Citations

Background Bicuspid aortic valve (BAV) is the most common congenital heart defect, associated with a lifetime morbidity burden exceeding 80% and an estimated 7% probability of transmission to first-degree relatives. Hence, echocardiographic screening of relatives is recommended. The aim of our study is to develop an artificial intelligence-enabled electrocardiogram (AI-ECG) model for the detection of BAV. Methods Between January 1, 1990, and June 30, 2023, the Mayo Clinic healthcare system database was searched for patients aged ≥18 years with echocardiograms performed for family history of BAV, abnormal auscultation, a BAV-related diagnosis, or to rule out the condition, and with an electrocardiogram (ECG) obtained within 6 months. Cases were patients with a confirmed BAV diagnosis, and controls were those with normal aortic valves. ECGs were randomly assigned into training (80%), internal validation (10%), and test (10%) groups. The AI-ECG model was built using Keras under TensorFlow. Results A total of 13,065 valid ECG–echocardiogram pairs were included (mean [SD] age, 58 [17] years; 44% women), with a median time of 1 day (interquartile range, 0–3 days) between studies. With a 45% BAV prevalence in the test group ( n = 1301), the area under the curve was 0.704, positive predictive value (PPV) 60.1%, and negative predictive value (NPV) 69.4%. At a prevalence of 7%, the NPV increased to 96.1%, but the PPV decreased to 12.2%. Conclusions In this retrospective cohort and using the current AI-ECG model, BAV detection showed modest diagnostic performance and limited standalone clinical utility, suggesting that ECG-only screening for BAV may require additional modalities or more refined phenotyping.

Background Diastolic dysfunction is common in patients with aortic stenosis and may influence outcomes following surgical aortic valve replacement. We aimed to examine the association of preoperative artificial intelligence (AI)‐generated diastolic function grades with early and late outcomes following aortic valve replacement and how postoperative progression influence prognosis. Methods We identified 5503 patients undergoing aortic valve replacement between 2000 and 2023. Diastolic function was assessed using a validated deep‐learning AI model applied to 12‐lead ECGs done preoperatively and on postoperative follow‐up. Diastolic grades were classified by AI into Grades 1 to 3. Longitudinal trend analyses and multivariable regression models were used to assess study end points. Results Among 5503 patients (mean age 72.4±10.8 years; 39% female), higher AI ECG diastolic grades were associated with greater comorbidity burden, including diabetes, renal disease, and heart failure. AI ECG diastolic Grade 3 was independently associated with higher in‐hospital mortality (odds ratio, 2.5; P =0.007) and other complications. At 5‐year follow‐up, patients with Grade 3 showed the least improvement in diastolic function by both ECG and echocardiography. Grades 2 and 3 diastolic function at baseline were independently associated with increased late mortality (hazard ratio, 1.3 and 2.45, respectively; both P <0.001). Additionally, lack of improvement in AI ECG diastolic grade by 1 year was also independently associated with late mortality. Conclusions AI ECG–derived diastolic function grades strongly correlates with early complications and long‐term mortality and diastolic progression after aortic valve replacement. AI ECG provides a powerful, noninvasive tool for risk stratification and longitudinal monitoring of patients with aortic stenosis.

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.

