Evaluating convolutional neural network-enhanced electrocardiography for hypertrophic cardiomyopathy detection in a specialized cardiovascular setting.

The efficacy of convolutional neural network (CNN)-enhanced electrocardiography (ECG) in detecting hypertrophic cardiomyopathy (HCM) and dilated HCM (dHCM) remains uncertain in real-world applications. This retrospective study analyzed data from 19,170 patients (including 140 HCM or dHCM) in the Shinken Database (2010-2017). We evaluated the sensitivity, positive predictive rate (PPR), and F1 score of CNN-enhanced ECG in a ''basic diagnosis'' model (total disease label) and a ''comprehensive diagnosis'' model (including disease subtypes). Using all-lead ECG in the "basic diagnosis" model, we observed a sensitivity of 76%, PPR of 2.9%, and F1 score of 0.056. These metrics improved in cases with a diagnostic probability of ≥ 0.9 and left ventricular hypertrophy (LVH) on ECG: 100% sensitivity, 8.6% PPR, and 0.158 F1 score. The ''comprehensive diagnosis'' model further enhanced these figures to 100%, 13.0%, and 0.230, respectively. Performance was broadly consistent across CNN models using different lead configurations, particularly when including leads viewing the lateral walls. While the precision of CNN models in detecting HCM or dHCM in real-world settings is initially low, it improves by targeting specific patient groups and integrating disease subtype models. The use of ECGs with fewer leads, especially those involving the lateral walls, appears comparably effective.

Drug-eluting stents versus coronary artery bypass grafting for left-main coronary artery disease.

OBJECTIVES: To compare follow-up outcomes after percutaneous coronary intervention with drug-eluting stents (DES-PCI) versus coronary artery bypass grafting (CABG) for left-main coronary artery disease (LMCAD), we performed a meta-analysis of randomized controlled trials (RCTs) and observational studies with propensity-score analysis. METHODS: MEDLINE, EMBASE, and the Cochrane Central Register of Controlled Trials were searched through November 2016. Eligible studies were RCTs or observational studies with propensity-score analysis of DES-PCI versus CABG enrolling patients with LMCAD and reporting ≥ 6-month mortality, myocardial infarction (MI), stroke, or repeat revascularization (RRV). Study-specific estimates were combined using inverse variance-weighted averages of logarithmic hazard ratios (HRs) in the random-effects model. RESULTS: We identified 5 RCTs and 17 observational studies with propensity-score analysis enrolling a total of 12,387 patients. Pooled analysis demonstrated a significant increase in a composite of death, MI, and RRV (with/without stroke) after DES-PCI (HR, 1.42; P < 0.00001); no significant difference in a composite of death and MI (with/without stroke); no significant differences in mortality and stroke; a strong trend toward an increase in MI after DES-PCI (HR, 1.44; P = 0.05); and significant increases in any (HR, 1.86; P < 0.00001), target-vessel (HR, 3.28; P < 0.00001), and target-lesion RRV (HR, 2.26; P = 0.003) after DES-PCI. CONCLUSIONS: When compared with CABG, DES-PCI for LMCAD was associated with increases in RRV and the composite of death, MI, and RRV (with/without stroke), despite no differences in mortality, MI, stroke, and the composite of death and MI (with/without stroke).

A meta-analysis of weekend admission and surgery for aortic rupture and dissection.

We performed a meta-analysis to determine whether weekend admission and surgery for ruptured abdominal/thoracic aortic aneurysm (RAAA/RTAA) and acute aortic dissection (AAD) is associated with increased mortality. MEDLINE and EMBASE were searched from January 1946 to December 2016 using PubMed and OVID. Eligible studies were prospective or retrospective, comparative or cohort studies enrolling patients admitting or undergoing surgery for RAAA/RTAA/AAD and reporting mortality after weekend (including holiday) versus weekday admission/surgery. Our search identified 11 studies including a total of 166,195 patients. A pooled analysis of 13 adjusted odds ratios (ORs), one adjusted hazard ratio, and one unadjusted OR from all 11 studies demonstrated a statistically significant 32% increase in mortality with weekend admission/surgery (OR, 1.32; 95% confidence interval (CI), 1.20 to 1.45; p < 0.00001). Despite possible publication bias disadvantageous to weekend admission/surgery based on funnel plot asymmetry, adjustment for the asymmetry using the trim-and-fill method did not alter the significant association of weekend admission/surgery with increased mortality (OR, 1.21; 95% CI, 1.09 to 1.34; p = 0.0006). In conclusion, weekend admission/surgery for ruptured abdominal/thoracic aortic aneurysm and acute aortic dissection (AAD) may be associated with increased mortality.

Perioperative depression or anxiety and postoperative mortality in cardiac surgery: a systematic review and meta-analysis.

