Infect Genet Evol. 2025 Jan 10:105713. doi: 10.1016/j.meegid.2025.105713. Online ahead of print.

ABSTRACT

Whole genome sequencing has been used to investigate the genomic diversity of M. tuberculosis in the northern and southern states of India, but information about the eastern part of the country is still limited. Through a sequencing-based strategy, this study seeks to comprehend the diversity and drug resistance pattern in the eastern region. A total of 102 M. tuberculosis isolates from North East (n = 54), and Odisha (n = 48) were sequenced along with 7 follow up isolates from Sikkim. The pre-XDR and XDR isolates diagnosed as per the NTEP diagnostic algorithm were subjected for phenotypic second-line liquid culture drug susceptibility testing in MGIT-960 system. After filtering out low quality isolates based on taxonomic classification and depth of coverage, variant calling was performed. We observed a high prevalence of multi-drug resistant TB (MDR-TB) lineage 2 (52/54) isolates in northeast whereas there was a mixed representation of lineage 1 (30/48) & lineage 3 (11/48) in Odisha. The MDR-TB isolates from Sikkim posed a high rate (51/53) of fluoroquinolone resistance and pairwise SNV distances (≤10) indicating possible local transmission events in the region. We observed occurrence of genetic variations in genes associated with bedaquiline and delamanid resistance. Our findings show the diversity of M. tuberculosis vary across the eastern regions, in north eastern states lineage 2 has a dominant presence while lineage 1 and 3 has mixed representation in Odisha. The high prevalence of fluoroquinolone resistance in north eastern region associated with variations in gyrA gene and may have been caused by local transmission events based on genomic similarities.

PMID:39800206 | DOI:10.1016/j.meegid.2025.105713

Mol Cells. 2025 Jan 10:100180. doi: 10.1016/j.mocell.2025.100180. Online ahead of print.

ABSTRACT

Recent advancements in fluorescence-based biosensor technologies have enabled more precise and accurate Förster Resonance Energy Transfer (FRET) imaging within Agrobacterium-mediated plant transformation systems. However, the application of FRET imaging in plant tissues remains hindered by significant challenges, particularly the time-intensive process of generating transgenic lines and the complications arising from tissue autofluorescence. In contrast, protoplast-based FRET imaging offers a rapid and efficient platform for functional screening and analysis, making it an essential tool for plant research. Nevertheless, conventional protoplast-based FRET approaches are often limited by background interference, inconsistent imaging conditions, and difficulties in quantitative analysis. Here, we present a systematic optimization of imaging conditions using the calcium biosensor D3cpv, addressing these limitations to improve both precision and efficiency in protoplast-based FRET imaging. This work serves as a practical guide for streamlining FRET imaging workflows and maximizing the utility of biosensors in plant cell studies.

PMID:39800051 | DOI:10.1016/j.mocell.2025.100180

Ceska Gynekol. 2024;89(6):501-503. doi: 10.48095/cccg2024501.

ABSTRACT

Estrogens are key hormones that play a vital role in the physiology of the reproductive system in women. However, their therapeutic use in hormonal treatment, contraception, and the treatment of hormone-dependent diseases may be associated with a number of side effects, especially on the liver. This article focuses on the mechanisms of action of estrogens and their potential hepatotoxic effects, as well as risk factors and possible differences between representatives.

PMID:39800550 | DOI:10.48095/cccg2024501

Eur J Hosp Pharm. 2025 Jan 12:ejhpharm-2024-004434. doi: 10.1136/ejhpharm-2024-004434. Online ahead of print.

NO ABSTRACT

PMID:39800473 | DOI:10.1136/ejhpharm-2024-004434

J Prev Alzheimers Dis. 2025 Jan;12(1):100002. doi: 10.1016/j.tjpad.2024.100002. Epub 2025 Jan 1.

ABSTRACT

BACKGROUND: Investigators conducting clinical trials have an ethical, scientific, and regulatory obligation to protect the safety of trial participants. Traditionally, safety monitoring includes manual review and coding of adverse event data by expert clinicians.

OBJECTIVES: Our study explores the use of natural language processing (NLP) and artificial intelligence (AI) methods to streamline and standardize clinician coding of adverse event data in Alzheimer's disease (AD) clinical trials.

