Deep learning
Arterial phase CT radiomics for non-invasive prediction of Ki-67 proliferation index in pancreatic solid pseudopapillary neoplasms
Abdom Radiol (NY). 2025 Apr 3. doi: 10.1007/s00261-025-04921-z. Online ahead of print.
ABSTRACT
BACKGROUND: This study aimed to preoperatively predict Ki-67 proliferation levels in patients with pancreatic solid pseudopapillary neoplasm (pSPN) using radiomics features extracted from arterial phase helical CT images.
METHODS: We retrospectively analyzed 92 patients (Ningbo Medical Center Lihuili Hospital: n = 64, Taizhou Central Hospital: n = 28) with pathologically confirmed pSPN from June 2015 to June 2023. Ki-67 positivity > 3% was considered high. Radiomics features were extracted using PyRadiomics, with patients from training cohort (n = 64) and validation cohort (n = 28). A radiomics signature was constructed, and a CT radiomics score (CTscore) was calculated. Deep learning models were employed for prediction, with early stopping to prevent overfitting.
RESULTS: Seven key radiomics features were selected via LASSO regression with cross-validation. The deep learning model demonstrated improved accuracy with demographics and CTscore, with key features such as Morphology and CTscore contributing significantly to predictive accuracy. The best-performing models, including GBM and deep learning algorithms, achieved high predictive performance with an AUC of up to 0.946 in the training cohort.
CONCLUSIONS: We developed a robust deep learning-based radiomics model using arterial phase CT images to predict Ki-67 levels in pSPN patients, identifying CTscore and Morphology as key predictors. This non-invasive approach has potential utility in guiding personalized preoperative treatment strategies.
CLINICAL TRIAL NUMBER: Not applicable.
PMID:40178588 | DOI:10.1007/s00261-025-04921-z
Free-breathing, Highly Accelerated, Single-beat, Multisection Cardiac Cine MRI with Generative Artificial Intelligence
Radiol Cardiothorac Imaging. 2025 Apr;7(2):e240272. doi: 10.1148/ryct.240272.
ABSTRACT
Purpose To develop and evaluate a free-breathing, highly accelerated, multisection, single-beat cine sequence for cardiac MRI. Materials and Methods This prospective study, conducted from July 2022 to December 2023, included participants with various cardiac conditions as well as healthy participants who were imaged using a 3-T MRI system. A single-beat sequence was implemented, collecting data for each section in one heartbeat. Images were acquired with an in-plane spatiotemporal resolution of 1.9 × 1.9 mm2 and 37 msec and reconstructed using resolution enhancement generative adversarial inline neural network (REGAIN), a deep learning model. Multibreath-hold k-space-segmented (4.2-fold acceleration) and free-breathing single-beat (14.8-fold acceleration) cine images were collected, both reconstructed with REGAIN. Left ventricular (LV) and right ventricular (RV) parameters between the two methods were evaluated with linear regression, Bland-Altman analysis, and Pearson correlation. Three expert cardiologists independently scored diagnostic and image quality. Scan and rescan reproducibility was evaluated in a subset of participants 1 year apart using the intraclass correlation coefficient (ICC). Results This study included 136 participants (mean age [SD], 54 years ± 15; 69 female, 67 male), 40 healthy and 96 with cardiac conditions. k-Space-segmented and single-beat scan times were 2.6 minutes ± 0.8 and 0.5 minute ± 0.1, respectively. Strong correlations (P < .001) were observed between k-space-segmented and single-beat cine parameters in both LV (r = 0.97-0.99) and RV (r = 0.89-0.98). Scan and rescan reproducibility of single-beat cine was excellent (ICC, 0.97-1.0). Agreement among readers was high, with 125 of 136 (92%) images consistently assessed as diagnostic and 133 of 136 (98%) consistently rated as having good image quality by all readers. Conclusion Free-breathing 30-second single-beat cardiac cine MRI yielded accurate biventricular measurements, reduced scan time, and maintained high diagnostic and image quality compared with conventional multibreath-hold k-space-segmented cine images. Keywords: MR-Imaging, Cardiac, Heart, Imaging Sequences, Comparative Studies, Technology Assessment Supplemental material is available for this article. © RSNA, 2025.
PMID:40178397 | DOI:10.1148/ryct.240272
CoupleVAE: coupled variational autoencoders for predicting perturbational single-cell RNA sequencing data
Brief Bioinform. 2025 Mar 4;26(2):bbaf126. doi: 10.1093/bib/bbaf126.
