Pub Date : 2026-02-05DOI: 10.1093/eurheartj/ehaf784.2773
G Casas, E Rodenas-Alesina, A Subira, F Gonzalez-Santorum, R Barriales, J Mirelis, E Villacorta, E Zorio, A Teis, G Pontone, M Rabbat, A Varga-Szemes, J Schwitter, I Ferreira-Gonzalez, J F Rodriguez-Palomares
Background Excessive trabeculation of the left ventricle (ETLV) is a controversial entity with inconsistent outcomes. Cardiovascular magnetic resonance (CMR) may help in risk stratification. Purpose To develop a CMR prediction model of major adverse cardiovascular events (MACE) in ETLV. Methods Retrospective longitudinal international study. A total of 589 patients with ETLV from 17 centres across Europe and North America were recruited, 398 (67%) were assigned to the development cohort and 191 (33%) to the validation cohort. Also, 197 patients with DCM were recruited as a control group. Core-lab CMR analysis was performed, including conventional parameters (LVEF, burden of LGE, etc.) and advanced variables (all chambers strain, hemodynamic forces, etc.). MACE was defined as a composite of heart failure, ventricular arrhythmias, systemic embolisms and all-cause death. The prediction model with the highest Harrell’s C was chosen. Candidate CMR variables were categorized and converted into a risk score. Patients were divided according to terciles of score punctuation. Results Among the development cohort, age was 43.4 (18.2) years and 45% were women. LVEF was 48.0 (14.1) and 12% exhibited LGE (Figure 1). During a median follow-up of 2.8 years (IQR 0.99 – 5.55) years, MACE occurred in 77 (19%) patients. Most CMR variables were associated with MACE in univariate analysis (Figure 1). The best prediction model resulted to be a combination of indexed LV end-diastolic volume (iLVEDV), LVEF, LGE >5% (of total myocardial mass), left atrium global longitudinal strain (LA-GLS) and lateral-septal hemodynamic forces (ls-HDF). The risk score had a Harrell’s C of 0.742 (Figure 2a) and an adequate calibration (slope 1.04). Patients in the intermediate-risk group had a HR 4.32 (2.37 – 7.86) for MACE compared with those at low risk, and patients in the high-risk group had a HR 7.49 (4.34 – 12.92) (Figure 2b). These results were replicated in the validation cohort (Harrell’s C 0.716), that also displayed a good calibration. However, the performance of the model was only modest when applied to a DCM control cohort (Harrell’s C 0.599). Conclusions We developed and validated a CMR-based risk score for precise stratification in patients with excessive trabeculation of the LV. The modest results in a control DCM group suggest a differential phenotype. Our results could be used for individualised management.Baseline CMR variables ROC and Kaplan Meier curves for MACE
背景：左心室过度小梁（ETLV）是一个有争议的实体，其结果不一致。心血管磁共振（CMR）可能有助于风险分层。目的建立ETLV患者主要不良心血管事件（MACE）的CMR预测模型。方法国际纵向回顾性研究。共招募了来自欧洲和北美17个中心的589例ETLV患者，其中398例（67%）被分配到发展队列，191例（33%）被分配到验证队列。同时，招募了197例DCM患者作为对照组。进行Core-lab CMR分析，包括常规参数（LVEF、LGE负荷等）和高级变量（各腔室应变、血流动力学力等）。MACE被定义为心力衰竭、室性心律失常、全身栓塞和全因死亡的综合症状。选择Harrell’s C值最高的预测模型。候选CMR变量被分类并转换为风险评分。根据评分标点符号的顺序对患者进行分组。结果在发展队列中，年龄为43.4（18.2）岁，45%为女性。LVEF为48.0(14.1)，12%表现为LGE（图1）。在中位随访2.8年（IQR 0.99 - 5.55）年期间，77例（19%）患者发生MACE。在单变量分析中，大多数CMR变量与MACE相关（图1）。结果表明，最佳预测模型是指数左室舒张末期容积（iLVEDV）、LVEF、LGE >；5%（心肌总质量），左心房总纵向应变（LA-GLS）和侧间隔血流动力学力（ls-HDF）。风险评分的Harrell’s C为0.742（图2a），并进行了适当的校准（斜率为1.04）。中危组患者与低危组相比，MACE的HR为4.32(2.37 - 7.86)，高危组患者的HR为7.49(4.34 - 12.92)（图2b）。这些结果在验证队列中被复制（Harrell’s C 0.716），也显示出良好的校准。然而，当应用于DCM对照队列时，该模型的性能仅为适度（Harrell 's C 0.599）。我们开发并验证了一种基于cmr的风险评分，用于左室过度小梁患者的精确分层。对照组DCM的适度结果表明存在差异表型。我们的结果可用于个性化管理。基线CMR变量ROC和Kaplan Meier曲线为MACE
{"title":"Development and validation of a new CMR risk prediction score in excessive trabeculation of the left ventricle","authors":"G Casas, E Rodenas-Alesina, A Subira, F Gonzalez-Santorum, R Barriales, J Mirelis, E Villacorta, E Zorio, A Teis, G Pontone, M Rabbat, A Varga-Szemes, J Schwitter, I Ferreira-Gonzalez, J F Rodriguez-Palomares","doi":"10.1093/eurheartj/ehaf784.2773","DOIUrl":"https://doi.org/10.1093/eurheartj/ehaf784.2773","url":null,"abstract":"Background Excessive trabeculation of the left ventricle (ETLV) is a controversial entity with inconsistent outcomes. Cardiovascular magnetic resonance (CMR) may help in risk stratification. Purpose To develop a CMR prediction model of major adverse cardiovascular events (MACE) in ETLV. Methods Retrospective longitudinal international study. A total of 589 patients with ETLV from 17 centres across Europe and North America were recruited, 398 (67%) were assigned to the development cohort and 191 (33%) to the validation cohort. Also, 197 patients with DCM were recruited as a control group. Core-lab CMR analysis was performed, including conventional parameters (LVEF, burden of LGE, etc.) and advanced variables (all chambers strain, hemodynamic forces, etc.). MACE was defined as a composite of heart failure, ventricular arrhythmias, systemic embolisms and all-cause death. The prediction model with the highest Harrell’s C was chosen. Candidate CMR variables were categorized and converted into a risk score. Patients were divided according to terciles of score punctuation. Results Among the development cohort, age was 43.4 (18.2) years and 45% were women. LVEF was 48.0 (14.1) and 12% exhibited LGE (Figure 1). During a median follow-up of 2.8 years (IQR 0.99 – 5.55) years, MACE occurred in 77 (19%) patients. Most CMR variables were associated with MACE in univariate analysis (Figure 1). The best prediction model resulted to be a combination of indexed LV end-diastolic volume (iLVEDV), LVEF, LGE >5% (of total myocardial mass), left atrium global longitudinal strain (LA-GLS) and lateral-septal hemodynamic forces (ls-HDF). The risk score had a Harrell’s C of 0.742 (Figure 2a) and an adequate calibration (slope 1.04). Patients in the intermediate-risk group had a HR 4.32 (2.37 – 7.86) for MACE compared with those at low risk, and patients in the high-risk group had a HR 7.49 (4.34 – 12.92) (Figure 2b). These results were replicated in the validation cohort (Harrell’s C 0.716), that also displayed a good calibration. However, the performance of the model was only modest when applied to a DCM control cohort (Harrell’s C 0.599). Conclusions We developed and validated a CMR-based risk score for precise stratification in patients with excessive trabeculation of the LV. The modest results in a control DCM group suggest a differential phenotype. Our results could be used for individualised management.Baseline CMR variables ROC and Kaplan Meier curves for MACE","PeriodicalId":11976,"journal":{"name":"European Heart Journal","volume":"7 1","pages":""},"PeriodicalIF":39.3,"publicationDate":"2026-02-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146121967","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2026-02-05DOI: 10.1093/eurheartj/ehaf784.3953
J J N Daems, M A Van Diepen, R Lagerweij, F E Van Leusden, D Kramarenko, E Poel, M H Moen, A J Nederveen, A Van Randen, N B Bijsterveld, M Van Den Boogaard, Y M Pinto, F W Asselbergs, A S Amin, H T Jorstad
Background Increased left ventricular (LV) wall thickness (LVWT) can be a hallmark of both an athlete’s heart and early HCM, making it challenging to distinguish exercise-induced cardiac remodelling (EICR) from early-stage HCM. Navigating this ‘grey zone’ remains a key challenge in sports cardiology. Purpose To determine cardiac magnetic resonance imaging (CMR) parameters differentiating physiological EICR from early-stage genotype positive HCM. Methods We conducted a cross-sectional analysis of genotype-positive borderline HCM patients (LVWT: women 9-16mm, men 11-16mm), and matched elite athletes in a 1:1 ratio. Elite athletes were propensity score distance matched to HCM patients based on sex, length and LVWT. Athlete were selected from ELITE, a prospective elite athlete cohorts which collects data from standardized cardiovascular screenings (>16 years, >10h/week training, competing at national or Olympic levels) in the Netherlands, including CMR (Siemens Avanto Fit 1.5T). HCM patients were selected from the Cardiogenetic Biobank of Amsterdam UMC, which systematically collects data from index HCM patients and those identified via cascade screening. Primary metrics of interest were segmented LVWT, cardiac volume/function parameters and ratios. Differences in LV hypertrophy (LVH) distribution were assessed with MANOVA and principal component (PC) analysis (PCA). LVWT risk score (sum(segment i loading x segment i LVWT (mm)) based on the loadings of the PCs. Multivariate backwards logistic regression was used to determine discriminating factors. Receiver operating characteristic (ROC) curves were plotted and areas under the curves (AUC) were calculated to evaluate sensitivity (Sens) and specificity (Spec). Results We included a total of 30 athletes and 30 HCM patients. Sex (female 32% vs 50%, p = .277), length (180 cm ± 10 vs 170 cm ± 13, p = .124) and LVWT (12 mm ± 1.6 vs 13 mm ± 2.0, p = .070) were comparable between groups. HCM patients were older than elite athletes (29 years ± 10 vs 47 years ± 16, p < .001). Differences in cardiac volume, function and ratios are shown in [Table 1]. MANOVA showed a difference in LVWT distribution across the 16 AHA segments between HCM patients and elite athletes (Pillai’s trace = .823, F(16,41) = 11.912, p < .001). The maximal LVWT to LV end-diastolic volume (LVEDV) ratio was highly differentiating as singular variable (AUC = .977; Sens = 91.7%; Spec = 92.6%). A multivariate logistic regression model including LVM/LVEDV ratio and the LVWT risk score also performed well (AUC = .987, Sens = 96.4%, Spec = 93.1%). Conclusion Athletes exhibit a distinct LVH pattern with increased LVWT and concomitantly higher LVEDV. The LVWT/LVEDV ratio strongly differentiates physiological adaptation from HCM. Moreover, LVH distribution across the 16 segments differs between athletes and HCM patients, and a multivariate model including the LVM/LVEDV ratio and LVWT risk score performed well to distinguish