DESIGN: Our quantitative retrospective study aimed to develop a gold standard AD adverse event data set, evaluate the predictive performance of NLP-based models to classify adverse events, and determine whether automated coding is more efficient, accurate, reliable, and consistent than clinician coding.

SETTING: Our study was conducted at the University of Southern California's Alzheimer's Therapeutic Research Institute (ATRI). ATRI serves as the clinical and data coordinating center for the Alzheimer's Clinical Trial Consortium (ACTC).

PARTICIPANTS: We collected demographic and adverse event data from eight completed clinical trials in participants (n=1920) with symptomatic AD conducted between 2005 and 2020.

MEASUREMENTS: Original expert clinician-confirmed codes were used for all model performance comparisons. F1 score was used as the primary model selection metric. Final classifier performance was evaluated using predictive accuracy. Clinician effort was measured in time to code, review, and confirm coded adverse events.

RESULTS: In a sample of 1000 adverse events, AI-based AE coding achieved higher accuracy (∼20% increase in accuracy) and was more cost-effective (∼80% cost reduction) than traditional clinician coding.

CONCLUSIONS: Our study results demonstrate how approaches that effectively combine AI and human expertise can improve the efficiency and quality of adverse event coding and clinical trial safety monitoring.

PMID:39800465 | DOI:10.1016/j.tjpad.2024.100002

J Emerg Nurs. 2025 Jan;51(1):20-26. doi: 10.1016/j.jen.2024.08.005.

ABSTRACT

Acute generalized exanthematous pustulosis is a severe cutaneous adverse reaction characterized by the rapid onset of nonfollicular, sterile pustules on an erythematous base, typically accompanied by fever (≥38 °C), neutrophilia (7.0 × 10⁹/L), and characteristic histopathological features. This case report presents the first documented instance of acute generalized exanthematous pustulosis after hyaluronic acid viscosupplementation. A 61-year-old female developed a pruritic, erythematous rash that rapidly evolved into generalized erythroderma with systemic manifestations after receiving intra-articular hyaluronic acid injections for knee osteoarthritis. Initial physical examination and diagnostic workup, including biopsy and blood tests, were performed to exclude other differential diagnoses such as generalized pustular psoriasis, subcorneal pustular dermatosis, and immunoglobulin A pemphigus. The temporal association with hyaluronic acid injections and the patient's positive response to treatment with systemic corticosteroids and antihistamines supported the definitive diagnosis of drug-induced acute generalized exanthematous pustulosis. The patient was managed with the withdrawal of the offending agent, and supportive care was provided. She did not require rehabilitation and experienced no adverse events during the recovery period. Follow-up visits confirmed the absence of recurrence and complete resolution of symptoms, with no lasting sequelae. This case underscores the importance of recognizing acute generalized exanthematous pustulosis' acute manifestations and potential triggers, even with treatments generally considered safe. ED personnel, including advanced practice registered nurses and other clinicians, must include acute generalized exanthematous pustulosis in their differential diagnoses of severe cutaneous disorders to initiate prompt and appropriate management. The development of atrial fibrillation during hospitalization in this patient raises questions about the systemic effects of acute generalized exanthematous pustulosis, suggesting an area for further research. Early detection and treatment of acute generalized exanthematous pustulosis are crucial for favorable outcomes, illustrating the vital role ED personnel play in managing this condition. Awareness of rare triggers such as hyaluronic acid is essential for preventing and effectively treating such severe adverse reactions.

PMID:39800444 | DOI:10.1016/j.jen.2024.08.005

J Immunother Cancer. 2025 Jan 11;13(1):e010565. doi: 10.1136/jitc-2024-010565.

ABSTRACT

BACKGROUND: SL-172154 is a hexameric fusion protein adjoining the extracellular domain of SIRPα to the extracellular domain of CD40L via an inert IgG4-derived Fc domain. In preclinical studies, a murine equivalent SIRPα-Fc-CD40L fusion protein provided superior antitumor immunity in comparison to CD47- and CD40-targeted antibodies. A first-in-human phase I trial of SL-172154 was conducted in patients with platinum-resistant ovarian cancer.

