Citations

The purpose of this study was to examine the validity of a Garmin device for estimating heart rate, energy expenditure, and step count during incremental treadmill exercise. Eighteen physically active males (n = 12; mean ± SD age = 22.8 ± 4.7 years; body mass = 83.8 ± 13.4 kg) and females (n = 6; 24.2 ± 5.6 years; 65.2 ± 9.6 kg) completed an incremental treadmill test with the Garmin Vivoactive 4 and criterion methods measuring heart rate, energy expenditure, and step count. Mean absolute percent error (MAPE) and Bland–Altman plots were used to assess accuracy. Acceptable accuracy was defined as MAPE < 5% for heart rate and <10% for energy expenditure and step count. MAPE (±SD) values were 13.0 (±10.1), 19.1 (±15.0), and 4.6 (±5.3)% for heart rate, energy expenditure, and step count, respectively. The Bland–Altman regression analyses illustrated proportional bias was present for heart rate (r = 0.591, p < 0.001) and step count (r = 0.516, p = 0.028), but not energy expenditure (r = 0.351, p = 0.153). These findings indicated that the Garmin Vivoactive 4 provided acceptable accuracy for step count but unacceptable accuracy for heart rate and energy expenditure during incremental treadmill exercise.

BACKGROUND: The Commission on Cancer (CoC) implemented 6 operative standards to reduce technical variation in cancer surgery. Although designed to guide surgeons in performing key oncologic techniques, these standards also provide a framework for teaching high-quality cancer surgery to trainees. We assessed surgical trainees’ familiarity with the CoC Operative Standards and the underlying principles of cancer surgery to identify knowledge gaps and opportunities for improved education. STUDY DESIGN: The American College of Surgeons Cancer Surgery Standards Program distributed an anonymous survey to general surgery residents and surgical fellows across the US. Questions addressed the CoC Operative Standards and cancer surgery principles, with scores calculated on the basis of correctness. Univariate logistic regression evaluated associations between trainee or program characteristics and knowledge. Trainees were also asked about their role in procedural documentation, given the synoptic operative report requirements of the standards. RESULTS: A total of 244 surveys were completed by 205 residents and 39 fellows. Fellows reported greater familiarity with the CoC Operative Standards (68% vs 24%, p < 0.001). Only 30% of trainees reported receiving a formal curriculum on the standards. Correct response rates were 30% for CoC Operative Standards questions vs 50% for cancer surgery principles, with minimal association between knowledge and trainee or program characteristics. Most residents (71%) reported contributing to operative documentation. CONCLUSIONS: Surgical trainees demonstrated limited knowledge of the CoC Operative Standards, highlighting the need for multifaceted educational strategies at the national and institutional levels. Incorporating education about the CoC Operative Standards into surgical training can help emphasize the importance of performing and documenting critical elements of cancer operations.

This joint Clinical Consensus Statement on the safety and feasibility of concomitant urogynecologic and gynecologic oncology surgical procedures reflects statements drafted by content experts from the American Urogynecologic Society and the Society of Gynecologic Oncology. The writing group used a Delphi process to evaluate statements developed from a structured literature search and assessed consensus. A total of 35 statements were assessed in the following categories: (1) background, (2) treatment options, (3) timing of concurrent surgery, (4) risk factors, and (5) value-based care. Of the 35 statements, 28 reached consensus on initial review. Three statements were omitted, and 4 statements were revised and reached consensus following second review utilizing the Delphi process.

Integrating artificial intelligence (AI) with healthcare data is rapidly transforming medical diagnostics and driving progress toward precision medicine. However, effectively leveraging multimodal data, particularly digital pathology whole slide images (WSIs) and genomic sequencing, remains a significant challenge due to the intrinsic heterogeneity of these modalities and the need for scalable and interpretable frameworks. Existing diagnostic models typically operate on unimodal data, overlooking critical cross-modal interactions that can yield richer clinical insights. We introduce MarbliX (Multimodal Association and Retrieval with Binary Latent Indexed matriX), a self-supervised framework that learns to embed WSIs and immunogenomic profiles into compact, scalable binary codes, termed “monogram.” By optimizing a triplet contrastive objective across modalities, MarbliX captures high-resolution patient similarity in a unified latent space, enabling efficient retrieval of clinically relevant cases and facilitating case-based reasoning. In lung cancer, MarbliX achieves 85%–89% across all evaluation metrics, outperforming histopathology (69%–71%) and immunogenomics (73%–76%). In kidney cancer, real-valued monograms yield the strongest performance (F1: 80%–83%, Accuracy: 87%–90%), with binary monograms slightly lower (F1: 78%–82%).