Background Pericarditis may be associated with significant morbidity and mortality and can be a clinically challenging entity to diagnose. There is a clinical need for automated artificial intelligence (AI) algorithms to act as a screening aid to assist with the early and accurate detection of pericarditis. Methods We developed p-Cal, a browser-based pericarditis risk calculator that fuses the patient clinical data and 12-lead ECG image to estimate risk for pericarditis. The ECG encoder was designed based on a pretrained MedCLIP Vision Transformer and random forest classifier was used for the clinical tabular data. These two modalities were fused at the decision-level using a meta classifier. Model reasoning for both modalities was assessed using advanced explainable AI techniques—attention maps from the transformer model were used to interpret ECG images, while SHapley Additive exPlanations (SHAP) force plots were applied to analyse the tabular data. Results A total of 6508 patients (mean age 54.8±16.1 years, 50.9% female) were used for training and internal validation of the AI model, of whom 902 patients had a confirmed diagnosis of pericarditis (13.8%) and the remainder were normal controls. On 1302 hold-out test patients, the Fusion Model demonstrated the best discriminatory performance, achieving an area under the curve (AUC) of 0.89, which surpassed the ECG (AUC = 0.85) and the clinical data only models (AUC = 0.74) performance. On external validation in the Medical Information Mart for Intensive Care (MIMIC) Database, the fusion model achieved an AUC of 0.81. Conclusion A late fusion AI model integrating tabular clinical electronic medical record data and ECG signal has robust predictive capability for pericarditis as a screening tool.

ABSTRACT Many real‐world patients with advanced biliary tract cancer (BTC) are excluded by TOPAZ‐1 criteria; this study assessed cisplatin and gemcitabine plus durvalumab (CGD) efficacy and safety in these patients. Data from 1358 patients with advanced BTC treated with first‐line CGD across 55 international centers were retrospectively analyzed. Patients were classified as TOPAZ‐1‐in (meeting all inclusion criteria) or TOPAZ‐1‐out (meeting ≥ 1 exclusion criterion). Primary endpoints were overall survival (OS) and progression‐free survival (PFS); secondary endpoints included objective response rate (ORR), disease control rate (DCR), and safety. Furthermore, OS and PFS of the TOPAZ‐1‐out cohort were compared with reconstructed survival data from the phase III trial. 912 (67.1%) patients were classified as TOPAZ‐1‐in , and 446 (32.9%) as TOPAZ‐1‐out . After a median follow‐up of 14.5 months (95% CI: 13.7–36.5), OS was 16.1 months in TOPAZ‐1‐in and 12.5 months in TOPAZ‐1‐out (HR 0.69, 95% CI: 14.6–16.5, p = 0.0001); PFS was 8.2 versus 6.5 months (HR 0.73, 95% CI: 7.1–8.1, p < 0.0001). Median OS in the TOPAZ‐1‐out cohort (12.5 months) was nearly identical to the phase III trial experimental arm (12.9 months; HR 1.15, p = 0.13); for PFS the HR was 1.12 (95% CI 0.93–1.42; p = 0.09). Among TOPAZ‐1‐out patients, active infection, elevated bilirubin, and ECOG PS > 1 were linked to poorer OS, while no detrimental impact emerged for ALT/AST abnormalities, corticosteroid use, renal or hematologic parameters, or prior surgery within 6 months, suggesting treatment effectiveness was broadly maintained across clinical subgroups. Furthermore, TOPAZ‐1‐out patients experienced no significant increase in adverse events compared with the TOPAZ‐1‐in group. Real‐world data suggest CGD may be effective in TOPAZ‐1‐out patients, with no evident increase in toxicity, supporting the potential expanded use of the regimen in clinical practice.

ABSTRACT Surgery after response to first‐line chemoimmunotherapy may offer a potential curative option for patients with locally advanced biliary tract cancer (BTC) initially considered unresectable. However, robust real‐world evidence in this setting remains limited. We retrospectively analyzed patients with locally advanced BTC treated with first‐line cisplatin, gemcitabine, and durvalumab (CGD) across 55 centers in 12 countries. Endpoints included overall survival (OS), progression‐free survival (PFS), disease‐free survival (DFS) among resected patients, objective response rate (ORR), and the prognostic impact of surgery. A multivariable logistic regression model was developed to predict surgical conversion using baseline clinical and laboratory variables. Among 1358 screened patients, 219 had locally advanced disease and were included in the analysis. Median follow‐up was 14.5 months. ORR was 34.6%, with median PFS and OS of 10.0 and 20.7 months, respectively. Twenty‐four patients (10.9%) underwent surgical resection after systemic therapy. Median OS was 22.5 months in resected patients versus 19.9 months in those who did not undergo surgery. Median DFS after resection was 15.8 months. Pathological lymph node involvement was independently associated with shorter DFS. Larger baseline tumor size correlated with higher odds of radiological response but not with OS. An exploratory elastic‐net model retained four baseline variables and showed moderate discriminative ability for surgical conversion, with an apparent AUC of 0.72 and an optimism‐corrected AUC of 0.65. First‐line CGD enabled secondary resection in approximately 11% of patients with locally advanced BTC. A simple baseline model may support patient selection for conversion surgery, although prospective validation is needed.