METHODS: SL-172154 was administered intravenously at 0.1, 0.3, 1.0, 3.0, and 10.0 mg/kg. Dose escalation followed a modified toxicity probability interval-2 design. Objectives included evaluation of safety, dose-limiting toxicity, recommended phase II dose, pharmacokinetic (PK) and pharmacodynamic (PD) parameters, and antitumor activity.

RESULTS: 27 patients (median age 66 years (range, 33-85); median of 4 prior systemic therapies (range, 2-9)) with ovarian (70%), fallopian tube (15%), or primary peritoneal (15%) cancer received SL-172154. Treatment-emergent adverse events (TEAEs) were reported for 27 patients (100%), with 24 (88.9%) having a drug-related TEAE and infusion-related reactions being the most common. 12 patients (44.4%) had grade 3/4 TEAEs, and half of these patients (22.2%) had a drug-related grade 3/4 TEAE. There were no fatal adverse events, and no TEAEs led to drug discontinuation. SL-172154 Cmax and area under the curve increased with dose with greater than proportional exposure noted at 3.0 and 10.0 mg/kg. CD47 and CD40 target engagement on CD4+ T cells and B cells, respectively, approached 100% by 3.0 mg/kg. Dose-dependent responses in multiple cytokines (eg, interleukin 12 (IL-12), IP-10) approached a plateau at ≥3.0 mg/kg. Paired tumor biopsies demonstrated a shift in macrophages from an M2- to an M1-dominant phenotype and increased infiltration of CD8 T cells. PK/PD modeling showed near maximal margination of B cells and a dose-dependent production of IL-12 nearing a plateau at >3.0 mg/kg. The best response was stable disease in 6/27 (22%) patients.

CONCLUSIONS: SL-172154 was tolerable as monotherapy and induced, dose-dependent, and cyclical immune cell activation, increases in multiple serum cytokines, and trafficking of CD40-positive B cells and monocytes following each infusion. The safety, PK, and PD activity support 3.0 mg/kg as a safe and pharmacologically active dose.

TRIAL REGISTRATION NUMBER: NCT04406623.

PMID:39800375 | DOI:10.1136/jitc-2024-010565

Int J Comput Assist Radiol Surg. 2025 Jan 12. doi: 10.1007/s11548-024-03306-9. Online ahead of print.

ABSTRACT

PURPOSE: This paper presents a deep learning approach to recognize and predict surgical activity in robot-assisted minimally invasive surgery (RAMIS). Our primary objective is to deploy the developed model for implementing a real-time surgical risk monitoring system within the realm of RAMIS.

METHODS: We propose a modified Transformer model with the architecture comprising no positional encoding, 5 fully connected layers, 1 encoder, and 3 decoders. This model is specifically designed to address 3 primary tasks in surgical robotics: gesture recognition, prediction, and end-effector trajectory prediction. Notably, it operates solely on kinematic data obtained from the joints of robotic arm.

RESULTS: The model's performance was evaluated on JHU-ISI Gesture and Skill Assessment Working Set dataset, achieving highest accuracy of 94.4% for gesture recognition, 84.82% for gesture prediction, and significantly low distance error of 1.34 mm with a prediction of 1 s in advance. Notably, the computational time per iteration was minimal recorded at only 4.2 ms.

CONCLUSION: The results demonstrated the excellence of our proposed model compared to previous studies highlighting its potential for integration in real-time systems. We firmly believe that our model could significantly elevate realms of surgical activity recognition and prediction within RAS and make a substantial and meaningful contribution to the healthcare sector.

PMID:39799528 | DOI:10.1007/s11548-024-03306-9

Bioinformatics. 2025 Jan 12:btaf019. doi: 10.1093/bioinformatics/btaf019. Online ahead of print.

ABSTRACT

MOTIVATION: The knowledge of protein stability upon residue variation is an important step for functional protein design and for understanding how protein variants can promote disease onset. Computational methods are important to complement experimental approaches and allow a fast screening of large datasets of variations.

RESULTS: In this work we present DDGemb, a novel method combining protein language model embeddings and transformer architectures to predict protein ΔΔG upon both single- and multi-point variations. DDGemb has been trained on a high-quality dataset derived from literature and tested on available benchmark datasets of single- and multi-point variations. DDGemb performs at the state of the art in both single- and multi-point variations.