背景：左心室壁厚（LVWT）增加可能是运动员心脏和早期HCM的标志，这使得区分运动诱导的心脏重构（EICR）和早期HCM具有挑战性。在这个“灰色地带”中穿行仍然是运动心脏病学的一个关键挑战。目的探讨区分生理性EICR与早期基因型阳性HCM的心脏磁共振成像（CMR）参数。方法我们对基因型阳性的交界型HCM患者（LVWT：女性9-16mm，男性11-16mm）和匹配的精英运动员进行了横断面分析，比例为1:1。基于性别、长度和LVWT，优秀运动员倾向评分距离与HCM患者相匹配。运动员从ELITE中选出，ELITE是一个前瞻性的精英运动员队列，收集来自荷兰标准化心血管筛查（16年，每周训练10小时，参加国家或奥运会水平的比赛）的数据，包括CMR（西门子Avanto Fit 1.5T）。HCM患者从阿姆斯特丹UMC的心脏遗传生物银行（Cardiogenetic Biobank）中选择，该银行系统地收集了指数HCM患者和通过级联筛选确定的HCM患者的数据。主要感兴趣的指标是LVWT分割、心脏容积/功能参数和比率。采用方差分析（MANOVA）和主成分分析（PCA）评估左室肥厚（LVH）分布的差异。LVWT风险评分（和（第i段加载x第i段LVWT (mm)））基于pc的加载。采用多元倒向逻辑回归确定判别因素。绘制受试者工作特征（ROC）曲线，计算曲线下面积（AUC），评估灵敏度（Sens）和特异性（Spec）。结果共纳入30名运动员和30名HCM患者。性别（女性32% vs 50%， p = .277）、体长（180 cm±10 vs 170 cm±13,p = .124）和LVWT （12 mm±1.6 vs 13 mm±2.0,p = .070）组间具有可比性。HCM患者比优秀运动员年龄大（29岁±10岁vs 47岁±16岁，p < .001）。心脏容积、功能和比值的差异见[表1]。方差分析显示HCM患者和优秀运动员在16个AHA节段LVWT分布上存在差异(Pillai’s trace = .823， F(16,41) = 11.912, p <；措施)。最大LVWT与左室舒张末期容积（LVEDV）之比作为单一变量具有高度差异性（AUC = 0.977; Sens = 91.7%; Spec = 92.6%）。包括LVM/LVEDV比值和LVWT风险评分在内的多变量logistic回归模型也表现良好（AUC = 0.987, Sens = 96.4%, Spec = 93.1%）。结论运动员表现出明显的LVH模式，LVWT增加，同时lvvedv升高。LVWT/LVEDV比值将生理适应与HCM区分开来。此外，运动员和HCM患者在16个节段的LVH分布不同，包括LVM/LVEDV比和LVWT风险评分在内的多变量模型可以很好地区分EICR和HCM。散点图和ROC曲线
{"title":"Navigating the grey zone: patterns of left ventricular hypertrophy to differentiate early genotype-positive hypertrophic cardiomyopathy from left ventricular wall thickness-matched elite athletes","authors":"J J N Daems, M A Van Diepen, R Lagerweij, F E Van Leusden, D Kramarenko, E Poel, M H Moen, A J Nederveen, A Van Randen, N B Bijsterveld, M Van Den Boogaard, Y M Pinto, F W Asselbergs, A S Amin, H T Jorstad","doi":"10.1093/eurheartj/ehaf784.3953","DOIUrl":"https://doi.org/10.1093/eurheartj/ehaf784.3953","url":null,"abstract":"Background Increased left ventricular (LV) wall thickness (LVWT) can be a hallmark of both an athlete’s heart and early HCM, making it challenging to distinguish exercise-induced cardiac remodelling (EICR) from early-stage HCM. Navigating this ‘grey zone’ remains a key challenge in sports cardiology. Purpose To determine cardiac magnetic resonance imaging (CMR) parameters differentiating physiological EICR from early-stage genotype positive HCM. Methods We conducted a cross-sectional analysis of genotype-positive borderline HCM patients (LVWT: women 9-16mm, men 11-16mm), and matched elite athletes in a 1:1 ratio. Elite athletes were propensity score distance matched to HCM patients based on sex, length and LVWT. Athlete were selected from ELITE, a prospective elite athlete cohorts which collects data from standardized cardiovascular screenings (&gt;16 years, &gt;10h/week training, competing at national or Olympic levels) in the Netherlands, including CMR (Siemens Avanto Fit 1.5T). HCM patients were selected from the Cardiogenetic Biobank of Amsterdam UMC, which systematically collects data from index HCM patients and those identified via cascade screening. Primary metrics of interest were segmented LVWT, cardiac volume/function parameters and ratios. Differences in LV hypertrophy (LVH) distribution were assessed with MANOVA and principal component (PC) analysis (PCA). LVWT risk score (sum(segment i loading x segment i LVWT (mm)) based on the loadings of the PCs. Multivariate backwards logistic regression was used to determine discriminating factors. Receiver operating characteristic (ROC) curves were plotted and areas under the curves (AUC) were calculated to evaluate sensitivity (Sens) and specificity (Spec). Results We included a total of 30 athletes and 30 HCM patients. Sex (female 32% vs 50%, p = .277), length (180 cm ± 10 vs 170 cm ± 13, p = .124) and LVWT (12 mm ± 1.6 vs 13 mm ± 2.0, p = .070) were comparable between groups. HCM patients were older than elite athletes (29 years ± 10 vs 47 years ± 16, p &lt; .001). Differences in cardiac volume, function and ratios are shown in [Table 1]. MANOVA showed a difference in LVWT distribution across the 16 AHA segments between HCM patients and elite athletes (Pillai’s trace = .823, F(16,41) = 11.912, p &lt; .001). The maximal LVWT to LV end-diastolic volume (LVEDV) ratio was highly differentiating as singular variable (AUC = .977; Sens = 91.7%; Spec = 92.6%). A multivariate logistic regression model including LVM/LVEDV ratio and the LVWT risk score also performed well (AUC = .987, Sens = 96.4%, Spec = 93.1%). Conclusion Athletes exhibit a distinct LVH pattern with increased LVWT and concomitantly higher LVEDV. The LVWT/LVEDV ratio strongly differentiates physiological adaptation from HCM. Moreover, LVH distribution across the 16 segments differs between athletes and HCM patients, and a multivariate model including the LVM/LVEDV ratio and LVWT risk score performed well to distinguish","PeriodicalId":11976,"journal":{"name":"European Heart Journal","volume":"28 1","pages":""},"PeriodicalIF":39.3,"publicationDate":"2026-02-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146122012","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2026-02-05DOI: 10.1093/eurheartj/ehaf784.152
C De Gori, A Aimo, V Castiglione, G Vergaro, F Pignatelli, M Muca, M Occhipinti, M Emdin, A Clemente
Background Photon-counting computed tomography (CT) is gaining prominence in cardiac imaging, but its potential for tissue characterization in patients with a hypertrophic phenotype remains to be explored, particularly in comparison with the current gold standard, cardiac magnetic resonance (CMR) imaging. Methods Consecutive patients with hypertrophic cardiomyopathy (HCM) or amyloid transthyretin cardiomyopathy (ATTR-CM) followed in the outpatient clinic of a referral center for cardiomyopathies were referred to a photon-counting CT as part of a research protocol. A subgroup of these patients underwent also a CMR scan within 3 months from the date of the CT scan, when deemed clinically indicated. Results Patients with HCM (n=22) were younger than those with ATTR-CM (n=22; p=0.001), while the percentages of men and women did not differ (p=0.966). On CT scan, patients with HCM had lower values of LV mass (p=0.028) and higher LV ejection fraction (p=0.023), while the extracellular volume (ECV) was lower in patients with HCM (p<0.001). The values of iodine density (a measure of the amount of contrast within the myocardial tissue) did not differ between HCM and ATTR-CM (p=0.101). A subgroup of patients (n=27; n=16 with HCM, n=11 with ATTR-CM) underwent also a CMR scan. Patients with HCM displayed tight correlations between LV mass values from CT and CMR (p<0.001, beta=0.869), and maximal wall thickness (p<0.001, beta=0.969), but no significant correlations between ECV values from the two techniques (p=0.275), or between iodine density and CMR-derived ECV (p=0.274). Patients with ATTR-CM showed significant correlations between LV mass values (p<0.001, beta=0.956) and maximal LV wall thickness (p<0.001, beta=0.990) from CT and CMR, but not between ECV values from the two techniques (p=0.139). Conversely, patients with ATTR-CM displayed a close correlation between iodine density and CMR-derived ECV (p=0.016, beta=0.894). Conclusions Photon-counting CT demonstrated strong agreement with CMR for structural parameters (LV mass, maximal wall thickness), supporting its utility for morphologic evaluation. Furthermore, photon-counting CT may hold promise for tissue characterization in patients with ATTR-CM, given the strong concordance between iodine density and CMR-derived ECV.
