Structural brain deficits associated with generalized anxiety disorder (GAD), panic disorder (PD), and obsessive-compulsive disorder (OCD) have been documented, but their integration within a unified framework remains unexplored. This study investigates, in anxiety and anxiety-related disorders, whether they share neurophysiological bases, whether structural changes (SCs) are linked to common genes, whether shared SCs co-occur with similar functional brain impairments, and whether brain morphometry can serve as biomarkers for diagnosis and treatment prediction. Participants included 100 individuals with GAD, 58 with PD, 45 with OCD, and 85 healthy controls, all drug-free. Structural and resting-state functional magnetic resonance imaging scans and clinical assessments were conducted before and after 4 weeks of paroxetine monotherapy. Analyses included voxel-based and surface-based morphometry; functional connectivity (FC) and Granger causality analysis (GCA) with shared SCs as regions of interest; associations between clinical assessments and neuroimaging metrics; associations between gene expression profiles and SCs; and machine learning. Cingulate atrophy (CA) emerged as a common SC, with disorder-specific atrophy in gray matter volume (GMV) and cortical surface. Transcriptome-neuroimaging correlations identified shared genetic associations with GMV alterations, with negatively correlated genes enriched in neurodevelopment and cellular growth regulation (ND-CGR). Cingulate GMV was positively correlated with cognitive performance in GAD and PD patients. FC and GCA showed CA disrupted networks governing emotional regulation and cognitive control, characterized by overactive top-down influence and reduced bottom-up feedback. Machine learning demonstrated strong performance in classification and treatment response prediction, with cingulate morphometry contributing significantly. CA is a shared neural substrate in GAD, PD, and OCD, linked to genetic disruptions in ND-CGR, cognitive impairments, and functional brain deficits. Cingulate morphometry holds promise as a biomarker for diagnosis and treatment response in these conditions.

Actionable findings requiring follow-up with additional imaging or other diagnostic procedures are frequently reported for a wide variety of radiology exams. Completion of recommended follow-up can lead to new diagnoses including cancer. However, recommended follow-up completion is inconsistent, particularly when follow-up is for findings unrelated to the initial reason for the exam. Follow-up recommendation tracking systems, using a combination of information technology tools and human navigators, can facilitate completion of recommended follow-up, but often require significant effort for manual chart review and direct communication with providers and patients. Artificial intelligence (AI), including large language models (LLMs) able to process vast and diverse unstructured text data, offer the opportunity to improve efficiency with data extraction and aggregation tasks, like those required for follow-up recommendation management. In this review article, we will review the key components of follow-up recommendation management systems: (1) identification of follow-up recommendations within radiology reports, (2) communication of these recommendations, (3) tracking of follow-up recommendations to completion, and (4) outcomes tracking. For each component, we will explore how AI can improve efficiency and expand capabilities of robust management systems that ensure the loop is closed for follow-up recommendations.

Breast cancer remains a significant global health concern, and machine learning algorithms and computer-aided detection systems have shown great promise in enhancing the accuracy and efficiency of mammography image analysis. However, there is a critical need for large, benchmark datasets for training deep learning models for breast cancer detection. In this work we developed Mammo-Bench, a large-scale benchmark dataset of mammography images, by collating data from six well-curated resources, viz., DDSM, INbreast, KAU-BCMD, CMMD, CDD-CESM and DMID. To ensure consistency across images from diverse sources while preserving clinically relevant features, a preprocessing pipeline that includes breast segmentation, pectoral muscle removal, and intelligent cropping is proposed. The dataset consists of 19,731 high-quality mammographic images from 6,500 patients across 6 countries and is one of the largest open-source mammography databases to the best of our knowledge. To show the efficacy of training on the large dataset, performance of ResNet101 architecture was evaluated on Mammo-Bench and the results compared by training independently on a few member datasets and an external dataset, VinDr-Mammo. An accuracy of 78.8% (with data augmentation of the minority classes) and 77.8% (without data augmentation) was achieved on the proposed benchmark dataset, compared to the other datasets for which accuracy varied from 25 - 69%. Noticeably, improved prediction of the minority classes is observed with the Mammo-Bench dataset. These results establish baseline performance and demonstrate Mammo-Benchs utility as a comprehensive resource for developing and evaluating mammography analysis systems.