AVAILABILITY: DDGemb is available as web server at https://ddgemb.biocomp.unibo.it. Datasets used in this study are available at https://ddgemb.biocomp.unibo.it/datasets.

PMID:39799516 | DOI:10.1093/bioinformatics/btaf019

Bioinformatics. 2025 Jan 12:btaf015. doi: 10.1093/bioinformatics/btaf015. Online ahead of print.

ABSTRACT

SUMMARY: With the increased reliance on multi-omics data for bulk and single cell analyses, the availability of robust approaches to perform unsupervised learning for clustering, visualization, and feature selection is imperative. We introduce nipalsMCIA, an implementation of multiple co-inertia analysis (MCIA) for joint dimensionality reduction that solves the objective function using an extension to Non-linear Iterative Partial Least Squares (NIPALS). We applied nipalsMCIA to both bulk and single cell datasets and observed significant speed-up over other implementations for data with a large sample size and/or feature dimension.

AVAILABILITY AND IMPLEMENTATION: nipalsMCIA is available as a Bioconductor package at https://bioconductor.org/packages/release/bioc/html/nipalsMCIA.html, and includes detailed documentation and application vignettes.

SUPPLEMENTARY MATERIALS: Supplementary Materials are available online.

PMID:39799512 | DOI:10.1093/bioinformatics/btaf015

Sci Rep. 2025 Jan 11;15(1):1690. doi: 10.1038/s41598-025-85900-8.

ABSTRACT

Primary aldosteronism (PA), characterized by autonomous aldosterone overproduction, is a major cause of secondary hypertension with significant cardiovascular complications. Current treatments mainly focus on symptom management rather than addressing underlying mechanisms. This study aims to discover novel therapeutic targets for PA using integrated bioinformatics and experimental validation approaches. We employed a systematic approach combining: gene identification through transcriptome-wide association studies (TWAS); causal inference using summary data-based Mendelian randomization (SMR) and two-sample Mendelian randomization (MR) analyses; additional analyses included phenome-wide association analysis, enrichment analysis, protein-protein interaction (PPI) networks, drug repurposing, molecular docking and clinical validation through aldosterone-producing adenomas (APAs) tissue. Through systematic screening and prioritization, we identified 163 PA-associated genes, of which seven emerged as potential drug targets: CEP104, HIP1, TONSL, ZNF100, SHMT1, and two long non-coding RNAs (AC006369.2 and MRPL23-AS1). SHMT1 was identified as the most promising target, showing significantly elevated expression in APAs compared to adjacent non-tumorous tissues. Drug repurposing analysis identified four potential SHMT1-targeting compounds (Mimosine, Pemetrexed, Leucovorin, and Irinotecan), supported by molecular docking studies. The integration of multiple bioinformatics methods and experimental validation successfully identified novel drug targets for hyperaldosteronism. SHMT1, in particular, represents a promising candidate for future therapeutic development. These findings provide new opportunities for developing causative treatments for PA, though further clinical validation is warranted.

PMID:39799159 | DOI:10.1038/s41598-025-85900-8

Pharmacol Res. 2025 Jan 9:107575. doi: 10.1016/j.phrs.2024.107575. Online ahead of print.

ABSTRACT

Statins are first-line treatments in the primary and secondary prevention of cardiovascular disease. Clinical studies show statins act independently of lipid-lowering mechanisms to decrease C-reactive protein (CRP), an inflammation marker. We aim to elucidate genetic loci associated with CRP statin response. CRP statin response is the change in log-CRP between off-treatment and on-treatment measurements. Cohort-level Genome-Wide Association Studies (GWAS) of CRP response were performed using 1000 Genomes imputed data, testing ~10 million common genetic variants. GWAS meta-analysis combined results from seven cohorts and clinical trials totalling 14,070 statin-treated individuals of European ancestry within the GIST consortium. Secondary analyses included statin-by-placebo interaction analyses, and lookups in African ancestry cohorts. Our GWAS identified two genome-wide significant (P<5e-8) loci: APOE and HNF1A for CRP statin response corrected for baseline CRP. The missense lead variant rs429358 at APOE, contributing to the APOE-E4 haplotype, is a risk locus for dyslipidaemia, Alzheimer's and coronary artery disease (CAD). The HNF1A locus is associated with diabetes, cholesterol levels, and CAD. Both loci are also associated with baseline CRP levels, and neither locus achieved a significant (P<0.05) result from the statin v. placebo interaction meta-analysis using randomized clinical trial data. However, the interaction result (P-int=0.09) for APOE was suggestive and possibly underpowered. The APOE-E4 signal may therefore be associated with both CRP and LDL-cholesterol statin response. Combined with suggestions in the literature that APOE also leads to differential statin benefit in Alzheimer's, the APOE locus warrants further investigation for potential genetic effects on healthcare with statin treatment.

