of 0.98 (0.97-0.99).

CONCLUSION: The RenoTrNet DL algorithm demonstrated high accuracy in detecting kidney trauma on CT scans, both in internal and external validation. By optimizing image segmentation and computational efficiency, this model has potential for clinical deployment, potentially aiding in trauma diagnosis in real-world clinical scenarios.

PMID:39869390 | DOI:10.1097/JS9.0000000000002221

J Chem Inf Model. 2025 Jan 27. doi: 10.1021/acs.jcim.4c01735. Online ahead of print.

ABSTRACT

Drug-drug interaction can lead to diminished therapeutic effects or increased toxicity, posing significant risks, especially in polypharmacy, and cytochrome P450 plays an indispensable role in this interaction. Cytochrome P450, responsible for the metabolism and detoxification of most drugs, metabolizes about 90% of Food and Drug Administration-approved drugs, making early detection of potential drug-drug interactions. Over the years, in-silico modeling has become a valuable tool for predicting drug-drug interactions. Still, conventional molecular descriptors focusing on the structural properties of drugs often overlook complex electronic interactions critical for accurate predictions. To address this, we implemented the Manifold Embedding of Molecular Surface (MEMS) approach, which retains the quantum mechanical characteristics of molecules. MEMS-generated electronic attributes were embedded and featurized for deep learning using the DeepSets architecture, where our models achieved high accuracy, particularly for cytochrome P450 enzyme 1A2 (CYP1A2), with F1 scores reaching up to 0.866. This study highlights the potential of integrating detailed electronic properties with deep learning to improve predictive models for drug-drug interactions, addressing the limitations of traditional molecular descriptors and machine-learning techniques.

PMID:39869197 | DOI:10.1021/acs.jcim.4c01735

J Mater Chem B. 2025 Jan 27. doi: 10.1039/d4tb02494j. Online ahead of print.

ABSTRACT

Tissue regeneration after a wound occurs through three main overlapping and interrelated stages namely inflammatory, proliferative, and remodelling phases, respectively. The inflammatory phase is key for successful tissue reconstruction and triggers the proliferative phase. The macrophages in the non-healing wounds remain in the inflammatory loop, but their phenotypes can be changed via interactions with nanofibre-based scaffolds mimicking the organisation of the native structural support of healthy tissues. However, the organisation of extracellular matrix (ECM) is highly complex, combining order and disorder, which makes it difficult to replicate. The possibility of predicting the desirable biomimetic geometry and chemistry of these nanofibre scaffolds would streamline the scaffold design process. Fifteen families of nanofibre scaffolds, electrospun from combinations of polyesters (polylactide, polyhydroxybutyrate), polysaccharides (polysucrose, carrageenan, cellulose), and polyester ether (polydioxanone) were investigated and analysed using machine learning (ML). The Random Forest model had the best performance (92.8%) in predicting inflammatory responses of macrophages on the nanoscaffolds using tumour necrosis factor-alpha as the output. CellProfiler proved to be an effective tool to process scanning electron microscopy (SEM) images of the macrophages on the scaffolds, successfully extracting various features and measurements related to cell phenotypes M0, M1, and M2. Deep learning modelling indicated that convolutional neural network models have the potential to be applied to SEM images to classify macrophage cells according to their phenotypes. The complex organisation of the nanofibre scaffolds can be analysed using graph theory (GT), revealing the underlying connectivity patterns of the nanofibres. Analysis of GT descriptors showed that the electrospun membranes closely mimic the connectivity patterns of the ECM. We conclude that ML-facilitated, GT-quantified engineering of cellular scaffolds has the potential to predict cell interactions, streamlining the pipeline for tissue engineering.

PMID:39869000 | DOI:10.1039/d4tb02494j

Anal Chem. 2025 Jan 27. doi: 10.1021/acs.analchem.4c04531. Online ahead of print.

