Citations

Agentic artificial intelligence marks a shift from large language models to autonomous, goal-driven clinical assistants. The framework presented in this report outlines implementation in radiology workflows and addresses cybersecurity, bias, and accountability challenges.

Abstract INTRODUCTION The genetic basis of sporadic early‐onset Alzheimer's disease (EOAD) remains largely unknown, prompting evaluation of late‐onset Alzheimer's disease (LOAD) polygenic risk in EOAD. METHODS A LOAD polygenic score (PGS) was calculated in the Longitudinal Early‐onset Alzheimer's Disease Study (LEADS) and Alzheimer's Disease Neuroimaging Initiative (ADNI) study and tested for associations with AD risk, cognitive performance, and imaging and fluid biomarkers. RESULTS Though PGS was elevated in LOAD and EOAD, it was not a significant predictor of EOAD adjusting for APOE ε4 carrier status and was not associated with age of EOAD onset ( p = 0.106) or with cognitive performance ( p = 0.417). In LEADS, greater LOAD PGS was associated with differences in neuroimaging and fluid biomarkers, including elevated synaptosomal‐associated protein 25 (SNAP‐25) ( p = 2.3 × 10 −5 ). DISCUSSION While LOAD polygenic risk contributed minimally to EOAD onset and cognitive dysfunction, PGS association with fluid biomarkers in LEADS suggests a role for LOAD polygenic risk in EOAD pathophysiology. Highlights LOAD PGSs were elevated in both LOAD and EOAD compared to controls; however, LOAD PGS did not significantly predict EOAD risk, age at onset, or cognitive performance independent of APOE ε4 in the LEADS. Higher LOAD PGS was associated with lower amyloid PET Centiloids (less brain amyloid deposition) as well as lower CSF biomarker Aβ42 in LEADS (proxy marker suggesting higher brain amyloid deposition) in LEADS; these contradictory findings support the need for larger studies to further investigate whether LOAD PGS is associated with increased amyloid deposition in EOAD. Higher LOAD PGS was also associated with higher levels of CSF synaptosomal‐associated protein 25 (SNAP‐25), a key component of the SNARE complex, suggesting that LOAD genetic factors may contribute to dysregulation of synaptic transmission and/or pathological protein aggregation in EOAD.

BACKGROUND AND OBJECTIVES: Unruptured intracranial aneurysms (UIAs) are increasingly detected and require careful management to prevent rupture. No externally validated score currently predicts procedural risk to guide treatment decisions. We developed and validated 2 predictive scores for complications after endovascular treatment (EVT) or neurosurgical treatment (NT) of UIA using routinely collected clinical and aneurysmal features. METHODS: We conducted a multicenter retrospective study including patients with UIA treated with EVT or NT across 15 neurovascular centers (2014-2024). Predictive models were built using multivariable logistic regression, with variables derived from Delphi consensus. The primary outcome was a composite safety end point: new neurological deficits, modified Rankin Scale (mRS) worsening (≥1 point or mRS 2-5 within 30 days), or procedural death. Internal validation used bootstrapping, and external validation was performed temporally and institutionally. Model performance was assessed using area under the receiver operating characteristic curve (AUROC) and calibration. Final scores, named Morbidity and Mortality Associated Risk in the Treatment of UIAs (MARTA)-EVT and MARTA-NT, were compared with existing models identified through systematic review. RESULTS: Among 2647 patients (1907 EVT and 740 NT), procedural complications occurred in 6.3% (EVT) and 12.8% (NT). Independent predictors included age, baseline mRS, aneurysm location, size, morphology, and procedural factors. MARTA-EVT (AUROC = 0.68, 95% CI = 0.57-0.78) and MARTA-NT (AUROC = 0.65, 95% CI = 0.54-0.77) showed moderate discrimination and good calibration. MARTA-EVT outperformed existing models; MARTA-NT performed similarly to SAFETEA. Predictive models are available open-source: https://martascoreapp.shinyapps.io/martascoreapp/. CONCLUSION: MARTA-EVT and MARTA-NT are validated tools for predicting procedural risks in UIA treatment and may support patient counseling and clinical decision making.

