This study aims to establish a dual-feature fusion model integrating radiomic features with deep learning features, utilizing single-modality pre-treatment lung CT image data to achieve early warning of brain metastasis (BM) risk within 2 years in EGFR-positive lung adenocarcinoma. After rigorous screening of 362 EGFR-positive lung adenocarcinoma patients with pre-treatment lung CT images, 173 eligible participants were ultimately enrolled in this study, including 93 patients with BM and 80 without BM. Radiomic features were extracted from manually segmented lung nodule regions, and a selection of features was used to develop radiomics models. For deep learning, ROI-level CT images were processed using several deep learning networks, including the novel vision mamba, which was applied for the first time in this context. A feature-level fusion model was developed by combining radiomic and deep learning features. Model performance was assessed using receiver operating characteristic (ROC) curves and decision curve analysis (DCA), with statistical comparisons of area under the curve (AUC) values using the DeLong test. Among the models evaluated, the fused vision mamba model demonstrated the best classification performance, achieving an AUC of 0.86 (95% CI: 0.82-0.90), with a recall of 0.88, F1-score of 0.70, and accuracy of 0.76. This fusion model outperformed both radiomics-only and deep learning-only models, highlighting its superior predictive accuracy for early BM risk detection in EGFR-positive lung adenocarcinoma patients. The fused vision mamba model, utilizing single CT imaging data, significantly enhances the prediction of brain metastasis within two years in EGFR-positive lung adenocarcinoma patients. This novel approach, combining radiomic and deep learning features, offers promising clinical value for early detection and personalized treatment.

<b>Aims/Background</b> Pulmonary embolism (PE) is a life-threatening condition with significant diagnostic challenges due to high rates of missed or delayed detection. Computed tomography pulmonary angiography (CTPA) is the current standard for diagnosing PE, however, demand for imaging places strain on healthcare systems and increases error rates. This systematic review aims to assess the diagnostic accuracy and clinical applicability of artificial intelligence (AI)-based models for PE detection on CTPA, exploring their potential to enhance diagnostic reliability and efficiency across clinical settings. <b>Methods</b> A systematic review was conducted in accordance with Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines. Excerpta Medica Database (EMBASE), Medical Literature Analysis and Retrieval System Online (MEDLINE), Cochrane, PubMed, and Google Scholar were searched for original articles from inception to September 2024. Articles were included if they reported successful AI integration, whether partial or full, alongside CTPA scans for PE detection in patients. <b>Results</b> The literature search identified 919 articles, with 745 remaining after duplicate removal. Following rigorous screening and appraisal aligned with inclusion and exclusion criteria, 12 studies were included in the final analysis. A total of three primary AI modalities emerged: convolutional neural networks (CNNs), segmentation models, and natural language processing (NLP), collectively used in the analysis of 341,112 radiographic images. CNNs were the most frequently applied modality in this review. Models such as AdaBoost and EmbNet have demonstrated high sensitivity, with EmbNet achieving 88-90.9% per scan and reducing false positives to 0.45 per scan. <b>Conclusion</b> AI shows significant promise as a diagnostic tool for identifying PE on CTPA scans, particularly when combined with other forms of clinical data. However, challenges remain, including ensuring generalisability, addressing potential bias, and conducting rigorous external validation. Variability in study methodologies and the lack of standardised reporting of key metrics complicate comparisons. Future research must focus on refining models, improving peripheral emboli detection, and validating performance across diverse settings to realise AI's potential fully.

<b>Aims/Background</b> Breast nodules are highly prevalent among women, and ultrasound is a widely used screening tool. However, single ultrasound examinations often result in high false-positive rates, leading to unnecessary biopsies. Artificial intelligence (AI) has demonstrated the potential to improve diagnostic accuracy, reducing misdiagnosis and minimising inter-observer variability. This study developed a deep learning-based AI model to evaluate its clinical utility in assisting sonographers with the Breast Imaging Reporting and Data System (BI-RADS) classification of breast nodules. <b>Methods</b> A retrospective analysis was conducted on 558 patients with breast nodules classified as BI-RADS categories 3 to 5, confirmed through pathological examination at The People's Hospital of Pingyang County between December 2019 and December 2023. The image dataset was divided into a training set, validation set, and test set, and a convolutional neural network (CNN) was used to construct a deep learning-based AI model. Patients underwent ultrasound examination and AI-assisted diagnosis. The receiver operating characteristic (ROC) curve was used to analyse the performance of the AI model, physician adjudication results, and the diagnostic efficacy of physicians before and after AI model assistance. Cohen's weighted Kappa coefficient was used to assess the consistency of BI-RADS classification among five ultrasound physicians before and after AI model assistance. Additionally, statistical analyses were performed to evaluate changes in BI-RADS classification results before and after AI model assistance for each physician. <b>Results</b> According to pathological examination, 765 of the 1026 breast nodules were benign, while 261 were malignant. The sensitivity, specificity, and accuracy of routine ultrasonography in diagnosing benign and malignant nodules were 80.85%, 91.59%, and 88.31%, respectively. In comparison, the AI system achieved a sensitivity of 89.36%, specificity of 92.52%, and accuracy of 91.56%. Furthermore, AI model assistance significantly improved the consistency of physicians' BI-RADS classification (<i>p</i> < 0.001). <b>Conclusion</b> A deep learning-based AI model constructed using ultrasound images can enhance the differentiation between benign and malignant breast nodules and improve classification accuracy, thereby reducing the incidence of missed and misdiagnoses.