ABSTRACT

In metabolomic analysis based on liquid chromatography coupled with mass spectrometry, detecting and quantifying intricate objects is a massive job. Current peak picking methods still cause high rates of incorrectly picked peaks to influence the reliability and reproducibility of results. To address these challenges, we developed QuanFormer, a deep learning method based on object detection designed to accurately quantify peak signals. Our algorithm combines the feature extraction capabilities of convolutional neural networks (CNNs) with the global computation capability of Transformer architecture. Data training in QuanFormer by using nearly 20,000 annotated regions-of-interest (ROIs) ensures unique prediction via bipartite matching, achieving 96.5% of the average precision value on the test set. Even without retraining, QuanFormer achieves over 90% accuracy in distinguishing true from false peaks. Performance was further analyzed using visualization techniques applied to the encoder and decoder layers. We also demonstrated that QuanFormer could correct retention time shifts for peak alignment and generally surpass the existing methods, including MZmine 3 and PeakDetective, to obtain a larger number of picked peaks and higher accurate quantification. Finally, we also carried out metabolomic analysis in a clinical cohort of breast cancer patients and utilized QuanFormer to detect and quantify the potential biomarkers. QuanFormer is open-source and available at https://github.com/LinShuhaiLAB/QuanFormer.

PMID:39868899 | DOI:10.1021/acs.analchem.4c04531

J Intern Med. 2025 Jan 27. doi: 10.1111/joim.20053. Online ahead of print.

ABSTRACT

BACKGROUND: Steatotic liver disease (SLD) is a potentially reversible condition but often goes unnoticed with the risk for end-stage liver disease.

PURPOSE: To opportunistically estimate SLD on lung screening chest computed tomography (CT) and investigate its prognostic value in heavy smokers participating in the National Lung Screening Trial (NLST).

MATERIAL AND METHODS: We used a deep learning model to segment the liver on non-contrast-enhanced chest CT scans of 19,774 NLST participants (age 61.4 ± 5.0 years; 41.2% female) at baseline and on the 1-year follow-up scan if no cancer was detected. SLD was defined as hepatic fat fraction (HFF) ≥5% derived from Hounsfield unit measures of the segmented liver. Participants with SLD were categorized as lean (body mass index [BMI] < 25 kg/m2) and overweight (BMI ≥ 25 kg/m2). The primary outcome was all-cause mortality. Cox proportional hazard regression assessed the association between (1) SLD and mortality at baseline and (2) the association between a change in HFF and mortality within 1 year.

RESULTS: There were 5.1% (1000/19,760) all-cause deaths over a median follow-up of 6 (range, 0.8-6) years. At baseline, SLD was associated with increased mortality in lean but not in overweight/obese participants as compared to participants without SLD (hazard ratio [HR] adjusted for risk factors: 1.93 [95% confidence interval 1.52-2.45]; p = 0.001). Individuals with an increase in HFF within 1 year had a significantly worse outcome than participants with stable HFF (HR adjusted for risk factors: 1.29 [1.01-1.65]; p = 0.04).

CONCLUSION: SLD is an independent predictor for long-term mortality in heavy smokers beyond known clinical risk factors.

PMID:39868889 | DOI:10.1111/joim.20053

J Magn Reson Imaging. 2025 Jan 27. doi: 10.1002/jmri.29720. Online ahead of print.

ABSTRACT

BACKGROUND: The spinal column is a frequent site for metastases, affecting over 30% of solid tumor patients. Identifying the primary tumor is essential for guiding clinical decisions but often requires resource-intensive diagnostics.

PURPOSE: To develop and validate artificial intelligence (AI) models using noncontrast MRI to identify primary sites of spinal metastases, aiming to enhance diagnostic efficiency.

STUDY TYPE: Retrospective.

POPULATION: A total of 514 patients with pathologically confirmed spinal metastases (mean age, 59.3 ± 11.2 years; 294 males) were included, split into a development set (360) and a test set (154).

FIELD STRENGTH/SEQUENCE: Noncontrast sagittal MRI sequences (T1-weighted, T2-weighted, and fat-suppressed T2) were acquired using 1.5 T and 3 T scanners.