Background While delivery of supplemental oxygen is a life-saving therapy, exposure to high oxygen, called hyperoxia, leads to increased intensive care unit mortality. Hyperoxia induces oxidant-mediated acute lung injury (ALI) and pulmonary cell death, called hyperoxic ALI (HALI). Elucidating molecular mechanisms in HALI could identify therapeutic targets in ALI. Methods In the current study, we examined in vivo effects of HALI on Beclin-1 (BCN1), which regulates autophagy, and modulation of BCN1 by epidermal growth factor receptor (EGFR). Effects of HALI on BCN1 and autophagy were examined in mice with genetically decreased EGFR (EGFR Wa5/+ ). Wildtype (WT) and EGFR Wa5/+ mice were exposed to 100% oxygen for 24–72 hours along with normoxia controls (eight groups; n=4–6/group), and analysis of pulmonary BCN1 and autophagy was completed. Results In WT, HALI led to increased BCN1 (59% increased total BCN1/β-Actin; p<0.01) in lung and alveolar epithelium (484% increased H-score; p<0.001). HALI led to decreased microtubule-associated protein 1B-light chain (LC3B)-II/-I ratios (43% decrease; p<0.05) and increased p62 (93% increase; p<0.05), suggesting reduced autophagic flux. In human alveolar type-II cells derived from induced pluripotent stem cells (AT2s iPSC ), HALI caused increased LDH release (130% increase; p<0.0001) and decreased LC3B-II/-I ratios (32% decrease; p<0.05). We previously showed that EGFR Wa5/+ mice are protected in HALI. In the current study, EGFR Wa5/+ mice showed increased pulmonary BCN1 (122% increase; p<0.05), reduced phosphorylated-(p-)/total BCN1 (47% decrease; p<0.05) and decreased LC3B-II/-I ratios (70% decrease; p<0.01) in HALI compared with WT. In addition, wortmannin (1 mg/kg), which decreases BCN1-mediated autophagy, increased mortality in HALI compared with vehicle control (n=10/group; 18% decreased hours survived; p<0.001). Conclusions These data delineate a novel cell death pathway in HALI involving BCN1 and EGFR with therapeutic potential.

INTRODUCTION: A Model for End-Stage Liver Disease-lactate (MELD-LA) score is an independent predictor of inpatient mortality and organ failure validated in more than 8,600 patients across diverse settings. However, its utility in sepsis, particularly when combined with lactate clearance (LA clearance), has not been fully evaluated. In this study, we examined the dynamic role of MELD-LA with 24-hour LA clearance as a predictor of 30-day mortality in critically ill patients with cirrhosis and sepsis. METHODS: We analyzed cirrhosis patients with sepsis admitted to the intensive care unit (ICU) in a large US healthcare system (20 hospitals) from 2014 to 2022, examining MELD-LA and LA clearance within 24 hours as predictors of 30-day mortality. We externally validated our findings in the Asian Pacific Association for the Study of the Liver-acute-on-chronic liver failure Research Consortium, a multinational Asian cohort (31 centers) from 2009 to 2019. RESULTS: A total of 3,879 cirrhosis patients with sepsis were admitted to the ICU (43.2% female, 20.5% Hispanic, 12.4% on dialysis; mean initial LA 4.5 mmol/L, MELD 20.8, MELD-LA 17.6). Survivors showed a median lactate change of −1.04 (35% clearance) compared with −0.4 (12%) for nonsurvivors. LA clearance within 24 hours was influenced by the presenting MELD-LA. In the North American cohort, ≥35% LA clearance was associated with improved 30-day survival (odds ratio 0.41, 95% CI 0.31–0.53; P < 0.001). Similarly, in the Asian Pacific Association for the Study of the Liver cohort (n = 1,259), ≥35% clearance was linked to a 60% reduction in mortality (odds ratio 0.42, 95% CI 0.28–0.62; P < 0.001). When stratified by LA clearance, patients with LA clearance of ≥35% had a higher percentage of early administration of resuscitative measures than those with LA clearance of < 10% especially regarding early antibiotic administration (82.5% vs 67.6%; P value <0.001). DISCUSSION: Early reduction in initial LA by one-third predicts 30-day mortality and may indicate treatment response in cirrhosis patients with sepsis in the ICU. Combining MELD-LA and LA clearance early enables better risk stratification and informs management.

INTRODUCTION: To perform a proof-of-concept study to evaluate the performance of individual and combined anorectal manometry and balloon expulsion testing (BET) parameters for diagnosing dyssynergic defecation (DD) using machine learning (ML). METHODS: A single gastrointestinal motility specialist assessed (2008–2018) 307 patients. A ML model was constructed on anorectal manometry/BET data. BET was weight-based in 235 patients and time-based in 72 patients. Missing data were imputed. Data were split 75% train/25% test with preprocessing performed. Five common ML models were trained, and performance was compared using receiver operating characteristic (ROC) area under the curve and accuracy based on 100 bootstrapped samples. The most accurate models were fine-tuned and evaluated on the test set. RESULTS: The median age was 40 years, body mass index 22.2 kg/m 2 , and 81% were women. Two-hundred and 22 (72.3%) patients had a clinical diagnosis of DD. Compared with patients without clinical diagnosis of DD, patients with DD had significantly lower rectoanal pressure gradient (RAG) (median −34.7 [IQR −57.1, −13.0] vs −24.6 [IQR −44.5, −2.4] mm Hg, P = 0.009) and more abnormal BET (59.5% vs 11.8%, P < 0.001). Four features were retained in the optimized model as predictors of DD: abnormal BET, greater resting anal pressure (RAP), and more negative RAG. The optimal parameter was abnormal BET ( P = 0.003), but combinations of 2 manometry results (RAP and RAG) yielded positive predictive values (PPVs) >80%. The optimized logistic regression model had an area under the curve of 0.878, with a sensitivity of 75.0% and specificity of 81.8% at a probability threshold of 0.704. DISCUSSION: Abnormal BET, or a combination of 2 manometric parameters (RAP and RAG), provides >80% positive predictive value for diagnosing DD in patients with chronic constipation.