PMID:39798939 | DOI:10.1016/j.phrs.2024.107575

J Orthop Surg Res. 2025 Jan 11;20(1):41. doi: 10.1186/s13018-024-05334-2.

ABSTRACT

PURPOSE: The study aimed to develop a deep learning model for rapid, automated measurement of full-spine X-rays in adolescents with Adolescent Idiopathic Scoliosis (AIS). A significant challenge in this field is the time-consuming nature of manual measurements and the inter-individual variability in these measurements. To address these challenges, we utilized RTMpose deep learning technology to automate the process.

METHODS: We conducted a retrospective multicenter diagnostic study using 560 full-spine sagittal plane X-ray images from five hospitals in Inner Mongolia. The model was trained and validated using 500 images, with an additional 60 images for independent external validation. We evaluated the consistency of keypoint annotations among different physicians, the accuracy of model-predicted keypoints, and the accuracy of model measurement results compared to manual measurements.

RESULTS: The consistency percentages of keypoint annotations among different physicians and the model were 90-97% within the 4-mm range. The model's prediction accuracies for key points were 91-100% within the 4-mm range compared to the reference standards. The model's predictions for 15 anatomical parameters showed high consistency with experienced physicians, with intraclass correlation coefficients ranging from 0.892 to 0.991. The mean absolute error for SVA was 1.16 mm, and for other parameters, it ranged from 0.22° to 3.32°. A significant challenge we faced was the variability in data formats and specifications across different hospitals, which we addressed through data augmentation techniques. The model took an average of 9.27 s to automatically measure the 15 anatomical parameters per X-ray image.

CONCLUSION: The deep learning model based on RTMpose can effectively enhance clinical efficiency by automatically measuring the sagittal plane parameters of the spine in X-rays of patients with AIS. The model's performance was found to be highly consistent with manual measurements by experienced physicians, offering a valuable tool for clinical diagnostics.

PMID:39799363 | DOI:10.1186/s13018-024-05334-2

BMC Bioinformatics. 2025 Jan 11;26(1):10. doi: 10.1186/s12859-025-06033-3.

ABSTRACT

Antimicrobial peptides (AMPs) have been widely recognized as a promising solution to combat antimicrobial resistance of microorganisms due to the increasing abuse of antibiotics in medicine and agriculture around the globe. In this study, we propose UniAMP, a systematic prediction framework for discovering AMPs. We observe that feature vectors used in various existing studies constructed from peptide information, such as sequence, composition, and structure, can be augmented and even replaced by information inferred by deep learning models. Specifically, we use a feature vector with 2924 values inferred by two deep learning models, UniRep and ProtT5, to demonstrate that such inferred information of peptides suffice for the task, with the help of our proposed deep neural network model composed of fully connected layers and transformer encoders for predicting the antibacterial activity of peptides. Evaluation results demonstrate superior performance of our proposed model on both balanced benchmark datasets and imbalanced test datasets compared with existing studies. Subsequently, we analyze the relations among peptide sequences, manually extracted features, and automatically inferred information by deep learning models, leading to observations that the inferred information is more comprehensive and non-redundant for the task of predicting AMPs. Moreover, this approach alleviates the impact of the scarcity of positive data and demonstrates great potential in future research and applications.

PMID:39799358 | DOI:10.1186/s12859-025-06033-3

Commun Biol. 2025 Jan 11;8(1):43. doi: 10.1038/s42003-025-07469-2.