ABSTRACT
With the rapid advances in single-cell sequencing technology, it is now feasible to conduct in-depth genetic analysis in individual cells. Study on the dynamics of single cells in response to perturbations is of great significance for understanding the functions and behaviors of living organisms. However, the acquisition of post-perturbation cellular states via biological experiments is frequently cost-prohibitive. Predicting the single-cell perturbation responses poses a critical challenge in the field of computational biology. In this work, we propose a novel deep learning method called coupled variational autoencoders (CoupleVAE), devised to predict the postperturbation single-cell RNA-Seq data. CoupleVAE is composed of two coupled VAEs connected by a coupler, initially extracting latent features for controlled and perturbed cells via two encoders, subsequently engaging in mutual translation within the latent space through two nonlinear mappings via a coupler, and ultimately generating controlled and perturbed data by two separate decoders to process the encoded and translated features. CoupleVAE facilitates a more intricate state transformation of single cells within the latent space. Experiments in three real datasets on infection, stimulation and cross-species prediction show that CoupleVAE surpasses the existing comparative models in effectively predicting single-cell RNA-seq data for perturbed cells, achieving superior accuracy.
PMID:40178283 | DOI:10.1093/bib/bbaf126
Data imbalance in drug response prediction: multi-objective optimization approach in deep learning setting
Brief Bioinform. 2025 Mar 4;26(2):bbaf134. doi: 10.1093/bib/bbaf134.
ABSTRACT
Drug response prediction (DRP) methods tackle the complex task of associating the effectiveness of small molecules with the specific genetic makeup of the patient. Anti-cancer DRP is a particularly challenging task requiring costly experiments as underlying pathogenic mechanisms are broad and associated with multiple genomic pathways. The scientific community has exerted significant efforts to generate public drug screening datasets, giving a path to various machine learning models that attempt to reason over complex data space of small compounds and biological characteristics of tumors. However, the data depth is still lacking compared to application domains like computer vision or natural language processing domains, limiting current learning capabilities. To combat this issue and improves the generalizability of the DRP models, we are exploring strategies that explicitly address the imbalance in the DRP datasets. We reframe the problem as a multi-objective optimization across multiple drugs to maximize deep learning model performance. We implement this approach by constructing Multi-Objective Optimization Regularized by Loss Entropy loss function and plugging it into a Deep Learning model. We demonstrate the utility of proposed drug discovery methods and make suggestions for further potential application of the work to achieve desirable outcomes in the healthcare field.
PMID:40178282 | DOI:10.1093/bib/bbaf134
DOMSCNet: a deep learning model for the classification of stomach cancer using multi-layer omics data
Brief Bioinform. 2025 Mar 4;26(2):bbaf115. doi: 10.1093/bib/bbaf115.
ABSTRACT
The rapid advancement of next-generation sequencing (NGS) technology and the expanding availability of NGS datasets have led to a significant surge in biomedical research. To better understand the molecular processes, underlying cancer and to support its development, diagnosis, prediction, and therapy; NGS data analysis is crucial. However, the NGS multi-layer omics high-dimensional dataset is highly complex. In recent times, some computational methods have been developed for cancer omics data interpretation. However, various existing methods face challenges in accounting for diverse types of cancer omics data and struggle to effectively extract informative features for the integrated identification of core units. To address these challenges, we proposed a hybrid feature selection (HFS) technique to detect optimal features from multi-layer omics datasets. Subsequently, this study proposes a novel hybrid deep recurrent neural network-based model DOMSCNet to classify stomach cancer. The proposed model was made generic for all four multi-layer omics datasets. To observe the robustness of the DOMSCNet model, the proposed model was validated with eight external datasets. Experimental results showed that the SelectKBest-maximum relevancy minimum redundancy-Boruta (SMB), HFS technique outperformed all other HFS techniques. Across four multi-layer omics datasets and validated datasets, the proposed DOMSCNet model outdid existing classifiers along with other proposed classifiers.
PMID:40178281 | DOI:10.1093/bib/bbaf115
Application of Deep Learning to Predict the Persistence, Bioaccumulation, and Toxicity of Pharmaceuticals
J Chem Inf Model. 2025 Apr 3. doi: 10.1021/acs.jcim.4c02293. Online ahead of print.
