Integration of Artificial Intelligence (AI), particularly deep learning, into medical imaging represents a profound shift in diagnostic medicine, moving from purely descriptive analysis to advanced predictive and prescriptive analytics. This Collection explores the rapid advancement of AI-driven tools in their specific fields such as oncology, cardiology, ophthalmology and so on, highlighting their potential to improve diagnostic accuracy, workflow efficiency, and personalized treatment planning. However, significant challenges remain, including the heterogeneity of medical image data, the "black box" nature of some intelligent models, and the critical hurdles of clinical integration and validation. The research presented here addresses these frontiers, showcasing innovations in algorithm development, explainable AI, and translational application. This Editorial synthesizes the contributions and outlines the essential collaborative pathway-uniting computer scientists, clinicians, and regulatory bodies-required to translate algorithmic promise into robust, trustworthy, and equitable clinical tools that genuinely improve patient care.
Pancreatic cancer is characterized by prolonged subclinical progression, molecular heterogeneity, and late clinical presentation, resulting in diagnosis predominantly at advanced stages. Current screening approaches lack sufficient sensitivity and scalability, underscoring the need for risk-adapted early detection strategies. Artificial intelligence (AI) offers a shift from reactive diagnosis toward proactive, precision-oriented screening. This review synthesizes recent advances in AI for the early screening and diagnosis of pancreatic cancer. We focus on how AI enables population-level and high-risk prediction, augments diagnostic assessment in patients with suspicious clinical, imaging, or molecular findings, and supports precision stratification through multimodal integration of radiologic imaging, circulating biomarkers, and longitudinal electronic health records (EHRs). Advances span three domains. In imaging, deep learning models-including convolutional neural networks, transformer architectures, and self-configuring segmentation frameworks-improve pancreas segmentation, lesion detection, and classification, with several systems demonstrating radiologist-level performance in retrospective multicenter studies. In biomarker discovery, machine learning approaches such as LASSO, random forest, and XGBoost facilitate high-dimensional feature selection from transcriptomic, metabolomic, and exosomal data, enabling composite diagnostic signatures beyond CA19-9. In longitudinal EHR analysis, temporal deep learning models identify latent disease trajectories and predict pancreatic cancer risk months to years before clinical diagnosis. Despite these advances, most models remain retrospectively validated and face limitations related to data heterogeneity, interpretability, and cross-population generalizability. AI strengthens early detection through multimodal integration, risk-adapted stratification, and data-driven clinical support aligned with precision medicine. Its near-term value lies in augmenting detection among high-risk populations rather than enabling universal screening or autonomous diagnosis. Prospective multicenter validation and improved model transparency are critical for translation into routine practice.
Endoscopic ultrasound (EUS)- guided drainage has emerged as a novel technique for managing pelvic abscesses. This single-center retrospective case series aims to assess the safety and efficacy of EUS-guided drainage in treating pelvic abscesses of varying etiologies from 2021 to the present. Consecutive patients with pelvic abscesses who underwent EUS-guided drainage were retrospectively reviewed. Etiologies included appendiceal abscess secondary to acute appendicitis (n = 1), pelvic abscesses resulting from anastomotic leaks following rectal cancer surgery (n = 2), and perianal abscesses associated with Crohn's disease (n = 7). The primary outcome was technical success and reduction in abscess cavity size, assessed via follow-up imaging. Clinical success was defined as significant reduction or complete resolution of the abscess cavity size on follow-up imaging at one-month post-procedure, accompanied by clinical symptom resolution and without the need for additional interventions. Secondary outcomes included post-procedural complications and resolution of the abscess without additional interventions. EUS-guided drainage was technically successful in all cases. The median reduction in abscess size was statistically significant (Mean SD: 24.1 ± 11.11, p < 0.05). During follow-up, imaging results confirmed significant reduction in the size of pelvic abscesses in 9 patients, except for one case at the 1-month post-procedure. None of the patients required further surgical intervention, and 2 cases recurrences were observed in the sixth- and tenth-months post-procedure. Additionally, no procedure-related complications were reported. EUS-guided drainage is a safe and effective therapeutic option for managing pelvic abscesses of various etiologies. Its efficacy, particularly in Crohn's disease-related cases, and the absence of complications in this cohort, suggest significant potential for broader clinical application.