Abstract Objectives To comprehensively evaluate the validity of International Classification of Diseases, Tenth Revision, Clinical Modification (ICD-10-CM) codes for both prevalent diagnoses and less common diseases, and to assess the performance of a large language model (LLM)-based system in validating these codes. Materials and Methods This retrospective study analyzed hospital admissions from the Medical Information Mart for Intensive Care IV (MIMIC-IV) database. We developed a validated LLM-based system using GPT-4o, refined through iterative prompt engineering, to assess ICD-10-CM code validity. We measured the positive predictive value (PPV) of ICD-10-CM codes, PPV of principal and secondary diagnoses, and the performance of an LLM-based system in code validation. Results Among 865 079 assigned codes, the PPV was 84.6% (95% CI, 84.5%-84.6%). Principal diagnoses had a PPV of 93.9% (95% CI, 93.7%-94.1%), while secondary diagnoses had a PPV of 83.8% (95% CI, 83.7%-83.9%). The LLM system demonstrated high performance in validating ICD codes, achieving 93.6% accuracy, 95.4% sensitivity, and 85.2% specificity. Among correctly assigned secondary diagnoses, the majority (67.9%) represented historical or baseline conditions, while 32.1% reflected active conditions that deviated from baseline status; 22.3% of these emerged after hospital admission. PPV decreases with later diagnosis positions, with the largest decline occurring between principal and secondary diagnoses. Discussion and Conclusion In this large-scale evaluation, ICD-10-CM codes exhibited generally high accuracy, though variability existed by position and condition type. A validated LLM system performed comparably to physician review and offers a scalable means to improve coding accuracy. These findings support the potential for integrating LLM-based auditing into routine workflows to strengthen the quality of administrative and research data.

Abstract Aims Heart failure (HF) continues to be a major cause of morbidity and mortality worldwide, placing a significant burden on healthcare systems. Differences in the outcomes of HF hospitalizations for adults in teaching vs. non-teaching hospitals in urban settings are uncertain. We evaluated outcomes of HF hospitalizations in patients hospitalized in urban teaching vs. non-teaching hospitals in the United States. Methods and results HF hospitalizations were abstracted from the 2016 through 2022 Nationwide Readmissions Database and stratified into urban teaching vs. urban non-teaching hospital settings. Propensity-score matching was used to control for baseline differences between teaching and non-teaching hospital cohorts. Logistic regression and lognormal models were estimated to assess differences in inpatient mortality, length of stay (LOS), total costs, complications, and 30-day and 90-day all-cause readmissions. A total of 7 558 299 weighted HF hospitalizations were included in the analysis, of which 76.3% involved urban teaching hospitalizations. Compared to urban non-teaching, HF hospitalizations in urban teaching hospitals were associated with significantly higher odds of inpatient mortality (OR 1.19, 95% confidence interval [CI]: 1.16–1.22), complications including cardiogenic shock, cardiac arrest, Intra-aortic balloon pump, cardiopulmonary resuscitation, extracorporeal membrane oxygenation and mechanical ventilation use as well as more extended hospital LOS, higher total costs, palliative care consultation and readmissions rates (all P < 0.001). 30-day and 90-day all-cause readmission rates were also statistically higher in teaching hospitals, though the difference was clinically non-significant. Conclusion HF hospitalizations in urban teaching hospitals were associated with higher inpatient mortality, complication rates, resource utilization, and palliative care consult rates than urban non-teaching hospitals.

Abstract Motivation: Epilepsy is a diverse group of neurological disorders affecting over 50 million people worldwide. While common epilepsy types are well studied, rare epilepsies—often severe and genetically complex—pose significant challenges in diagnosis, research, and treatment. Accurate and interoperable etiology and disease classifications are critical for improving data sharing, supporting clinical decision-making, and advancing rare disease research. Results: To enhance the accuracy of epilepsy-related disease concept representation within the Mondo Disease Ontology (Mondo), we conducted a series of expert-driven workshops in collaboration with the team from the Rare Disease Cures Accelerator-Data and Analytics Platform (RDCA-DAP). Specialists in epileptology, genetics, neurodevelopment, biomedical ontology, and patient community advocates systematically reviewed and revised the epilepsy hierarchy in Mondo, aligning it with the International League Against Epilepsy (ILAE) classification system. These updates include reclassification of epilepsy subtypes, including syndromes, age-related epilepsies, and developmental epileptic encephalopathies, resulting in a more granular, standardized, and clinically relevant structure. Mondo now offers an enhanced framework for integrating epilepsy data across resources, enabling improved interoperability and facilitating rare disease research and data curation, with continued efforts underway to further refine and expand this integration.

Abstract Patent foramen ovale (PFO) has a prevalence of approximately 25%. Its association with pyogenic brain abscess (PBA) is unclear. We reviewed adults with PBA and known PFO status from 1 January 2009 through 31 December 2021 at Mayo Clinic. High PFO prevalence (40.5%) was seen. PBA patients with PFO had distinct abscess characteristics but unchanged 1-year, all-cause mortality.

