Heliyon. 2025 Jan 9;11(3):e41850. doi: 10.1016/j.heliyon.2025.e41850. eCollection 2025 Feb 15.

ABSTRACT

BACKGROUND: As one of the cancers with the highest incidence and mortality rates worldwide, the timeliness and accuracy of cell type diagnosis in lung cancer are crucial for patients' treatment decisions. This study aims to develop a novel deep learning model to provide efficient, accurate, and cost-effective auxiliary diagnosis for the pathological types of lung cancer cells.

METHOD: This paper introduces a model named PortNet, designed to significantly reduce the model's parameter size and achieve lightweight characteristics without compromising classification accuracy. We incorporated 1 × 1 convolutional blocks into the Depthwise Separable Convolution architecture to further decrease the model's parameter count. Additionally, the integration of the Squeeze-and-Excitation self-attention module enhances feature representation without substantially increasing the number of parameters, thereby maintaining high predictive performance.

RESULT: Our tests demonstrated that PortNet significantly reduces the total parameter count to 2,621,827, which is over a fifth smaller compared to some mainstream CNN models, marking a substantial advancement for deployment in portable devices. We also established widely-used traditional models as benchmarks to illustrate the efficacy of PortNet. In external tests, PortNet achieved an average accuracy (ACC) of 99.89 % and Area Under the Curve (AUC) of 99.27 %. During five-fold cross-validation, PortNet maintained an average ACC of 99.51 % ± 1.50 % and F1 score of 99.50 % ± 1.51 %, showcasing its lightweight capability and exceptionally high accuracy. This presents a promising opportunity for integration into hospital systems to assist physicians in diagnosis.

CONCLUSION: This study significantly reduces the parameter count through an innovative model structure while maintaining high accuracy and stability, demonstrating outstanding performance in lung cancer cell classification tasks. The model holds the potential to become an efficient, accurate, and cost-effective auxiliary diagnostic tool for pathological classification of lung cancer in the future.

PMID:39931476 | PMC:PMC11808607 | DOI:10.1016/j.heliyon.2025.e41850

Ophthalmol Sci. 2024 Nov 28;5(2):100659. doi: 10.1016/j.xops.2024.100659. eCollection 2025 Mar-Apr.

ABSTRACT

PURPOSE: Recent studies utilized ocular images and deep learning (DL) to predict refractive error and yielded notable results. However, most studies did not address biases from imbalanced datasets or conduct external validations. To address these gaps, this study aimed to integrate the deep imbalanced regression (DIR) technique into ResNet and Vision Transformer models to predict refractive error from retinal photographs.

DESIGN: Retrospective study.

SUBJECTS: We developed the DL models using up to 103 865 images from the Singapore Epidemiology of Eye Diseases Study and the United Kingdom Biobank, with internal testing on up to 8067 images. External testing was conducted on 7043 images from the Singapore Prospective Study and 5539 images from the Beijing Eye Study. Retinal images and corresponding refractive error data were extracted.

METHODS: This retrospective study developed regression-based models, including ResNet34 with DIR, and SwinV2 (Swin Transformer) with DIR, incorporating Label Distribution Smoothing and Feature Distribution Smoothing. These models were compared against their baseline versions, ResNet34 and SwinV2, in predicting spherical and spherical equivalent (SE) power.

MAIN OUTCOME MEASURES: Mean absolute error (MAE) and coefficient of determination were used to evaluate the models' performances. The Wilcoxon signed-rank test was performed to assess statistical significance between DIR-integrated models and their baseline versions.

RESULTS: For prediction of the spherical power, ResNet34 with DIR (MAE: 0.84D) and SwinV2 with DIR (MAE: 0.77D) significantly outperformed their baseline-ResNet34 (MAE: 0.88D; P < 0.001) and SwinV2 (MAE: 0.87D; P < 0.001) in internal test. For prediction of the SE power, ResNet34 with DIR (MAE: 0.78D) and SwinV2 with DIR (MAE: 0.75D) consistently significantly outperformed its baseline-ResNet34 (MAE: 0.81D; P < 0.001) and SwinV2 (MAE: 0.78D; P < 0.05) in internal test. Similar trends were observed in external test sets for both spherical and SE power prediction.

CONCLUSIONS: Deep imbalanced regressed-integrated DL models showed potential in addressing data imbalances and improving the prediction of refractive error. These findings highlight the potential utility of combining DL models with retinal imaging for opportunistic screening of refractive errors, particularly in settings where retinal cameras are already in use.

