Search Results (279)

Background: Chest X-ray (CXR) foundation models can predict patient demographic categories (sex, age, race) from images alone by linear probing, but whether encoded attributes drive finding prediction has not been tested at scale. Methods: On MIMIC-CXR (230,697 images, 60,518 patients), we measured attribute dependence (AUROC drop after residualizing an attribute from a frozen embedding) across 24 patient attributes (four demographics and 20 ICD-coded comorbidities), 10 thoracic findings, and 6 overlap-free foundation models ($n=1440$ triplets), with 3 additional CXR-pretrained models (RAD-DINO, CheXzero, CheSS) for encoding and fairness analyses. Dependence was regressed on attribute-finding odds ratios (ORs), encoding strength, and model-level factors. Results: Encoding and dependence dissociated. Sex (AUROC 0.942) contributed $<0.001$; race (0.83) contributed $0.0015$ (rank 14/24); heart failure (0.774) showed the largest dependence ($0.018$). $\left|\log \right(\mathrm{OR}\left)\right|$ explained 50.6% of dependence variance ($\beta =0.029$, $p<{10}^{-15}$); model factors added no detectable contribution ($\Delta R^2=0.000$, $n=6$). Residualizing the top three high-OR attributes reduced AUROC by 0.026 without narrowing sex or age subgroup gaps (minimum detectable effect size (MDES) = 0.0019). Across 9 models, four-category race subgroup gaps (mean 0.069) were 30–75× larger than race residualization drops (mean 0.0015); CheXzero showed the same decoupling. Conclusions: Encoding, residualization-sensitive dependence, and subgroup bias are three separable quantities on the same model. Pre-deployment audits on inpatient-skewed cohorts can prioritize attributes by local OR; jointly residualizing race and its cardiac correlates does not narrow the race subgroup gap, which instead tracks group-wise finding base rates. Cross-institutional transfer remains open: no public CXR cohort currently links comorbidity electronic health records for external validation of the OR-dependence relationship. Full article

Background/Objectives: Chest X-ray (CXR) images are a widely used first-line screening tool for pulmonary tuberculosis (TB) detection but are difficult to interpret, which has increased demand for an automated screening tool. Deep-learning-based computer-aided diagnosis systems have demonstrated a classification performance comparable to that of trained radiologists, but they rely on dense annotations such as lesion-level or pixel-level labels, which are costly and difficult to obtain in routine clinical workflows. We developed CPM-XNet, an annotation-efficient framework for lesion-annotation-free downstream TB classification in CXR images. Methods: CPM-XNet incorporates a compressing–projecting mask (CPM) to provide soft lung-aware modulation while preserving global contextual information. The CPM-modulated images are then used for downstream classification with multiple convolutional neural network backbones and a vision transformer baseline. Results: Experiments were conducted using an internal hospital dataset and public TB datasets, and CPM-XNet showed improved performance compared with baseline models trained on unmodulated images. In a repeated-seed evaluation of the main ResNet-101 configuration on the Tung cohort, CPM-ResNet101 showed higher and more stable performance than the non-CPM counterpart and demonstrated significant paired improvement using McNemar’s exact test. An ablation analysis indicated that CPM modulation was the main contributor to performance improvement while data augmentation and the classifier architecture further influenced the overall robustness. Conclusions: CPM-XNet provides an annotation-efficient strategy for lesion-annotation-free downstream TB classification in CXR images. The findings support preliminary technical feasibility, although larger, naturally imbalanced, cross-institutional validation is required before clinical deployment can be inferred. Full article