Near-infrared fluorescence (NIRF) can deliver high-contrast, video-rate, non-contact imaging of tumor-targeted contrast agents with the potential to guide surgeries excising solid tumors. However, it has been met with skepticism for wide-margin excision due to sensitivity and resolution limitations at depths larger than ~ 5 mm in tissue. To address this limitation, fast-sweep photoacoustic-ultrasound (PAUS) imaging is proposed to complement NIRF. In an exploratory in vitro feasibility study using dark-red bovine muscle tissue, we observed that PAUS scanning can identify tozuleristide, a clinical stage investigational imaging agent, at a concentration of 20 µM from the background at depths estimated to be of up to ~ 34 mm, highly extending the capabilities of NIRF alone. The capability of spectroscopic PAUS imaging was tested by direct injection of 20 µM tozuleristide into bovine muscle tissue at a depth of ~ 8 mm. Experimental results demonstrate that multi-point laser fluence compensation and strong clutter suppression enabled by the unique capabilities of the fast-sweep approach greatly improve spectroscopic accuracy and the PA detection limit and strongly reduce image artifacts. Thus, the complementary NIRF-PAUS approach can be promising for comprehensive pre- (with PA) and intra- (with NIRF) operative solid tumor detection and wide-margin excision in optically guided solid tumor surgery.
Ferritins are vital macromolecules that have been widely used in a number of biotechnological fields. Ferritin-based hybrid nanoparticles, composed of different types of subunits and conjugates, represent a next generation of tools, which can significantly enhance their efficiency and expand the range of existing applications. This review outlines the application landscape of these hybrids in developing recombinant vaccines, drug delivery and imaging systems. We highlight the increasing trend towards the development of ferritin-based mosaic vaccines and some of them are already in the first or second phases of clinical studies. In comparison, drug delivery research, which is mostly focused on cancer theranostics, to our knowledge, has not progressed beyond the preclinical stage. Herein, we describe the key limitations and challenges of ferritin-based drug delivery systems development, suggest strategies that address these limitations and discuss promising future research directions. We conclude that engineered ferritin hybrids hold significant potential as useful tools for immunology, theranostics and other biomedical applications.
Cyanuric chloride is a highly reactive, widely recognized compound in medicinal chemistry, enabling rapid and selective nucleophilic substitution reactions at its three chlorine positions. In the present study, explore the structural advantages of cyanuric chloride to develop a new DOTA-linked triazine-based scaffold for PSMA. Two scaffolds, abbreviated as PSMA-C1D and PSMA-C2D, were successfully synthesized with good yields and evaluated their properties through molecular docking, in vitro studies, radiolabelling, physicochemical properties, and internalization studies. The initial screening revealed that, the PSMA-C1D had greater potential as a PSMA-targeted imaging agent than PSMA-C2D. In vitro cytotoxicity assays further indicated good biocompatibility at imaging-relevant concentrations. The molecular docking demonstrated strong site-specific binding of PSMA-C1D to the PSMA active pocket (ΔG = - 10.2 kcal/mol), with interactions closely resembling the co-crystallized ligand. The radiolabelling of PSMA-C1D with Ga-68 shows high yield with > 95% radiochemical purity, excellent stability in multiple biological media, and high apparent molar activity (508 GBq/µmol). The tracer shows hydrophilicity (logD7.4 = - 2.76 ± 0.02), low %PPB (18 ± 5.4), and has a nanomolar affinity (Kd = 0.38 nM), with the percentage of bound internalization in LNCaP cells was 15 ± 2.9% incubation for 1 h. The study highlights the value of cyanuric chloride as a modular chemical hub for the design and linking of radiopharmaceuticals. It identifies [68Ga]Ga-PSMA-C1D as a promising, efficiently synthesizable, and highly PSMA-specific PET radiotracer for imaging prostate cancer.