FINANCIAL DISCLOSURES: Proprietary or commercial disclosure may be found in the Footnotes and Disclosures at the end of this article.

PMID:39931359 | PMC:PMC11808727 | DOI:10.1016/j.xops.2024.100659

Data Brief. 2025 Jan 21;59:111311. doi: 10.1016/j.dib.2025.111311. eCollection 2025 Apr.

ABSTRACT

A comprehensive Hog plum leaf disease dataset is greatly needed for agricultural research, precision agriculture, and efficient management of disease. It will find applications toward the formulation of machine learning models for early detection and classification of disease, thus reducing dependency on manual inspections and timely interventions. Such a dataset provides a benchmark for training and testing algorithms, further enhancing automated monitoring systems and decision-support tools in sustainable agriculture. It enables better crop management, less use of chemicals, and more focused agronomical practices. This dataset will contribute to the global research being carried out for the advancement of disease-resistant plant strategy development and efficient management practices for better agricultural productivity along with sustainability. These images have been collected from different regions of Bangladesh. In this work, two classes were used: 'Healthy' and 'Insect hole', representing different stages of disease progression. The augmentation techniques that involve flipping, rotating, scaling, translating, cropping, adding noise, adjusting brightness, adjusting contrast, and scaling expanded a dataset of 3782 images to 20,000 images. These have formed very robust deep learning training sets, hence better detection of the disease.

PMID:39931093 | PMC:PMC11808602 | DOI:10.1016/j.dib.2025.111311

Front Oncol. 2025 Jan 27;15:1507592. doi: 10.3389/fonc.2025.1507592. eCollection 2025.

ABSTRACT

Cervical cancer remains a significant global health concern, characterized by high morbidity and mortality rates. High-dose-rate brachytherapy (HDR-BT) is a critical component of cervical cancer treatment, requiring precise and efficient treatment planning. However, the process is labor-intensive, heavily reliant on operator expertise, and prone to variability due to factors such as applicator shifts and organ filling changes. Recent advancements in artificial intelligence (AI), particularly in medical image processing, offer significant potential for automating and standardizing treatment planning in HDR-BT. This review examines the progress and challenge of AI applications in HDR-BT treatment planning, focusing on automatic segmentation, applicator reconstruction, dose calculation, and plan optimization. By addressing current limitations and exploring future directions, this paper aims to guide the integration of AI into clinical practice, ultimately improving treatment accuracy, reducing preparation time, and enhancing patient outcomes.

PMID:39931087 | PMC:PMC11808022 | DOI:10.3389/fonc.2025.1507592

BMC Oral Health. 2025 Feb 10;25(1):216. doi: 10.1186/s12903-025-05577-3.

ABSTRACT

BACKGROUND: The aim of this study was to evaluate the efficacy of detecting dental caries under fixed dental prostheses (FDPs) through the analysis of panoramic radiographs utilizing convolutional neural network (CNN) based You Only Look Once (YOLO) models. Deep learning algorithms can analyze datasets of dental images, such as panoramic radiographs to accurately identify and classify carious lesions. Using artificial intelligence, specifically deep learning methods, may help practitioners to detect and diagnose caries using radiograph images.

METHODS: The panoramic radiographs of 1004 patients, who had FDPs on their teeth and met the inclusion criteria, were divided into 904 (90%) images as training dataset and 100 (10%) images as the test dataset. Following the attainment of elevated detection scores with YOLOv7, regions of interest (ROIs) containing FDPs were automatically detected and cropped by the YOLOv7 model. In the second stage, 2467 cropped images were divided into 2248 (91%) images as the training dataset and 219 (9%) images as the test dataset. Caries under the FDPs were detected using both the YOLOv7 and the improved YOLOv7 (YOLOv7 + CBAM) models. The performance of the deep learning models used in the study was evaluated using recall, precision, F1, and mean average precision (mAP) scores.

RESULTS: In the first stage, the YOLOv7 model achieved 0.947 recall, 0.966 precision, 0.968 mAP and 0.956 F1 scores in detecting the FDPs. In the second stage the YOLOv7 model achieved 0.791 recall, 0.837 precision, 0.800 mAP and 0.813 F1 scores in detecting the caries under the FDPs, while the YOLOv7 + CBAM model achieved 0.827 recall, 0.834 precision, 0.846 mAP, and 0.830 F1 scores.