{"title":"Photon-counting computed tomography for tissue characterization in patients with hypertrophic cardiomyopathy or amyloidosis","authors":"C De Gori, A Aimo, V Castiglione, G Vergaro, F Pignatelli, M Muca, M Occhipinti, M Emdin, A Clemente","doi":"10.1093/eurheartj/ehaf784.152","DOIUrl":"https://doi.org/10.1093/eurheartj/ehaf784.152","url":null,"abstract":"Background Photon-counting computed tomography (CT) is gaining prominence in cardiac imaging, but its potential for tissue characterization in patients with a hypertrophic phenotype remains to be explored, particularly in comparison with the current gold standard, cardiac magnetic resonance (CMR) imaging. Methods Consecutive patients with hypertrophic cardiomyopathy (HCM) or amyloid transthyretin cardiomyopathy (ATTR-CM) followed in the outpatient clinic of a referral center for cardiomyopathies were referred to a photon-counting CT as part of a research protocol. A subgroup of these patients underwent also a CMR scan within 3 months from the date of the CT scan, when deemed clinically indicated. Results Patients with HCM (n=22) were younger than those with ATTR-CM (n=22; p=0.001), while the percentages of men and women did not differ (p=0.966). On CT scan, patients with HCM had lower values of LV mass (p=0.028) and higher LV ejection fraction (p=0.023), while the extracellular volume (ECV) was lower in patients with HCM (p&lt;0.001). The values of iodine density (a measure of the amount of contrast within the myocardial tissue) did not differ between HCM and ATTR-CM (p=0.101). A subgroup of patients (n=27; n=16 with HCM, n=11 with ATTR-CM) underwent also a CMR scan. Patients with HCM displayed tight correlations between LV mass values from CT and CMR (p&lt;0.001, beta=0.869), and maximal wall thickness (p&lt;0.001, beta=0.969), but no significant correlations between ECV values from the two techniques (p=0.275), or between iodine density and CMR-derived ECV (p=0.274). Patients with ATTR-CM showed significant correlations between LV mass values (p&lt;0.001, beta=0.956) and maximal LV wall thickness (p&lt;0.001, beta=0.990) from CT and CMR, but not between ECV values from the two techniques (p=0.139). Conversely, patients with ATTR-CM displayed a close correlation between iodine density and CMR-derived ECV (p=0.016, beta=0.894). Conclusions Photon-counting CT demonstrated strong agreement with CMR for structural parameters (LV mass, maximal wall thickness), supporting its utility for morphologic evaluation. Furthermore, photon-counting CT may hold promise for tissue characterization in patients with ATTR-CM, given the strong concordance between iodine density and CMR-derived ECV.","PeriodicalId":11976,"journal":{"name":"European Heart Journal","volume":"89 1","pages":""},"PeriodicalIF":39.3,"publicationDate":"2026-02-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146122135","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2026-02-05DOI: 10.1093/eurheartj/ehaf784.4453
T Kaihara, K Seki, S Hattori, S Kou, J Kim, Y Cho, K Sasaki, Y Akashi
Background The integration of artificial intelligence (AI) and 12-lead ECGs is an important focus in digital cardiology, and the evidence is a growing focus. Recently, a smartphone app has enabled the capture and analysis of 12-lead ECGs. In this study, our AI-based app captures 12-lead ECGs and extracts ECG rhythms and digital biomarkers. Purpose This study will evaluate the accuracy of detecting moderate pulmonary hypertension (PH) by 12-lead ECG imaging using this app in Japanese patients. Methods A cross-sectional study was conducted on patients who underwent Swan-Ganz catheterization (SGC) from January 2020 to August 2024 at a medical institution in Japan. The app was used to extract 10 digital biomarkers, including the "PH score" (0-100 points), which estimates electrocardiographic rhythm and pulmonary arterial pressure (PAP) elevation. Right ventricular systolic pressure (RVSP) estimated by transthoracic echocardiogram (TTE) was also analysed in the same patients. Results Among 726 patients, 457 met inclusion criteria (exclusion criteria: cardiac surgery, pacemaker rhythm, no ECG within 3 days after SGC, etc.). The mean age was 65 ± 16 years, and 59% were male; moderate PH (mean PAP > 40 mmHg) diagnosed at SGC was 7.2%. The AUC-ROC (Area Under the Receiver Operating Characteristic Curve) of the "PH score" calculated by the app was 0.84 (95% CI: 0.77-0.91, p < 0.001) (Figure). The "PH score" threshold of 35 points (maximum Youden index) resulted in a sensitivity of 79% and a specificity of 81%; the AUC-PR (Area Under the Precision-Recall Curve) was 0.407 (95% CI: 0.171-0.662) The AUC-ROC of RVSP by TTE was 0.94 (95% CI: 0.89-0.99, p < 0.001) (Figure), and the RVSP threshold of 49 mmHg (maximum Youden index) achieved a sensitivity of 87% and specificity of 86%. Finally, PH estimation by RVSP outperformed the "PH score" calculated by the app (DeLong's p = 0.010). Conclusion Moderate PH can be predicted from a 12-lead ECG using an AI-powered smartphone app. Although RVSP is superior to "PH Score" in estimating moderate PH, the app has the potential to identify potentially lethal PH with good AUC-ROC in settings where cardiologists and echocardiography are not available. Additionally, it could also be integrated with electronic medical records.