ABSTRACT
This study investigates the application of a deep learning (DL) model, specifically a message-passing neural network (MPNN) implemented through Chemprop, to predict the persistence, bioaccumulation, and toxicity (PBT) characteristics of compounds, with a focus on pharmaceuticals. We employed a clustering strategy to provide a fair assessment of the model performances. By applying the generated model to a set of pharmaceutically relevant molecules, we aim to highlight potential PBT chemicals and extract PBT-relevant substructures. These substructures can serve as structural flags, alerting drug designers to potential environmental issues from the earliest stages of the drug discovery process. Incorporating these findings into pharmaceutical development workflows is expected to drive significant advancements in creating more environmentally friendly drug candidates while preserving their therapeutic efficacy.
PMID:40178174 | DOI:10.1021/acs.jcim.4c02293
Early Colon Cancer Prediction from Histopathological Images Using Enhanced Deep Learning with Confidence Scoring
Cancer Invest. 2025 Apr 3:1-19. doi: 10.1080/07357907.2025.2483302. Online ahead of print.
ABSTRACT
Colon Cancer (CC) arises from abnormal cell growth in the colon, which severely impacts a person's health and quality of life. Detecting CC through histopathological images for early diagnosis offers substantial benefits in medical diagnostics. This study proposes NalexNet, a hybrid deep-learning classifier, to enhance classification accuracy and computational efficiency. The research methodology involves Vahadane stain normalization for preprocessing and Watershed segmentation for accurate tissue separation. The Teamwork Optimization Algorithm (TOA) is employed for optimal feature selection to reduce redundancy and improve classification performance. Furthermore, the NalexNet model is structured with convolutional layers and normal and reduction cells, ensuring efficient feature representation and high classification accuracy. Experimental results demonstrate that the proposed model achieves a precision of 99.9% and an accuracy of 99.5%, significantly outperforming existing models. This study contributes to the development of an automated and computationally efficient CC classification system, which has the potential for real-world clinical implementation, aiding pathologists in early and accurate diagnosis.
PMID:40178023 | DOI:10.1080/07357907.2025.2483302
CMV2U-Net: A U-shaped network with edge-weighted features for detecting and localizing image splicing
J Forensic Sci. 2025 Apr 3. doi: 10.1111/1556-4029.70033. Online ahead of print.
ABSTRACT
The practice of cutting and pasting portions of one image into another, known as "image splicing," is commonplace in the field of image manipulation. Image splicing detection using deep learning has been a hot research topic for the past few years. However, there are two problems with the way deep learning is currently implemented: first, it is not good enough for feature fusion, and second, it uses only simple models for feature extraction and encoding, which makes the models vulnerable to overfitting. To tackle these problems, this research proposes CMV2U-Net, an edge-weighted U-shaped network-based image splicing forgery localization approach. An initial step is the development of a feature extraction module that can process two streams of input images simultaneously, allowing for the simultaneous extraction of semantically connected and semantically agnostic features. One characteristic is that a hierarchical fusion approach has been devised to prevent data loss in shallow features that are either semantically related or semantically irrelevant. This approach implements a channel attention mechanism to monitor manipulation trajectories involving multiple levels. Extensive trials on numerous public datasets prove that CMV2U-Net provides high AUC and F1 in localizing tampered regions, outperforming state-of-the-art techniques. Noise, Gaussian blur, and JPEG compression are post-processing threats that CMV2U-Net has successfully resisted.
PMID:40177991 | DOI:10.1111/1556-4029.70033
Deep Learning-Powered Colloidal Digital SERS for Precise Monitoring of Cell Culture Media
Nano Lett. 2025 Apr 3. doi: 10.1021/acs.nanolett.5c01071. Online ahead of print.
ABSTRACT
Maintaining consistent quality in biomanufacturing is essential for producing high-quality complex biologics. Yet, current process analytical technologies (PAT) often fall short in achieving rapid and accurate monitoring of small-molecule critical process parameters and critical quality attributes. Surface-enhanced Raman spectroscopy (SERS) holds great promise but faces challenges like intensity fluctuations, compromising reproducibility. Herein, we propose a deep learning-powered colloidal digital SERS platform. This innovation converts SERS spectra into binary "ON/OFF" signals based on defined intensity thresholds, which allows single-molecule event visualization and reduces false positives. Through integration with deep learning, this platform enables detection of a broad range of analytes, unlimited by the lack of characteristic SERS peaks. Furthermore, we demonstrate its accuracy and reproducibility for studying AMBIC 1.1 mammalian cell culture media. These results highlight its rapidity, accuracy, and precision, paving the way for widespread adoption and scale-up as a novel PAT tool in biomanufacturing and diagnostics.