CONCLUSION: The use of deep learning models to detect dental caries under FDPs by analyzing panoramic radiographs has shown promising results. The study highlights that panoramic radiographs with appropriate image features can be used in combination with a detection system supported by deep learning methods. In the long term, our study may allow for accurate and rapid diagnoses that significantly improve the preservation of teeth under FDPs.

PMID:39930440 | DOI:10.1186/s12903-025-05577-3

BMC Med Imaging. 2025 Feb 10;25(1):44. doi: 10.1186/s12880-025-01566-8.

ABSTRACT

OBJECTIVES: This study aimed to investigate the diagnostic test accuracy of MRI-based radiomics studies for predicting EGFR mutation in brain metastasis originating from lung cancer.

METHODS: This meta-analysis, conducted following PRISMA guidelines, involved a systematic search in PubMed, Embase, and Web of Science up to November 3, 2024. Eligibility criteria followed the PICO framework, assessing population, intervention, comparison, and outcome. The RQS and QUADAS-2 tools were employed for quality assessment. A Bayesian model determined summary estimates, and statistical analysis was conducted using R and STATA software.

RESULTS: Eleven studies consisting of nine training and ten validation cohorts were included in the meta-analysis. In the training cohorts, MRI-based radiomics showed robust predictive performance for EGFR mutations in brain metastases, with an AUC of 0.90 (95% CI: 0.82-0.93), sensitivity of 0.84 (95% CI: 0.80-0.88), specificity of 0.86 (95% CI: 0.80-0.91), and a diagnostic odds ratio (DOR) of 34.17 (95% CI: 19.16-57.49). Validation cohorts confirmed strong performance, with an AUC of 0.91 (95% CI: 0.69-0.95), sensitivity of 0.79 (95% CI: 0.73-0.84), specificity of 0.88 (95% CI: 0.83-0.93), and a DOR of 31.33 (95% CI: 15.50-58.3). Subgroup analyses revealed notable trends: the T1C + T2WI sequences and 3.0 T scanners showed potential superiority, machine learning-based radiomics and manual segmentation exhibited higher diagnostic accuracy, and PyRadiomics emerged as the preferred feature extraction software.

CONCLUSION: This meta-analysis suggests that MRI-based radiomics holds promise for the non-invasive prediction of EGFR mutations in brain metastases of lung cancer.

PMID:39930347 | DOI:10.1186/s12880-025-01566-8

BMC Cancer. 2025 Feb 10;25(1):232. doi: 10.1186/s12885-025-13628-9.

ABSTRACT

BACKGROUND: Biochemical recurrence (BCR) occurs in 20%-40% of men with prostate cancer (PCa) who undergo radical prostatectomy. Predicting which patients will experience BCR in advance helps in formulating more targeted prostatectomy procedures. However, current preoperative recurrence prediction mainly relies on the use of the Gleason grading system, which omits within-grade morphological patterns and subtle histopathological features, leaving a significant amount of prognostic potential unexplored.

METHODS: We collected and selected a total of 1585 prostate biopsy images with tumor regions from 317 patients (5 Whole Slide Images per patient) to develop a deep learning system for predicting BCR of PCa before prostatectomy. The Inception_v3 neural network was employed to train and test models developed from patch-level images. The multiple instance learning method was used to extract whole slide image-level features. Finally, patient-level artificial intelligence models were developed by integrating deep learning -generated pathology features with several machine learning algorithms.

RESULTS: The BCR prediction system demonstrated great performance in the testing cohort (AUC = 0.911, 95% Confidence Interval: 0.840-0.982) and showed the potential to produce favorable clinical benefits according to Decision Curve Analyses. Increasing the number of WSIs for each patient improves the performance of the prediction system. Additionally, the study explores the correlation between deep learning -generated features and pathological findings, emphasizing the interpretative potential of artificial intelligence models in pathology.

CONCLUSIONS: Deep learning system can use biopsy samples to predict the risk of BCR in PCa, thereby formulating targeted treatment strategies.

PMID:39930342 | DOI:10.1186/s12885-025-13628-9

Med Phys. 2025 Feb 10. doi: 10.1002/mp.17685. Online ahead of print.

ABSTRACT

BACKGROUND: Sparse-view computed tomography (CT) reduces radiation exposure but suffers from severe artifacts caused by insufficient sampling and data scarcity, which compromise image fidelity. Recent advancements in deep learning (DL)-based methods for inverse problems have shown promise for CT reconstruction but often require high-quality paired datasets and lack interpretability.