人工智能（AI）与12导联心电图的整合是数字心脏病学的一个重要焦点，其证据也越来越受到关注。最近，一款智能手机应用程序可以捕获和分析12导联心电图。在这项研究中，我们基于人工智能的应用程序捕获12导联心电图，并提取ECG节律和数字生物标志物。目的本研究将评估该应用程序在日本患者中使用12导联心电图成像检测中度肺动脉高压（PH）的准确性。方法对2020年1月至2024年8月在日本某医疗机构行Swan-Ganz导管（SGC）的患者进行横断面研究。该应用程序用于提取10个数字生物标志物，包括“PH值”（0-100分），用于估计心电图节律和肺动脉压（PAP）升高。同时分析了经胸超声心动图（TTE）测量的右心室收缩压（RVSP）。结果726例患者中，457例符合纳入标准（排除标准：心脏手术、起搏器节律、SGC后3天内无心电图等）。平均年龄65±16岁，男性占59%；中度PH（平均PAP 40 mmHg）在SGC诊断为7.2%。应用程序计算的“PH评分”的AUC-ROC （Receiver Operating Characteristic Curve下面积）为0.84 (95% CI: 0.77-0.91, p < 0.001)（图）。“PH评分”阈值为35分（最大约登指数），敏感性为79%，特异性为81%；精确度-召回曲线下面积（AUC-PR）为0.407 (95% CI: 0.171-0.662)， TTE检测RVSP的AUC-ROC为0.94 (95% CI: 0.89-0.99, p < 0.001)（图），RVSP阈值为49 mmHg（最大约登指数），灵敏度为87%，特异性为86%。最后，RVSP估计的PH值优于应用程序计算的“PH值”（DeLong’s p = 0.010）。结论：使用人工智能智能手机应用程序可以从12导联心电图中预测中度PH值。尽管RVSP在估计中度PH值方面优于“PH评分”，但在没有心脏病专家和超声心动图的情况下，该应用程序具有良好的AUC-ROC识别潜在致命PH值的潜力。此外，它还可以与电子医疗记录集成。
{"title":"Can an AI-powered smartphone app estimate moderate pulmonary hypertension by taking a 12-lead ECG of Japanese patients?","authors":"T Kaihara, K Seki, S Hattori, S Kou, J Kim, Y Cho, K Sasaki, Y Akashi","doi":"10.1093/eurheartj/ehaf784.4453","DOIUrl":"https://doi.org/10.1093/eurheartj/ehaf784.4453","url":null,"abstract":"Background The integration of artificial intelligence (AI) and 12-lead ECGs is an important focus in digital cardiology, and the evidence is a growing focus. Recently, a smartphone app has enabled the capture and analysis of 12-lead ECGs. In this study, our AI-based app captures 12-lead ECGs and extracts ECG rhythms and digital biomarkers. Purpose This study will evaluate the accuracy of detecting moderate pulmonary hypertension (PH) by 12-lead ECG imaging using this app in Japanese patients. Methods A cross-sectional study was conducted on patients who underwent Swan-Ganz catheterization (SGC) from January 2020 to August 2024 at a medical institution in Japan. The app was used to extract 10 digital biomarkers, including the \"PH score\" (0-100 points), which estimates electrocardiographic rhythm and pulmonary arterial pressure (PAP) elevation. Right ventricular systolic pressure (RVSP) estimated by transthoracic echocardiogram (TTE) was also analysed in the same patients. Results Among 726 patients, 457 met inclusion criteria (exclusion criteria: cardiac surgery, pacemaker rhythm, no ECG within 3 days after SGC, etc.). The mean age was 65 ± 16 years, and 59% were male; moderate PH (mean PAP &gt; 40 mmHg) diagnosed at SGC was 7.2%. The AUC-ROC (Area Under the Receiver Operating Characteristic Curve) of the \"PH score\" calculated by the app was 0.84 (95% CI: 0.77-0.91, p &lt; 0.001) (Figure). The \"PH score\" threshold of 35 points (maximum Youden index) resulted in a sensitivity of 79% and a specificity of 81%; the AUC-PR (Area Under the Precision-Recall Curve) was 0.407 (95% CI: 0.171-0.662) The AUC-ROC of RVSP by TTE was 0.94 (95% CI: 0.89-0.99, p &lt; 0.001) (Figure), and the RVSP threshold of 49 mmHg (maximum Youden index) achieved a sensitivity of 87% and specificity of 86%. Finally, PH estimation by RVSP outperformed the \"PH score\" calculated by the app (DeLong's p = 0.010). Conclusion Moderate PH can be predicted from a 12-lead ECG using an AI-powered smartphone app. Although RVSP is superior to \"PH Score\" in estimating moderate PH, the app has the potential to identify potentially lethal PH with good AUC-ROC in settings where cardiologists and echocardiography are not available. Additionally, it could also be integrated with electronic medical records.","PeriodicalId":11976,"journal":{"name":"European Heart Journal","volume":"48 1","pages":""},"PeriodicalIF":39.3,"publicationDate":"2026-02-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146122195","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2026-02-05DOI: 10.1093/eurheartj/ehaf784.3518
B Lutterbey, E Aaseth, S Halvorsen, J Gravning
Background Improved prediction of future cardiovascular (CV) events in young adults is of importance, as preventive strategies might be warranted also in young adults. The Pathobiological Determinants of Atherosclerosis in Youth (PDAY) risk score was originally developed based on post-mortem assessment of atherosclerosis in the abdominal aorta and the coronary arteries in individuals aged 15-34 years who died accidentally. Later, a revised score has been launched. The PDAY risk score is also shown to predict clinical CV events among participants in the CARDIA cohort (18-30 years of age) during 30 years of follow-up. Purpose To apply the original and the revised PDAY risk score on a contemporary cohort of 30-year-old Norwegians and validate their respective predictive value by assessing the association with clinical CV events during long-term follow-up. Methods In the year 2000, all inhabitants of our city, Norway, born in 1969 or 1970 were invited to participate in a prospective cohort study (Oslo Health Study). The PDAY risk score includes eight risk factors (age, sex, non-high-density lipoprotein (HDL) cholesterol, HDL cholesterol, smoking, blood pressure, obesity, and hyperglycaemia) and was calculated according to Table 1 for the 5842 participants (mean age 31 years, 56% women). The revised PDAY risk score, additionally including positive family history of CV disease and putting less weight on female sex, was also calculated. The occurrence of CV events during 22 years follow-up was obtained through linkage to Norwegian health registries. The primary outcome was a composite of CV death, non-fatal myocardial infarction, non-fatal ischemic stroke, coronary revascularization and hospitalization due to unstable angina. A one standard deviation (SD) increase in PDAY risk score, with the mean as reference, was set to explore if increased risk of CV events was observed with increasing PDAY risk score. Model discrimination was evaluated with C-statistics for the original and revised PDAY risk score, respectively. Results The PDAY risk score ranged from 13 to 39 points (mean 17.8 ± 4.2). One SD increase in points from mean at baseline was associated with a 3.15-fold increase in the incidence of CV events (HR, 95% CI 2.14-4.63). Unadjusted C-statistic was 0.686 (95% CI 0.634-0.738). When comparing the revised PDAY risk score with the original, prediction of future CV events was not significantly improved (C-statistic 0.707 [95% CI 0.654-0.759], P-value 0.0661), as illustrated in Figure 1. Conclusion In our contemporary cohort of 30-year-old individuals, the PDAY risk score was modestly associated with the risk of CV events during long-term follow-up. No difference was found between the original and the revised version. Thus, improved prediction tools for future CV events among young adults are still needed.Table 1 – PDAY risk score Figure 1 - AUC for CV events
背景：提高对年轻人未来心血管（CV）事件的预测是很重要的，因为预防策略也可能在年轻人中得到保证。青年动脉粥样硬化的病理生物学决定因素（PDAY）风险评分最初是基于对15-34岁意外死亡个体的腹主动脉和冠状动脉动脉粥样硬化的尸检评估而开发的。后来，修订后的分数已经发布。PDAY风险评分也被证明可以预测CARDIA队列（18-30岁）参与者在30年随访期间的临床CV事件。目的将原始和修订后的PDAY风险评分应用于30岁挪威人的当代队列，并通过评估长期随访期间与临床CV事件的关联来验证其各自的预测价值。方法在2000年，我们邀请所有1969年或1970年出生的挪威城市居民参加一项前瞻性队列研究（奥斯陆健康研究）。PDAY风险评分包括8个风险因素（年龄、性别、非高密度脂蛋白（HDL）胆固醇、高密度脂蛋白胆固醇、吸烟、血压、肥胖和高血糖），并根据表1计算5842名参与者（平均年龄31岁，56%为女性）。还计算了修订后的PDAY风险评分，此外还包括心血管疾病阳性家族史和女性体重减轻。通过与挪威健康登记处的联系，获得了22年随访期间CV事件的发生情况。主要结局是CV死亡、非致死性心肌梗死、非致死性缺血性卒中、冠状动脉血运重建术和不稳定心绞痛住院的综合结果。设置PDAY风险评分增加1个标准差（SD），以平均值为参考，探讨PDAY风险评分增加是否会导致心血管事件风险增加。分别用c统计量对原始和修订后的PDAY风险评分进行模型判别。结果PDAY风险评分范围为13 ~ 39分（平均17.8±4.2分）。基线时每增加一个SD点，CV事件发生率增加3.15倍（HR, 95% CI 2.14-4.63）。未经校正的c统计量为0.686 （95% CI 0.634-0.738）。将修订后的PDAY风险评分与原始评分进行比较，对未来CV事件的预测没有显著提高（c统计量0.707 [95% CI 0.654-0.759]， p值0.0661），如图1所示。结论：在我们的当代30岁个体队列中，PDAY风险评分与长期随访期间心血管事件的风险有一定的相关性。原来的版本和修改后的版本没有区别。因此，仍需要改进预测年轻人心血管事件的工具。表1 - PDAY风险评分图1 - CV事件的AUC