Abstract Objective Recent studies suggest a unique disease profile in elderly patients with primary aldosteronism (PA), who have a high prevalence of comorbidities and low clinical cure rate post-adrenalectomy. This study aims to investigate histopathology and genetic characteristics of PA in late adulthood. Methods Formalin-fixed, paraffin-embedded adrenal tissue sections from 114 patients with PA who underwent unilateral adrenalectomy at age 60 or older were used. Aldosterone synthase (CYP11B2) immunohistochemistry (IHC) was performed for histopathologic analysis. Somatic DNA variants were identified in aldosterone-producing tissue following CYP11B2 IHC-guided sequencing. Results Histopathologic analysis revealed that 56 adrenals contained an aldosterone-producing nodule (APN) (49%), 44 contained an aldosterone-producing adenoma (APA) (39%), 7 contained multiple APN (6%), and 7 contained only non-functioning adenomas [NFA, of which 6 had at least one aldosterone-producing micronodule (APM) in the adjacent tissue] (6%). NFAs were also identified alongside APN/APA in 24 cases. An APN was the most common histologic subtype in men (51/88, 58%), while APA was predominant in women (18/26, 69%). Of 67 APNs with successful sequencing, the majority had variants in CACNA1D (40/67, 60%). KCNJ5 variants were the most common alteration in APA (14/44, 32%). Conclusions Older individuals with lateralized PA exhibit distinct genetic and histologic causes of disease with sex differences in mutation prevalence and adrenal histopathology. In APN, the most frequently altered gene is CACNA1D, while KCNJ5 variants are most common in APA. Unlike young patients with PA, the presence of NFA is a frequent observation in late adulthood individuals with PA.

Abstract Objectives To develop and validate an agent-based Large Language Model (LLM) system for extracting structured data from breast cancer synoptic pathology reports and assess the performance gap between synthetic and real-world validation. Materials and Methods We developed a modular artificial intelligence (AI) agent-based framework employing sequential specialized LLMs for parsing pathology reports and extracting structured data. We normalized College of American Pathologists (CAP) cancer protocols into 8 sections, 86 subsections, and 229 discrete fields. Seven leading LLMs (gemini-2.5-pro, llama3.3-70b, phi4-14b, deepseek-r1 14B/70B, gemma3-27b, gemini-2.0-flash-lite) were validated using dual evaluation: synthetic validation (864 controlled test cases) and real-world ground truth (6651 annotated fields from 90 pathology reports). Results Synthetic validation demonstrated strong performance (accuracy: 93.8%-99.0%). Real-world evaluation revealed field extraction recall ranging from 61.8% to 87.7%, demonstrating a substantial “reality gap” with performance drops of 11-32 percentage points. The gemini-2.5-pro model achieved the highest real-world recall (87.7%). Model size did not predict performance: the 14B-parameter deepseek-r1 (77.6%) outperformed its 70B-parameter counterpart (70.4%). Discussion The substantial performance degradation from synthetic to real-world data underscores the complexity of authentic clinical documentation. Smaller models can achieve competitive or superior recall, reducing computational costs. With even the best models missing 12%-38% of annotated fields, mandatory human verification is essential for clinical deployment. Conclusion While LLM-based extraction systems show promise for pathology data extraction, synthetic validation alone provides false confidence. Rigorous real-world ground truth evaluation with expert annotation is essential before clinical deployment. These systems are best positioned as screening tools with mandatory human oversight rather than autonomous decision-making systems.

Coronary artery disease (CAD) remains the leading cause of cardiovascular morbidity and mortality worldwide, with plaque composition and morphology being as key determinants of disease progression and clinical outcomes. Accurate plaque characterization is essential for risk stratification and therapeutic decision-making, yet conventional image interpretation is limited by inter-observer variability and time-intensive workflows. Artificial intelligence (AI) models have emerged as a transformative tool for automated coronary plaque analysis across multiple imaging modalities. AI-driven models demonstrate high diagnostic accuracy for plaque detection, segmentation, quantification, and vulnerability assessment. Integration of AI-derived imaging biomarkers with clinical risk scores can further enhance prediction of major adverse cardiovascular events and supports personalized management. These advances position AI-enhanced imaging as a powerful adjunct for both invasive and non-invasive evaluation of CAD. Despite its promise, important barriers to widespread clinical adoption remain, including data heterogeneity, algorithmic bias, limited model transparency, insufficient prospective validation, regulatory challenges, and incomplete integration into clinical workflows. Addressing these challenges will be essential to ensure safe, generalizable, and cost-effective implementation of AI in routine cardiovascular care.