PURPOSETo prospectively evaluate the feasibility and performance of expedited screening mammogram interpretation for women identified as high-risk by a deep learning risk model. METHODS AND MATERIALSThis HIPAA-compliant, IRB-approved prospective controlled study was conducted at an urban safety-net facility. The Mirai breast cancer risk model was retrospectively calibrated on 114,229 local mammograms (2006-2023) to identify the top 10% of 1-year breast cancer risk scores. During the prospective study (12/2024-6/2025), Mirai 1-year risk scores were generated in real time. On enrollment days, high-risk women were approached for consent and offered immediate interpretation of their screening exam. Patients assessed as BI-RADS 0 were offered same-day diagnostic evaluation when feasible. Outcomes included feasibility of immediate interpretation, time to screening result (Ts), diagnostic evaluation (Td), and biopsy (Tb), as well as cancer detection rate (CDR). Comparisons were made with high-risk controls on non-enrollment days. RESULTSAmong 4,145 screening mammograms, Mirai flagged 525 (12.7%) as high-risk; 973 (23.5%) were performed on enrollment days with 115 (11.8%) flagged as high-risk. Of 100 women who consented, 94% received immediate reads. Thirty-one were assessed as BI-RADS 0; 30 underwent diagnostic imaging (26 same day). Thirteen biopsies yielded 6 malignant (4 invasive, 2 DCIS), 2 high-risk, and 5 benign lesions. The CDR in high-risk expedited women was 60/1,000 (95% CI, 22.3-126.0) compared with 2.3/1,000 (95% CI, 0.3-8.4) in non-high-risk women (odds ratio 27.1; p<0.001). Median Ts, Td, and Tb were significantly shorter in expedited patients versus high-risk controls (13.0 min vs 191.9 min; 1.3 hrs vs 852.8 hrs; 20.1 vs 59.0 days; all p<0.001). For screen-detected cancers, expedited interpretation reduced mean Ts, Td, and Tb by 99.1%, 99.1%, and 87.2%, respectively. CONCLUSIONIntegrating an AI risk model into mammography workflow is feasible and enables same-day evaluation for high-risk women. This approach markedly shortens time to diagnostic imaging and biopsy to provide timely breast cancer care.

Computed tomography imaging, a non-invasive tool, is used around the globe by medical professionals to identify and diagnose lung cancer; a lethal disease with high rates of occurrence and mortality globally. Radiomics extracted from medical images, including computed tomography, in tandem with machine learning frameworks has received considerable focus and research for lung nodule identification.This investigation can help out clinicians to reach radiomics-based better and quicker decision support system for treatments and early diagnosis. However, it is still foggy and unclear which radiomics feature(s) to use for the prediction of pulmonary nodule. Consequently, this work is offered with an endeavor to efficiently apply machine learning techniques and radiomics to classify CT pulmonary nodules. Lung Image Data Consortium (LIDC), containing 1018 CT cancer cases, is put to use. The Wavelet Packet Transform is used in conjunction with geometrical features, gray level run length matrix, gray level co-occurrence method and gray level difference method techniques to extract radiomics. Two techniques, boosted and bagged ensemble classification trees, are employed to choose an apposite set of features. The categorization of nodules as malignant or benign is assessed by the utilization of cutting-edge machine learning models: Support Vector Machines, Boosted Classification Ensemble Tree, Decision Trees, Bagged Classification Ensemble Tree, RUSBoosted Ensemble Trees, Subspace Discriminant Ensemble and Subspace KNN Ensemble. The findings reveal that the Ensemble Subspace KNN gives best AUROC (93.4%), accuracy (88.3%) and F1-score (85.2%) using BACET feature selection method. The best sensitivity is produced by FGSVM (97.1%). RUSBOCET gives best precision and specificity of 93.4% and 83.1% respectively. Lung Cancer remains the most common and deadly type of cancer. Early detection of lung lesions and nodules is crucial in the fight against lung cancer. The purpose of this study was to investigate radiomics based on geometrical, texture, and Daubechies WPT texture features for quantitative CT image analysis. The LIDC database was used in this study. Geometrical features, texture features based on three statistical methodologies (GLCM, GLDM GLRLM) and Daubechies WPT texture features are retrieved from the nodules. Using the ensemble EFS, BOCET and BACET, pertinent features were identified. Lastly, various cutting-edge ML classifiers were used to classify LC as malignant or benign. The out-turn shows that, using BACET EFS, Ensemble Subspace KNN gives best AUROC (93.4%), accuracy (88.3%) and F1-score (85.2%). FGSVM yields the best sensitivity of 97.1%. RUSBOCET gives best precision and best specificity of 93.4% and 83.1% respectively. Therefore, the methodology can be applied with efficacy to the CT based PN classification. Thus, the result can assist medical professionals in making better decisions and interventions.

