Citations

ABSTRACT Background Family history of pancreatic cancer (PC), pathogenic germline variants (PGVs), and increased intrapancreatic fat (IPF) are individually associated with increased PC risk. Whether IPF is elevated in high‐risk individuals (HRIs) with PC family history or PGVs is unknown. Although MRI is the preferred modality for IPF measurement, region‐of‐interest–based approaches are prone to sampling bias because pancreatic fat is heterogeneously distributed. We aimed to compare IPF among HRIs and matched non‐HRI controls using an Artificial intelligence (AI)‐based tool for whole pancreatic IPF measurement applied to magnetic resonance imaging (MRI). Methods We developed and validated a deep‐learning pipeline for IPF measurement that incorporates automated pancreas segmentation on T1‐weighted MRI images registered to the corresponding iterative decomposition of water and fat with echo asymmetry and least squares estimation images to compute IPF using proton density fat fraction (PDFF). This pipeline was used to quantify IPF and assess IPF distribution patterns in 85 HRIs and 170 non‐HRIs matched 1:2 for age, sex, BMI, and diabetes status. A surface erosion approach was developed to minimize the impact of peripancreatic fat on IPF measurement. Linear models adjusting for matching variables were used to assess associations. Results Model‐expert agreement for whole‐pancreas PDFF was excellent (intraclass correlation coefficient 0.99; 95% CI, 0.98–0.99 mean absolute error 1.4%). Mean PDFF did not differ between HRIs and controls (25.8% [11.5] vs. 26.9% [11.4]; p = 0.32). PDFF increased with age ( p = 0.015), male sex ( p < 0.001), presence of diabetes ( p = 0.009), BMI ( p < 0.001). IPF distribution patterns did not differ by risk group. Conclusion An AI‐based approach provides accurate, scalable, and objective assessment of IPF quantity and distribution. IPF quantity and spatial distribution do not appear to differ based on germline or familial predisposition to PC.

Objective Antinuclear antibody (ANA) indirect immunofluorescence (IIF) is the reference standard but depends on subjective pattern interpretation. We tested whether artificial intelligence (AI) applied to ANA IIF images could identify specific autoantibodies. Methods We assembled a cohort of ANA‐positive patients (human epithelial type 2 IIF titer ≥1:80) with anti–double‐stranded DNA (dsDNA), anti‐Smith (Sm) antibody, ribonucleoprotein (RNP), anti‐Sjögren's disease‐related antigen A (SSA), anti‐Sjögren's disease‐related antigen B (SSB), anti‐topoisomerase I (Scl‐70), or centromere antibody (CMA) testing within 90 days. The development cohort (December 12, 2016–December 31, 2022) included the first ANA test per patient. Two AI architectures were trained for each autoantibody. Performance was assessed using area under the receiver operating characteristic curve (AUROC), sensitivity, specificity, and accuracy; comparisons with IIF patterns were summarized as pattern‐antibody agreement. A prospective cohort (January 1, 2023–August 20, 2024), including a dsDNA assay transition, provided temporal validation. Results Among 512,959 patients, 47,890 met inclusion criteria (mean age 55 [SD 17.6], 78% female). Models for CMA, dsDNA, Sm, SSA, and SSB achieved AUROC ≥0.80 (CMA highest: AUROC 0.95, accuracy 94.2%, specificity 95.3%). Compared with associated IIF patterns, dsDNA and Sm models more often classified antibody‐negative cases as negative (dsDNA 88.2% vs 54.7%; Sm 82.6% vs 60.3%) while similarly classifying antibody‐positive cases. Performance for RNP (AUROC 0.68) and Scl‐70 (AUROC 0.59) was lower. In the prospective cohort (n = 21,427), performance was similar; dsDNA discrimination remained stable across assays (AUROC ~0.85). Conclusion AI applied to ANA IIF images identified specific autoantibodies and often improved classification of antibody‐negative cases versus simplified pattern‐based comparators. These findings support its potential to standardize ANA interpretation and triage or reduce confirmatory testing. image

Cardiovascular-kidney-metabolic (CKM) syndrome is a complex disease that affects up to 90% of US adults. This work characterized the renal transcriptomic landscape of a novel model of CKM in aging swine. The cardiovascular, renal, and metabolic parameters in pigs are very similar to those in humans, and our work may guide future studies aimed at modulating specific age-related renal abnormalities in CKM to slow disease progression and reduce mortality.