ASSESSMENT: Two models were evaluated for identifying primary sites of spinal metastases: the expert-derived features (EDF) model using radiologist-identified imaging features and a ResNet50-based deep learning (DL) model trained on noncontrast MRI. Performance was assessed using accuracy, precision, recall, F1 score, and the area under the receiver operating characteristic curve (ROC-AUC) for top-1, top-2, and top-3 indicators.

STATISTICAL TESTS: Statistical analyses included Shapiro-Wilk, t tests, Mann-Whitney U test, and chi-squared tests. ROC-AUCs were compared via DeLong tests, with 95% confidence intervals from 1000 bootstrap replications and significance at P < 0.05.

RESULTS: The EDF model outperformed the DL model in top-3 accuracy (0.88 vs. 0.69) and AUC (0.80 vs. 0.71). Subgroup analysis showed superior EDF performance for common sites like lung and kidney (e.g., kidney F1: 0.94 vs. 0.76), while the DL model had higher recall for rare sites like thyroid (0.80 vs. 0.20). SHapley Additive exPlanations (SHAP) analysis identified sex (SHAP: -0.57 to 0.68), age (-0.48 to 0.98), T1WI signal intensity (-0.29 to 0.72), and pathological fractures (-0.76 to 0.25) as key features.

DATA CONCLUSION: AI techniques using noncontrast MRI improve diagnostic efficiency for spinal metastases. The EDF model outperformed the DL model, showing greater clinical potential.

PLAIN LANGUAGE SUMMARY: Spinal metastases, or cancer spreading to the spine, are common in patients with advanced cancer, often requiring extensive tests to determine the original tumor site. Our study explored whether artificial intelligence could make this process faster and more accurate using noncontrast MRI scans. We tested two methods: one based on radiologists' expertise in identifying imaging features and another using a deep learning model trained to analyze MRI images. The expert-based method was more reliable, correctly identifying the tumor site in 88% of cases when considering the top three likely diagnoses. This approach may help doctors reduce diagnostic time and improve patient care.

LEVEL OF EVIDENCE: 3 TECHNICAL EFFICACY: Stage 2.

PMID:39868626 | DOI:10.1002/jmri.29720

Acta Crystallogr D Struct Biol. 2025 Feb 1. doi: 10.1107/S2059798325000592. Online ahead of print.

ABSTRACT

Predicting the 3D structure of RNA is a significant challenge despite ongoing advancements in the field. Although AlphaFold has successfully addressed this problem for proteins, RNA structure prediction raises difficulties due to the fundamental differences between proteins and RNA, which hinder its direct adaptation. The latest release of AlphaFold, AlphaFold3, has broadened its scope to include multiple different molecules such as DNA, ligands and RNA. While the AlphaFold3 article discussed the results for the last CASP-RNA data set, the scope of its performance and the limitations for RNA are unclear. In this article, we provide a comprehensive analysis of the performance of AlphaFold3 in the prediction of 3D structures of RNA. Through an extensive benchmark over five different test sets, we discuss the performance and limitations of AlphaFold3. We also compare its performance with ten existing state-of-the-art ab initio, template-based and deep-learning approaches. Our results are freely available on the EvryRNA platform at https://evryrna.ibisc.univ-evry.fr/evryrna/alphafold3/.

PMID:39868559 | DOI:10.1107/S2059798325000592

Anal Chem. 2025 Jan 27. doi: 10.1021/acs.analchem.4c05913. Online ahead of print.