We performed a systematic review and meta-analysis to determine whether perioperative depression and anxiety are associated with increased postoperative mortality in patients undergoing cardiac surgery. MEDLINE and EMBASE were searched through January 2017 using PubMed and OVID, to identify observational studies enrolling patients undergoing cardiac surgery and reporting relative risk estimates (RREs) (including odds, hazard, or mortality ratios) of short term (30 days or in-hospital) and/or late all-cause mortality for patients with versus without perioperative depression or anxiety. Study-specific estimates were combined using inverse variance-weighted averages of logarithmic RREs in the random-effects models. Our search identified 16 eligible studies. In total, the present meta-analysis included data on 236,595 patients undergoing cardiac surgery. Pooled analysis demonstrated that perioperative depression was significantly associated with increased both postoperative early (RRE, 1.44; 95% confidence interval [CI] 1.01-2.05; p = 0.05) and late mortality (RRE, 1.44; 95% CI 1.24-1.67; p < 0.0001), and that perioperative anxiety significantly correlated with increased postoperative late mortality (RRE, 1.81; 95% CI 1.20-2.72; p = 0.004). The relation between anxiety and early mortality was reported in only one study and not statistically significant. In the association of depression with late mortality, there was no evidence of significant publication bias and meta-regression indicated that the effects of depression are not modulated by the duration of follow-up. In conclusion, perioperative depression and anxiety may be associated with increased postoperative mortality in patients undergoing cardiac surgery.

Association of Hypertension with Abdominal Aortic Aneurysm Expansion.

BACKGROUND: Hypertension is positively associated with abdominal aortic aneurysm (AAA) presence, which supports a hypothesis that hypertension may also be positively associated with AAA expansion. To determine whether hypertension is associated with AAA expansion, we reviewed currently available studies with a systematic literature search and meta-analytic estimate. METHODS: Databases including MEDLINE and EMBASE were searched through July 2015 using Web-based search engines (PubMed and OVID). Studies considered for inclusion met the following criteria: the study population was AAA patients with and without hypertension, and outcomes included data regarding AAA expansion. For each study, expansion rates in both the hypertensive and nonhypertensive groups were used to generate standardized mean differences (SMDs) and 95% confidence intervals (CIs). RESULTS: Of 614 potentially relevant publications screened initially, we identified 20 eligible studies including data on 6,619 AAA patients. No individual study indicated a statistically significant (positive or negative) association of hypertension with AAA expansion rates. A pooled analysis of all the 20 studies demonstrated that hypertension was not associated with AAA expansion rates in the fixed-effect model (SMD 0.03, 95% CI -0.01 to 0.17, P = 0.19). There was no evidence of significant publication bias. CONCLUSIONS: Hypertension is not associated with AAA expansion. Further investigations would be required to elucidate why hypertension is not associated with AAA expansion despite its positive association with AAA presence.

Associations of coronary and peripheral artery disease with presence, expansion, and rupture of abdominal aortic aneurysm - a grin without a cat!

Both coronary and peripheral artery disease are representative atherosclerotic diseases, which are also known to be positively associated with presence of abdominal aortic aneurysm. It is still controversial, however, whether coronary and peripheral artery disease are positively associated with expansion and rupture as well as presence of abdominal aortic aneurysm. In the present article, we overviewed epidemiological evidence, i. e. meta-analyses, regarding the associations of coronary and peripheral artery disease with presence, expansion, and rupture of abdominal aortic aneurysm through a systematic literature search. Our exhaustive search identified seven meta-analyses, which suggest that both coronary and peripheral artery disease are positively associated with presence of abdominal aortic aneurysm, may be negatively associated with expansion of abdominal aortic aneurysm, and might be unassociated with rupture of abdominal aortic aneurysm.

A meta-analysis of monthly variation in occurrence of abdominal aortic aneurysm rupture.

BACKGROUND: We performed a meta-analysis to assess the presence of an annual rhythmic variability of ruptured abdominal aortic aneurysm (RAAA) onset. PATIENTS AND METHODS: Eligible studies were observational studies enrolling patients with RAAA and reporting monthly incidence of RAAA. Study-specific estimates, i. e. monthly incidence of RAAA, were combined using the random-effects model. Chronobiological analysis was performed by applying a partial Fourier series to pooled monthly incidence by using the weighted least-squares method. RESULTS: We identified 14 eligible studies enrolling a total of 3,798 patients with RAAA. Pooled monthly incidence of RAAA was 8.7 % in January, 7.7 % in February, 8.7 % in March, 7.3 % in April, 7.8 % in May, 7.2 % in June, 7.0 % in July, 7.0 % in August, 8.1 % in September, 8.8 % in October, 8.4 % in November, and 8.3 % in December. Chronobiological analysis identified a significant (p = 0.0020) annual pattern in the occurrence of RAAA with a peak in December to January and a nadir in June to July. Pooled analysis demonstrated significantly more incidence in December than in June (p = 0.03) as well as in January than in July (p = 0.05). CONCLUSIONS: A significant annual pattern in the occurrence of RAAA with a peak in December to January and a nadir in June to July was identified with significantly more incidence in December than in June and in January than in July.

Simple renal cyst and abdominal aortic aneurysm.