Study Design: Retrospective. Objective: To examine the long-term perioperative changes in lumbar bone quality, paraspinal muscle size and fat content, and facet degeneration after hemipelvectomy (HP) and hip disarticulation (HD). Summary of Background Data: Emerging evidence shows that transfemoral (TFA) and transtibial (TTA) amputations may accelerate lumbar modic changes and precipitate axial back pain. TFA and TTA amputation studies demonstrate that the severity of postamputation lower back pain is associated with the amputation type (TFA > TTA) and the degree of alteration in lumbopelvic kinematics. Methods: All patients who underwent HP or HD for malignancy or infection at an academic center between 2003 and 2022 were retrospectively identified. Preoperative to postoperative changes in multifidus, erector spinae, quadratus lumborum, and psoas cross-sectional area (CSA), and Hounsfield units (HU) were calculated on CT for the amputated and contralateral side at the level of the L3–4 disc space. Paired univariate analysis was utilized to compare preoperative and postoperative outcomes. Results: A total of 16 patients (8 HP, 8 HD) met the inclusion/exclusion criteria. The preoperative to postoperative changes in HUs in the L1, L3, and L5 vertebral bodies were −27.0 ( P =0.004), −51.5 ( P =0.002), and −48.1 ( P =0.001), respectively. The decrease in L1 HU was greater for prosthesis nonusers (user: −16.50, nonuser: −51.32, P =0.021). L5–S1 segment facet degeneration increased postoperatively ( P =0.030, n=7 progressed). There were significant preoperative to postoperative changes in psoas CSA (preop: 1063.4, postop: 586.5, P =0.005) and psoas HU (preop: 50.95, postop: 33.55, P <0.001) on the amputated but not the contralateral side. Conclusion: Hind-quarter amputations were associated with decreased lumbar bone quality, progression of facet degeneration, and loss of amputated side psoas muscle tone. Daily prosthesis use was moderately protective for lumbar bone quality following amputation. HP but not HD was associated with postoperative loss of amputated side quadratus lumborum tone. Level of Evidence: Level III.

Abstract Cognitive impairment is among the most common non-motor symptoms in parkinsonism and Parkinson’s disease, ranging from subjective complaints to dementia. Although several predictive factors have been described, their prognostic relevance remains uncertain. This study aims to examine the predictive association between clinical features and the risk of dementia and determine the incidence of dementia in a population-based cohort of incident parkinsonism cases. Patients with parkinsonism onset between 1991 and 2020 were identified using the Rochester Epidemiology Project records-linkage system. Subjects with cognitive impairment preceding motor symptoms onset or concomitant Alzheimer’s disease were excluded. Diagnoses of dementia and mild cognitive impairment were established using neuropsychological testing or brief cognitive screening. Development of dementia was assessed using a Cox proportional hazards regression model. Associations between dementia and each risk factor were individually estimated in adjusted Cox models. All risk factors were also included in a least absolute shrinkage and selection operator (LASSO) penalized multivariable Cox model. A total of 1,104 patients were included: 153 (13.9%) developed mild cognitive impairment and 543 (49.2%) developed dementia during follow-up. The cumulative incidence of dementia was 30.3% at 5 years, 48.6% at 10 years, 61.4% at 15 years, and 69.3% at 20 years, indicating that a considerable portion of patients remained free of dementia after a prolonged follow-up. The majority of patients with dementia were males (61.5%), and older age at disease onset was associated with risk of dementia (hazard ratio [HR] = 1.32, P &lt; 0.001 per 5-year increment). In the penalized model, dementia risk was significantly higher among individuals presenting with impaired postural reflexes (HR = 1.26, P = 0.014), consistent with a poorer prognosis in akinetic-rigid-predominant parkinsonism. Eye movement abnormalities including reduced ocular range of motion (HR = 1.86, P &lt; 0.001), and smooth pursuit alterations (HR = 1.34, P = 0.024) were also associated with dementia. Non-motor symptoms, including REM sleep behavior disorder (HR = 1.51, P &lt; 0.001), orthostatic hypotension (HR = 1.30, P = 0.009), and constipation (HR = 1.24, P = 0.022), were linked to dementia risk, supporting a potential involvement of non-dopaminergic and central autonomic pathways in cognitive decline. These findings were largely confirmed in the Parkinson’s disease/Parkinson’s disease with dementia group, in which 77 (13.5%) developed mild cognitive impairment and 249 (43.5%) developed dementia. Among these patients, smooth pursuit alterations, reduced ocular range of motion, orthostatic hypotension, and constipation were associated with dementia.

