Automatic radiology report generation has gained increasing attention for its potential to assist in clinical reporting and reduce the workload of radiologists. Existing 3D radiology report generation methods employ multi-modal foundation model to encode volume-text inputs and produce diagnosis reports, while they ignore the characteristics of 3D volumes including much background regions and suffer from generating hallucinations, especially in medical domain that contains many uncommon professional terms. In this paper, we aim to efficiently adapt the pre-trained foundation model to specific 3D radiology report generation, and present a Prior-drIven Prompting with diagnosis-oriented retrieval-Augmentation (PIPA) framework. In PIPA, we design a Prior-drIven Prompting (PIP) strategy to exploit diagnostic knowledge from input volumes and a Diagnosis-oriented volume-report retrievalaugmentation Generation (DIG) module to explore beneficial knowledge from external database. Specifically, in PIP, to take full advantage of the patient's clinical information, e.g., age and symptoms, and the possible disease information, e.g., brain tumor, edema, we formulate them as the patient and disease priors to mine clinical relevant knowledge. Furthermore, we propose utilizing visual and textual embeddings as queries to retrieve similar external data by devising a diagnosis-oriented retrieval-augmentation scheme for leveraging more report resources as references for LLM to produce accuracy outcomes. With PIP and DIG, PIPA integrates clinical priors and external data to learn effective diagnostic representations for high-quality report generation. We evaluate the framework on both public and in-house 3D medical datasets with corresponding reports, demonstrating its strong performance in generating accurate diagnosis reports. Source codes have been published at https://github.com/CUHK-AIM-Group/PIPA/tree/ main.

Heart failure remains a leading cause of potentially avoidable hospital admissions throughout the United States and a significant driver of unnecessary, avoidable costs within value-based care programs. In a baseline population of 3233 hospital emergency department (ED) visits for heart failure, which occurred over 12 months ending in June 2023 across six WellSpan Health acute care hospitals, 2868 (88.7%) resulted in inpatient admissions. Among those admissions, 2535 (96.0%) were deemed potentially avoidable by the U.S. Centers for Medicare and Medicaid Services (CMS) Prevention Quality Indicator (PQI) 08 heart failure admission quality indicator. Collectively, these admissions totaled US$27,618,825 in potentially avoidable total costs of care. Through process improvements that modified triaging algorithms related to heart failure exacerbations where shortness of breath or lower extremity edema were present, 92.2% of triaged patients with heart failure were managed in an ambulatory setting, with 84.7% avoiding an ED visit within 24 hours. The 12% cost reduction goal was almost achieved just 1 year into the intervention, with an 11.2% reduction in potentially avoidable admissions (CMS PQI 08, a core quality measure) and a US$3,352,248 reduction in total costs of care. This novel intervention executed with Lean management principles was a pragmatic process improvement that can be readily replicated by other systems to reduce unnecessary, avoidable referrals of mild to moderate heart failure exacerbations to hospital EDs.

Endovascular thrombectomy (EVT) has transformed the treatment of acute ischemic stroke (AIS). However, a substantial proportion of AIS patients experience poor outcomes despite successful recanalization, often due to severe neurological deterioration or life-threatening complications. Early identification of these high-risk patients remains a major unmet need. In this study, we developed and validated machine-learning (ML) models that integrate automated quantitative brain arterial morphology and collateral grading with demographic, clinical, laboratory, and imaging variables to predict major post-EVT complications and early neurological outcomes. Using a prospectively collected database of 727 AIS patients that underwent EVT, we developed ML models to incorporate patient-specific vascular morphometry with conventional clinical, laboratory, and imaging data to predict emergence of early neurological deterioration (END), symptomatic intracranial hemorrhage (sICH), malignant brain edema (MBE) requiring surgical decompression, and neurogenic respiratory failure and dysphagia requiring tracheostomy/gastrostomy (TC/PEG). Our analysis of morphological features, including increased tortuosity and reduced vessel diameter, showed strong associations with complications. Morphology-informed (MI) models consistently outperformed baseline-clinical (BC) models for patients with END (AUROC 0.81 for MI model vs. 0.73 for BC), sICH (AUROC 0.68 MI vs. 0.56 BC model), MBE (AUROC 0.67 MI model vs. 0.56 BC), or patients who underwent TC/PEG (AUROC 0.66MI vs. 0.58 BC model). Statistical testing confirmed significant AUROC improvements for END, sICH and mRS (p < 0.05), Finally, patient-specific calibrated probability profiles enabled individualized, multidimensional risk stratification, revealing distinct complication-specific risk patterns across patients. These findings demonstrate that cerebrovascular structure-an often overlooked yet physiologically fundamental determinant of ischemic injury and reperfusion dynamics-provides significant predictive information that is not captured by standard clinical or visual imaging assessments. Automated vascular segmentation and collateral grading techniques enable rapid and objective integration of cerebrovascular metrics into prognostic models, offering a scalable tool for precision risk stratification, supporting earlier intervention, targeted monitoring, and improved post-EVT management.

