This study aimed to design a web-based application for computer-aided diagnosis (CADx) of intraosseous jaw lesions, and assess its diagnostic accuracy. In this diagnostic test study, a web-based application was designed for CADx of 19 types of intraosseous jaw lesions. To assess its diagnostic accuracy, clinical and radiographic information of 95 cases with confirmed histopathological diagnosis of intraosseous jaw lesions were retrieved from hospital archives and published literature and imported to the application by a senior dental student. The top-N accuracy, kappa value, and Brier score were calculated, and the sensitivity, specificity, positive (PPV) and negative (NPV) predictive values, and the area under the receiver operating characteristic (ROC) curve (AUC) were calculated separately for each lesion according to DeLong et al. In assessment of top-N accuracy, the designed application gave a correct differential diagnosis in 93 cases (97.89%); the correct diagnosis was at the top of the list of differential diagnoses in 78 cases (82.10%); these values were 85 (89.47%) and 67 (70.52%) for an oral radiologist. The kappa value was 0.53. The Brayer score for the prevalence match was 0.18, and the pattern match was 0.15. The results highlighted the optimally high diagnostic accuracy of the designed application, indicating that it may be reliably used for CADx of intraosseous jaw lesions, if given accurate data.

Vision-grounded medical report generation aims to produce clinically accurate descriptions of medical images, anchored in explicit visual evidence to improve interpretability and facilitate integration into clinical workflows. However, existing methods often rely on separately trained detection modules that require extensive expert annotations, introducing high labeling costs and limiting generalizability due to pathology distribution bias across datasets. To address these challenges, we propose Self-Supervised Anatomical Consistency Learning (SS-ACL) -- a novel and annotation-free framework that aligns generated reports with corresponding anatomical regions using simple textual prompts. SS-ACL constructs a hierarchical anatomical graph inspired by the invariant top-down inclusion structure of human anatomy, organizing entities by spatial location. It recursively reconstructs fine-grained anatomical regions to enforce intra-sample spatial alignment, inherently guiding attention maps toward visually relevant areas prompted by text. To further enhance inter-sample semantic alignment for abnormality recognition, SS-ACL introduces a region-level contrastive learning based on anatomical consistency. These aligned embeddings serve as priors for report generation, enabling attention maps to provide interpretable visual evidence. Extensive experiments demonstrate that SS-ACL, without relying on expert annotations, (i) generates accurate and visually grounded reports -- outperforming state-of-the-art methods by 10\% in lexical accuracy and 25\% in clinical efficacy, and (ii) achieves competitive performance on various downstream visual tasks, surpassing current leading visual foundation models by 8\% in zero-shot visual grounding.

While deep convolutional neural networks (DCNNs) have achieved remarkable performance in chest X-ray interpretation, their success typically depends on access to large-scale, expertly annotated datasets. However, collecting such data in real-world clinical settings can be difficult because of limited labeling resources, privacy concerns, and patient variability. In this study, we applied a multimodal Transformer pretrained on free-text reports and their paired CXRs to evaluate the effectiveness of this method in settings with limited labeled data. Our dataset consisted of more than 1 million CXRs, each accompanied by reports from board-certified radiologists and 31 structured labels. The results indicated that a linear model trained on embeddings from the pretrained model achieved AUCs of 0.907 and 0.903 on internal and external test sets, respectively, using only 128 cases and 384 controls; the results were comparable those of DenseNet trained on the entire dataset, whose AUCs were 0.908 and 0.903, respectively. Additionally, we demonstrated similar results by extending the application of this approach to a subset annotated with structured echocardiographic reports. Furthermore, this multimodal model exhibited excellent small sample learning capabilities when tested on external validation sets such as CheXpert and ChestX-ray14. This research significantly reduces the sample size necessary for future artificial intelligence advancements in CXR interpretation.