PMID:40177940 | DOI:10.1021/acs.nanolett.5c01071
ChatGPT for speech-impaired assistance
Disabil Rehabil Assist Technol. 2025 Apr 3:1-3. doi: 10.1080/17483107.2025.2483300. Online ahead of print.
ABSTRACT
Background: Speech and language impairments, though often used interchangeably, are two very distinct types of challenges. A speech impairment may lead to impaired ability to produce speech sounds whilst communication may be affected due to lack of fluency or articulation of words. Consequently this may affect a person's ability to articulate may affect academic achievement, social development and progress in life. ChatGPT (Generative Pretrained Transformer) is an open access AI (Artificial Intelligence) tool developed by Open AI® based on Large language models (LLMs) with the ability to respond to human prompts to generate texts using Supervised and Unsupervised Machine Learning (ML) Algorithms. This article explores the current role and future perspectives of ChatGPT AI Tool for Speech-Impaired Assistance.
Methods: A cumulative search strategy using databases of PubMed, Google Scholar, Scopus and grey literature was conducted to generate this narrative review.
Results: A spectrum of Enabling Technologies for Speech & Language Impairment have been explored. Augmentative and Alternative Communication technology (AAC), Integration with Neuroprosthesis technology and Speech therapy applications offer considerable potential to aid speech and language impaired individuals.
Conclusion: Current applications of AI, ChatGPT and other LLM's offer promising solutions in enhancing communication in people affected by Speech and Language impairment. However, further research and development is required to ensure affordability, accessibility and authenticity of these AI Tools in clinical Practice.
PMID:40177878 | DOI:10.1080/17483107.2025.2483300
Age-sex-specific burden of urological cancers attributable to risk factors in China and its provinces, 1990-2021, and forecasts with scenarios simulation: a systematic analysis for the Global Burden of Disease Study 2021
Lancet Reg Health West Pac. 2025 Mar 18;56:101517. doi: 10.1016/j.lanwpc.2025.101517. eCollection 2025 Mar.
ABSTRACT
BACKGROUND: As global aging intensifies, urological cancers pose increasing health and economic burdens. In China, home to one-fifth of the world's population, monitoring the distribution and determinants of these cancers and simulating the effects of health interventions are crucial for global and national health.
METHODS: With Global Burden of Disease (GBD) China database, the present study analyzed age-sex-specific patterns of incidence, prevalence, mortality, disability-adjusted life years (DALYs), years lived with disability (YLDs), and years of life lost (YLLs) in China and its 34 provinces as well as the association between gross domestic product per capita (GDPPC) and these patterns. Importantly, a multi-attentive deep learning pipeline (iTransformer) was pioneered to model the spatiotemporal patterns of urological cancers, risk factors, GDPPC, and population, to provide age-sex-location-specific long-term forecasts of urological cancer burdens, and to investigate the impacts of risk-factor-directed interventions on their future burdens.
