Citations

Abstract Background Stroke is a leading cause of mortality and disability globally. However, information about stroke burden in the Philippines is limited. We sought to analyze stroke burden in the Philippines from 1990 to 2023. Methods Incidence, prevalence, mortality, and disability-adjusted life-years (DALYs) estimates from the Global Burden of Disease Study 2023 data were used as indicators to analyze the burden of stroke by sex and age. Temporal trends in both crude and age-standardized rates were analyzed using joinpoint regression analysis. Results In 2023, stroke incidence was estimated at 156.2 (95% uncertainty interval [UI]: 140.8–175.4) thousand, prevalence at 1.2 (95% UI: 1.2–1.4) million, mortality at 72.2 (95% UI: 63.2–83.0) thousand, and DALYs at 2.1 (95% UI: 1.8–2.3) million. High systolic blood pressure was the leading contributor to risk-attributable stroke mortality and DALYs. Since 1990, age-standardized rates declined significantly, whereas crude rates increased markedly. Compared with women, men had a higher fatal burden and consistently exhibited a higher age-standardized burden. Although older adults (≥ 55 years) had the highest stroke burden and achieved reductions in stroke incidence and fatal outcomes, both fatal and non-fatal burdens consistently increased among young adults (35–54 years). Conclusion While age-standardized rates have improved, the rising crude burden and shift towards younger adults present significant public health challenges. These trends highlight the pressing need for aggressive and targeted risk factor control, sustained risk monitoring, and strengthened acute and post-stroke care to mitigate the growing health burden of stroke in the Philippines.

Background Predictive models employing machine learning algorithms are increasingly being used in clinical decision making, and improperly calibrated models can result in systematic harm. We sought to investigate the impact of class imbalance correction, a commonly applied preprocessing step in machine learning model development, on calibration and modelled clinical decision making in a large real-world context. Methods A histogram boosted gradient classifier was trained on a highly imbalanced national dataset of >1.8 million patients undergoing surgery, to predict the risk of 90-day mortality and complications after surgery. Class imbalance correction strategies including random oversampling, synthetic minority oversampling technique, random under-sampling, and cost-sensitive learning were compared to the natural distribution ('natural'). Models were tested and compared with classification metrics, calibration plots, decision curve analysis, and simulated clinical impact analysis. Results The natural model demonstrated high performance (AUROC 0.94, 95% CI 0.94--0.95 for mortality; 0.84, 95% CI 0.84--0.85 for complications) and calibration (log loss 0.05, 95% CI 0.04--0.05 for mortality; 0.23, 95% CI 0.23--0.24 for complications). Class imbalance mitigation (CSL, ROS, RUS, and SMOTE) did not improve AUROC or AUPRC but increased recall and F1 scores at the expense of precision and accuracy. However, these methods severely compromised model calibration, leading to significant over-prediction of risks (up to a 62.8 % increase) as further evidenced by increased log loss across all mitigation techniques. Decision curve analysis and clinical scenario testing confirmed that the natural model provided the highest net benefit. Conclusion Class imbalance correction methods result in significant miscalibration, leading to possible harm whenused for clinical decision making.

OBJECTIVES: To evaluate and compare the ability of three popular open-source artificial intelligence platforms to diagnose common trauma-related fractures using radiologic imaging. METHODS: Design: Retrospective diagnostic performance comparison study. Setting: Publicly accessible online radiologic imaging databases. Patient Selection Criteria: Five common orthopedic trauma fractures were assessed: ankle, tibial plateau, intertrochanteric, femoral neck, and humerus. Radiographs and computed tomography (CT) images were collected. Images were randomly selected from confirmed diagnoses on Radiopaedia.org. Outcome Measures and Comparisons: ChatGPT 5, Grok 3, and Claude 4.5 Sonnet were queried with each image. Diagnostic accuracy, sensitivity, specificity, positive and negative predictive values, and performance by modality (X-ray vs. CT) were assessed. The reference standard was the expert-verified diagnosis provided by Radiopaedia.org, limited to cases labeled with a “diagnosis certain” tag. RESULTS: Each model was provided with 30 radiographs and 20 CT images whenever possible. ChatGPT 5, Grok 3, and Claude 4.5 Sonnet accurately diagnosed diseased images in 26.8%, 18.8%, and 22.4% of cases, respectively. By fracture type, ChatGPT 5 demonstrated the highest correct classification rates for ankle (10%), femoral neck (38%), humerus (40%), and tibial plateau (44%) fractures. Grok 3 demonstrated the highest correct classification rate for intertrochanteric fractures (6%). Overall sensitivities were 0.267, 0.187, and 0.223 for ChatGPT 5, Grok 3, and Claude 4.5 Sonnet, respectively. ChatGPT 5 and Grok 3 outperformed Claude 4.5 Sonnet (both P < 0.001). No modality-based performance differences were observed for any model. CONCLUSIONS: Among the publicly available large language models evaluated for radiologic interpretation of orthopedic trauma imaging, ChatGPT 5 demonstrated the highest overall diagnostic accuracy, followed by Claude 4.5 Sonnet and Grok 3. Despite relative variation between the models, overall diagnostic accuracy for fracture detection was low across all platforms (<27%). In their baseline forms, these publicly accessible large language models are not recommended for radiologic imaging interpretation. LEVEL OF EVIDENCE: Diagnostic Level III. See Instructions for Authors for a complete description of levels of evidence.