{"title":"Validation of the pathobiological determinants of atherosclerosis in youth risk score in a Norwegian cohort of young adults","authors":"B Lutterbey, E Aaseth, S Halvorsen, J Gravning","doi":"10.1093/eurheartj/ehaf784.3518","DOIUrl":"https://doi.org/10.1093/eurheartj/ehaf784.3518","url":null,"abstract":"Background Improved prediction of future cardiovascular (CV) events in young adults is of importance, as preventive strategies might be warranted also in young adults. The Pathobiological Determinants of Atherosclerosis in Youth (PDAY) risk score was originally developed based on post-mortem assessment of atherosclerosis in the abdominal aorta and the coronary arteries in individuals aged 15-34 years who died accidentally. Later, a revised score has been launched. The PDAY risk score is also shown to predict clinical CV events among participants in the CARDIA cohort (18-30 years of age) during 30 years of follow-up. Purpose To apply the original and the revised PDAY risk score on a contemporary cohort of 30-year-old Norwegians and validate their respective predictive value by assessing the association with clinical CV events during long-term follow-up. Methods In the year 2000, all inhabitants of our city, Norway, born in 1969 or 1970 were invited to participate in a prospective cohort study (Oslo Health Study). The PDAY risk score includes eight risk factors (age, sex, non-high-density lipoprotein (HDL) cholesterol, HDL cholesterol, smoking, blood pressure, obesity, and hyperglycaemia) and was calculated according to Table 1 for the 5842 participants (mean age 31 years, 56% women). The revised PDAY risk score, additionally including positive family history of CV disease and putting less weight on female sex, was also calculated. The occurrence of CV events during 22 years follow-up was obtained through linkage to Norwegian health registries. The primary outcome was a composite of CV death, non-fatal myocardial infarction, non-fatal ischemic stroke, coronary revascularization and hospitalization due to unstable angina. A one standard deviation (SD) increase in PDAY risk score, with the mean as reference, was set to explore if increased risk of CV events was observed with increasing PDAY risk score. Model discrimination was evaluated with C-statistics for the original and revised PDAY risk score, respectively. Results The PDAY risk score ranged from 13 to 39 points (mean 17.8 ± 4.2). One SD increase in points from mean at baseline was associated with a 3.15-fold increase in the incidence of CV events (HR, 95% CI 2.14-4.63). Unadjusted C-statistic was 0.686 (95% CI 0.634-0.738). When comparing the revised PDAY risk score with the original, prediction of future CV events was not significantly improved (C-statistic 0.707 [95% CI 0.654-0.759], P-value 0.0661), as illustrated in Figure 1. Conclusion In our contemporary cohort of 30-year-old individuals, the PDAY risk score was modestly associated with the risk of CV events during long-term follow-up. No difference was found between the original and the revised version. Thus, improved prediction tools for future CV events among young adults are still needed.Table 1 – PDAY risk score Figure 1 - AUC for CV events","PeriodicalId":11976,"journal":{"name":"European Heart Journal","volume":"35 1","pages":""},"PeriodicalIF":39.3,"publicationDate":"2026-02-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146121866","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2026-02-05DOI: 10.1093/eurheartj/ehaf784.4086
T Caller, T Hasin, E Sharon, R Loutati, Y Yarkoni, N Naftali-Shani, T Itkin, J Leor, E Maor
Introduction Cardiovascular diseases (CVD) are associated with an increased risk of cancer. However, the population-attributable fraction (PAF) of CVD to cancer and the association between CVD and second-primary cancer, cancer-related hospitalizations, or death weren’t described. Methods We analyzed data from 109,204 adults, free of cancer at baseline, with valid echocardiography examination between 2000-2024. Cancer, hospitalization, and mortality data were obtained from Israel's National Registries and institutional records. CVD was defined as any significant structural or clinical heart/vascular disease. This definition includes heart failure, ischemic heart disease, atrial fibrillation, valvular diseases and stroke. We used ICD9 codes to identify 2nd primary cancer. To avoid the detection of asymptomatic cancer, we implicated a 3-month blanking period at the start of the follow-up. Then, we used Cox regression and Poisson regression to assess the link between CVD and cancer incidence, hospitalizations, and death and to adjust for age, sex, BMI, smoking, and eGFR. Results During a median follow-up of 6.0 ± 4.4 years, 4441 (4%) patients developed cancer; of them, 192 patients developed 2nd primary cancer. CVD was associated with an increased multivariable-adjusted risk of cancer (HR=1.53, 95%CI: 1.43-1.63). Moreover, CVD contributed 13% of the PAF for cancer, second only to aging and smoking (Figure 1A). CVD was also associated with a 10% higher incidence of second primary cancer (IRR=1.10, 95%CI: 1.04-1.15, Figure 1B), which is the development of another cancer type. Furthermore, CVD was also associated with a 2-fold increase in cancer-related hospitalizations in patients with cancer (1.64 vs 3.83 hospitalizations per year, IRR=2.05, 95%CI: 1.86-2.26, p<0.001, Figure 1C). Finally, we used multivariable-adjusted interaction analysis to assess the link between CVD, cancer, and mortality. We found a significant interaction between CVD and cancer, indicating that the increased risk of death associated with cancer twice fold in patients with CVD (HR=2.05, 95%CI: 1.85-2.23, p<0.001). Using the Kaplan-Meier method, we demonstrated that patients with concomitant CVD and cancer suffer from worse prognoses compared to cancer patients without CVD (Figure 1D). Conclusion We show, for the first time, that CVD is a major contributor to the burden of cancer. CVD was linked to a higher risk of cancer, second-primary cancer, a high population-attributable fraction for cancer, more cancer-related hospitalization, and increased mortality. Recognizing this association may enhance cancer prevention, early diagnosis, and treatment for patients with CVD.
{"title":"Cardiovascular disease is a significant contributor to the incidence of cancer, second-primary cancer, and cancer-related hospitalization","authors":"T Caller, T Hasin, E Sharon, R Loutati, Y Yarkoni, N Naftali-Shani, T Itkin, J Leor, E Maor","doi":"10.1093/eurheartj/ehaf784.4086","DOIUrl":"https://doi.org/10.1093/eurheartj/ehaf784.4086","url":null,"abstract":"Introduction Cardiovascular diseases (CVD) are associated with an increased risk of cancer. However, the population-attributable fraction (PAF) of CVD to cancer and the association between CVD and second-primary cancer, cancer-related hospitalizations, or death weren’t described. Methods We analyzed data from 109,204 adults, free of cancer at baseline, with valid echocardiography examination between 2000-2024. Cancer, hospitalization, and mortality data were obtained from Israel's National Registries and institutional records. CVD was defined as any significant structural or clinical heart/vascular disease. This definition includes heart failure, ischemic heart disease, atrial fibrillation, valvular diseases and stroke. We used ICD9 codes to identify 2nd primary cancer. To avoid the detection of asymptomatic cancer, we implicated a 3-month blanking period at the start of the follow-up. Then, we used Cox regression and Poisson regression to assess the link between CVD and cancer incidence, hospitalizations, and death and to adjust for age, sex, BMI, smoking, and eGFR. Results During a median follow-up of 6.0 ± 4.4 years, 4441 (4%) patients developed cancer; of them, 192 patients developed 2nd primary cancer. CVD was associated with an increased multivariable-adjusted risk of cancer (HR=1.53, 95%CI: 1.43-1.63). Moreover, CVD contributed 13% of the PAF for cancer, second only to aging and smoking (Figure 1A). CVD was also associated with a 10% higher incidence of second primary cancer (IRR=1.10, 95%CI: 1.04-1.15, Figure 1B), which is the development of another cancer type. Furthermore, CVD was also associated with a 2-fold increase in cancer-related hospitalizations in patients with cancer (1.64 vs 3.83 hospitalizations per year, IRR=2.05, 95%CI: 1.86-2.26, p&lt;0.001, Figure 1C). Finally, we used multivariable-adjusted interaction analysis to assess the link between CVD, cancer, and mortality. We found a significant interaction between CVD and cancer, indicating that the increased risk of death associated with cancer twice fold in patients with CVD (HR=2.05, 95%CI: 1.85-2.23, p&lt;0.001). Using the Kaplan-Meier method, we demonstrated that patients with concomitant CVD and cancer suffer from worse prognoses compared to cancer patients without CVD (Figure 1D). Conclusion We show, for the first time, that CVD is a major contributor to the burden of cancer. CVD was linked to a higher risk of cancer, second-primary cancer, a high population-attributable fraction for cancer, more cancer-related hospitalization, and increased mortality. Recognizing this association may enhance cancer prevention, early diagnosis, and treatment for patients with CVD.","PeriodicalId":11976,"journal":{"name":"European Heart Journal","volume":"58 1","pages":""},"PeriodicalIF":39.3,"publicationDate":"2026-02-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146122015","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2026-02-05DOI: 10.1093/eurheartj/ehaf784.4364
P M Croon, L S Dhingra, D Biswas, E K Oikonomou, R Khera