FINDINGS: From 1990 to 2021, the incidence and prevalence of urological cancers in China has increased, leading to 266,887 new cases (95% confidence interval: 205,304-346,033) and 159,506,067 (12,236,0000-207,447,070) cases in 2021, driven primarily by males aged 55+ years. In 2021, Taiwan, Beijing, and Zhejiang had the highest age-standardized incidence rate (ASIR) and age-standardized prevalence rates of urological cancer in China, highlighting significant regional disparities in the disease burden. Conversely, the national age-standardized mortality rate (ASMR) has declined from 6.5 (5.1-7.8) per 100,000 population in 1990 to 5.6 (4.4-7.2) in 2021, notably in Jilin [-166.7% (-237 to -64.6)], Tibet [-135.4% (-229.1 to 4.4)], and Heilongjiang [-118.5% (-206.5 to -4.6)]. Specifically, the national ASMR for bladder and testicular cancers reduced by -32.1% (-47.9 to 1.9) and -31.1% (-50.2 to 7.2), respectively, whereas prostate and kidney cancers rose by 7.9% (-18.4 to 43.6) and 9.2% (-12.2 to 36.5). Age-standardized DALYs, YLDs, and YLLs for urological cancers were consistent with ASMR. Males suffered higher burdens of urological cancers than females in all populations, except those aged <5 years. Regionally and provincially, high GDPPC provinces have the highest burden of prostate cancer, while the main burden in other provinces is bladder cancer. The main risk factors for urological cancers in 2021 are smoking [accounting for 55.1% (42.7-67.4)], high body mass index [13.9% (5.3-22.4)], and high fasting glycemic index [5.9% (-0.8 to 13.4)] for both males and females, with smoking remarkably affecting males and high body mass index affecting females. Between 2022 and 2040, the ASIR of urological cancers increased from 10.09 (9.19-10.99) to 14.42 (14.30-14.54), despite their ASMR decreasing. Notably, prostate cancer surpassed bladder cancer as the primary subcategory, with those aged 55+ years showing the highest increase in ASIR, highlighting the aging-related transformation of the urological cancer burden. Following the implementation of targeted interventions, smoking control achieved the greatest reduction in urological cancer burden, mainly affecting male bladder cancer (-45.8% decline). In females, controlling smoking and high fasting plasma glucose reduced by 5.3% and 5.8% ASMR in urological cancers. Finally, the averaged mean-square-Percentage-Error, absolute-Percentage-Error, and root-mean-square Logarithmic-Error of the forecasting model are 0.54 ± 0.22, 1.51 ± 1.26, and 0.15 ± 0.07, respectively, indicating that the model performs well.
INTERPRETATION: Urological cancers exhibit an aging trend, with increased incidence rates among the population aged 55+ years, making prostate cancer the most burdensome subcategory. Moreover, urological cancer burden is imbalanced by age, sex, and province. Based on our findings, authorities and policymakers could refine or tailor population-specific health strategies, including promoting smoking cessation, weight reduction, and blood sugar control.
FUNDING: Bill & Melinda Gates Foundation.
PMID:40177596 | PMC:PMC11964562 | DOI:10.1016/j.lanwpc.2025.101517
The promise and limitations of artificial intelligence in CTPA-based pulmonary embolism detection
Front Med (Lausanne). 2025 Mar 19;12:1514931. doi: 10.3389/fmed.2025.1514931. eCollection 2025.
ABSTRACT
Computed tomography pulmonary angiography (CTPA) is an essential diagnostic tool for identifying pulmonary embolism (PE). The integration of AI has significantly advanced CTPA-based PE detection, enhancing diagnostic accuracy and efficiency. This review investigates the growing role of AI in the diagnosis of pulmonary embolism using CTPA imaging. The review examines the capabilities of AI algorithms, particularly deep learning models, in analyzing CTPA images for PE detection. It assesses their sensitivity and specificity compared to human radiologists. AI systems, using large datasets and complex neural networks, demonstrate remarkable proficiency in identifying subtle signs of PE, aiding clinicians in timely and accurate diagnosis. In addition, AI-powered CTPA analysis shows promise in risk stratification, prognosis prediction, and treatment optimization for PE patients. Automated image interpretation and quantitative analysis facilitate rapid triage of suspected cases, enabling prompt intervention and reducing diagnostic delays. Despite these advancements, several limitations remain, including algorithm bias, interpretability issues, and the necessity for rigorous validation, which hinder widespread adoption in clinical practice. Furthermore, integrating AI into existing healthcare systems requires careful consideration of regulatory, ethical, and legal implications. In conclusion, AI-driven CTPA-based PE detection presents unprecedented opportunities to enhance diagnostic precision and efficiency. However, addressing the associated limitations is critical for safe and effective implementation in routine clinical practice. Successful utilization of AI in revolutionizing PE care necessitates close collaboration among researchers, medical professionals, and regulatory organizations.
PMID:40177281 | PMC:PMC11961422 | DOI:10.3389/fmed.2025.1514931
Construction of a predictive model for the efficacy of anti-VEGF therapy in macular edema patients based on OCT imaging: a retrospective study
Front Med (Lausanne). 2025 Mar 19;12:1505530. doi: 10.3389/fmed.2025.1505530. eCollection 2025.
ABSTRACT
BACKGROUND: Macular edema (ME) is an ophthalmic disease that poses a serious threat to human vision. Anti-vascular endothelial growth factor (anti-VEGF) therapy has become the first-line treatment for ME due to its safety and high efficacy. However, there are still cases of refractory macular edema and non-responding patients. Therefore, it is crucial to develop automated and efficient methods for predicting therapeutic outcomes.