Airway foreign body aspiration remains a potentially life-threatening emergency, predominantly affecting children under 5 years and adults over 65 years. This review synthesizes current evidence on diagnostic strategies, bronchoscopic extraction techniques, procedural outcomes, complication management, and emerging technologies in airway foreign body management. Multidetector computed tomography with three-dimensional reconstruction has significantly improved diagnostic accuracy, achieving sensitivity of 98-99% for radiopaque objects and 85-92% for radiolucent materials. Flexible bronchoscopy has evolved from a diagnostic tool to a first-line therapeutic modality, with recent pediatric meta-analyses demonstrating 87% success rates and adult series showing comparable outcomes to rigid bronchoscopy for appropriately selected cases. Rigid bronchoscopy maintains superiority in asphyxiating presentations, and for large (>1.5 cm), sharp, or severely impacted foreign bodies. Technological innovations including robotic-assisted bronchoscopy, electromagnetic navigation systems, and artificial intelligence-powered imaging analysis are enhancing procedural precision and safety. Successful airway foreign body management requires individualized, multidisciplinary approaches integrating patient clinical status, foreign body characteristics, and institutional expertise. Success depends on appropriate bronchoscopic modality selection, comprehensive preprocedural planning, availability of specialized retrieval instruments, and readiness to manage potential complications. Integration of advanced imaging, simulation-based training protocols, and telemedicine consultation are becoming essential components of contemporary practice.

Given the global rise of MASLD, which impacts approximately one-third of the population, there is a need for earlier diagnosis and effective treatment strategies to avoid long-term hepatic cardiovascular and renal complications. This review summarizes the recent advances in noninvasive diagnosis and new pharmacological agents approved for MASLD. The main step forward in diagnostics is a step away from invasive biopsy and emphasis on noninvasive methods including serum biomarkers (e.g. CK-18 and FGF21), imaging (e.g. MRI-PDFF and US-FLI), combination of the two and use of artificial intelligence and machine learning models, for early detection and risk stratification of MASLD and MASH. Multiomics approaches, such as metabolomics and lipidomics, reveal disease-specific signatures, and may help with phenotypic classification of MASLD. Personalized management for MASLD include gut microbiota modulation and point-of-care devices for rapid diagnosis. Novel therapies include THR β agonists, GLP-1/dual GLP-1/GIP agonists, FXR agonists and FGF analogues, which show promise in reducing hepatic fat and fibrosis. These findings enable earlier MASLD diagnosis and tailored interventions, improving clinical outcomes in primary care and resource-limited settings. Future research should focus on validating cost-effective tools, and developing combination therapies to address the multifaceted nature of MASLD.