The detection of dental caries is typically overlooked on panoramic radiographs. This study aims to leverage deep learning to identify multistage caries on panoramic radiographs. The panoramic radiographs were confirmed with the gold standard bitewing radiographs to create a reliable ground truth. The dataset of 500 panoramic radiographs with corresponding bitewing confirmations was labelled by an experienced and calibrated radiologist for 1,792 caries from 14,997 teeth. The annotations were stored using the annotation and image markup standard to ensure consistency and reliability. The deep learning system employed a two-model approach: YOLOv5 for tooth detection and Attention U-Net for segmenting caries. The system achieved impressive results, demonstrating strong agreement with dentists for both caries counts and classifications (enamel, dentine, and pulp). However, some discrepancies exist, particularly in underestimating enamel caries. While the model occasionally overpredicts caries in healthy teeth (false positive), it prioritizes minimizing missed lesions (false negative), achieving a high recall of 0.96. Overall performance surpasses previously reported values, with an F1-score of 0.85 and an accuracy of 0.93 for caries segmentation in posterior teeth. The deep learning approach demonstrates promising potential to aid dentists in caries diagnosis, treatment planning, and dental education.

The limited availability of annotated data presents a major challenge for applying deep learning methods to medical image analysis. Few-shot learning methods aim to recognize new classes from only a small number of labeled examples. These methods are typically studied under the standard few-shot learning setting, where all classes in a task are new. However, medical applications such as pathology classification from chest X-rays often require learning new classes while simultaneously leveraging knowledge of previously known ones, a scenario more closely aligned with generalized few-shot classification. Despite its practical relevance, few-shot learning has been scarcely studied in this context. In this work, we present MetaChest, a large-scale dataset of 479,215 chest X-rays collected from four public databases. MetaChest includes a meta-set partition specifically designed for standard few-shot classification, as well as an algorithm for generating multi-label episodes. We conduct extensive experiments evaluating both a standard transfer learning approach and an extension of ProtoNet across a wide range of few-shot multi-label classification tasks. Our results demonstrate that increasing the number of classes per episode and the number of training examples per class improves classification performance. Notably, the transfer learning approach consistently outperforms the ProtoNet extension, despite not being tailored for few-shot learning. We also show that higher-resolution images improve accuracy at the cost of additional computation, while efficient model architectures achieve comparable performance to larger models with significantly reduced resource requirements.

Adolescent idiopathic scoliosis (AIS) is a common spinal deformity affecting approximately 2.2% of boys and 4.8% of girls worldwide. The Cobb angle serves as the gold standard for AIS severity assessment, yet traditional manual measurements suffer from significant observer variability, compromising diagnostic accuracy. Despite prior automation attempts, existing methods use simplified spinal models and predetermined curve patterns that fail to address clinical complexity. We present a novel deep learning framework for AIS assessment that simultaneously predicts both superior and inferior endplate angles with corresponding midpoint coordinates for each vertebra, preserving the anatomical reality of vertebral wedging in progressive AIS. Our approach combines an HRNet backbone with Swin-Transformer modules and biomechanically informed constraints for enhanced feature extraction. We employ Singular Value Decomposition (SVD) to analyze angle predictions directly from vertebral morphology, enabling flexible detection of diverse scoliosis patterns without predefined curve assumptions. Using 630 full-spine anteroposterior radiographs from patients aged 10-18 years with rigorous dual-rater annotation, our method achieved 83.45% diagnostic accuracy and 2.55{\deg} mean absolute error. The framework demonstrates exceptional generalization capability on out-of-distribution cases. Additionally, we introduce the Vertebral Wedging Index (VWI), a novel metric quantifying vertebral deformation. Longitudinal analysis revealed VWI's significant prognostic correlation with curve progression while traditional Cobb angles showed no correlation, providing robust support for early AIS detection, personalized treatment planning, and progression monitoring.