Background: Radiology underpins diagnosis and treatment across pediatrics, yet most artificial intelligence (AI) tools are developed for adults and validated on adult datasets only. Of more than 200 AI systems cleared by the United States (U.S.) Food and Drug Administration (FDA), only about 3% include pediatric validation. Because children differ from adults in anatomy, physiology, pathology, epidemiology, and imaging protocols, adult-trained models often perform sub-optimally in pediatric settings. Methods: A narrative review of peer-reviewed literature from 2000 to 2025 was conducted using PubMed, MEDLINE, Google Scholar, and Scopus. Studies involving AI applications in pediatric X-ray, ultrasound, computed tomography (CT), magnetic resonance imaging (MRI), echocardiography, and point-of-care ultrasound with quantitative performance metrics were included. Findings were synthesized by imaging modality, clinical task, and differences between high-income countries (HICs) and low- and middle-income countries (LMICs). Results: AI demonstrated strong performance across multiple pediatric imaging tasks. In X-ray interpretation, AI detected fractures with area under the curve (AUC) values up to 0.96 (sensitivity, 90.8%; specificity, 88.7%). Pneumonia classification achieved 76.5% accuracy, and foreign body aspiration detection showed 95.3% specificity in HICs. In ultrasound, AI improved junior sonographers’ detection of intussusception (AUC 0.857 to 0.966) and reduced scan time by more than 50%. AI-assisted bone age estimation achieved a mean error of 0.39 years. In echocardiography, AI-derived ejection fraction showed excellent agreement with experts’ interclass correlation coefficient (ICC 0.983), and AI support improved atrioventricular septal defect detection (84.4% to 86.5%). In MRI, the use of AI enhanced lesion detection and supported quantitative analysis. Deep-learning models trained on routine T1- and T2-weighted sequences predicted liver stiffness across multi-site datasets, while advanced neuroimaging pipelines improved the identification of subtle epileptogenic lesions that are often missed on conventional pediatric MRI. However, adult-trained models showed limited generalizability to children. Still, excluding children under the age of two years improved the reading accuracy of pediatric chest X-rays (CXRs) by adult-trained models from 88% to 97%. AI faces challenges beyond the development of age-specific models. Substantial heterogeneity, limited pediatric-specific datasets, and unresolved medicolegal responsibility further restrict adoption worldwide. Challenges are amplified in LMICs, where unstable electricity, limited radiology resources, weak digital infrastructure, and scarce pediatric providers limit implementation. Additionally, many large language models underperform and lack inclusive algorithms suitable for pediatric radiology in many LMICs. Conclusions: AI can enhance diagnostic accuracy, efficiency, and access to pediatric imaging, particularly in resource-limited settings, through task-shifting and decision support. However, it cannot replace pediatric radiologists as of today. Safe adoption requires pediatric-specific model development, standardized validation metrics, diverse datasets that include LMIC populations, stronger digital infrastructure, robust radiologist training in AI capabilities, and the establishment of clear guidelines and medicolegal policies. Full article

Acute hypoxemic respiratory failure (AHRF) represents one of the most common and clinically challenging indications for invasive mechanical ventilation in the intensive care unit, characterized by profound etiological heterogeneity that demands accurate diagnosis to guide treatment. While clinical history, physical examination, and laboratory data remain essential, they are often insufficient to reliably discriminate among conditions such as acute respiratory distress syndrome (ARDS), cardiogenic pulmonary edema, and pneumonia—particularly in mechanically ventilated patients. Lung imaging has therefore emerged as an indispensable complement to clinical assessment. In this narrative review, we systematically describe the physical principles, clinical applications, and limitations of the imaging modalities currently available in critical care: chest X-ray (CXR), computed tomography (CT), lung ultrasound (LUS), electrical impedance tomography (EIT), and positron emission tomography (PET). CXR remains the most widely used bedside tool but is constrained by low sensitivity and significant interobserver variability. CT is the gold standard for morphological and quantitative lung phenotyping, enabling the assessment of recruitability, baby lung characterization, and the identification of complications, but requires patient transport and exposes patients to ionizing radiation. LUS offers real-time, bedside evaluation of aeration with high diagnostic accuracy for pneumothorax and pleural effusion, and is increasingly integrated into revised ARDS diagnostic criteria. EIT enables continuous, radiation-free monitoring of regional ventilation distribution and positive end-expiratory pressure (PEEP)-guided titration directly at the bedside. While PET provides unparalleled quantification of regional inflammation and ventilation-perfusion mismatch, it currently remains a purely investigative research tool. Finally, we discuss emerging technological and AI-driven advances—including dual-energy CT, next-generation EIT, and deep learning algorithms—that are poised to transform lung imaging from a passive diagnostic tool into an active, personalized guide to respiratory management. Full article