ABSTRACT

Timely and accurate detection of trace mycotoxins in agricultural products and food is significant for ensuring food safety and public health. Herein, a deep learning-assisted and entropy-driven catalysis (EDC)-Argonaute powered fluorescence single-particle aptasensing platform was developed for ultrasensitive detection of fumonisin B1 (FB1) using single-stranded DNA modified with biotin and red fluorescence-encoded microspheres as a signal probe and streptavidin-conjugated magnetic beads as separation carriers. The binding of aptamer with FB1 releases the trigger sequence to mediate EDC cycle to produce numerous 5'-phosphorylated output sequences, which can be used as the guide DNA to activate downstream Thermus thermophilus Argonaute (TtAgo) for cleaving the signal probe, resulting in increased number of fluorescence microspheres remaining in the final reaction supernatant after magnetic separation. Subsequently, through fast and accurate counting of red bright particles in the captured confocal fluorescence images from the supernatant via a YOLOv9 deep learning model, the sensitive and specific detection of FB1 could be realized. This approach has a limit of detection (LOD) of 0.89 pg/mL with a linear range from 1 pg/mL to 100 ng/mL, and satisfactory recovery (87.2-113.5%) in real food samples indicates its practicality. The integration of the aptamer and EDC with TtAgo broadens the target range of Argonaute and enhances sensitivity. Furthermore, incorporating deep learning significantly improves the analytical efficiency of single-particle detection. This work provides a promising analytical strategy in biosensing and promotes the application of fluorescence single-particle detection in food safety monitoring.

PMID:39868471 | DOI:10.1021/acs.analchem.4c05913

Front Oncol. 2025 Jan 10;14:1469427. doi: 10.3389/fonc.2024.1469427. eCollection 2024.

ABSTRACT

PURPOSE: To create a system to enable the identification of histological variants of bladder cancer in a simple, efficient, and noninvasive manner.

MATERIAL AND METHODS: In this multicenter diagnostic study, we retrospectively collected basic information and CT images about the patients concerned from three hospitals. An interactive deep learning-based bladder cancer image segmentation framework was constructed using the Swin UNETR algorithm for further features extraction. Radiomic features and deep learning features were extracted for further stacking ensemble system construction. The segmentation model' performance was assessed by using Dice Similarity (Dice) metrics, Intersection Over Union (IOU), Sensitivity (SEN) and Specificity (SPE). To evaluate the system's performance, we used the Receiver Operating Characteristics (ROC) curve, the Accuracy Score (ACC) and Decision Curve Analysis (DCA).

RESULTS: 410 patients from one hospital were included in the training set, while 60 patients from two other hospitals were included in the test set. A total of 50 features comprising 46 radiomic features and 4 deep learning features were finally retained for further stacking ensemble model building. The interactive segmentation model and system exhibited excellent performance in both training (Dice = 0.78, IOU = 0.65, SEN = 0.83, SPE = 1.00, AUC = 0.940, ACC = 0.868) and testing datasets (Dice = 0.80, IOU = 0.67, SEN = 0.89, SPE = 1.00, AUC = 0.905, ACC = 0.900).

CONCLUSION: We successfully constructed a stacking ensemble machine learning model for early, non-invasive identification of histological variants in bladder cancer which will help urologists make clinical decisions.

PMID:39868365 | PMC:PMC11757263 | DOI:10.3389/fonc.2024.1469427

iScience. 2024 Dec 20;28(1):111658. doi: 10.1016/j.isci.2024.111658. eCollection 2025 Jan 17.

ABSTRACT

The regulation of gene expression relies on the coordinated action of transcription factors (TFs) at enhancers, including both activator and repressor TFs. We employed deep learning (DL) to dissect HepG2 enhancers into positive (PAR), negative (NAR), and neutral activity regions. Sharpr-MPRA and STARR-seq highlight the dichotomy impact of NARs and PARs on modulating and catalyzing the activity of enhancers, respectively. Approximately 22% of HepG2 enhancers, termed "repressive impact enhancers" (RIEs), are predominantly populated by NARs and transcriptional repression motifs. Genes flanking RIEs exhibit a stage-specific decline in expression during late development, suggesting RIEs' role in trimming enhancer activities. About 16.7% of human NARs emerge from neutral rhesus macaque DNA. This gain of repressor binding sites in RIEs is associated with a 30% decrease in the average expression of flanking genes in humans compared to rhesus macaque. Our work reveals modulated enhancer activity and adaptable gene regulation through the evolutionary dynamics of TF binding sites.

PMID:39868043 | PMC:PMC11761325 | DOI:10.1016/j.isci.2024.111658

iScience. 2024 Dec 20;28(1):111638. doi: 10.1016/j.isci.2024.111638. eCollection 2025 Jan 17.