METHODS: We have developed a predictive model for the surgical efficacy in ME patients based on deep learning and optical coherence tomography (OCT) imaging, aimed at predicting the treatment outcomes at different time points. This model innovatively introduces group convolution and multiple convolutional kernels to handle multidimensional features based on traditional attention mechanisms for visual recognition tasks, while utilizing spatial pyramid pooling (SPP) to combine and extract the most useful features. Additionally, the model uses ResNet50 as a pre-trained model, integrating multiple knowledge through model fusion.
RESULTS: Our proposed model demonstrated the best performance across various experiments. In the ablation study, the model achieved an F1 score of 0.9937, an MCC of 0.7653, an AUC of 0.9928, and an ACC of 0.9877 in the test conducted on the first day after surgery. In comparison experiments, the ACC of our model was 0.9930 and 0.9915 in the first and the third months post-surgery, respectively, with AUC values of 0.9998 and 0.9996, significantly outperforming other models. In conclusion, our model consistently exhibited superior performance in predicting outcomes at various time points, validating its excellence in processing OCT images and predicting postoperative efficacy.
CONCLUSION: Through precise prediction of the response to anti-VEGF therapy in ME patients, deep learning technology provides a revolutionary tool for the treatment of ophthalmic diseases, significantly enhancing treatment outcomes and improving patients' quality of life.
PMID:40177270 | PMC:PMC11961644 | DOI:10.3389/fmed.2025.1505530
Measurement-guided therapeutic-dose prediction using multi-level gated modality-fusion model for volumetric-modulated arc radiotherapy
Front Oncol. 2025 Mar 19;15:1468232. doi: 10.3389/fonc.2025.1468232. eCollection 2025.
ABSTRACT
OBJECTIVES: Radiotherapy is a fundamental cancer treatment method, and pre-treatment patient-specific quality assurance (prePSQA) plays a crucial role in ensuring dose accuracy and patient safety. Artificial intelligence model for measurement-free prePSQA have been investigated over the last few years. While these models stack successive pooling layers to carry out sequential learning, directly splice together different modalities along channel dimensions and feed them into shared encoder-decoder network, which greatly reduces the anatomical features specific to different modalities. Furthermore, the existing models simply take advantage of low-dimensional dosimetry information, meaning that the spatial features about the complex dose distribution may be lost and limiting the predictive power of the models. The purpose of this study is to develop a novel deep learning model for measurement-guided therapeutic-dose (MDose) prediction from head and neck cancer radiotherapy data.
METHODS: The enrolled 310 patients underwent volumetric-modulated arc radiotherapy (VMAT) were randomly divided into the training set (186 cases, 60%), validation set (62 cases, 20%), and test set (62 cases, 20%). The effective prediction model explicitly integrates the multi-scale features that are specific to CT and dose images, takes into account the useful spatial dose information and fully exploits the mutual promotion within the different modalities. It enables medical physicists to analyze the detailed locations of spatial dose differences and to simultaneously generate clinically applicable dose-volume histograms (DVHs) metrics and gamma passing rate (GPR) outcomes.
RESULTS: The proposed model achieved better performance of MDose prediction, and dosimetric congruence of DVHs, GPR with the ground truth compared with several state-of-the-art models. Quantitative experimental predictions show that the proposed model achieved the lowest values for the mean absolute error (37.99) and root mean square error (4.916), and the highest values for the peak signal-to-noise ratio (52.622), structural similarity (0.986) and universal quality index (0.932). The predicted dose values of all voxels were within 6 Gy in the dose difference maps, except for the areas near the skin or thermoplastic mask indentation boundaries.
CONCLUSIONS: We have developed a feasible MDose prediction model that could potentially improve the efficiency and accuracy of prePSQA for head and neck cancer radiotherapy, providing a boost for clinical adaptive radiotherapy.
PMID:40177241 | PMC:PMC11961879 | DOI:10.3389/fonc.2025.1468232
A flexible transoral swab sampling robot system with visual-tactile fusion approach
Front Robot AI. 2025 Mar 19;12:1520374. doi: 10.3389/frobt.2025.1520374. eCollection 2025.