Background Artificial intelligence (AI) applications for electrocardiograms (ECGs) have been proposed for the detection and prediction of a range of specific structural and functional cardiac abnormalities. To better define the clinical utility as diagnostic and predictive tools, we sought to explore the specificity of the cross-sectional and longitudinal phenotypic associations of several AI-ECG tools. Purpose To systematically evaluate the cross-sectional and longitudinal phenotypic associations of 6 AI-ECG models across a US-based tertiary care hospital, 4 community hospitals, an outpatient medical network, and the UK Biobank. Methods We deployed 6 AI-ECG models on ECG images, including five validated models for the detection of left ventricular systolic dysfunction (LVSD), aortic stenosis (AS), mitral regurgitation (MR), left ventricular hypertrophy (LVH), a composite model for structural heart disease (SHD), and a negative control AI-ECG model for biological sex. Diagnosis codes from the electronic health records were transformed into phenotype codes and a phenome-wide association study (PheWAS) was conducted. We assessed the association of AI-ECG-probabilities with clinical phenotypes, (i) cross-sectionally using age/sex-adjusted logistic regression, and (ii) longitudinally for new-onset CV diseases in age/sex-adjusted Cox regression. Results The study included 265,187 individuals (mean age 59±18 years, 146,090 [55%] women) across sites, with one random ECG per person. Each of the 5 AI-ECG models had differentially stronger association with cardiovascular phenotypes compared with other phenotype groups, which was not observed for the AI-ECG model for sex, which was most strongly associated with non-CV phenotypes (Figure 1). Each of the AI-ECG models was significantly associated with their target phenotype, but they also exhibited similar or stronger associations with a broad range of other cardiovascular phenotypes. For instance, the AI-ECG model for AS was more strongly associated with heart failure NOS (OR 3.2, p <10⁻³⁰⁰) than with aortic valve disease (OR 2.7, p <10⁻²⁵⁹). Each of the models had similar strong cross-phenotype associations (Figure 2A). For predicting future disease, AI-ECG models had strong non-specific associations with a broad range of CV phenotypes, spanning both intended and related phenotypes (Figure 2B). These findings were consistent across models and cohorts. Conclusion Despite AI-ECG being developed to detect specific cardiovascular conditions, they are non-specific and detect a range of CV abnormalities and predict the occurrence of a range of adverse CV outcomes. These findings suggest that several AI-ECG models best serve as general biomarkers of CV health rather than dichotomous diagnostic or predictive tools.figure 1 Figure 2
人工智能（AI）在心电图（ECGs）中的应用已被提出用于检测和预测一系列特定的结构和功能心脏异常。为了更好地定义作为诊断和预测工具的临床效用，我们试图探索几种AI-ECG工具的横断面和纵向表型关联的特异性。目的系统评估美国一家三级医院、4家社区医院、一个门诊医疗网络和英国生物银行的6种AI-ECG模型的横断面和纵向表型关联。方法将6个AI-ECG模型应用于心电图图像，包括5个经验证的左室收缩功能障碍（LVSD）、主动脉瓣狭窄（AS）、二尖瓣反流（MR）、左室肥厚（LVH）检测模型、1个结构性心脏病（SHD）复合模型和1个阴性对照AI-ECG生物性别模型。将电子健康记录中的诊断代码转换为表型代码，并进行全表型关联研究（PheWAS）。我们评估了ai - ecg概率与临床表型的关联，(i)使用年龄/性别调整的logistic回归进行横断面分析，（ii）使用年龄/性别调整的Cox回归对新发CV疾病进行纵向分析。结果该研究纳入了265,187例个体（平均年龄59±18岁，146,090例[55%]女性），随机每人一次心电图。与其他表型组相比，5种AI-ECG模型中的每一种模型与心血管表型的相关性都有差异，这在性别的AI-ECG模型中没有观察到，它与非cv表型的相关性最强（图1）。每种AI-ECG模型都与其目标表型显著相关，但它们也与广泛的其他心血管表型表现出相似或更强的相关性。例如，AS的AI-ECG模型与心力衰竭NOS （OR 3.2, p <10⁻³⁰⁰）的相关性比与主动脉瓣疾病（OR 2.7, p <10⁻²）的相关性更强。每种模型都具有相似的强交叉表型关联（图2A）。为了预测未来的疾病，AI-ECG模型与广泛的CV表型具有很强的非特异性关联，包括预期表型和相关表型（图2B）。这些发现在不同的模型和队列中是一致的。尽管AI-ECG被用于检测特定的心血管疾病，但它们是非特异性的，只能检测一系列CV异常并预测一系列不良CV结局的发生。这些发现表明，几种AI-ECG模型最适合作为心血管健康的一般生物标志物，而不是二元诊断或预测工具
{"title":"Specificity of artificial intelligence-enhanced electrocardiography for diagnosis and prediction of cardiovascular disorders","authors":"P M Croon, L S Dhingra, D Biswas, E K Oikonomou, R Khera","doi":"10.1093/eurheartj/ehaf784.4364","DOIUrl":"https://doi.org/10.1093/eurheartj/ehaf784.4364","url":null,"abstract":"Background Artificial intelligence (AI) applications for electrocardiograms (ECGs) have been proposed for the detection and prediction of a range of specific structural and functional cardiac abnormalities. To better define the clinical utility as diagnostic and predictive tools, we sought to explore the specificity of the cross-sectional and longitudinal phenotypic associations of several AI-ECG tools. Purpose To systematically evaluate the cross-sectional and longitudinal phenotypic associations of 6 AI-ECG models across a US-based tertiary care hospital, 4 community hospitals, an outpatient medical network, and the UK Biobank. Methods We deployed 6 AI-ECG models on ECG images, including five validated models for the detection of left ventricular systolic dysfunction (LVSD), aortic stenosis (AS), mitral regurgitation (MR), left ventricular hypertrophy (LVH), a composite model for structural heart disease (SHD), and a negative control AI-ECG model for biological sex. Diagnosis codes from the electronic health records were transformed into phenotype codes and a phenome-wide association study (PheWAS) was conducted. We assessed the association of AI-ECG-probabilities with clinical phenotypes, (i) cross-sectionally using age/sex-adjusted logistic regression, and (ii) longitudinally for new-onset CV diseases in age/sex-adjusted Cox regression. Results The study included 265,187 individuals (mean age 59±18 years, 146,090 [55%] women) across sites, with one random ECG per person. Each of the 5 AI-ECG models had differentially stronger association with cardiovascular phenotypes compared with other phenotype groups, which was not observed for the AI-ECG model for sex, which was most strongly associated with non-CV phenotypes (Figure 1). Each of the AI-ECG models was significantly associated with their target phenotype, but they also exhibited similar or stronger associations with a broad range of other cardiovascular phenotypes. For instance, the AI-ECG model for AS was more strongly associated with heart failure NOS (OR 3.2, p &lt;10⁻³⁰⁰) than with aortic valve disease (OR 2.7, p &lt;10⁻²⁵⁹). Each of the models had similar strong cross-phenotype associations (Figure 2A). For predicting future disease, AI-ECG models had strong non-specific associations with a broad range of CV phenotypes, spanning both intended and related phenotypes (Figure 2B). These findings were consistent across models and cohorts. Conclusion Despite AI-ECG being developed to detect specific cardiovascular conditions, they are non-specific and detect a range of CV abnormalities and predict the occurrence of a range of adverse CV outcomes. These findings suggest that several AI-ECG models best serve as general biomarkers of CV health rather than dichotomous diagnostic or predictive tools.figure 1 Figure 2","PeriodicalId":11976,"journal":{"name":"European Heart Journal","volume":"398 1","pages":""},"PeriodicalIF":39.3,"publicationDate":"2026-02-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146122019","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2026-02-05DOI: 10.1093/eurheartj/ehaf784.4011
S E Alexander, E J Thompson, K A Bolam, E J Howden
Background Cardiovascular disease (CVD) in a leading cause of disease burden and mortality globally. Sex differences exist regarding CVD incidence, but the reasons for this are unclear. Further, women with polycystic ovary syndrome and chronically high levels of testosterone consistently present with worse cardiometabolic profiles than their normoandrogenic counterparts. But whether this is a direct consequence of higher levels of testosterone is unclear. Aim This systematic review aimed to examine current evidence regarding the associations between testosterone and CVD incidence in women. Method MEDLINE complete, Embase and CINAHL complete were systematically searched in September 2024. Articles that investigated the relationship between plasma testosterone concentrations and overt CVD in pre- or post-menopausal women were included. Results Forty-two studies were included in the final review. Outcome measures included the occurrence of a major adverse cardiac event, CVD mortality and morbidity, coronary artery disease/coronary heart disease, atherosclerotic disease, heart failure, myocardial infarction and stroke. Eight studies were case-control studies, 13 were observational cross-sectional studies, 19 were prospective cohort studies and two were retrospective cohort studies. Of the prospective cohort studies, the median follow-up length was 11.5 years (range 2.4-19.2 years). Thirty-nine articles included endogenous plasma testosterone concentrations as an exposure, one article assessed the safety of exogenous testosterone administration and two studies included genetically predicted testosterone concentrations from large-scale genome-wide associative studies (GWAS). Sixteen studies combined pre- and post-menopausal women, 24 examined post-menopausal women only, one study was conducted in peri-menopausal women, and one study did not specify the menopausal status or age of the participants. Conclusion Discrepancies existed between studies regarding the associations between testosterone and CVD outcomes. This may be due to the heterogeneity of cohorts, differences in endpoint definitions or testosterone analytic techniques. Further, no studies examined pre-menopausal cohorts alone. This highlights the need for well-controlled studies using gold standard testosterone analytic techniques. There is also an unmet requirement for knowledge regarding the effect of testosterone on CVD outcomes in pre-menopausal women.