This study identifies amide proton transfer (APT) imaging radiomic features that correlate with glioblastoma survival and explore the relationship of combined radiomic data and biomarkers such as age and MGMT methylation status to survival prognosis. In a retrospective study of 66 glioblastoma patients (2022-2023), SlicerBrats was used for automatic tumor segmentation and 504 radiomic features were extracted from APT images. A two-step feature reduction process was applied: (1) Pearson correlation reduced the features to 59 by mitigating multicollinearity, and (2) a Cox model with best subset selection identified the optimal features using the lowest Akaike Information Criterion (AIC). The Goodness-of-Fit (GOF) test was used to evaluate model adequacy. Bootstrapping with 1000 iterations was used to compute the confidence interval for the concordance index (C-index). Univariate Cox models assessed individual radiomics, while models with multiple features evaluated radiomics alone, with age, MGMT methylation, and both. Kaplan-Meier and log-rank tests were performed to stratify survival groups. The radiomics-only model showed the highest C-index (0.74), AIC (191.35), and GOF (p < 0.005). Adding age maintained the C-index (0.74) but raised AIC to 193.16 (p = 0.01). The radiomics + MGMT model achieved a C-index of 0.73 and the lowest AIC (190.15). Kaplan-Meier curves and log-rank tests confirmed all models significantly distinguished survival groups (p < 0.05). APT-derived radiomic features, especially from edema and necrosis, correlate with glioblastoma survival time. Adding MGMT methylation improved predictive accuracy, achieving the lowest partial AIC and a C-index of 0.73. Further validation in larger cohorts is recommended.

Lymphatic vessel dysfunction is increasingly recognized in metabolic diseases, contributing to edema, dyslipidemia, and tissue lipid accumulation. Type 2 diabetes has been experimentally linked to lymphatic vascular defects characterized by enhanced permeability resulting from low nitric oxide (NO) bioavailability. To examine the effects of this dysfunction, we developed a mathematical model of fluid exchange between a collecting lymphatic vessel and the surrounding tissue. The lymph flow is modeled using the Stokes equations for an incompressible, viscous fluid, while the collecting vessel is represented by a thin Koiter shell model. The main novelty of our approach is that the interstitium is described as poroelastic, and modeled using the Biot equations, which capture both porous medium flow and the displacement. Lymph movement within the vessel is regulated by biological factors, specifically calcium ion (Ca 2 +) concentration and NO production, which are included to modulate the pumping mechanism. Using this model, we quantify lymph flow and fluid exchange for both wild-type (WT) and diabetic vessels. Our findings illustrate how variations in NO production rate, Young's modulus, and hydraulic conductivity impact leakage and pumping efficiency. The results demonstrate that reduced NO production in diabetic conditions leads to increased fluid leakage and altered pulsation frequency, while increased vessel stiffness further compromises lymphatic function.

Seatbelt-related breast trauma is an uncommon consequence of motor vehicle collisions. Most injuries present as localized contusion or hematoma, whereas progressive breast hematomas requiring surgical intervention are exceedingly rare, particularly in patients without coagulation abnormalities. Obesity may further alter trauma biomechanics and influence hematoma progression. A 49-yearold woman with obesity (Body Mass Index: 38 kg/m2) presented to the emergency department one hour after a motor vehicle collision with left breast swelling, firmness, and ecchymosis. Body temperature, blood pressure, and heart rate were within normal limits. Contrast-enhanced computed tomography (CT) demonstrated a 62×52 mm breast hematoma without evidence of arterial extravasation. Conservative treatment with compression therapy was continued for three days; however, pain and breast tension progressively increased, and ultrasonography demonstrated mild enlargement of the hematoma. On the fourth day, minimally invasive surgical evacuation was performed, yielding a large volume of organized hematoma. Postoperatively, symptoms improved rapidly. At two-week follow-up, breast symmetry was fully restored with no residual deformity. This case represents a rare presentation of seatbelt-related breast trauma: a progressive breast hematoma without active bleeding that ultimately required surgical evacuation. Obesity may have contributed to the failure of conservative management and earlier clinical progression. Early recognition of symptom worsening is essential to optimize cosmetic and functional outcomes.