Case: A 4-year-old boy with a history of posterior fossa ependymoma resection and radiation treatment developed postlaminectomy kyphosis at the occipito-cervical junction. Cervical radiographs showed 46° kyphosis between C2 and C4. A 3-stage surgical approach was done: halo-gravity traction, anterior C2 to C4 fusion with a fibular allograft, and posterior occiput-C4 instrumented fusion. Halo traction reduced kyphosis by over 50%. At 1-year follow-up, the patient showed stable fixation and successful fusion with modest limitation in cervical motion. Conclusion: The approach effectively corrected the deformity and successfully achieved fusion despite previous radiation treatment and bone deficiency in a very young, skeletally immature patient.

Integration of the high-risk human papillomavirus 16 (HPV16) genome into the host chromosome, frequently driven by microhomology-mediated end joining (MMEJ), is a critical step in the carcinogenesis of HPV-associated tumors. However, the mechanisms by which viral oncoproteins manipulate the error-prone MMEJ pathway remain poorly defined. Here, we demonstrate that the HPV16 E6 oncoprotein upregulates MMEJ to facilitate viral genome integration. This heightened MMEJ activity is driven by a marked increase in the protein levels of DNA polymerase theta (PolΘ), a central enzyme of the MMEJ pathway. Mechanistically, we show that the elevation of PolΘ levels in response to HPV16 E6 expression is dependent on the host E3 ubiquitin ligase UBE3A/E6AP but is independent of p53 degradation. E6 redirects UBE3A to enhance the ubiquitination and degradation of RAD23A, a shuttle protein required for delivering polyubiquitinated PolΘ to the proteasome. Consequently, the loss of functional RAD23A phenocopies the effect of HPV16 E6, leading to PolΘ protein stabilization and increased MMEJ activity. By elucidating the E6-UBE3A-RAD23A-PolΘ axis, our findings reveal a mechanism through which HPV manipulates the host DNA repair machinery to promote its integration and oncogenic potential.

This slide presentation uses real-world examples from diverse modern systems to categorize digital radiographic artifacts by likely cause and outlines investigation and mitigation strategies to improve troubleshooting, artifact prevention, and patient care.

Immune checkpoint inhibitors (ICIs) can cause immune-related adverse events (irAEs), with acute interstitial nephritis (ICI-AIN) being the most common irAE. While the exact mechanism remains unclear, upregulation of IFN-γ and TNF-α pathways has been implicated. This study used a humanized chimeric PD-1/PD-L1 mouse model to assess renal effects of ICIs, alone or combined with proinflammatory cytokines, and to test if selective TNF-α blockade could prevent ICI-AIN. Mice were randomly divided into 4 experimental groups: Control, ICI-Only, ICI-Cytokines (ICI-Cyt), and ICI-Block (ICI-TNF-α blockade). Renal function and cytokine profiles were assessed, while kidney tissue was analyzed using microscopy and single-cell RNA-seq. Histology revealed increased renal infiltration of CD4 + /CD8 + T cells in ICI-treated groups and decreased TNF-α expression following TNF-α blockade. Additionally, kidney tissue ELISA demonstrated reduced IFN-γ levels following TNF-α blockade. Plasma IL-6, MCP-1, and TNF-α were lower in ICI-Block mice. Single-cell RNA-seq revealed shifts in immune cell populations and genes of interest including Bcl2a1 , Icos , Il18r1 , Ccr2 , and Jaml . This humanized model replicates ICI-AIN key features, revealing a synergistic role of ICIs and proinflammatory cytokines. TNF-α blockade demonstrated protective effects, supporting its potential role in mitigating the risk of ICI-AIN.