ABSTRACT

Biomedical research increasingly relies on three-dimensional (3D) cell culture models and artificial-intelligence-based analysis can potentially facilitate a detailed and accurate feature extraction on a single-cell level. However, this requires for a precise segmentation of 3D cell datasets, which in turn demands high-quality ground truth for training. Manual annotation, the gold standard for ground truth data, is too time-consuming and thus not feasible for the generation of large 3D training datasets. To address this, we present a framework for generating 3D training data, which integrates biophysical modeling for realistic cell shape and alignment. Our approach allows the in silico generation of coherent membrane and nuclei signals, that enable the training of segmentation models utilizing both channels for improved performance. Furthermore, we present a generative adversarial network (GAN) training scheme that generates not only image data but also matching labels. Quantitative evaluation shows superior performance of biophysical motivated synthetic training data, even outperforming manual annotation and pretrained models. This underscores the potential of incorporating biophysical modeling for enhancing synthetic training data quality.

PMID:39799275 | DOI:10.1038/s42003-025-07469-2

NPJ Precis Oncol. 2025 Jan 11;9(1):11. doi: 10.1038/s41698-024-00778-5.

ABSTRACT

Cervical cancer remains the fourth most common cancer among women worldwide. This study proposes an end-to-end deep learning framework to predict consensus molecular subtypes (CMS) in HPV-positive cervical squamous cell carcinoma (CSCC) from H&E-stained histology slides. Analysing three CSCC cohorts (n = 545), we show our Digital-CMS scores significantly stratify patients by both disease-specific (TCGA p = 0.0022, Oslo p = 0.0495) and disease-free (TCGA p = 0.0495, Oslo p = 0.0282) survival. In addition, our extensive tumour microenvironment analysis reveals differences between the two CMS subtypes, with CMS-C1 tumours exhibit increased lymphocyte presence, while CMS-C2 tumours show high nuclear pleomorphism, elevated neutrophil-to-lymphocyte ratio, and higher malignancy, correlating with poor prognosis. This study introduces a potentially clinically advantageous Digital-CMS score derived from digitised WSIs of routine H&E-stained tissue sections, offers new insights into TME differences impacting patient prognosis and potential therapeutic targets, and identifies histological patterns serving as potential surrogate markers of the CMS subtypes for clinical application.

PMID:39799271 | DOI:10.1038/s41698-024-00778-5

NPJ Digit Med. 2025 Jan 12;8(1):23. doi: 10.1038/s41746-024-01418-9.

ABSTRACT

The 12-lead electrocardiogram (ECG) is inexpensive and widely available. Whether conditions across the human disease landscape can be detected using the ECG is unclear. We developed a deep learning denoising autoencoder and systematically evaluated associations between ECG encodings and ~1,600 Phecode-based diseases in three datasets separate from model development, and meta-analyzed the results. The latent space ECG model identified associations with 645 prevalent and 606 incident Phecodes. Associations were most enriched in the circulatory (n = 140, 82% of category-specific Phecodes), respiratory (n = 53, 62%) and endocrine/metabolic (n = 73, 45%) categories, with additional associations across the phenome. The strongest ECG association was with hypertension (p < 2.2×10-308). The ECG latent space model demonstrated more associations than models using standard ECG intervals, and offered favorable discrimination of prevalent disease compared to models comprising age, sex, and race. We further demonstrate how latent space models can be used to generate disease-specific ECG waveforms and facilitate individual disease profiling.

PMID:39799251 | DOI:10.1038/s41746-024-01418-9

Sci Rep. 2025 Jan 11;15(1):1670. doi: 10.1038/s41598-025-85632-9.