Background/Objectives: Lung cancer survival depends on early detection; however, in the Philippines, high radiologist workloads and the anatomical complexity of chest X-rays (CXRs) contribute to missed pulmonary nodules and false-negative diagnoses. This study aims to develop an enhanced deep learning model to improve nodule classification and localization sensitivity. Methods: We propose RNNet-MST, an extension of ResNet-50 that incorporates Multi-Scale Transformer blocks for global context modeling and a custom spatial attention mechanism for attention-based weak localization of disease-relevant regions. The model was trained and evaluated on the NODE21 chest X-ray dataset and compared with a baseline ResNet-50 using classification metrics, with attention maps used for weak localization analysis. Results: RNNet-MST demonstrated consistent improvements over the baseline ResNet-50 across evaluated metrics. Mean Nodule Recall improved from 88.02 ± 1.92% to 91.55 ± 1.41%, reducing false negatives. Mean Test Precision reached 90.46 ± 0.99%, and mean Nodule F1-Score improved to 90.99 ± 0.39%. On the isolated small-nodule subset, RNNet-MST achieved a 12.3% improvement in sensitivity over the baseline. Conclusions: The integration of multi-scale transformer features improved classification sensitivity, while the attention mechanism provided weak localization cues that aligned more closely with annotated nodule regions than the baseline. RNNet-MST shows potential as a diagnostic support tool, warranting further validation on larger and more diverse clinical datasets to reduce perceptual errors and facilitate early lung cancer detection in resource-constrained settings. Full article

In this work, we propose ARTEMIS, a novel and highly interpretable deep learning pipeline for the automatic classification of Chest X-ray (CXR) and Computed Tomography (CT) images into different categories related to important clinical outcomes: COVID-19 infection, Community-Acquired Pneumonia (CAP) cases, and Normal cases. Unlike existing models based on the static feature enhancement step, ARTEMIS proposes a learnable preprocessing component that dynamically adapts the image contrast and sharpness in training mode, facilitating adaptive optimization. Our hybrid network combines EfficientNet-B0 backbone with built-in SE attention with the optional lightweight Transformer encoder block to jointly learn local radiological features and global relationships between pixels. Comprehensive experiments have been conducted on five different datasets, which comprise four publicly available ones and one novel CT dataset annotated by radiologists, including X-ray and CT modalities. Experimental results show strong robustness and generalization with macro F1-scores greater than 96% on public datasets and 99.39% accuracy on our new CT dataset. To interpret the decision-making process, Grad-CAM++ is employed to generate class-discriminative saliency maps; the highlighted regions are systematically validated against established radiological criteria by a board-certified radiologist, confirming that model decisions are grounded in clinically meaningful pulmonary findings rather than imaging artifacts. Full article

The accurate detection of dental caries is often hindered by modality-specific imaging challenges, such as illumination artifacts in intra-oral photographs and low lesion contrast in radiographs. This study proposes a comprehensive framework comprising three key components: (1) HybridAugment+, an entropy-guided adaptive augmentation strategy that applies stronger transformations to low-information images; (2) DBAttNet, a dual-branch attention network featuring illumination–reflection aware attention (IRAA) for photographs and contrast–frequency-aware attention (CFA) for radiographs; and (3) a CAM-based explainability method, selected through a systematic evaluation of five advanced techniques. This study utilized two datasets derived from public sources, comprising 639 intra-oral photographs (481 caries, 158 healthy) and 456 radiographs (268 caries, 188 healthy). These were annotated by two dentists, with established inter-rater reliability (κ = 0.82 for photographs, κ = 0.79 for radiographs). The experimental results demonstrate that HybridAugment+ improved performance over conventional augmentation by up to 8.72% on photographs and 7.67% on radiographs. Furthermore, DBAttNet achieved F1-scores of 97.90% on photographs and 95.72% on radiographs, outperforming ResNet50, InceptionV3, MSDNet, DCANet, and ARM-Net. A comparative evaluation identified XGrad-CAM as the most suitable explainability method, with optimal visualization thresholds of 30% for photographs and 20% for radiographs. Generalization experiments on ophthalmology (APTOS 2019, Messidor-2) and chest radiography datasets (Kermany CXR, NIH ChestX-ray14) demonstrated consistent performance gains over domain-specific methods (DT-Net, ConvNeXt-Tiny). These results confirm that the core design principles effectively transfer to other modalities facing analogous imaging challenges. Full article