PURPOSE: This paper aims to advance the field of CT reconstruction by introducing a novel unsupervised deep learning method. It builds on the foundation of Deep Radon Prior (DRP), which utilizes an untrained encoder-decoder network to extract implicit features from the Radon domain, and leverages Neural Architecture Search (NAS) to optimize network structures.

METHODS: We propose a novel unsupervised deep learning method for image reconstruction, termed NAS-DRP. This method leverages reinforcement learning-based NAS to explore diverse architectural spaces and integrates reinforcement learning with data inconsistency in the Radon domain. Building on previous DRP research, NAS-DRP utilizes an untrained encoder-decoder network to extract implicit features from the Radon domain. It further incorporates insights from studies on Deep Image Prior (DIP) regarding the critical impact of upsampling layers on image quality restoration. The method employs NAS to search for the optimal network architecture for upsampling unit tasks, while using Recurrent Neural Networks (RNNs) to constrain the optimization process, ensuring task-specific improvements in sparse-view CT image reconstruction.

RESULTS: Extensive experiments demonstrate that the NAS-DRP method achieves significant performance improvements in multiple CT image reconstruction tasks. The proposed method outperforms traditional reconstruction methods and other DL-based techniques in terms of both objective metrics (PSNR, SSIM, and LPIPS) and subjective visual quality. By automatically optimizing network structures, NAS-DRP effectively enhances the detail and accuracy of reconstructed images while minimizing artifacts.

CONCLUSIONS: NAS-DRP represents a significant advancement in the field of CT image reconstruction. By integrating NAS with deep learning and leveraging Radon domain-specific adaptations, this method effectively addresses the inherent challenges of sparse-view CT imaging. Additionally, it reduces the cost and complexity of data acquisition, demonstrating substantial potential for broader application in medical imaging. The evaluation code will be available at https://github.com/fujintao1999/NAS-DRP/.

PMID:39930320 | DOI:10.1002/mp.17685

J Med Syst. 2025 Feb 11;49(1):22. doi: 10.1007/s10916-025-02156-5.

ABSTRACT

This study aims to apply a multi-modal approach of the deep learning method for survival prediction in patients with non-small-cell lung cancer (NSCLC) using CT-based radiomics. We utilized two public data sets from the Cancer Imaging Archive (TCIA) comprising NSCLC patients, 420 patients and 516 patients for Lung 1 training and Lung 2 testing, respectively. A 3D convolutional neural network (CNN) survival was applied to extract 256 deep-radiomics features for each patient from a CT scan. Feature selection steps are used to choose the radiomics signatures highly associated with overall survival. Deep-radiomics and traditional-radiomics signatures, and clinical parameters were fed into the DeepSurv neural network. The C-index was used to evaluate the model's effectiveness. In the Lung 1 training set, the model combining traditional-radiomics and deep-radiomics performs better than the single parameter models, and models that combine all three markers (traditional-radiomics, deep-radiomics, and clinical) are most effective with C-index 0.641 for Cox proportional hazards (Cox-PH) and 0.733 for DeepSurv approach. In the Lung 2 testing set, the model combining traditional-radiomics, deep-radiomics, and clinical obtained a C-index of 0.746 for Cox-PH and 0.751 for DeepSurv approach. The DeepSurv method improves the model's prediction compared to the Cox-PH, and models that combine all three parameters with the DeepSurv have the highest efficiency in training and testing data sets (C-index: 0.733 and 0.751, respectively). DeepSurv CT-based deep-radiomics method outperformed Cox-PH in survival prediction of patients with NSCLC patients. Models' efficiency is increased when combining multi parameters.

PMID:39930275 | DOI:10.1007/s10916-025-02156-5

Commun Biol. 2025 Feb 10;8(1):209. doi: 10.1038/s42003-025-07479-0.