Purpose The limited availability of bronchoscopy images makes image synthesis particularly interesting for training deep learning models. Robust image translation across different domains-virtual bronchoscopy, phantom as well as in vivo and ex vivo image data-is pivotal for clinical applications. Methods This paper proposes BronchoGAN introducing anatomical constraints for image-to-image translation being integrated into a conditional GAN. In particular, we force bronchial orifices to match across input and output images. We further propose to use foundation model-generated depth images as intermediate representation ensuring robustness across a variety of input domains establishing models with substantially less reliance on individual training datasets. Moreover, our intermediate depth image representation allows to easily construct paired image data for training. Results Our experiments showed that input images from different domains (e.g., virtual bronchoscopy, phantoms) can be successfully translated to images mimicking realistic human airway appearance. We demonstrated that anatomical settings (i.e., bronchial orifices) can be robustly preserved with our approach which is shown qualitatively and quantitatively by means of improved FID, SSIM and dice coefficients scores. Our anatomical constraints enabled an improvement in the Dice coefficient of up to 0.43 for synthetic images. Conclusion Through foundation models for intermediate depth representations and bronchial orifice segmentation integrated as anatomical constraints into conditional GANs, we are able to robustly translate images from different bronchoscopy input domains. BronchoGAN allows to incorporate public CT scan data (virtual bronchoscopy) in order to generate large-scale bronchoscopy image datasets with realistic appearance. BronchoGAN enables to bridge the gap of missing public bronchoscopy images.

The peripheral immune system is essential for maintaining central nervous system homeostasis. This study investigates the effects of peripheral immune markers on accelerated brain aging and dementia using brain-predicted age difference based on neuroimaging. By leveraging data from the UK Biobank, Cox regression was used to explore the relationship between peripheral immune markers and dementia, and multivariate linear regression to assess associations between peripheral immune biomarkers and brain structure. Additionally, we established a brain age prediction model using Simple Fully Convolutional Network (SFCN) deep learning architecture. Analysis of the resulting brain-Predicted Age Difference (PAD) revealed relationships between accelerated brain aging, peripheral immune markers, and dementia. During the median follow-up period of 14.3 years, 4, 277 dementia cases were observed among 322, 761 participants. Both innate and adaptive immune markers correlated with dementia risk. NLR showed the strongest association with dementia risk (HR = 1.14; 95% CI: 1.11-1.18, P<0.001). Multivariate linear regression revealed significant associations between peripheral immune markers and brain regional structural indices. Utilizing the deep learning-based SFCN model, the estimated brain age of dementia subjects (MAE = 5.63, r2 = - 0.46, R = 0.22) was determined. PAD showed significant correlation with dementia risk and certain peripheral immune markers, particularly in individuals with positive brain age increment. This study employs brain age as a quantitative marker of accelerated brain aging to investigate its potential associations with peripheral immunity and dementia, highlighting the importance of early intervention targeting peripheral immune markers to delay brain aging and prevent dementia.

Recent research has suggested that neuroinflammation may be important in the pathogenesis of neurodegenerative diseases. Free-water diffusion (FWD) has been proposed as a non-invasive neuroimaging-based biomarker for neuroinflammation. Free-water maps were generated using diffusion MRI data in 367 patients from the Ontario Neurodegenerative Disease Research Initiative (108 Alzheimer's Disease/Mild Cognitive Impairment, 42 Frontotemporal Dementia, 37 Amyotrophic Lateral Sclerosis, 123 Parkinson's Disease, and 58 vascular disease-related Cognitive Impairment). The ability of FWD to predict neuroinflammation and neurodegeneration from biofluids was estimated using plasma glial fibrillary-associated protein (GFAP) and neurofilament light chain (NfL), respectively. Recursive Feature Elimination (RFE) performed the strongest out of all feature selection algorithms used and revealed regional specificity for areas that are the most important features for predicting GFAP over NfL concentration. Deep learning models using selected features and demographic information revealed better prediction of GFAP over NfL. Based on feature selection and deep learning methods, FWD was found to be more strongly related to GFAP concentration (measure of astrogliosis) over NfL (measure of neuro-axonal damage), across neurodegenerative disease groups, in terms of predictive performance. Non-invasive markers of neurodegeneration such as MRI structural imaging that can reveal neurodegeneration already exist, while non-invasive markers of neuroinflammation are not available. Our results support the use of FWD as a non-invasive neuroimaging-based biomarker for neuroinflammation.