Background Randomized trials have shown equivocal outcomes between a direct aspiration as first-pass technique (ADAPT) and stent retriever (SR) stroke thrombectomy, yet whether specific patient subgroups benefit differentially from one technique over the other remains unknown. Methods In this retrospective multicenter cohort study, patients with large vessel occlusion were included, and those who underwent ADAPT versus SR first-pass thrombectomy (either as standalone or combined with aspiration) were compared. The primary outcome was a composite of first-pass effect (FPE, modified Thrombolysis In Cerebral Infarction (mTICI) 2c or better with one attempt) with no symptomatic intracranial hemorrhage (sICH). A causal forest model was used to estimate the conditional average treatment effect (CATE) for each patient, with larger CATEs indicating that ADAPT is expected to be more advantageous. Patients were stratified by CATEs, and outcomes were compared between treatment groups. Results Among 6700 included patients, ADAPT (n=3695, 55%) was associated with a higher rate of FPE with no sICH (39.9% vs 26.9%; adjusted OR (aOR) 2.07, 95% CI 1.77 to 2.41, P<0.001). The causal forest identified significant treatment effect heterogeneity (interaction P<0.001), where ADAPT was more advantageous in more recent years and among patients with M1 occlusions presenting in earlier time windows. Low-CATE patients showed no statistically significant difference between techniques (aOR 0.91, 95% CI 0.57 to 1.45, P=0.70), while medium- and high-CATE patients demonstrated progressively larger ADAPT advantages (aOR 2.06, 95% CI 1.67 to 2.55, and aOR 2.99, 95% CI 2.24 to 3.98, respectively, both P<0.001). Conclusions ADAPT may be superior to SR for a majority of patients, particularly in recent years, and among patients with M1 occlusions presenting in earlier time windows.

ABSTRACT Aging reshapes the architecture of human skeletal muscle, yet objective tissue‐level markers that capture this process remain limited. We combined large‐scale histology with deep learning to identify reproducible features of muscle aging and to test their biological relevance. We analyzed 974 hematoxylin–eosin whole‐slide images from a population resource using a dual‐attention convolutional neural network and an independent Mask R‐CNN model to quantify nuclear size and density, verified by manual review. The classifier distinguished young from aged muscle with high accuracy (AUC 0.91; accuracy 86.2%), and attention maps consistently highlighted nuclear enlargement and spatial disorganization as salient features. Nuclear diameter increased with age (Spearman's ρ = 0.71, p < 0.0001) across automated and manual measurements. Transcriptomes matched to the same donors showed that samples with larger nuclei were enriched for pathways related to chromatin remodeling, proteostasis, cellular senescence, mitochondrial activity, and telomere regulation, whereas smaller nuclei aligned with anti‐inflammatory and DNA repair programs. External pediatric inflammatory myopathies exhibited nuclear enlargement comparable to aged muscle, suggesting inflammation‐related premature histologic aging. These findings identify nuclear enlargement as a robust, quantifiable feature that integrates structural and molecular signatures of muscle aging. The proposed deep learning–based nuclear morphometry provides a scalable framework for tissue‐level aging biomarkers and suggests a potential “muscle aging clock” applicable to both physiological aging and disease states.