Diagnostic pathology has long relied on the morphological interpretation of hematoxylin and eosin (H&E)-stained tissues to guide diagnosis and assess prognostic features. While pathologists intuitively recognize spatial patterns and architectural organization, these assessments remain largely qualitative and difficult to quantify systematically. Immunohistochemistry and immunofluorescence have introduced molecular specificity but are limited in multiplexing capacity, whereas bulk genomic and transcriptomic assays provide high molecular depth but lose spatial context by averaging signals across heterogeneous cell populations. Recent advances in spatial proteomics-including mass spectrometry-based imaging and cyclic immunofluorescence-now enable multiplexed, single-cell protein analysis within intact tissue architecture. These technologies have revealed complex immune and stromal microenvironments, spatially organized biomarkers predictive of therapeutic response, and molecular gradients underlying disease progression. By integrating histological and molecular information, spatial proteomics bridges traditional microscopy with high-dimensional omics, allowing quantitative, spatially resolved insights into tissue organization and disease mechanisms. This review summarizes recent developments in multiplexed spatial proteomics from both scientific and pathological perspectives, highlighting how these technologies extend beyond morphology to quantify histologic patterns, refine biomarker discovery, and facilitate clinical translation. The review also examines translational challenges and barriers to clinical implementation, including costs, standardization requirements, and workflow integration.
Deep learning methods have made great progress in the automatic segmentation of nasopharyngeal carcinoma, but challenges remain. Computer-aided automatic segmentation of nasopharyngeal cancer primary area is of great significance for automatic outlining of nasopharyngeal cancer target areas and accurate prediction of responsiveness and prognosis of metastatic lymph nodes in the neck after radiotherapy. In this paper, we use deep learning methods to construct an automatic segmentation network for gross target volume of nasopharynx, combine clinical factors and radiomics features to establish a radiomics nomogram model, which will then predict the final outcome of metastatic lymph nodes that have not achieved complete remission after radical radiotherapy. Clinical and IMRT radiotherapy plan CT data were retrospectively collected from 69 patients who received intensity-modulated radiation therapy between July 2014 and December 2016. These patients exhibited residual metastatic lymph node lesions without residual primary lesions on the first follow-up MRI and had continuous follow-up records. The median follow-up was 53 months (IQR 39.75-62.37), with 30 patients eventually regressing and 39 patients persisting or progressing. The ct images of 69 radiotherapy plans were randomly divided into training and test sets according to 8:2, and a fusion attention-based model was trained for automatic nasopharyngeal carcinoma segmentation. Based on the unet framework, a fusion attention model was proposed, and a 2·5 d convolutional neural network was used to deal with the anisotropy. An improved channel and spatial attention module is fused in the codec 4 layer to enable the network to focus on small targets. 2d interlaced sparse self-attention module is extended to 3d to better extract the feature information of the tumor target area and solve the problem of low contrast between the target area and the surrounding soft tissues, thus optimizing the overall segmentation effect. The performance of the segmentation model was evaluated using the mean dice coefficient, relative volume error (RVE), average symmetric surface distance (ASSD) and hausdorff distance (HD), using the target area of the primary lesion of nasopharyngeal carcinoma manually outlined by a senior radiation therapy specialist as the gold standard. Radiomics features were extracted using the pyradiomics package, and the classification performance of the radiomics model was assessed by the area under the curve of the receiver operating curve (ROC). The average dice coefficient, RVE, ASSD and HD of our model for nasopharyngeal carcinoma were 75.05%, 14.63%, 2.224 mm, and 8.75 mm, respectively, which were 11.01%, 26.34%, 3.101 mm, and 52.58 mm better than the baseline 3dunet model. The radiomic features were an effective predictor of tumor outcome in nasopharyngeal carcinoma, with the highest area under the receiver operating characteristic curve (AUC) of 0.892 for the radiomic nomogram in the training set and 0.825 for the radiomic model in the test set. The fused attention-based segmentation network for nasopharyngeal carcinoma can effectively and reliably segment the region of the primary nasopharyngeal carcinoma, and the radiomic nomogram can effectively predict the response after treatment.