ABSTRACT

Quantifying emesis in Suncus murinus (S. murinus) has traditionally relied on direct observation or reviewing recorded behaviour, which are laborious, time-consuming processes that are susceptible to operator error. With rapid advancements in deep learning, automated animal behaviour quantification tools with high accuracy have emerged. In this study, we pioneere the use of both three-dimensional convolutional neural networks and self-attention mechanisms to develop the Automatic Emesis Detection (AED) tool for the quantification of emesis in S. murinus, achieving an overall accuracy of 98.92%. Specifically, we use motion-induced emesis videos as training datasets, with validation results demonstrating an accuracy of 99.42% for motion-induced emesis. In our model generalisation and application studies, we assess the AED tool using various emetics, including resiniferatoxin, nicotine, copper sulphate, naloxone, U46619, cyclophosphamide, exendin-4, and cisplatin. The prediction accuracies for these emetics are 97.10%, 100%, 100%, 97.10%, 98.97%, 96.93%, 98.91%, and 98.41%, respectively. In conclusion, employing deep learning-based automatic analysis improves efficiency and accuracy and mitigates human bias and errors. Our study provides valuable insights into the development of deep learning neural network models aimed at automating the analysis of various behaviours in S. murinus, with potential applications in preclinical research and drug development.

PMID:39930110 | DOI:10.1038/s42003-025-07479-0

Eur Spine J. 2025 Feb 10. doi: 10.1007/s00586-025-08708-0. Online ahead of print.

ABSTRACT

PURPOSE: Cervical spine injuries in the elderly (defined as individuals aged 65 years and older) are increasing, often resulting from falls and minor trauma. Prognosis varies widely, influenced by multiple factors. This study aimed to develop a deep-learning-based predictive model for post-injury outcomes.

METHODS: This study analyzed a nationwide dataset from the Japan Association of Spine Surgeons with Ambition, comprising 1512 elderly patients (aged 65 years and older) with cervical spine injuries from 2010 to 2020. Deep learning predictive models were constructed for residence, mobility, and the American Spinal Injury Association Impairment Scale (AIS). The model's performance was compared with that of a traditional statistical analysis.

RESULTS: The deep-learning model predicted the residence and AIS outcomes with varying accuracies. The highest accuracy was observed in predicting residence one year post-injury. The model also identified that the AIS score at discharge was significantly predicted by upper extremity trauma, mobility, and elbow extension strength. The deep learning model highlighted factors, such as upper extremity trauma, that were not considered significant in the traditional statistical analysis.

CONCLUSION: Our deep learning-based model offers a novel method for predicting outcomes following cervical spine injuries in the elderly population. The model is highly accurate and provides additional insights into potential prognostic factors. Such models can improve patient care and individualize future interventions.

PMID:39930051 | DOI:10.1007/s00586-025-08708-0

Sci Rep. 2025 Feb 10;15(1):4938. doi: 10.1038/s41598-025-89230-7.

ABSTRACT

Skin cancer represents a significant global health concern, where early and precise diagnosis plays a pivotal role in improving treatment efficacy and patient survival rates. Nonetheless, the inherent visual similarities between benign and malignant lesions pose substantial challenges to accurate classification. To overcome these obstacles, this study proposes an innovative hybrid deep learning model that combines ConvNeXtV2 blocks and separable self-attention mechanisms, tailored to enhance feature extraction and optimize classification performance. The inclusion of ConvNeXtV2 blocks in the initial two stages is driven by their ability to effectively capture fine-grained local features and subtle patterns, which are critical for distinguishing between visually similar lesion types. Meanwhile, the adoption of separable self-attention in the later stages allows the model to selectively prioritize diagnostically relevant regions while minimizing computational complexity, addressing the inefficiencies often associated with traditional self-attention mechanisms. The model was comprehensively trained and validated on the ISIC 2019 dataset, which includes eight distinct skin lesion categories. Advanced methodologies such as data augmentation and transfer learning were employed to further enhance model robustness and reliability. The proposed architecture achieved exceptional performance metrics, with 93.48% accuracy, 93.24% precision, 90.70% recall, and a 91.82% F1-score, outperforming over ten Convolutional Neural Network (CNN) based and over ten Vision Transformer (ViT) based models tested under comparable conditions. Despite its robust performance, the model maintains a compact design with only 21.92 million parameters, making it highly efficient and suitable for model deployment. The Proposed Model demonstrates exceptional accuracy and generalizability across diverse skin lesion classes, establishing a reliable framework for early and accurate skin cancer diagnosis in clinical practice.

PMID:39930026 | DOI:10.1038/s41598-025-89230-7

Sci Rep. 2025 Feb 10;15(1):4940. doi: 10.1038/s41598-025-89544-6.