This study evaluates four large language models' (LLMs) ability to answer common patient questions preceding transarterial periarticular embolization (TAPE), computed tomography (CT)-guided high-dose-rate (HDR) brachytherapy, and bleomycin electrosclerotherapy (BEST). The goal is to evaluate their potential to enhance clinical workflows and patient comprehension, while also assessing associated risks. Thirty-five TAPE, 34 CT-HDR brachytherapy, and 36 BEST related questions were presented to ChatGPT-4o, DeepSeek-V3, OpenBioLLM-8b, and BioMistral-7b. The LLM-generated responses were independently assessed by two board-certified radiologists. Accuracy was rated on a 5-point Likert scale. Statistics compared LLM performance across question categories for patient-education suitability. DeepSeek-V3 attained the highest mean scores for BEST [4.49 (± 0.77)] and CT-HDR [4.24 (± 0.81)] and demonstrated comparable performance to ChatGPT-4o for TAPE-related questions (DeepSeek-V3 [4.20 (± 0.77)] vs. ChatGPT-4o [4.17 (± 0.64)]; p = 1.000). In contrast, OpenBioLLM-8b (BEST 3.51 (± 1.15), CT-HDR 3.32 (± 1.13), TAPE 3.34 (± 1.16)) and BioMistral-7b (BEST 2.92 (± 1.35), CT-HDR 3.03 (± 1.06), TAPE 3.33 (± 1.28)) performed significantly worse than DeepSeek-V3 and ChatGPT-4o across all procedures. Preparation/Planning was the only category without statistically significant differences across all three procedures. DeepSeek-V3 and ChatGPT-4o excelled on TAPE, BEST, and CT-HDR brachytherapy questions, indicating potential to enhance patient education in interventional radiology, where complex but minimally invasive procedures often are explained in brief consultations. However, OpenBioLLM-8b and BioMistral-7b exhibited more frequent inaccuracies, suggesting that LLMs cannot replace comprehensive clinical consultations yet. Patient feedback and clinical workflow implementation should validate these findings.

For over two decades, medical imaging modalities have played crucial roles in clinical diagnosis. Extracting comprehensive information from a single modality often proves challenging for ensuring clinical accuracy. Consequently, multi-modal medical image fusion methods integrate images from diverse modalities into a single fused image, enhancing information quality and diagnostic reliability. In recent years, deep learning for multi-modal medical image segmentation has emerged as a vibrant research area, yielding promising outcomes. This paper conducts a thorough survey and comparative analysis of advancements in deep learning techniques for multi-modal medical image segmentation from 2019 to 2025. It aims to provide a comprehensive overview of deep learning-based approaches and fusion strategies for integrating information from different imaging modalities. Additionally, the survey highlights how various deep learning models enhance segmentation accuracy and reliability. Common challenges in medical image segmentation are discussed, along side current research trends in the field.

To develop and rigorously validate radiomics-based predictive models using postoperative intravoxel incoherent motion diffusion-weighted imaging (IVIM-DWI) MRI for the early, noninvasive assessment of impaired renal allograft function (IRF) in kidney transplant recipients. This retrospective study included 97 kidney transplant recipients (mean age, 36.77 ± 10.71 years), categorized into normal or impaired renal function groups based on an estimated glomerular filtration rate (eGFR) cutoff of 60 ml/min/1.73 m². Patients were randomly assigned to training (n = 68) or validation (n = 29) groups. Postoperative IVIM-DWI MRI with 11 b-values was performed on a 3T scanner, generating parametric maps (apparent diffusion coefficient (ADC), slow diffusion coefficient (D_slow), fast diffusion coefficient (D_fast), perfusion fraction (PF)). Whole-graft 3D manual segmentation was used to extract 1604 radiomic features per dataset. Feature selection was performed through analysis of variance (ANOVA), Relief, and recursive feature elimination (RFE), followed by classification using ten machine learning algorithms, including auto-encoder (AE) and naïve Bayes (NB). Performance was evaluated using receiver operating characteristic (ROC) analysis, with area under the curve (AUC), accuracy, sensitivity, and specificity as metrics. Radiomics models based on IVIM-derived parametric maps (ADC, D_slow, D_fast, PF) achieved superior diagnostic performance, with a validation AUC of 0.790 (95% confidence interval (CI) 0.607-0.937) using ANOVA-based feature selection and AE classification, and a training AUC of 0.770. Integrative models combining multi-b-value DWI and IVIM maps further enhanced predictive power, achieving a validation AUC of 0.790 (95% CI 0.600-0.951) and a training AUC of 0.816, utilizing 16 features selected via ANOVA and classified with the NB algorithm. AE and NB classifiers consistently exhibited the strongest discriminative performance across all model configurations. Notably, the median histogram intensity from the D_slow map was the most influential feature for predicting impaired renal function. This study is the first to comprehensively compare the predictive performance of radiomics models based on IVIM-DWI, including both single b-value DWI and IVIM parametric maps, for early assessment of renal allograft dysfunction. The integrative use of multi-b-value DWI and IVIM imaging markedly improves diagnostic accuracy, demonstrating a robust noninvasive framework for early detection of renal allograft dysfunction.