High-risk stage II colorectal cancer (CRC) shows heterogeneous outcomes despite adjuvant chemotherapy. We developed and validated an interpretable multimodal deep learning model integrating clinical data, serum biomarkers, and venous-phase CT to predict 5-year CRC-specific mortality in high-risk stage II CRC. This retrospective, multicenter cohort included 778 high-risk stage II CRC patients from three centers, all treated with adjuvant chemotherapy and with complete preoperative clinical, biomarker, and venous-phase CT data. Patients were split into a development cohort (Centers A + B, n = 720) and an external testing cohort (Center C, n = 58). A multimodal model combining numerical (clinical + biomarker) and imaging (CT) inputs was developed and internally validated using tenfold cross-validation in the development cohort and evaluated in the external cohort. Interpretability was assessed using SHAP and Grad-CAM. In the development cohort, the multimodal model showed superior discrimination (AUC 0.89; 95% CI, 0.87-0.91) versus numerical-only (AUC 0.76) and imaging-only (AUC 0.69). In the external testing cohort (9/58 CRC-specific deaths), the multimodal model achieved an AUC of 0.88 (95% CI, 0.76-0.96). SHAP and Grad-CAM consistently highlighted age, CA125, and tumor regions on CT as key contributors. This interpretable multimodal approach, using routine clinical, biomarker, and CT data, improves 5-year mortality risk stratification in high-risk stage II CRC and may inform risk-adapted surveillance and clinical decision support; prospective validation is warranted before treatment modification.
Achondroplasia (ACH) is the most common skeletal dysplasia characterized by disproportionate short stature due to impaired endochondral ossification. One of the most critical and potentially fatal complications of ACH is foramen magnum and upper cervical canal stenosis. Compression at the cervicomedullary junction may lead to myelopathy, hypotonia, developmental delay, and central sleep apnea. Early detection and timely surgical intervention are essential to prevent permanent neurological injury. This retrospective study evaluated 15 pediatric patients with ACH (9 girls, 6 boys; age range 3-42 months, mean 17.2 months) who underwent foramen magnum decompression and C1 laminectomy at Marmara University Neurosurgery Department between 2016 and 2025. All patients underwent comprehensive neurological and radiological evaluation, including MRI and 3D CT of the craniovertebral junction, and were classified by the Achondroplasia Foramen Magnum Score (AFMS). Nine patients had AFMS level 4 stenosis and six had level 3. The anteroposterior diameter of the foramen magnum ranged from 4.03 to 11.03 mm, with an area between 17.40 and 105.16 mm2. Presenting symptoms included motor delay (n = 4), respiratory disturbances or central apnea (n = 4), and macrocephaly (n = 3). Postoperative imaging confirmed adequate decompression in all patients. Neurological and respiratory improvement occurred in all patients except one with persistent hypotonia. One patient died early postoperatively due to recurrent pneumonia and sepsis. Complications were minimal. Foramen magnum decompression with C1 laminectomy is a safe and effective procedure for infants and children with achondroplasia presenting with cervicomedullary compression. Early radiological and neurological evaluation, particularly with AFMS, facilitates accurate surgical decision-making and improves outcomes.
暂无摘要（点击查看详情）
Attention-Deficit/Hyperactivity Disorder (ADHD) is a prevalent neurodevelopmental disorder that imposes significant personal and societal burdens. Traditional diagnostic approaches, which rely on behavioral assessments, are susceptible to subjectivity and variability, underscoring the need for objective and automated diagnostic tools. This study develops an ADHD-specific, biologically informed multi-stream deep learning framework for pediatric brain MRI classification, in which a Vision Transformer (ViT) and an Enhanced Convolutional Neural Network (ECNN) are integrated with Raw MRI, Phase Spectrum Transform (PST), and Quantile Histogram Equalization with Denoising (QHED) representations to capture complementary global and local neuroanatomical characteristics. The architecture leverages complementary modeling capacities by combining global contextual representations from ViT with localized discriminative features extracted by ECNN across a biologically informed multi-stream preprocessing strategy, including Raw MRI to preserve global anatomy, Phase Spectrum Transform (PST) to highlight cortical boundary irregularities, and Quantile Histogram Equalization with Denoising (QHED) to enhance subtle gray-white matter contrasts. Experimental evaluations conducted on a stratified pediatric MRI dataset demonstrated that the proposed ViT+ECNN model achieved a classification accuracy of 99.4%, precision of 99.3%, recall of 99.5%, and an F1-score of 0.99, substantially outperforming standalone ViT and ECNN configurations. These findings indicate that hybrid transformer-convolutional models can substantially enhance diagnostic accuracy and offer a promising approach for supporting early identification and intervention in ADHD.