ABSTRACT

This paper introduces a novel method for spleen segmentation in ultrasound images, using a two-phase training approach. In the first phase, the SegFormerB0 network is trained to provide an initial segmentation. In the second phase, the network is further refined using the Pix2Pix structure, which enhances attention to details and corrects any erroneous or additional segments in the output. This hybrid method effectively combines the strengths of both SegFormer and Pix2Pix to produce highly accurate segmentation results. We have assembled the Spleenex dataset, consisting of 450 ultrasound images of the spleen, which is the first dataset of its kind in this field. Our method has been validated on this dataset, and the experimental results show that it outperforms existing state-of-the-art models. Specifically, our approach achieved a mean Intersection over Union (mIoU) of 94.17% and a mean Dice (mDice) score of 96.82%, surpassing models such as Splenomegaly Segmentation Network (SSNet), U-Net, and Variational autoencoder based methods. The proposed method also achieved a Mean Percentage Length Error (MPLE) of 3.64%, further demonstrating its accuracy. Furthermore, the proposed method has demonstrated strong performance even in the presence of noise in ultrasound images, highlighting its practical applicability in clinical environments.

PMID:39799236 | DOI:10.1038/s41598-025-85632-9

Sci Rep. 2025 Jan 11;15(1):1736. doi: 10.1038/s41598-024-84193-7.

ABSTRACT

Deep learning (DL) methods have demonstrated remarkable effectiveness in assisting with lung cancer risk prediction tasks using computed tomography (CT) scans. However, the lack of comprehensive comparison and validation of state-of-the-art (SOTA) models in practical settings limits their clinical application. This study aims to review and analyze current SOTA deep learning models for lung cancer risk prediction (malignant-benign classification). To evaluate our model's general performance, we selected 253 out of 467 patients from a subset of the National Lung Screening Trial (NLST) who had CT scans without contrast, which are the most commonly used, and divided them into training and test cohorts. The CT scans were preprocessed into 2D-image and 3D-volume formats according to their nodule annotations. We evaluated ten 3D and eleven 2D SOTA deep learning models, which were pretrained on large-scale general-purpose datasets (Kinetics and ImageNet) and radiological datasets (3DSeg-8, nnUnet and RadImageNet), for their lung cancer risk prediction performance. Our results showed that 3D-based deep learning models generally perform better than 2D models. On the test cohort, the best-performing 3D model achieved an AUROC of 0.86, while the best 2D model reached 0.79. The lowest AUROCs for the 3D and 2D models were 0.70 and 0.62, respectively. Furthermore, pretraining on large-scale radiological image datasets did not show the expected performance advantage over pretraining on general-purpose datasets. Both 2D and 3D deep learning models can handle lung cancer risk prediction tasks effectively, although 3D models generally have superior performance than their 2D competitors. Our findings highlight the importance of carefully selecting pretrained datasets and model architectures for lung cancer risk prediction. Overall, these results have important implications for the development and clinical integration of DL-based tools in lung cancer screening.

PMID:39799226 | DOI:10.1038/s41598-024-84193-7

Sci Rep. 2025 Jan 11;15(1):1726. doi: 10.1038/s41598-025-85866-7.

ABSTRACT

Network security is crucial in today's digital world, since there are multiple ongoing threats to sensitive data and vital infrastructure. The aim of this study to improve network security by combining methods for instruction detection from machine learning (ML) and deep learning (DL). Attackers have tried to breach security systems by accessing networks and obtaining sensitive information.Intrusion detection systems (IDSs) are one of the significant aspect of cybersecurity that involve the monitoring and analysis, with the intention of identifying and reporting of dangerous activities that would help to prevent the attack.Support Vector Machine (SVM), K-Nearest Neighbors (KNN), Random Forest (RF), Decision Tree (DT), Long Short-Term Memory (LSTM), and Artificial Neural Network (ANN) are the vector figures incorporated into the study through the results. These models are subjected to various test to established the best results on the identification and prevention of network violation. Based on the obtained results, it can be stated that all the tested models are capable of organizing data originating from network traffic. thus, recognizing the difference between normal and intrusive behaviors, models such as SVM, KNN, RF, and DT showed effective results. Deep learning models LSTM and ANN rapidly find long-term and complex pattern in network data. It is extremely effective when dealing with complex intrusions since it is characterised by high precision, accuracy and recall.Based on our study, SVM and Random Forest are considered promising solutions for real-world IDS applications because of their versatility and explainability. For the companies seeking IDS solutions which are reliable and at the same time more interpretable, these models can be promising. Additionally, LSTM and ANN, with their ability to catch successive conditions, are suitable for situations involving nuanced, advancing dangers.

PMID:39799225 | DOI:10.1038/s41598-025-85866-7