ABSTRACT

Accurate space object pose estimation is crucial for various space tasks, including 3D reconstruction, satellite navigation, rendezvous and docking maneuvers, and collision avoidance. Many previous studies, however, often presuppose the availability of the space object's computer-aided design model for keypoint matching and model training. This work proposes a generalized pose estimation pipeline that is independent of 3D models and applicable to both instance- and category-level scenarios. The proposed framework consists of three parts based on deep learning approaches to accurately estimate space objects pose. First, a keypoints extractor is proposed to extract sub-pixel-level keypoints from input images. Then a multichannel matching network with triple loss is designed to obtain the matching pairs of keypoints in the body reference system. Finally, a pose graph optimization algorithm with a dynamic keyframes pool is designed to estimate the target pose and reduce long-term drifting pose errors. A space object dataset including nine different types of non-cooperative targets with 11,565 samples is developed for model training and evaluation. Extensive experimental results indicate that the proposed method demonstrates robust performance across various challenging conditions, including different object types, diverse illumination scenarios, varying rotation rates, and different image resolutions. To verify the demonstrated approach, the model is compared with several state-of-the-art approaches and shows superior estimation results. The mAPE and mMS scores of the proposed approach reach 0.63° and 0.767, respectively.

PMID:39930024 | DOI:10.1038/s41598-025-89544-6

Sci Rep. 2025 Feb 10;15(1):4931. doi: 10.1038/s41598-025-88451-0.

ABSTRACT

In the case of neonates, especially low birth weight preterm and high-risk infants, portable X-rays are frequently used. However, the image quality of portable X-rays is significantly lower compared to standard adult or pediatric X-rays, leading to considerable challenges in identifying abnormalities. Although attempts have been made to introduce deep learning to address these image quality issues, the poor quality of the images themselves hinders the training of deep learning models, further emphasizing the need for image enhancement. Additionally, since neonates have a high cell division rate and are highly sensitive to radiation, increasing radiation exposure to improve image quality is not a viable solution. Therefore, it is crucial to enhance image quality through preprocessing before training deep learning models. While various image enhancement methods have been proposed, Contrast Limited Adaptive Histogram Equalization (CLAHE) has been recognized as an effective technique for contrast-based image improvement. However, despite extensive research, the process of setting CLAHE's hyperparameters still relies on a brute force, manual approach, making it inefficient. To address this issue, we propose a method called Bayesian Optimization CLAHE(BO-CLAHE), which leverages Bayesian optimization to automatically select the optimal hyperparameters for X-ray images used in diagnosing lung diseases in preterm and high-risk neonates. The images enhanced by BO-CLAHE demonstrated superior performance across several classification models, with particularly notable improvements in diagnosing Transient Tachypnea of the Newborn (TTN). This approach not only reduces radiation exposure but also contributes to the development of AI-based diagnostic tools, playing a crucial role in the early diagnosis and treatment of preterm and high-risk neonates.

PMID:39929905 | DOI:10.1038/s41598-025-88451-0

Comput Biol Med. 2025 Feb 9;187:109785. doi: 10.1016/j.compbiomed.2025.109785. Online ahead of print.

ABSTRACT

Knee Osteoarthritis (KOA), a leading cause of disability worldwide, is challenging to detect early due to subtle radiographic indicators. Diverse, extensive datasets are needed but are challenging to compile because of privacy, data collection limitations, and the progressive nature of KOA. However, a model capable of projecting genuine radiographs into different OA stages could augment data pools, enhance algorithm training, and offer pre-emptive prognostic insights. In this study, we developed a Cycle-Consistent Adversarial Network (CycleGAN) to generate synthetic past and future stages of KOA on any genuine radiograph. The model's effectiveness was validated through its impact on a KOA specialized Convolutional Neural Network (CNN). Transformations towards synthetic future disease states resulted in 83.76% of none-to-doubtful stage images being classified as moderate-to-severe stages, while retroactive transformations led to 75.61% of severe-stage images being classified as none-to-doubtful stages. Similarly, transformations from mild stages achieved 76.00% correct classification towards future stages and 69.00% for past stages. The CycleGAN demonstrated an exceptional ability to expand the knee joint space and eliminate bone-outgrowths (osteophytes), key radiographic indicators of disease progression. These results signify a promising potential for enhancing diagnostic models, data augmentation, and educational and prognostic uses. Nevertheless, further refinement, validation, and a broader evaluation process encompassing both CNN-based assessments and expert medical feedback are emphasized for future research and development.