Large-scale chest X-ray (CXR) datasets often rely on report-derived or weak labels, introducing missing and incorrect annotations that can degrade downstream models and limit trust. We study training-free mislabel detection in multi-label CXRs by scoring neighborhood label consistency in a fixed embedding space. Using the NIH Chest X-ray Kaggle sample (5606 CXRs), we extract intermediate CoAtNet features and obtain 64-dimensional embeddings with a frozen CoAtNet backbone and a lightweight refinement head. On top of these embeddings, we compare kNN consistency baselines with distance weighting and label-set similarity against LPV-DW-CS, clustered prototype voting weighted by distance and cluster support. We evaluate three synthetic label-noise regimes with review budgets matched to the corruption rate: random single-label (5% and 20%), boundary-noise (20% corruption within the lowest-density 20% subset), and disjoint-label replacement (20% within that subset). LPV-DW-CS yields the highest downstream macro-AUROC after filtering top-ranked samples (up to 0.8860), while kNN variants achieve higher Recall@budget at the same review rates (up to 99.44%). An image-only expert Likert review of top-ranked real samples finds substantial label-set inconsistencies (54.1% for LPV-DW-CS-280-A; 60.5% for KNN-DW-LSS), supporting neighborhood-consistency ranking as a practical, training-free tool for targeted dataset auditing. Full article

COVID-19 still poses a global public health challenge, exerting pressure on radiology services. Chest X-ray (CXR) imaging is widely used for respiratory assessment due to its accessibility and cost-effectiveness. However, its interpretation is often challenging because of subtle radiographic features and inter-observer variability. Although recent deep learning (DL) approaches have shown strong performance in automated CXR classification, their black-box nature limits interpretability. This study proposes an explainable deep learning framework for COVID-19 detection from chest X-ray images. The framework incorporates anatomically guided preprocessing, including lung-region isolation, contrast-limited adaptive histogram equalization (CLAHE), bone suppression, and feature enhancement. A novel four-channel input representation was constructed by combining lung-isolated soft-tissue images with frequency-domain opacity maps, vessel enhancement maps, and texture-based features. Classification was performed using a modified Xception-based convolutional neural network, while Gradient-weighted Class Activation Mapping (Grad-CAM) was employed to provide visual explanations and enhance interpretability. The framework was evaluated on the publicly available COVID-19 Radiography Database, achieving an accuracy of 95.3%, an AUC of 0.983, and a Matthews Correlation Coefficient of approximately 0.83. Threshold optimisation improved sensitivity, reducing missed COVID-19 cases while maintaining high overall performance. Explainability analysis showed that model attention was primarily focused on clinically relevant lung regions. Full article