Perivascular epithelioid cell tumors (PEComas) are rare mesenchymal tumors composed of cells exhibiting an epithelioid morphology. These cells typically arrange around small blood vessels (perivascular spaces) and display dual differentiation characteristics of smooth muscle cells and melanocytes. Diagnosis is challenging due to the absence of specific symptoms or tumor markers. This case features a young male patient with a large hepatic PEComa, whose imaging findings resemble those of hepatocellular carcinoma. We have detailed the entire process from diagnosis to treatment to aid in differential diagnosis and surgical planning. A 31-year-old male patient with no prior medical history underwent a routine health examination 20 days prior to presentation. Although the patient was asymptomatic, ultrasound revealed an incidental hepatic lesion measuring 58 × 50 × 45 mm (maximum diameter 58 mm, or 5.8 cm). The screening center suspected a hemangioma. Subsequently, he presented to our hospital. Comprehensive imaging studies, including ultrasound, computed tomography (CT), and magnetic resonance imaging (MRI), revealed a 58 mm-diameter space-occupying lesion in segments V and VIII of the right hepatic lobe. Imaging findings initially raised suspicion for hepatocellular carcinoma. To minimize surgical trauma and preserve liver function, our team discussed surgical approaches and ultimately decided on a laparoscopic partial hepatectomy. During the procedure, we obtained a specimen for pathological examination. The final histopathological analysis confirmed the diagnosis of a PEComa with undetermined malignant potential. The patient recovered smoothly postoperatively and was successfully discharged. PEComa has an insidious onset and is rare. Early diagnosis is often challenging, and imaging studies typically show no highly specific findings. Clinical diagnosis frequently relies on biopsy. In terms of treatment, radical resection (R0 resection, i.e., negative margins) represents the definitive therapeutic approach.
Positron emission tomography (PET) instrumentation has seen significant advances over the last 20 years. In particular, substantial improvement in electronics and scalability have allowed new clinical scanners to combine high gamma stopping power and patient geometrical coverage in the total body PET paradigm. Similarly, spatial resolution has almost reached its physical limit imposed by the uncertainty caused by the positron range and acollinearity of emitted annihilation gamma rays. Time-of-flight (ToF) offers an improvement to effective sensitivity, proportional to the reverse square of detector timing resolution. In contrast to the aforementioned, this is a development direction that has not yet been sufficiently harnessed, with state of the art remaining far from the physical limit of ToF resolution. Nevertheless, existing scanners offer a glimpse of the advantages of ToF for imaging and diagnosis. In this review, analyse the limiting factors; describe the motivation for enhancing detector ToF capabilities; explore the physical mechanisms related to ToF application; describe the state-of-the-art in clinical, prototyping and laboratory stages; offer insights on emerging approaches and their capabilities to provide scalable, and cost-effective ToF; and finally envision the future of medicine, once ultraToF of the order of 10 ps has become the standard in clinically deployed scanners.
Neuroimaging studies have revealed altered functional connectome dynamics in autism spectrum disorder (ASD) and linked these alterations to clinical symptoms. However, most studies have emphasized population-level contrasts, leaving interindividual variability in connectome dynamics and its structural underpinnings poorly understood. To address this gap, we analyzed resting-state functional and structural MRI data from 939 male participants (440 with ASD, 499 typically developing controls) across 18 sites in the Autism Brain Imaging Data Exchange (ABIDE). Whole-brain functional state dynamics was characterized using five leading activity modes and their expressions via eigen-microstate analysis. Age-related trajectories of mode expressions were constructed for typically developing controls using normative modeling, enabling quantification of individual-level deviations in functional dynamics. Compared with controls, ASD individuals showed greater interindividual variability in functional deviation profiles. Unsupervised clustering of these profiles identified two robust ASD subtypes with distinct mode-specific dysfunctions. One subtype primarily involved the visual, default-mode, frontoparietal, and dorsal attention networks, whereas the other subtype primarily involved the somatomotor, visual, frontoparietal, and ventral attention networks. These subtypes were clinically dissociable, differing in restricted and repetitive behaviors and social impairments, and exhibited mode-specific brain-symptom associations. Furthermore, the subtypes exhibited distinct cortical thickness alterations, and individual subtype membership was predicted with high accuracy (83%) using a random forest classifier based on cortical thickness. The main findings were replicated in an independent cohort outside ABIDE. This study delineates two reproducible and clinically dissociable ASD subtypes and links functional connectome dynamics to structural substrates, offering novel insights into the neurobiological basis behind ASD heterogeneity.