PMID:39929004 | DOI:10.1016/j.compbiomed.2025.109785

J Med Internet Res. 2025 Feb 10;27:e60888. doi: 10.2196/60888.

ABSTRACT

BACKGROUND: Although catheter ablation (CA) is currently the most effective clinical treatment for atrial fibrillation, its variable therapeutic effects among different patients present numerous problems. Machine learning (ML) shows promising potential in optimizing the management and clinical outcomes of patients undergoing atrial fibrillation CA (AFCA).

OBJECTIVE: This scoping review aimed to evaluate the current scientific evidence on the application of ML for managing patients undergoing AFCA, compare the performance of various models across specific clinical tasks within AFCA, and summarize the strengths and limitations of ML in this field.

METHODS: Adhering to the PRISMA-ScR (Preferred Reporting Items for Systematic Reviews and Meta-Analyses extension for Scoping Reviews) guidelines, relevant studies published up to October 7, 2023, were searched from PubMed, Web of Science, Embase, the Cochrane Library, and ScienceDirect. The final included studies were confirmed based on inclusion and exclusion criteria and manual review. The PROBAST (Prediction model Risk Of Bias Assessment Tool) and QUADAS-2 (Quality Assessment of Diagnostic Accuracy Studies-2) methodological quality assessment tools were used to review the included studies, and narrative data synthesis was performed on the modeled results provided by these studies.

RESULTS: The analysis of 23 included studies showcased the contributions of ML in identifying potential ablation targets, improving ablation strategies, and predicting patient prognosis. The patient data used in these studies comprised demographics, clinical characteristics, various types of imaging (9/23, 39%), and electrophysiological signals (7/23, 30%). In terms of model type, deep learning, represented by convolutional neural networks, was most frequently applied (14/23, 61%). Compared with traditional clinical scoring models or human clinicians, the model performance reported in the included studies was generally satisfactory, but most models (14/23, 61%) showed a high risk of bias due to lack of external validation.

CONCLUSIONS: Our evidence-based findings suggest that ML is a promising tool for improving the effectiveness and efficiency of managing patients undergoing AFCA. While guiding data preparation and model selection for future studies, this review highlights the need to address prevalent limitations, including lack of external validation, and to further explore model generalization and interpretability.

PMID:39928932 | DOI:10.2196/60888

PLoS One. 2025 Feb 10;20(2):e0317304. doi: 10.1371/journal.pone.0317304. eCollection 2025.

ABSTRACT

The World Natural Heritage is a rare and irreplaceable natural landscape recognized by all mankind, with outstanding significance and universal value. Among them, the World Heritage Karst sites(WHKs) holds an important position due to its special natural beauty and aesthetic value. In the field of landscape evaluation, interdisciplinary and interdisciplinary cooperation using different methods has always been a research focus. However, there is still a gap in the evaluation of natural landscape aesthetic value based on UGC(User Generated Content) data and deep learning models. This article is based on a public perspective, using social media UGC data, crawling images and texts as data sources, and combining SegFormer deep learning models, ArcGIS spatial analysis, natural Language Processing Technology (NLP) and other methods to conduct quantitative research on aesthetic value. Research has found that: (1) Huangguoshu Scenic Area has an excellent natural environment, and landscape elements with high naturalness (vegetation, water) are more attractive to tourists, with diverse landscape combinations; (2) There is no complete positive correlation between tourist sentiment bias, landscape diversity, and vegetation coverage. Emphasis is placed on the aesthetic perception path from bottom to top, from the surface to the inside. The comprehensive emotional value is 14.35, and the emotional values are all positively distributed. The distribution density and extreme value of positive emotions are greater than those of negative emotions; (3) The emotional bias of tourists is directly related to visual sensitivity, showing a synchronous trend of change. The visual sensitivity of the Great Waterfall and Dishuitan areas is relatively high, mostly at I-II level sensitivity. This method enhances the data source channel, which is conducive to obtaining the correct tourist evaluation orientation. In traditional subjective landscape evaluation, rational parameter indicators are added to reduce the probability of error, provide data support for its natural beauty description, break through the time and space limitations of aesthetic evaluation, and provide scientific reference for quantifying the aesthetic value of other heritage sites.

PMID:39928674 | DOI:10.1371/journal.pone.0317304

PLoS Comput Biol. 2025 Feb 10;21(2):e1012286. doi: 10.1371/journal.pcbi.1012286. Online ahead of print.