Background/Objectives: Pediatric thoracic trauma requires prompt stabilization and timely imaging; however, actual sequencing and escalation triggers are infrequently delineated at the pathway level. The aim of this study was to analyze imaging pathways observed in routine clinical practice at our institution and to outline a preliminary escalation framework integrating injury mechanism, clinical severity, and initial ultrasound findings. Methods: A retrospective cohort study was conducted at the “Louis Țurcanu” Clinical Emergency Hospital for Children, Timișoara, Romania, including 66 children admitted with primary thoracic trauma between January 2022 and December 2024. Clinical trajectory markers (transfer-in, ICU admission, length of stay) and imaging utilization/sequencing (FAST, CXR, CT, MRI/CTA) were extracted. We divided injuries into two groups: bony (like fractures of the clavicle or scapula) and non-bony. CT escalation was characterized as a chest CT conducted upon admission. Fisher’s exact and Mann–Whitney U tests were used for comparative analyses. Results: FAST was done on all patients but was infrequently positive. Imaging followed heterogeneous but structured patterns, most commonly FAST with CXR, with or without CT. A large group of them had CT scans without first having any X-rays. CT escalation was associated with fracture-pattern injuries and higher-acuity trajectories (transfer-in and ICU admission), as well as prolonged hospital stays. Pathway-level assessment demonstrated that CT escalation effectively captured bony injury patterns, whereas FAST proficiently sorted ICU-level trajectories. Conclusions: Pediatric thoracic trauma imaging functioned as a selective escalation system: FAST served as a universal bedside entry step, and CT operated as an injury pattern- and acuity-linked severity gate. Making this escalation logic clear may help with standardization while still protecting against radiation. Full article

Background: This paper investigates the use of the Segment Anything Model (SAM) for chest X-ray (CXR) image segmentation, with a focus on improving its performance using low-rank adaptation (LoRA). Methods: We evaluate three versions of SAM: two zero-shot methods (using coordinate and bounding box prompts) and a fine-tuned SAM using LoRA. To support these approaches, we also trained two standard convolutional neural networks (CNNs), U-Net and DeepLabv3+, to generate draft lung segmentations that serve as input prompts for the SAM methods. Our fine-tuning approach uses LoRA to add lightweight trainable adapters within the Transformer blocks of the SAM, allowing only a small subset of parameters to be updated. The rest of the SAM remains frozen, helping preserve its pre-trained knowledge while reducing memory and computational needs. We tested all models on a dataset of CXR images labeled for COVID-19, viral pneumonia, and normal cases. Results: Results show that fine-tuned SAM with LoRA outperforms zero-shot SAM methods and CNN baselines in terms of segmentation accuracy and efficiency. Conclusions: This demonstrates the potential of combining LoRA with SAM for practical and effective medical image segmentation. Full article

Chest X-ray (CXR) imaging remains the most widely used radiological modality for assessing pulmonary and cardiothoracic disease, yet its interpretation is inherently constrained by tissue superposition, subtle radiographic findings and marked inter-observer variability. Recent advances in artificial intelligence (AI) have driven significant progress in automated CXR analysis, supported by large public datasets, evolving annotation strategies and increasingly expressive deep learning architectures. This review presents a comprehensive synthesis of approaches for pulmonary abnormality detection, encompassing convolutional neural networks, transformers, multimodal and vision–language models and self-supervised representation learning. We critically discuss their strengths, limitations and vulnerability to label noise, domain shift and shortcut learning. In parallel, we examine dataset properties, annotation practices, robustness challenges, explainability methods and the heterogeneity of evaluation protocols that hinder fair comparison and clinical translation. Building on these observations, the review identifies key future directions, including foundation models, multimodal integration, federated and domain-generalized training, longitudinal modeling, synthetic data generation and standardized clinical evaluation frameworks. By integrating methodological and clinical perspectives, this work offers an up-to-date reference for researchers and clinicians and outlines a roadmap toward reliable, interpretable and clinically deployable AI systems for chest radiography. Full article