Wound management, especially for infected and diabetic ulcers, poses a persistent clinical challenge, largely attributed to sustained inflammation, bacterial colonization, and compromised tissue regeneration. Herein, a multifunctional hydrogel (CBOS) with integrated therapeutic properties was developed through Schiff base reaction between betaine-modified chitosan (CSBT) and oxidized sodium alginate (OSA), with poly(vinyl alcohol) and sodium borate incorporated to enhance network stability. The resulting CBOS hydrogel exhibited a uniform porous microstructure, robust viscoelasticity, high swelling capacity, and appropriate water vapor transmission. Comprehensive biocompatibility assessments confirmed its negligible cytotoxicity and applicable hemocompatibility. The hydrogel demonstrated broad-spectrum tissue adhesion and potent hemostatic efficacy in murine tail incision and liver hemorrhage models. Moreover, CBOS displayed antioxidant activity by scavenging multiple free radicals and reducing intracellular reactive oxygen species in H2O2-stressed fibroblasts. Antibacterial evaluations revealed potent activity against S. aureus, with molecular docking studies suggesting multitarget interactions with key bacterial proteins, including AgrA, ClpP, and FtsZ. In vitro, CBOS significantly enhances fibroblast migration. In vivo, the hydrogel accelerated healing in both S. aureus-infected full-thickness wounds and diabetic chronic wounds, as evidenced by reduced bacterial burden, attenuated inflammatory responses, enhanced collagen deposition, and promoted neovascularization. With its integrated hemostatic, antioxidant, antibacterial, and pro-regenerative properties, the CBOS hydrogel offers a viable and attractive therapeutic approach for complex wound tissue repair.

Additive manufacturing offers significant opportunities for the development of blood-contacting devices. However, traditional biocompatible materials suitable for material extrusion (ME) often exhibit limited leak-tightness, restricting their application in this field. Acrylonitrile butadiene styrene (ABS) is a widely used thermoplastic valued for its impact resistance, durability, and compatibility with ME technology, making it a promising candidate for functional biomedical components and blood-contacting applications. Although some ABS filaments are marketed as medical grade and assessed for general biocompatibility, their hemocompatibility remains largely untested. This study aims to evaluate the hemocompatibility of ME 3D-printed ABS and to investigate whether acetone post-processing affects its blood-contact properties. Dimensional accuracy, surface roughness, PBS uptake, swelling, and wettability were characterized, and hemocompatibility was assessed through hemolysis, coagulation, and platelet and leukocyte interaction assays before and after acetone treatment, using two ABS filaments, one of which was marketed as medical grade. Hemolysis remained below the hemolytic threshold, indicating that ABS is non-hemolytic. Similarly, clot masses and platelet and leukocyte adhesion were significantly lower than positive controls. However, coagulation assays performed without anticoagulants indicated a time-dependent tendency toward thrombogenesis under static conditions. Acetone post-treatment improved surface smoothness and layer adhesion without adversely affecting hemocompatibility. No significant differences were observed between filament types or between acetone-treated and untreated groups. These findings suggest that 3D-printed ABS is suitable for in vitro and lab-scale blood-contacting applications, particularly when post-processed with acetone, although careful design considerations remain necessary to minimize thrombotic risks.

Olecranon bursitis is a common clinical condition that is usually traumatic, inflammatory, or bacterial in origin. Fungal involvement of the olecranon bursa is exceedingly rare and may present a diagnostic challenge because of its indolent course and nonspecific clinical and radiological features. A 63-year-old female with diabetes mellitus presented with a gradually progressive, painless swelling over the posterior aspect of the left elbow of 1 year duration. Ultrasonography revealed a thick-walled cystic lesion with internal echoes and septations overlying the olecranon process, while plain radiography showed a posterior soft-tissue density lesion with tiny calcific specks. The lesion was surgically excised. Gross examination showed a cystic mass containing thick greenish purulent material. Histopathological examination revealed a fibrocollagenous cyst wall with dense mixed inflammatory infiltrates and fibrinous necroinflammatory material containing scattered fungal hyphae. Gomori methenamine silver stain highlighted fungal elements, confirming the diagnosis of mycotic olecranon bursitis. Fungal bursitis should be considered in the differential diagnosis of chronic, nonresolving olecranon bursitis, particularly in patients with underlying risk factors such as diabetes mellitus. Careful histopathological examination with special stains plays a critical role in establishing the definitive diagnosis and guiding appropriate management.

We report the case of a 36-year-old male patient with an 18-year history of Behçet's disease who was referred to our department in October 2025 for dyspnea and dysphagia. Flexible laryngoscopy revealed a partial epiglottic defect with supraglottic mucosal hyperplasia and edema. Microlaryngoscopic carbon dioxide (CO2) laser surgery of the larynx was performed, with dyspnea disappearing after the operation. No recurrence of dyspnea was observed during the 3-month follow-up.