The adnexal region in females presents complex imaging due to the menstrual cycle. Accurate diagnosis is crucial for effective tumor treatment. This study aims to assess the clinical utility of abnormal adnexal uptake on 68Ga-FAPI PET/CT for early and precise lesion characterization. This study retrospectively analyzed all female patients with abnormal adnexal uptake on 68Ga-FAPI PET/CT at our institution from November 2021 to June 2024. Semiquantitative analysis of PET/CT imaging parameters was performed, combined with serum tumor markers and immunohistochemical markers. Pathological findings or imaging follow-up ≥ 6 months served as the gold standard for evaluating the diagnostic performance of 68Ga-FAPI PET/CT. The study included 121 female patients with a mean age of 53.8 ± 12.2 years (18-80 years). A total of 184 adnexal lesions with abnormal uptake were identified. Pathology/follow-up confirmed 82.6% as malignancies, comprising 84 primary and 16 metastatic cases. Additionally, 2 borderline tumors and 19 benign lesions were detected. The positive predictive value was 83.5%. SUVmax differed significantly among primary malignant, metastatic, and benign lesions (12.52 ± 5.41 vs. 9.78 ± 3.39 vs. 5.52 ± 4.17). In the subset of patients with available pathological specimens, SUVmax showed weak to moderate positive correlations with Ki-67 (r = 0.361, p < 0.001) and p53 (r = 0.419, p < 0.001). Notably, the mean SUVmax in the Ki67 > 20% group was significantly higher than in the Ki67 ≤ 20% group (p < 0.001). ROC analysis showed an AUC of 0.85 of SUVmax alone for diagnosing malignant adnexal lesions, increasing to 0.89 when combined with tumor markers. 68Ga-FAPI PET/CT demonstrates high diagnostic performance for ovarian lesions. Among pathologically confirmed cases, SUVmax correlates with proliferative activity and malignant potential, supporting its role in diagnostic optimization.
Residual cardiovascular risk persists despite intensive statin therapy in patients with established atherosclerotic cardiovascular disease (CVD). Omega-3 fatty acids, particularly high-dose eicosapentaenoic acid (EPA), have been proposed as adjunctive therapy, yet trial results conflict, likely due to formulation differences. We conducted a formulation-focused meta-analysis to determine whether high-dose EPA-dominant supplementation reduces cardiovascular events and to quantify the impact of mixed EPA/docosahexaenoic acid (DHA) regimens on efficacy. Following Preferred Reporting Items for Systematic reviews and Meta-Analyses (PRISMA) 2020 guidelines, we searched MEDLINE, Embase, CENTRAL, and trial registries through May 2025 for randomized controlled trials, including placebo-controlled and open-label designs, of high-dose EPA-dominant omega-3 (≥ 1.8 g/day; ≥ 50% EPA) in adults with established CVD or other high-risk settings. Six trials (n = 42,738; 31-85% male) were eligible. Random-effects models generated pooled risk ratios (RRs), with I2 assessing heterogeneity; sensitivity analyses excluded mixed EPA/DHA formulations. Imaging surrogate outcomes were summarized narratively when study modalities were not directly comparable. EPA-based therapy significantly reduced hospitalizations for unstable angina (RR 0.75, 95% CI 0.66-0.87; I2 = 0%). Overall effects on recurrent myocardial infarction and revascularization were not statistically significant, but both became significant after exclusion of STRENGTH, the only mixed EPA/DHA cardiovascular outcomes trial. No significant effect was observed for ischemic stroke, cardiovascular death, or high-sensitivity C-reactive protein (hs-CRP). CHERRY and EVAPORATE both suggested attenuation of plaque progression, but these imaging studies were not pooled because intravascular ultrasound and coronary computed tomography angiography-derived measures were not directly comparable. High-dose EPA-dominant therapy was associated with fewer unstable angina hospitalizations, and formulation appeared to modify clinical benefit. Among blinded, placebo-controlled, cardiovascular outcomes trials, 4 g/day icosapent ethyl is the only formulation independently associated with reduced cardiovascular events. Larger formulation-specific trials are needed to clarify the roles of purified EPA, mixed EPA/DHA regimens, and patient selection. PROSPERO identifier number: CRD420251063069.