Sharing research code in an open access version-controlled repository offers significant benefits for both science as a whole and for individual researchers. In this article, we focus on this practice, which is fully aligned with the NIH’s Gold Standard Science (GSS) program as well as FAIR (findable, accessible, interoperable, reusable) and TRUST (transparency, responsibility, user focus, sustainability, technology) principles. Gold Standard Science supports open science by emphasizing transparency, reproducibility, and the use of best practices that enable others to verify and extend research. Pairing a research article’s cited data snapshot with a versioned, environment-specific code release, deposited in a companion code repository, ensures that, upon submission to a medical journal, readers and reviewers can directly verify results. An executable and updatable companion code repository complements, rather than replaces, established research data repositories. When code underlying medical research results is made openly available, then other scientists can inspect, run, and validate analyses. These activities enhance reproducibility, which is a core aim of GSS. Shared code also facilitates collaborative innovation by allowing researchers to extend the utility of the code to new datasets and applications. For researchers, code sharing can increase visibility, credibility, and citation impact. Demonstrating transparency through shared executable and updatable code builds trust with journal readers, peer reviewers, funders, and peers. Shared code in an open access repository signals adherence to high standards of scientific integrity and attracts opportunities for collaboration. A researcher who shares code receives recognition as a leader in reproducible, trustworthy research consistent with NIH’s GSS principles.

Background/Objectives: Dizziness is a symptom of many disorders across a wide range of etiologies. If dizzy patients are seen for vestibular evaluation with an audiologist and no vestibular reason for the patient’s dizziness is found, the medical referral pathway can become convoluted. This can leave patients feeling discouraged and unable to manage their symptoms. Clinically symptomatic chronic respiratory alkalosis (CSCRA) is an acid–base disorder that typically presents with dizziness but is unfamiliar to practitioners in vestibular and balance care settings. Methods: In a retrospective chart review deemed exempt by the Mayo Clinic Institutional Review Board, 74 patients at Mayo Clinic Arizona were included. All had consultations with both Audiology and Aerospace Medicine to assess their dizzy symptoms. Results: After completing vestibular testing, arterial blood gas (ABG) testing, and a functional test developed at Mayo Clinic Arizona called the Capnic Challenge test, 40% of patients were found to have CSCRA contributing to their dizzy symptoms. Many of these patients also had common comorbidities of CSCRA, like postural orthostatic tachycardia syndrome (POTS), migraines, and sleep apnea. Fewer than one-fourth of these patients had measurable vestibulopathies causing their dizziness. Half of the patients referred by the vestibular audiologist to Aerospace Medicine had a diagnosis of CSCRA. Conclusions: Assessment for CSCRA should be considered as a next step for patients presenting with dizziness without a vestibular component. Being aware of the prevalence of CSCRA and its comorbidities may help balance providers offer quality interprofessional referrals and improve patient quality of life.

Background: ALG13-CDG is an X-linked N-linked glycosylation disorder caused by pathogenic variants in the glycosyltransferase ALG13, leading to severe neurological manifestations. Despite the clear CNS involvement, the impact of ALG13 dysfunction on human brain glycosylation and neurodevelopment remains unknown. We hypothesize that ALG13-CDG causes brain-specific hypoglycosylation that disrupts neurodevelopmental pathways and contributes directly to cortical network dysfunction. Methods: We generated iPSC-derived human cortical organoids (hCOs) from individuals with ALG13-CDG to define the impact of hypoglycosylation on cortical development and function. Electrophysiological activity was assessed using MEA recordings and integrated with multiomic profiling, including scRNA-seq, proteomics, glycoproteomics, N-glycan imaging, lipidomics, and metabolomics. X-inactivation status was evaluated in both iPSCs and hCOs. Results: ALG13-CDG hCOs showed reduced glycosylation of proteins involved in ECM organization, neuronal migration, lipid metabolism, calcium homeostasis, and neuronal excitability. These pathway disruptions were supported by proteomic and scRNA-seq data and included altered intercellular communication. Trajectory analyses revealed mistimed neuronal maturation with early inhibitory and delayed excitatory development, indicating an E/I imbalance. MEA recordings demonstrated early network hypoactivity with reduced firing rates, immature burst structure, and shortened axonal projections, while transcriptomic and proteomic signatures suggested emerging hyperexcitability. Altered lipid and GlcNAc metabolism, along with skewed X-inactivation, were also observed. Conclusions: Our study reveals that ALG13-CDG is a disorder of brain-specific hypoglycosylation that disrupts key neurodevelopmental pathways and destabilizes cortical network function. Through integrated multiomic and functional analyses, we identify early network hypoactivity, mistimed neuronal maturation, and evolving E/I imbalance that progresses to compensatory hyperexcitability, providing a mechanistic basis for seizure vulnerability. These findings redefine ALG13-CDG as disorders of cortical network instability, offering a new framework for targeted therapeutic intervention.