Osteoporosis is an under-screened musculoskeletal disorder that results in diminished quality of life and significant burden to the healthcare system. We aimed to evaluate the ability of Rho, an artificial intelligence (AI) tool, to prospectively identify patients at-risk for low bone mineral density (BMD) from standard x-rays, its adoption rate by radiologists, and acceptance by primary care providers (PCPs). Patients ≥50 years were recruited when undergoing an x-ray of a Rho-eligible body part for any clinical indication. Questionnaires were completed at baseline and 6-month follow-up, and PCPs of "Rho-Positive" patients (those likely to have low BMD) were asked for feedback. Positive predictive value (PPV) was calculated in patients who returned within 6 months for a DXA. Of 1145 patients consented, 987 had x-rays screened by Rho, and 655 were flagged as Rho-Positive. Radiologists included this finding in 524 (80%) of reports. Of all Rho-Positive patients, 125 had a DXA within 6 months; Rho had a 74% PPV for DXA T-Score <-1. From 51 PCP responses, 78% found Rho beneficial. Of 389 patients with follow-up questionnaire data, a greater proportion of Rho-Positive versus -negative patients had discussed bone health with their PCP since study start (36% vs 18%, <i>P</i> < .001), or were newly diagnosed with osteoporosis (11% vs 5%; <i>P</i> = .03). By identifying patients at-risk of low BMD, with acceptability of reporting by radiologists and generally positive feedback from PCPs, Rho has the potential to improve low screening rates for osteoporosis by leveraging existing x-ray data.

Deep learning models are increasingly used for radiographic analysis, but their reliability is challenged by the stochastic noise inherent in clinical imaging. A systematic, cross-task understanding of how different noise types impact these models is lacking. Here, we evaluate the robustness of state-of-the-art convolutional neural networks (CNNs) to simulated quantum (Poisson) and electronic (Gaussian) noise in two key chest X-ray tasks: semantic segmentation and pulmonary disease classification. Using a novel, scalable noise injection framework, we applied controlled, clinically-motivated noise severities to common architectures (UNet, DeepLabV3, FPN; ResNet, DenseNet, EfficientNet) on public datasets (Landmark, ChestX-ray14). Our results reveal a stark dichotomy in task robustness. Semantic segmentation models proved highly vulnerable, with lung segmentation performance collapsing under severe electronic noise (Dice Similarity Coefficient drop of 0.843), signifying a near-total model failure. In contrast, classification tasks demonstrated greater overall resilience, but this robustness was not uniform. We discovered a differential vulnerability: certain tasks, such as distinguishing Pneumothorax from Atelectasis, failed catastrophically under quantum noise (AUROC drop of 0.355), while others were more susceptible to electronic noise. These findings demonstrate that while classification models possess a degree of inherent robustness, pixel-level segmentation tasks are far more brittle. The task- and noise-specific nature of model failure underscores the critical need for targeted validation and mitigation strategies before the safe clinical deployment of diagnostic AI.

Medical imaging foundation models must adapt over time, yet full retraining is often blocked by privacy constraints and cost. We present a continual learning framework that avoids storing patient exemplars by pairing class conditional diffusion replay with Elastic Weight Consolidation. Using a compact Vision Transformer backbone, we evaluate across eight MedMNIST v2 tasks and CheXpert. On CheXpert our approach attains 0.851 AUROC, reduces forgetting by more than 30\% relative to DER\texttt{++}, and approaches joint training at 0.869 AUROC, while remaining efficient and privacy preserving. Analyses connect forgetting to two measurable factors: fidelity of replay and Fisher weighted parameter drift, highlighting the complementary roles of replay diffusion and synaptic stability. The results indicate a practical route for scalable, privacy aware continual adaptation of clinical imaging models.

Fracture detection plays a critical role in medical imaging analysis, traditional fracture diagnosis relies on visual assessment by experienced physicians, however the speed and accuracy of this approach are constrained by the expertise. With the rapid advancements in artificial intelligence, deep learning models based on the YOLO framework have been widely employed for fracture detection, demonstrating significant potential in improving diagnostic efficiency and accuracy. This study proposes an improved YOLO-based model, termed Fracture-YOLO, which integrates novel Critical-Region-Selector Attention (CRSelector) and Scale-Aware (ScA) heads to further enhance detection performance. Specifically, the CRSelector module utilizes global texture information to focus on critical features of fracture regions. Meanwhile, the ScA module dynamically adjusts the weights of features at different scales, enhancing the model's capacity to identify fracture targets at multiple scales. Experimental results demonstrate that, compared to the baseline model, Fracture-YOLO achieves a significant improvement in detection precision, with mAP50 and mAP50-95 increasing by 4 and 3, surpassing the baseline model and achieving state-of-the-art (SOTA) performance.