{"title":"Exploring the relationship between testosterone and cardiovascular disease in women","authors":"S E Alexander, E J Thompson, K A Bolam, E J Howden","doi":"10.1093/eurheartj/ehaf784.4011","DOIUrl":"https://doi.org/10.1093/eurheartj/ehaf784.4011","url":null,"abstract":"Background Cardiovascular disease (CVD) in a leading cause of disease burden and mortality globally. Sex differences exist regarding CVD incidence, but the reasons for this are unclear. Further, women with polycystic ovary syndrome and chronically high levels of testosterone consistently present with worse cardiometabolic profiles than their normoandrogenic counterparts. But whether this is a direct consequence of higher levels of testosterone is unclear. Aim This systematic review aimed to examine current evidence regarding the associations between testosterone and CVD incidence in women. Method MEDLINE complete, Embase and CINAHL complete were systematically searched in September 2024. Articles that investigated the relationship between plasma testosterone concentrations and overt CVD in pre- or post-menopausal women were included. Results Forty-two studies were included in the final review. Outcome measures included the occurrence of a major adverse cardiac event, CVD mortality and morbidity, coronary artery disease/coronary heart disease, atherosclerotic disease, heart failure, myocardial infarction and stroke. Eight studies were case-control studies, 13 were observational cross-sectional studies, 19 were prospective cohort studies and two were retrospective cohort studies. Of the prospective cohort studies, the median follow-up length was 11.5 years (range 2.4-19.2 years). Thirty-nine articles included endogenous plasma testosterone concentrations as an exposure, one article assessed the safety of exogenous testosterone administration and two studies included genetically predicted testosterone concentrations from large-scale genome-wide associative studies (GWAS). Sixteen studies combined pre- and post-menopausal women, 24 examined post-menopausal women only, one study was conducted in peri-menopausal women, and one study did not specify the menopausal status or age of the participants. Conclusion Discrepancies existed between studies regarding the associations between testosterone and CVD outcomes. This may be due to the heterogeneity of cohorts, differences in endpoint definitions or testosterone analytic techniques. Further, no studies examined pre-menopausal cohorts alone. This highlights the need for well-controlled studies using gold standard testosterone analytic techniques. There is also an unmet requirement for knowledge regarding the effect of testosterone on CVD outcomes in pre-menopausal women.","PeriodicalId":11976,"journal":{"name":"European Heart Journal","volume":"69 1","pages":""},"PeriodicalIF":39.3,"publicationDate":"2026-02-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146122189","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2026-02-05DOI: 10.1093/eurheartj/ehaf784.3564
J Yang, M Li, Y Zhang, J Ge
Background Atherosclerotic cardiovascular disease remains the leading cause of morbidity and mortality worldwide, with subclinical carotid plaque serving as a critical determinant of long-term cardiovascular risk. Inflammation plays a central role in the process of atherogenesis. Although the associations between elevated leukocyte counts and cardiovascular risk have been documented, the dynamic nature of inflammatory responses necessitates trajectory-based analysis. Furthermore, the differential impact of leukocyte trajectory patterns on major adverse cardiovascular events (MACE) among individuals with and without carotid plaque remains unclear. Purposes To identify distinct leukocyte trajectory patterns and evaluate their differential associations with MACE risk in individuals stratified by carotid plaque status. Methods Data were obtained from the ARIC study, including participants with complete leukocyte count measurements across three visits (1987–1995) and baseline carotid ultrasound assessment. The primary outcome was defined as MACE, including myocardial infarction, stroke, and cardiovascular death. Follow-up extended from Visit 2 through 2019, censored at first MACE occurrence. Group-based trajectory modeling identified distinct leukocyte trajectory patterns. Multivariable Cox proportional hazard models and survival analyses were applied to assess MACE risk. Results Among 2,342 participants (mean age 54.68 years; 53.97% male), 838 (35.8%) had carotid plaque at baseline. Individuals with plaques were more likely to be smokers, have higher cholesterol, higher glucose, and cardiovascular disease history (P < .05). Four leukocyte trajectories were identified (Figure 1 A and B): low-stable, moderate-increasing, moderate-decreasing, and high-stable. Over a median follow-up of 25.8 years, MACE occurred in 33.3% of participants with carotid plaque and 16.42% without plaque. After adjusting for confounders, the high-stable group showed a significant association with MACE incidence in individuals with carotid plaque (HR 1.43, 1.05-1.73, P = .03), compared to the low-stable pattern (P < .05) (Figure 1C and Figure 2A). Among individuals without carotid plaque, elevated risk with moderate-decreasing (HR 1.67, 1.13-2.83, P < .05) and high-stable (HR 2.02, 1.30-3.14, P < .05) trajectories were observed (Figure 1D). Survival curves revealed temporal divergence: moderate-decreasing trajectories exhibited early risk elevation, while moderate-increasing trajectories manifested late risk (Figure 2B). Conclusion The high-stable leukocyte trajectory pattern reflects sustained inflammatory exposure and serves as an independent risk factor for MACE irrespective of carotid plaque status. Fluctuations of leukocytes predict cardiovascular risk in individuals without carotid plaque. These findings underscore the prognostic value of longitudinal inflammatory monitoring for refined risk stratification.
背景：动脉粥样硬化性心血管疾病仍然是世界范围内发病率和死亡率的主要原因，亚临床颈动脉斑块是长期心血管风险的关键决定因素。炎症在动脉粥样硬化形成过程中起着核心作用。尽管白细胞计数升高与心血管风险之间的关联已被证实，但炎症反应的动态性质需要基于轨迹的分析。此外，在有和没有颈动脉斑块的个体中，白细胞轨迹模式对主要不良心血管事件（MACE）的不同影响仍不清楚。目的确定不同的白细胞轨迹模式，并评估其与颈动脉斑块状态分层个体MACE风险的差异关联。方法从ARIC研究中获得数据，包括三次就诊（1987-1995）的完整白细胞计数测量和基线颈动脉超声评估。主要终点定义为MACE，包括心肌梗死、卒中和心血管死亡。随访时间从第2期延长至2019年，在第一次MACE发生时进行审查。基于组的轨迹建模确定了不同的白细胞轨迹模式。采用多变量Cox比例风险模型和生存分析评估MACE风险。结果在2342名参与者中（平均年龄54.68岁，53.97%为男性），838名（35.8%）在基线时有颈动脉斑块。有斑块的个体更有可能是吸烟者、高胆固醇、高血糖和心血管疾病史（P < 0.05）。确定了四种白细胞轨迹（图1 A和B）：低稳定，中等增加，中等减少和高稳定。在中位25.8年的随访中，颈动脉斑块患者的MACE发生率为33.3%，无斑块患者为16.42%。在调整混杂因素后，与低稳定组相比，高稳定组与颈动脉斑块患者的MACE发生率显著相关（HR 1.43, 1.05-1.73, P = 0.03）。05)（图1C和图2A）。在没有颈动脉斑块的人群中，风险升高后呈中度降低（HR 1.67, 1.13-2.83, P <）。高稳定(HR 2.02, 1.30-3.14, P <；05)观察到轨迹（图1D）。生存曲线显示出时间差异：中度降低的轨迹表现出早期风险升高，而中度增加的轨迹表现出晚期风险（图2B）。结论高稳定的白细胞轨迹模式反映了持续的炎症暴露，是与颈动脉斑块状态无关的MACE的独立危险因素。白细胞的波动预测无颈动脉斑块个体的心血管风险。这些发现强调了纵向炎症监测对精细风险分层的预后价值。
{"title":"Leukocyte trajectories and cardiovascular events in individuals with and without carotid plaque: a community cohort study","authors":"J Yang, M Li, Y Zhang, J Ge","doi":"10.1093/eurheartj/ehaf784.3564","DOIUrl":"https://doi.org/10.1093/eurheartj/ehaf784.3564","url":null,"abstract":"Background Atherosclerotic cardiovascular disease remains the leading cause of morbidity and mortality worldwide, with subclinical carotid plaque serving as a critical determinant of long-term cardiovascular risk. Inflammation plays a central role in the process of atherogenesis. Although the associations between elevated leukocyte counts and cardiovascular risk have been documented, the dynamic nature of inflammatory responses necessitates trajectory-based analysis. Furthermore, the differential impact of leukocyte trajectory patterns on major adverse cardiovascular events (MACE) among individuals with and without carotid plaque remains unclear. Purposes To identify distinct leukocyte trajectory patterns and evaluate their differential associations with MACE risk in individuals stratified by carotid plaque status. Methods Data were obtained from the ARIC study, including participants with complete leukocyte count measurements across three visits (1987–1995) and baseline carotid ultrasound assessment. The primary outcome was defined as MACE, including myocardial infarction, stroke, and cardiovascular death. Follow-up extended from Visit 2 through 2019, censored at first MACE occurrence. Group-based trajectory modeling identified distinct leukocyte trajectory patterns. Multivariable Cox proportional hazard models and survival analyses were applied to assess MACE risk. Results Among 2,342 participants (mean age 54.68 years; 53.97% male), 838 (35.8%) had carotid plaque at baseline. Individuals with plaques were more likely to be smokers, have higher cholesterol, higher glucose, and cardiovascular disease history (P &lt; .05). Four leukocyte trajectories were identified (Figure 1 A and B): low-stable, moderate-increasing, moderate-decreasing, and high-stable. Over a median follow-up of 25.8 years, MACE occurred in 33.3% of participants with carotid plaque and 16.42% without plaque. After adjusting for confounders, the high-stable group showed a significant association with MACE incidence in individuals with carotid plaque (HR 1.43, 1.05-1.73, P = .03), compared to the low-stable pattern (P &lt; .05) (Figure 1C and Figure 2A). Among individuals without carotid plaque, elevated risk with moderate-decreasing (HR 1.67, 1.13-2.83, P &lt; .05) and high-stable (HR 2.02, 1.30-3.14, P &lt; .05) trajectories were observed (Figure 1D). Survival curves revealed temporal divergence: moderate-decreasing trajectories exhibited early risk elevation, while moderate-increasing trajectories manifested late risk (Figure 2B). Conclusion The high-stable leukocyte trajectory pattern reflects sustained inflammatory exposure and serves as an independent risk factor for MACE irrespective of carotid plaque status. Fluctuations of leukocytes predict cardiovascular risk in individuals without carotid plaque. These findings underscore the prognostic value of longitudinal inflammatory monitoring for refined risk stratification.","PeriodicalId":11976,"journal":{"name":"European Heart Journal","volume":"22 1","pages":""},"PeriodicalIF":39.3,"publicationDate":"2026-02-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146122190","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2026-02-05DOI: 10.1093/eurheartj/ehaf784.1987
T Zimmermann, I Strebel, P Lopez-Ayala, S Knecht, L Kirsten, E Kaplan, A T Champetier, F Mahfoud, J Boeddinghaus, C Mueller
Background Accurate and timely diagnosis of acute myocardial infarction (AMI) remains a challenge in clinical practice. While the 12-lead electrocardiogram (ECG) is an essential tool for identifying AMI, manual interpretation by healthcare professionals is skill-dependent and only identifies a minority of patients with clear signs of acute ischemia requiring urgent intervention. Automated ECG analysis using artificial intelligence has the potential to overcome these limitations and enhance patient care. Purpose To train and validate an AI-powered 12-lead ECG-only model for the detection of AMI and its subtypes on three large and high-quality datasets. Methods A convolutional neural network was trained on digital 12-lead ECG data of hospitalized patients (n=178’682, from 10/2021 to 09/2024, Figure 1) with discharge diagnoses as labels. Utilizing transfer-learning, the model was fine-tuned and internally validated on a 80% (n=6’721) / 20% (n=1’645) split of a prospective single-center cohort of adult chest-pain patients presenting to the Emergency Department (ED) (01/2019-01/2022). External validation was performed in a large prospective international multicenter study (04/2006-06/2018) of patients presenting to the ED with suspected AMI (n=3’839). Central adjudication of the final diagnosis (including AMI subtypes) was performed by two independent cardiologists using all clinical information and serial cardiac troponin concentrations according to the Fourth Universal Definition of Myocardial Infarction. The primary outcome was a diagnosis of ST-segment Elevation Myocardial Infarction (STEMI) and Non-STEMI (NSTEMI). Secondary outcomes included the differentiation between NSTEMI type 1 (due to atherothrombosis) and type 2 (due to oxygen supply-demand mismatch) and a diagnosis of Occlusion Myocardial Infarction (OMI). Results Internal validation showed good performance with an AUC of 0.96 (95%-CI 0.94-0.97) for STEMI and 0.82 (95%-CI 0.79-0.85) for NSTEMI. Results were similar in the external validation cohort, with an AUC of 0.97 (95%-CI 0.96-0.98) for STEMI and 0.80 (95%-CI 0.78-0.82) for NSTEMI (Figure 2). The model showed higher discrimination for NSTEMI type 1 than type 2 in both internal (type 1: AUC 0.82, 95%-CI 0.79-0.85, type 2: AUC 0.77, 95%-CI 0.70-0.83) and external validation (type 1: AUC 0.80, 95%-CI 0.78-0.83, type 2: AUC 0.69, 95%-CI 0.64-0.74). Secondary analysis revealed a very high diagnostic accuracy for OMI with an AUC of 0.90 (95%-CI 0.88-0.92). Overall calibration was good for STEMI (intercept -1.21, slope 1.1), NSTEMI (intercept 0.43, slope 0.85) and OMI (intercept 0.02, slope 0.85). Conclusion Our model showed very high diagnostic accuracy for STEMI and OMI and high accuracy for NSTEMI. Based on 12-lead ECG data only, the model more accurately identified NSTEMI type 1 compared to NSTEMI type 2. Whether care guided by our model can improve the early diagnosis of AMI requires prospective evaluation.
{"title":"Training and validation of an ECG-based deep-learning model for the early diagnosis of acute myocardial infarction","authors":"T Zimmermann, I Strebel, P Lopez-Ayala, S Knecht, L Kirsten, E Kaplan, A T Champetier, F Mahfoud, J Boeddinghaus, C Mueller","doi":"10.1093/eurheartj/ehaf784.1987","DOIUrl":"https://doi.org/10.1093/eurheartj/ehaf784.1987","url":null,"abstract":"Background Accurate and timely diagnosis of acute myocardial infarction (AMI) remains a challenge in clinical practice. While the 12-lead electrocardiogram (ECG) is an essential tool for identifying AMI, manual interpretation by healthcare professionals is skill-dependent and only identifies a minority of patients with clear signs of acute ischemia requiring urgent intervention. Automated ECG analysis using artificial intelligence has the potential to overcome these limitations and enhance patient care. Purpose To train and validate an AI-powered 12-lead ECG-only model for the detection of AMI and its subtypes on three large and high-quality datasets. Methods A convolutional neural network was trained on digital 12-lead ECG data of hospitalized patients (n=178’682, from 10/2021 to 09/2024, Figure 1) with discharge diagnoses as labels. Utilizing transfer-learning, the model was fine-tuned and internally validated on a 80% (n=6’721) / 20% (n=1’645) split of a prospective single-center cohort of adult chest-pain patients presenting to the Emergency Department (ED) (01/2019-01/2022). External validation was performed in a large prospective international multicenter study (04/2006-06/2018) of patients presenting to the ED with suspected AMI (n=3’839). Central adjudication of the final diagnosis (including AMI subtypes) was performed by two independent cardiologists using all clinical information and serial cardiac troponin concentrations according to the Fourth Universal Definition of Myocardial Infarction. The primary outcome was a diagnosis of ST-segment Elevation Myocardial Infarction (STEMI) and Non-STEMI (NSTEMI). Secondary outcomes included the differentiation between NSTEMI type 1 (due to atherothrombosis) and type 2 (due to oxygen supply-demand mismatch) and a diagnosis of Occlusion Myocardial Infarction (OMI). Results Internal validation showed good performance with an AUC of 0.96 (95%-CI 0.94-0.97) for STEMI and 0.82 (95%-CI 0.79-0.85) for NSTEMI. Results were similar in the external validation cohort, with an AUC of 0.97 (95%-CI 0.96-0.98) for STEMI and 0.80 (95%-CI 0.78-0.82) for NSTEMI (Figure 2). The model showed higher discrimination for NSTEMI type 1 than type 2 in both internal (type 1: AUC 0.82, 95%-CI 0.79-0.85, type 2: AUC 0.77, 95%-CI 0.70-0.83) and external validation (type 1: AUC 0.80, 95%-CI 0.78-0.83, type 2: AUC 0.69, 95%-CI 0.64-0.74). Secondary analysis revealed a very high diagnostic accuracy for OMI with an AUC of 0.90 (95%-CI 0.88-0.92). Overall calibration was good for STEMI (intercept -1.21, slope 1.1), NSTEMI (intercept 0.43, slope 0.85) and OMI (intercept 0.02, slope 0.85). Conclusion Our model showed very high diagnostic accuracy for STEMI and OMI and high accuracy for NSTEMI. Based on 12-lead ECG data only, the model more accurately identified NSTEMI type 1 compared to NSTEMI type 2. Whether care guided by our model can improve the early diagnosis of AMI requires prospective evaluation.","PeriodicalId":11976,"journal":{"name":"European Heart Journal","volume":"47 1","pages":""},"PeriodicalIF":39.3,"publicationDate":"2026-02-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146122373","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
京公网安备 11010802042870号
京ICP备2023020795号-1
扫码关注我们
求助内容：
应助结果提醒方式：