Metaphyseal anadysplasia 1, which includes Spondyloepimetaphyseal dysplasia Missouri type, is a rare autosomal dominant skeletal dysplasia characterized by short stature, mild limb deformities, and transient metaphyseal irregularities that typically resolve with age. The condition is caused by heterozygous missense variants in the MMP13 gene, encoding matrix metalloproteinase 13, a key enzyme in endochondral ossification and extracellular matrix remodeling. Pathogenic variants in MMP13 are exceedingly rare, with only a few families reported. We report two siblings, aged 3 and 1 years, in Sweden, presenting with clinical and radiological features consistent with Metaphyseal anadysplasia 1. Their father, of Syrian origin, exhibited short stature and mild femoral bowing. Genetic analysis revealed a novel heterozygous missense variant c.217T>C, p.(Ser73Pro) in MMP13, inherited from the affected father, and located within the same MMP13 domain as previously reported patients. The family pedigree demonstrates multiple affected individuals with short stature and bowed femurs, consistent with autosomal dominant inheritance. Radiographic imaging of father confirmed persistent but mild skeletal abnormalities. This report expands the genotypic spectrum of Metaphyseal anadysplasia 1 and suggests a putative mutational hotspot in exon 2. It further emphasizes the importance of thorough clinical, radiological, and genetic evaluation in families with short stature and metaphyseal irregularities and a clinical long-term follow-up is proposed with regular radiographic monitoring.
Breast ultrasound imaging is widely used for the early detection of breast cancer due to its accessibility and effectiveness, particularly in dense breast tissues. However, its diagnostic performance is often affected by operator dependency, speckle noise, low contrast, and variability in data quality. Although deep learning methods have shown promise in automated tumor segmentation and classification, their clinical applicability remains limited due to challenges such as small and imbalanced datasets, inconsistent annotations, and the lack of integrated learning strategies. In this work, we propose a Multi-Task U-Net framework that jointly performs lesion segmentation and tumor classification by leveraging shared feature representations. The proposed method incorporates a deterministic oversampling strategy for handling class imbalance, a prediction-refinement module to ensure consistency between segmentation and classification outputs, and an attention-guided feature learning mechanism to enhance lesion localization. Additionally, a curated version of the BUSI dataset is constructed by removing duplicate and inconsistent samples to ensure reliable evaluation. The proposed model achieves a Dice score of up to 0.81 in comparative evaluation, along with classification accuracy of up to 0.96-0.98, demonstrating improved performance over baseline methods. The consistent performance across both segmentation and classification tasks indicates good generalization capability despite dataset limitations. Finally, the proposed multi-task framework provides an effective and reliable solution for automated breast cancer detection in ultrasound images and shows strong potential for clinical application.
Posterior glenoid dysplasia (PGD) has been described as an adaptive osseous change in young throwing athletes exposed to repetitive mechanical stress. However, prior studies have lacked consistent definitions, reliable classifications, or sport-specific analysis. This study aimed to analyze and classify PGD in young athletes across various sports and compare its proportion and severity between symptomatic baseball players and athletes from other sports. We retrospectively reviewed 10 years of shoulder CT imaging data from 568 young athletes presenting with shoulder pain: 418 baseball players, 50 overhead athletes, and 100 athletes from other sports. Dysplasia of posterior to posteroinferior glenoid, bone defect size, and additional bone formation were assessed. We established a five-type classification system: Type 0 (normal sharp rim), Type 1 (smooth rounded rim), Type 2 (triangular defect), Type 3 (retreated joint surface with posterior bulging), and Type 4 (new bone formation). PGD was identified in 85.4% of baseball players, 32.0% of overhead athletes, and 9.0% of non-overhead athletes (p<0.001). Type 2 was most common in baseball players (42.6%), and advanced morphological PGD (Types 3-4) was present in 13.4% of baseball players but was absent in other groups. PGD was more commonly observed among overhead athletes, particularly baseball players, within this symptomatic imaging cohort. Advanced morphologic types were identified predominantly in baseball players. These findings provide a structured framework for describing posterior glenoid morphology in throwing athletes and may facilitate future prospective studies investigating clinical relevance and natural history.