ABSTRACT
A significant number of individuals have been affected by pandemic diseases, such as COVID-19 and seasonal influenza. Nucleic acid testing is a common method for identifying infected patients. However, manual sampling methods require the involvement of numerous healthcare professionals. To address this challenge, we propose a novel transoral swab sampling robot designed to autonomously perform nucleic acid sampling using a visual-tactile fusion approach. The robot comprises a series-parallel hybrid flexible mechanism for precise distal posture adjustment and a visual-tactile perception module for navigation within the subject's oral cavity. The series-parallel hybrid mechanism, driven by flexible shafts, enables omnidirectional bending through coordinated movement of the two segments of the bendable joint. The visual-tactile perception module incorporates a camera to capture oral images of the subject and recognize the nucleic acid sampling point using a deep learning method. Additionally, a force sensor positioned at the distal end of the robot provides feedback on contact force as the swab is inserted into the subject's oral cavity. The sampling robot is capable of autonomously performing transoral swab sampling while navigating using the visual-tactile perception algorithm. Preliminary experimental trials indicate that the designed robot system is feasible, safe, and accurate for sample collection from subjects.
PMID:40177224 | PMC:PMC11961991 | DOI:10.3389/frobt.2025.1520374
Developing predictive models for opioid receptor binding using machine learning and deep learning techniques
Exp Biol Med (Maywood). 2025 Mar 19;250:10359. doi: 10.3389/ebm.2025.10359. eCollection 2025.
ABSTRACT
Opioids exert their analgesic effect by binding to the µ opioid receptor (MOR), which initiates a downstream signaling pathway, eventually inhibiting pain transmission in the spinal cord. However, current opioids are addictive, often leading to overdose contributing to the opioid crisis in the United States. Therefore, understanding the structure-activity relationship between MOR and its ligands is essential for predicting MOR binding of chemicals, which could assist in the development of non-addictive or less-addictive opioid analgesics. This study aimed to develop machine learning and deep learning models for predicting MOR binding activity of chemicals. Chemicals with MOR binding activity data were first curated from public databases and the literature. Molecular descriptors of the curated chemicals were calculated using software Mold2. The chemicals were then split into training and external validation datasets. Random forest, k-nearest neighbors, support vector machine, multi-layer perceptron, and long short-term memory models were developed and evaluated using 5-fold cross-validations and external validations, resulting in Matthews correlation coefficients of 0.528-0.654 and 0.408, respectively. Furthermore, prediction confidence and applicability domain analyses highlighted their importance to the models' applicability. Our results suggest that the developed models could be useful for identifying MOR binders, potentially aiding in the development of non-addictive or less-addictive drugs targeting MOR.
PMID:40177220 | PMC:PMC11961360 | DOI:10.3389/ebm.2025.10359
Global trends in artificial intelligence applications in liver disease over seventeen years
World J Hepatol. 2025 Mar 27;17(3):101721. doi: 10.4254/wjh.v17.i3.101721.
ABSTRACT
BACKGROUND: In recent years, the utilization of artificial intelligence (AI) technology has gained prominence in the field of liver disease.
AIM: To analyzes AI research in the field of liver disease, summarizes the current research status and identifies hot spots.
METHODS: We searched the Web of Science Core Collection database for all articles and reviews on hepatopathy and AI. The time spans from January 2007 to August 2023. We included 4051 studies for further collection of information, including authors, countries, institutions, publication years, keywords and references. VOS viewer, CiteSpace, R 4.3.1 and Scimago Graphica were used to visualize the results.
RESULTS: A total of 4051 articles were analyzed. China was the leading contributor, with 1568 publications, while the United States had the most international collaborations. The most productive institutions and journals were the Chinese Academy of Sciences and Frontiers in Oncology. Keywords co-occurrence analysis can be roughly summarized into four clusters: Risk prediction, diagnosis, treatment and prognosis of liver diseases. "Machine learning", "deep learning", "convolutional neural network", "CT", and "microvascular infiltration" have been popular research topics in recent years.
CONCLUSION: AI is widely applied in the risk assessment, diagnosis, treatment, and prognosis of liver diseases, with a shift from invasive to noninvasive treatment approaches.
PMID:40177211 | PMC:PMC11959664 | DOI:10.4254/wjh.v17.i3.101721
Conditioning generative latent optimization for sparse-view computed tomography image reconstruction
J Med Imaging (Bellingham). 2025 Mar;12(2):024004. doi: 10.1117/1.JMI.12.2.024004. Epub 2025 Apr 1.