Importance Timely identification of aortic stenosis (AS) is essential for appropriate clinical management, yet screening remains limited by dependence on comprehensive echocardiography and trained imaging personnel. Objective To develop and validate a deep learning algorithm for detection of moderate or greater AS and prospectively evaluate its performance using artificial intelligence (AI)–guided focused cardiac ultrasound (FoCUS) acquired by novice operators. Design, Setting, and Participants This diagnostic study included retrospective algorithm development and validation and prospective evaluation of AI-guided FoCUS across Mayo Clinic sites in the Midwest, Arizona, and Florida. The model was developed using 6753 patients and evaluated in internal validation (n = 852), internal test (n = 844), and validation (n = 1912) cohorts. Performance was assessed on FoCUS acquired by experienced sonographers (n = 602) and prospectively by novice operators (n = 1302). The retrospective model development and validation cohorts comprised studies performed from January 2005 through September 2022. Prospective study was conducted in 2 enrollment periods from June to August 2024 and from June to September 2025. Participants from both periods were combined to comprise the final prospective cohort. Exposure AI-guided FoCUS acquisition and automated deep learning–based assessment for detection of moderate or greater AS. Main Outcomes and Measures Detection of moderate or greater AS. Performance was assessed using area under the receiver operating characteristic curve (AUROC), sensitivity, specificity, and positive predictive value. Results The model demonstrated excellent discrimination in the internal test cohort (AUROC, 0.99; 95% CI, 0.98-1.00) and geographically distinct validation cohorts in Arizona (AUROC, 0.99; 95% CI, 0.97-1.00) and Florida (AUROC, 0.99; 95% CI, 0.96-1.00). Among FoCUS examinations acquired by experienced sonographers, sensitivity was 95% (95% CI, 82-99) and specificity was 97% (95% CI, 95-98). In the prospective novice-operator cohort, 1258 of 1302 examinations (96.6%) were suitable for automated analysis. Sensitivity was 93% (95% CI, 82-99) and specificity was 96% (95% CI, 95-97). Expert review of AI-positive and uninterpretable examinations increased the positive predictive value from 49.4% to 91.1%, with sensitivity of 85.4%. Conclusions and Relevance A deep learning algorithm accurately detected moderate or greater AS across validation cohorts in this study. In prospective evaluation, novice operators were able to acquire AI-guided FoCUS examinations that enabled accurate detection of moderate or greater AS. These findings support a scalable strategy that may expand access to AS detection in settings with limited echocardiography resources.

Health care organizations are rapidly adopting artificial intelligence (AI)-enabled tools, yet nurse leaders often lack a clear framework to evaluate, govern, and safely scale these technologies. This article presents a practical, system-level AI governance model grounded in risk mitigation, interdisciplinary partnership, and nursing leadership decision rights commonly enacted through nursing professional governance structures. We outline core components of an effective governance structure, including clinical usefulness, safety and reliability, fairness and inclusiveness, transparency and explainability, privacy and security, and organizational accountability. Real-world examples illustrate how nurse executives can guide responsible AI adoption to reduce cognitive load, improve documentation efficiency, and enhance patient outcomes. The framework equips nursing leaders with actionable strategies and practical workflow checklist to lead AI transformation in complex health care environments.