ABSTRACT

There is an important challenge in systematically interpreting the internal representations of deep neural networks (DNNs). Existing techniques are often less effective for non-tabular tasks, or they primarily focus on qualitative, ad-hoc interpretations of models. In response, this study introduces a cognitive science-inspired, multi-dimensional quantification and visualization approach that captures two temporal dimensions of model learning: the "information-processing trajectory" and the "developmental trajectory." The former represents the influence of incoming signals on an agent's decision-making, while the latter conceptualizes the gradual improvement in an agent's performance throughout its lifespan. Tracking the learning curves of DNNs enables researchers to explicitly identify the model appropriateness of a given task, examine the properties of the underlying input signals, and assess the model's alignment (or lack thereof) with human learning experiences. To illustrate this method, we conducted 750 runs of simulations on two temporal tasks: gesture detection and sentence classification, showcasing its applicability across different types of deep learning tasks. Using four descriptive metrics to quantify the mapped learning curves-start, end - start, max, tmax-, we identified significant differences in learning patterns based on data sources and class distinctions (all p's < .0001), the prominent role of spatial semantics in gesture learning, and larger information gains in language learning. We highlight three key insights gained from mapping learning curves: non-monotonic progress, pairwise comparisons, and domain distinctions. We reflect on the theoretical implications of this method for cognitive processing, language models and representations from multiple modalities.

PMID:39928655 | DOI:10.1371/journal.pcbi.1012286

PLoS One. 2025 Feb 10;20(2):e0311250. doi: 10.1371/journal.pone.0311250. eCollection 2025.

ABSTRACT

The healthcare industry is generating a massive volume of data, promising a potential goldmine of information that can be extracted through machine learning (ML) techniques. The Intensive Care Unit (ICU) stands out as a focal point within hospitals and provides a rich source of data for informative analyses. This study examines the cardiac surgery ICU, where the vital topic of patient ventilation takes center stage. In other words, ventilator-supported breathing is a fundamental need within the ICU, and the limited availability of ventilators in hospitals has become a significant issue. A crucial consideration for healthcare professionals in the ICU is prioritizing patients who require ventilators immediately. To address this issue, we developed a prediction model using four ML and deep learning (DL) models-LDA, CatBoost, Artificial Neural Networks (ANN), and XGBoost-that are combined in an ensemble model. We utilized Simulated Annealing (SA) and Genetic Algorithm (GA) to tune the hyperparameters of the ML models constructing the ensemble. The results showed that our approach enhanced the sensitivity of the tuned ensemble model to 85.84%, which are better than the results of the ensemble model without hyperparameter tuning and those achieved using AutoML model. This significant improvement in model performance underscores the effectiveness of our hybrid approach in prioritizing the need for ventilators among ICU patients.

PMID:39928609 | DOI:10.1371/journal.pone.0311250

PLoS One. 2025 Feb 10;20(2):e0317396. doi: 10.1371/journal.pone.0317396. eCollection 2025.

ABSTRACT

In recent years, the challenge of imbalanced data has become increasingly prominent in machine learning, affecting the performance of classification algorithms. This study proposes a novel data-level oversampling method called Cluster-Based Reduced Noise SMOTE (CRN-SMOTE) to address this issue. CRN-SMOTE combines SMOTE for oversampling minority classes with a novel cluster-based noise reduction technique. In this cluster-based noise reduction approach, it is crucial that samples from each category form one or two clusters, a feature that conventional noise reduction methods do not achieve. The proposed method is evaluated on four imbalanced datasets (ILPD, QSAR, Blood, and Maternal Health Risk) using five metrics: Cohen's kappa, Matthew's correlation coefficient (MCC), F1-score, precision, and recall. Results demonstrate that CRN-SMOTE consistently outperformed the state-of-the-art Reduced Noise SMOTE (RN-SMOTE), SMOTE-Tomek Link, and SMOTE-ENN methods across all datasets, with particularly notable improvements observed in the QSAR and Maternal Health Risk datasets, indicating its effectiveness in enhancing imbalanced classification performance. Overall, the experimental findings indicate that CRN-SMOTE outperformed RN-SMOTE in 100% of the cases, achieving average improvements of 6.6% in Kappa, 4.01% in MCC, 1.87% in F1-score, 1.7% in precision, and 2.05% in recall, with setting SMOTE's neighbors' number to 5.

PMID:39928607 | DOI:10.1371/journal.pone.0317396