As artificial intelligence (AI) becomes increasingly prevalent within interventional radiology (IR) research and clinical practice, steps must be taken to ensure the robustness of novel technological systems presented in peer-reviewed journals. This report introduces comprehensive standards and an evaluation checklist (iCARE) that covers the application of modern AI methods in IR-specific contexts. The iCARE checklist encompasses the full "code-to-clinic" pipeline of AI development, including dataset curation, pre-training, task-specific training, explainability, privacy protection, bias mitigation, reproducibility, and model deployment. The iCARE checklist aims to support the development of safe, generalizable technologies for enhancing IR workflows, the delivery of care, and patient outcomes.

Clinical decision-making is inherently complex, time-sensitive, and prone to error. AI-enabled clinical decision support systems (CDSS) offer promising solutions by leveraging large datasets to provide evidence-based recommendations. These systems range from rule-based and knowledge-based to increasingly AI-driven approaches. However, key challenges persist, particularly concerning data quality, seamless integration into clinical workflows, and clinician trust and acceptance. Ethical and legal considerations, especially data privacy, are also paramount.AI-CDSS have demonstrated success in fields like radiology (e.g., pulmonary nodule detection, mammography interpretation) and cardiology, where they enhance diagnostic accuracy and improve patient outcomes. Looking ahead, chat and voice interfaces powered by large language models (LLMs) could support shared decision-making (SDM) by fostering better patient engagement and understanding.To fully realize the potential of AI-CDSS in advancing efficient, patient-centered care, it is essential to ensure their responsible development. This includes grounding AI models in domain-specific data, anonymizing user inputs, and implementing rigorous validation of AI-generated outputs before presentation. Thoughtful design and ethical oversight will be critical to integrating AI safely and effectively into clinical practice.

The primary aim of this in vitro study was to compare methods for generating 3D-printed replicas through virtual segmentation, utilizing artificial intelligence (AI) or manual processes, by assessing accuracy in terms of volumetric and linear discrepancies. The secondary aims were the assessment of time efficiency with both segmentation methods, and the effect of post-processing on 3D-printed replicas. Thirty teeth were scanned through Cone Beam Computed Tomography (CBCT), capturing the region of interest from human subjects. DICOM files underwent virtual segmentation through both AI and manual methods. Replicas were fabricated with a stereolithography 3D printer. After surface scanning of pre-processed replicas and extracted teeth, STL files were superimposed to compare linear and volumetric differences using the extracted teeth as the reference. Post-processed replicas were scanned to assess the effect of post-processing on linear and volumetric changes. AI-driven segmentation resulted in statistically significant mean linear and volumetric differences of -0.709mm (SD 0.491, P< 0.001) and -4.70%, respectively. Manual segmentation showed no statistically significant differences in mean linear, -0.463mm (SD 0.335, P<0.001) and volumetric (-1.20%) measures. Comparing manual and AI-driven segmentations, AI-driven segmentation displayed mean linear and volumetric differences of -0.329mm (SD 0.566, p=0.003) and -2.23%, respectively. Additionally, AI segmentation reduced the mean time by 21.8 minutes. When comparing post-processed to pre-processed replicas, there was a volumetric reduction of -4.53% and a mean linear difference of -0.151mm (SD 0.564, p=0.042). Both segmentation methods achieved acceptable accuracy, with manual segmentation slightly more accurate but AI-driven segmentation more time-efficient. Continuous improvement in AI offers the potential for increased accuracy, efficiency, and broader application in the future.

To compare the effectiveness of expert-designed machine learning models and code-free automated machine learning (AutoML) models in classifying optical coherence tomography (OCT) images for detecting age-related macular degeneration (AMD) and distinguishing between its dry and wet forms. Custom models were developed by an artificial intelligence expert using the EfficientNet V2 architecture, while AutoML models were created by an ophthalmologist utilizing LobeAI with transfer learning via ResNet-50 V2. Both models were designed to differentiate normal OCT images from AMD and to also distinguish between dry and wet AMD. The models were trained and tested using an 80:20 split, with each diagnostic group containing 500 OCT images. Performance metrics, including sensitivity, specificity, accuracy, and F1 scores, were calculated and compared. The expert-designed model achieved an overall accuracy of 99.67% for classifying all images, with F1 scores of 0.99 or higher across all binary class comparisons. In contrast, the AutoML model achieved an overall accuracy of 89.00%, with F1 scores ranging from 0.86 to 0.90 in binary comparisons. Notably lower recall was observed for dry AMD vs. normal (0.85) in the AutoML model, indicating challenges in correctly identifying dry AMD. While the AutoML models demonstrated acceptable performance in identifying and classifying AMD cases, the expert-designed models significantly outperformed them. The use of advanced neural network architectures and rigorous optimization in the expert-developed models underscores the continued necessity of expert involvement in the development of high-precision diagnostic tools for medical image classification.