ABSTRACT
PURPOSE: The issue of delivered doses during computed tomography (CT) scans encouraged sparser sets of X-ray projection, severely degrading reconstructions from conventional methods. Although most deep learning approaches benefit from large supervised datasets, they cannot generalize to new acquisition protocols (geometry, source/detector specifications). To address this issue, we developed a method working without training data and independently of experimental setups. In addition, our model may be initialized on small unsupervised datasets to enhance reconstructions.
APPROACH: We propose a conditioned generative latent optimization (cGLO) in which a decoder reconstructs multiple slices simultaneously with a shared objective. It is tested on full-dose sparse-view CT for varying projection sets: (a) without training data against Deep Image Prior (DIP) and (b) with training datasets of multiple sizes against state-of-the-art score-based generative models (SGMs). Peak signal-to-noise ratio (PSNR) and structural SIMilarity (SSIM) metrics are used to quantify reconstruction quality.
RESULTS: cGLO demonstrates better SSIM than SGMs (between + 0.034 and + 0.139 ) and has an increasing advantage for smaller datasets reaching a + 6.06 dB PSNR gain. Our strategy also outperforms DIP with at least a + 1.52 dB PSNR advantage and peaks at + 3.15 dB with fewer angles. Moreover, cGLO does not create artifacts or structural deformations contrary to DIP and SGMs.
CONCLUSIONS: We propose a parsimonious and robust reconstruction technique offering similar to better performances when compared with state-of-the-art methods regarding full-dose sparse-view CT. Our strategy could be readily applied to any imaging reconstruction task without any assumption about the acquisition protocol or the quantity of available data.
PMID:40177097 | PMC:PMC11961077 | DOI:10.1117/1.JMI.12.2.024004
Accurate V2X traffic prediction with deep learning architectures
Front Artif Intell. 2025 Mar 18;8:1565287. doi: 10.3389/frai.2025.1565287. eCollection 2025.
ABSTRACT
Vehicle-to-Everything (V2X) communication promises to revolutionize road safety and efficiency. However, challenges in data sharing and network reliability impede its full realization. This paper addresses these challenges by proposing a novel Deep Learning (DL) approach for traffic prediction in V2X environments. We employ Bidirectional Long Short-Term Memory (BiLSTM) networks and compare their performance against other prominent DL architectures, including unidirectional LSTM and Gated Recurrent Unit (GRU). Our findings demonstrate that the BiLSTM model exhibits superior accuracy in predicting traffic patterns. This enhanced prediction capability enables more efficient resource allocation, improved network performance, and enhanced safety for all road users, reducing fuel consumption, decreased emissions, and a more sustainable transportation system.
PMID:40176965 | PMC:PMC11962783 | DOI:10.3389/frai.2025.1565287
Automated Sleep Staging in Epilepsy Using Deep Learning on Standard Electroencephalogram and Wearable Data
J Sleep Res. 2025 Apr 3:e70061. doi: 10.1111/jsr.70061. Online ahead of print.
ABSTRACT
Automated sleep staging on wearable data could improve our understanding and management of epilepsy. This study evaluated sleep scoring by a deep learning model on 223 night-sleep recordings from 50 patients measured in the hospital with an electroencephalogram (EEG) and a wearable device. The model scored the sleep stage of every 30-s epoch on the EEG and wearable data, and we compared the output with a clinical expert on 20 nights, each for a different patient. The Bland-Altman analysis examined differences in the automated staging in both modalities, and using mixed-effect models, we explored sleep differences between patients with and without seizures. Overall, we found moderate accuracy and Cohen's kappa on the model scoring of standard EEG (0.73 and 0.59) and the wearable (0.61 and 0.43) versus the clinical expert. F1 scores also varied between patients and the modalities. The sensitivity varied by sleep stage and was very low for stage N1. Moreover, sleep staging on the wearable data underestimated the duration of most sleep macrostructure parameters except N2. On the other hand, patients with seizures during the hospital admission slept more each night (6.37, 95% confidence interval [CI] 5.86-7.87) compared with patients without seizures (5.68, 95% CI 5.24-6.13), p = 0.001, but also spent more time in stage N2. In conclusion, wearable EEG and accelerometry could monitor sleep in patients with epilepsy, and our approach can help automate the analysis. However, further steps are essential to improve the model performance before clinical implementation. Trial Registration: The SeizeIT2 trial was registered in clinicaltrials.gov, NCT04284072.
PMID:40176726 | DOI:10.1111/jsr.70061