ABSTRACT

Over the last decade, Hippo signaling has emerged as a major tumor-suppressing pathway. Its dysregulation is associated with abnormal expression of YAP1 and TEAD-family genes. Recent works have highlighted the role of YAP1/TEAD activity in several cancers and its potential therapeutic implications. Therefore, identifying patients with a dysregulated Hippo pathway is key to enhancing treatment impact. Although recent studies have derived RNA-seq-based signatures, there remains a need for a reproducible and cost-effective method to measure the pathway activation. In recent years, deep learning applied to histology slides have emerged as an effective way to predict molecular information from a data modality available in clinical routine. Here, we trained models to predict YAP1/TEAD activity from H&E-stained histology slides in multiple cancers. The robustness of our approach was assessed in seven independent validation cohorts. Finally, we showed that histological markers of disease aggressiveness were associated with dysfunctional Hippo signaling.

PMID:39868035 | PMC:PMC11758823 | DOI:10.1016/j.isci.2024.111638

Pak J Med Sci. 2025 Jan;41(1):225-229. doi: 10.12669/pjms.41.1.9171.

ABSTRACT

OBJECTIVE: To design a deep learning-based model for early screening of diabetic retinopathy, predict the condition, and provide interpretable justifications.

METHODS: The experiment's model structure is designed based on the Vision Transformer architecture which was initiated in March 2023 and the first version was produced in July 2023 at Affiliated Hospital of Hangzhou Normal University. We use the publicly available EyePACS dataset as input to train the model. Using the trained model, we predict whether a given patient's fundus images indicate diabetic retinopathy and provide the relevant affected areas as the basis for the judgement.

RESULTS: The model was validated using two subsets of the IDRiD dataset. Our model not only achieved good results in terms of detection accuracy, reaching around 0.88, but also performed comparably to similar models annotated for affected areas in predicting the affected regions.

CONCLUSION: Utilizing image-level annotations, we implemented a method for detecting diabetic retinopathy through deep learning and provided interpretable justifications to assist clinicians in diagnosis.

PMID:39867796 | PMC:PMC11755306 | DOI:10.12669/pjms.41.1.9171

J Pathol Inform. 2024 Dec 11;16:100416. doi: 10.1016/j.jpi.2024.100416. eCollection 2025 Jan.

ABSTRACT

BACKGROUND: Traditional liver fibrosis staging via percutaneous biopsy suffers from sampling bias and variable inter-pathologist agreement, highlighting the need for more objective techniques. Deep learning models for disease staging from medical images have shown potential to decrease diagnostic variability, with recent weakly supervised learning strategies showing promising results even with limited manual annotation.

PURPOSE: To study the clustering-constrained attention multiple instance learning (CLAM) approach for staging liver fibrosis on trichrome whole slide images (WSIs) of children and young adults.

METHODS: This is an ethics board approved retrospective study utilizing 217 trichrome WSI from pediatric liver biopsies for model development and testing. Two pediatric pathologists scored WSI using two liver fibrosis staging systems, METAVIR and Ishak. Cases were then secondarily categorized into either high- or low-stage liver fibrosis and used for model development. The CLAM pipeline was used to develop binary classification models for histological liver fibrosis. Model performance was evaluated using area under the curve (AUC), accuracy, sensitivity, specificity, and Cohen's Kappa.

RESULTS: The CLAM models showed strong diagnostic performance, with sensitivities up to 0.76 and AUCs up to 0.92 for distinguishing low- and high-stage fibrosis. The agreement between model predictions and average pathologist scores was moderate to substantial (Kappa: 0.57-0.69), whereas pathologist agreement on the METAVIR and Ishak scoring systems was only fair (Kappa: 0.39-0.46).

CONCLUSIONS: CLAM pipeline showed promise in detecting features important for differentiating low- and high-stage fibrosis from trichrome WSI based on the results, offering a promising objective method for liver fibrosis detection in children and young adults.

PMID:39867463 | PMC:PMC11760786 | DOI:10.1016/j.jpi.2024.100416