Automated radiology report generation has become a prominent research topic in medical multimodal learning. However, most existing approaches primarily focus on single-image interpretation and rarely address the task of tracking disease progression across longitudinal chest X-rays. This task presents two major challenges: accurately localizing pathological changes between temporally paired images, and effectively translating visual difference representations into clinically meaningful textual descriptions. To address these challenges, we propose MRID (Modeling Radiological Image Differences for Disease Progression Reasoning), a multi-task self-supervised framework that follows a pretraining–finetuning paradigm. MRID leverages multiple complementary self-supervised objectives to jointly achieve (1) intra-modal spatial alignment of organs and pathological regions across image pairs, and (2) cross-modal semantic alignment between visual difference representations and radiology report embeddings. Furthermore, we introduce a simple yet effective data augmentation strategy to alleviate the imbalance of disease progression categories. Extensive experiments conducted on the Longitudinal-MIMIC and MS-CXR-T datasets demonstrate that MRID effectively captures fine-grained disease progression patterns. In addition, the proposed framework achieves competitive performance on single-image radiology report generation, further highlighting its strong capability in modeling chest X-ray semantics. Full article

Chest X-ray (CXR) imaging is one of the most widely used techniques for screening and diagnosing pulmonary diseases. However, discerning subtle structural changes, such as small nodules, disordered pulmonary textures, tiny cavities, pleural thickening, or spiculation, is difficult using low-resolution images. Acquiring high-resolution CXRs typically requires higher radiation doses, posing a risk to patients. We propose a chest X-ray image super-resolution algorithm based on generative adversarial networks (GAN). Through adversarial training, our approach generates high-resolution CXRs with enhanced details and improved realism. We further incorporate a CSA hybrid attention module into the network, strengthening its ability to capture fine structures and improve texture fidelity. Moreover, we remove the pooling layer from the channel attention module to overcome limitations in super-resolution, thereby preserving spatial information more effectively. Experiments demonstrate our method’s superior performance and robustness, achieving a PSNR of 37.91 and SSIM of 0.9108 on the internal test set while consistently outperforming other methods on previously unseen external clinical datasets. After adversarial training, the method attains optimal visual performance, with LPIPS reduced to 0.0915, and the visual effect improved by 36.4% compared to low-resolution images. Ablation studies further verify the contribution of the proposed method to enhancing super-resolution capability. Overall, results indicate that the proposed method can obtain high-quality chest X-rays images from simulated low-quality inputs. Full article

Background/Objectives: Interpretation of chest X-rays (CXRs) requires accurate identification of lesion presence, diagnosis, location, size, and number to be considered complete. However, the effectiveness of large language models with vision capabilities (LLMs) in performing these tasks remains uncertain. This study aimed to evaluate the image-only interpretation performance of LLMs in the absence of clinical information. Methods: A total of 247 CXRs covering 13 diagnostic categories, including pulmonary edema, cardiomegaly, lobar pneumonia, and other conditions, were evaluated using Gemini 1.0, Gemini 1.5 Pro, GPT-4 Turbo, and GPT-4o. The text outputs generated by the LLMs were evaluated at two levels: (1) primary diagnosis accuracy across the 13 predefined diagnostic categories, and (2) identification of key imaging features described in the generated text. Primary diagnosis accuracy was assessed based on whether the model correctly identified the target diagnostic category and was classified as fully correct, partially correct, or incorrect according to predefined clinical criteria. Non-diagnostic imaging features, such as posteroanterior and anteroposterior (PA/AP) views, side markers, foreign bodies, and devices, were recorded and analyzed separately rather than being incorporated into the primary diagnostic scoring. Results: When fully and partially correct responses were treated as successful detections, vLLMs showed higher sensitivity for large, bilateral, multiple lesions and prominent devices, including acute pulmonary edema, lobar pneumonia, multiple malignancies, massive pleural effusions, and pacemakers, all of which demonstrated statistically significant differences across categories in chi-square analyses. Feature descriptions varied among models, especially in PA/AP views and side markers, though central lines were partially recognized. Across the entire dataset, Gemini 1.5 Pro achieved the highest overall detection rate, followed by Gemini 1.0, GPT-4o, and GPT-4 Turbo. Conclusions: Although LLMs were able to identify certain diagnoses and key imaging features, their limitations in detecting small lesions, recognizing laterality, reasoning through differential diagnoses, and using domain-specific expressions indicate that CXR interpretation without textual cues still requires further improvement. Full article