OBJECTIVE: To summarize the association of simple renal cyst (SRC) with abdominal aortic aneurysm (AAA), we reviewed currently available clinical studies with a systematic literature search and meta-analytic evaluation. METHODS: To identify all case-control studies evaluating the association of SRC with AAA, databases including MEDLINE and Embase were searched through April 2015 using web-based search engines (PubMed and Ovid). For each study, data regarding SRC prevalence in both the AAA and control groups were used to generate unadjusted odds ratios (ORs) and 95% confidence intervals. When an adjusted OR (by the use of multivariable logistic regression) was available, we preferentially abstracted the adjusted OR rather than an unadjusted OR. RESULTS: Of 139 potentially relevant articles screened initially, 5 eligible case-control studies enrolling a total of 2897 participants were identified and included. A pooled analysis of seven estimates from the five studies demonstrated a statistically significant 2.54-fold prevalence of SRC in patients with AAA relative to subjects without AAA (OR, 2.54; 95% confidence interval, 1.93-3.34; P < .00001). CONCLUSIONS: Our meta-analytic evaluation demonstrated 2.5-fold prevalence of SRC in patients with AAA relative to subjects without AAA, which suggests that SRC is associated with AAA.

Association of peripheral artery disease with abdominal aortic aneurysm growth.

OBJECTIVE: To determine whether an association of peripheral artery disease (PAD) with abdominal aortic aneurysm (AAA) growth is positive, none, or negative, a meta-analysis of all available studies was performed through a systematic literature search. METHODS: MEDLINE and Embase databases were searched until July 2015 by the use of PubMed and Ovid Web-based search engines. The search terms were enlargement, expansion, growth, or progression; and abdominal aortic aneurysm. Studies that fulfilled the following criteria were included: (1) AAA patients with PAD and those without PAD as the population and (2) data of AAA growth as the outcomes. By means of growth rates in patients with and without PAD for each study, we generated standardized mean differences (SMDs) and their 95% confidence intervals (CIs). RESULTS: Of 612 initially screened publications that were potentially relevant, 11 eligible studies that reported the correlation between PAD and AAA growth were identified, and a total of 4573 patients with AAA were included. A pooled analysis of the 11 studies demonstrated a statistically significant association of PAD with lower growth rates of AAA (SMD, -0.18; 95% CI, -0.25 to -0.11; P < .00001). In addition, separately combining seven adjusted effect estimates did not substantively alter the pooled estimate (SMD, -0.17; 95% CI, -0.25 to -0.09; P < .0001). No significant publication bias was observed. CONCLUSIONS: PAD is likely negatively associated with AAA growth. Further research is required to explain why PAD negatively correlates with AAA growth despite a positive correlation with AAA presence.

Coronary artery disease and abdominal aortic aneurysm growth.

To determine whether coronary artery disease (CAD) is associated with abdominal aortic aneurysm (AAA) growth, we performed a meta-analysis of currently available studies. Databases including MEDLINE and EMBASE were searched through October 2015 using PubMed and OVID. Search terms included enlargement, expansion, growth, or progression; rate or rates; and abdominal aortic aneurysm Studies considered for inclusion met the following criteria: the design was unrestricted; the study population was AAA patients with and without CAD; and outcomes included data regarding AAA growth. For each study, growth rates in both the CAD and non-CAD groups were used to generate standardized mean differences (SMDs) and 95% confidence intervals (CIs). Of 664 potentially relevant publications screened initially, we identified 20 eligible studies including data on a total of 7238 AAA patients. A pooled analysis of all 20 studies demonstrated a statistically significant association of CAD with slower AAA growth rates (i.e. a significantly negative association of CAD with AAA growth) in the fixed-effect model (SMD, -0.06 [-0.0592]; 95% CI, -0.12 [-0.1157] to -0.00 [-0.0027]; p = 0.04). There was minimal between-study heterogeneity (p = 0.16) and a statistically non-significant association of CAD with slower AAA growth rates (i.e. a non-significantly negative association of CAD with AAA growth) in the pooled result from random-effects modeling (SMD, -0.06; 95% CI, -0.13 to 0.01; p = 0.12). In conclusion, CAD may be negatively associated with AAA growth.

No association of chronic obstructive pulmonary disease with abdominal aortic aneurysm growth.

Chronic obstructive pulmonary disease (COPD) is independently and positively associated with abdominal aortic aneurysm (AAA) presence. The aim of the present article was to determine whether COPD is associated with AAA growth. We reviewed currently available studies with a systematic literature search and meta-analytic estimate. Databases including MEDLINE and EMBASE were searched through July 2015 using Web-based search engines (PubMed and OVID). Studies considered for inclusion met the following criteria: the study population was AAA patients with and without COPD; and outcomes included data regarding AAA growth. For each study, growth rates in both the COPD and non-COPD groups were used to generate standardized mean differences (SMDs) and 95 % confidence intervals (CIs). Of 614 potentially relevant publications screened initially, we identified 10 eligible studies including data on a total of 2663 AAA patients. None of them demonstrated a statistically significant (positive or negative) association of COPD with AAA growth rates. A pooled analysis demonstrated that COPD is not associated with AAA growth rates (SMD, 0.04; 95 % CI, -0.07 to 0.15; p = 0.50). There was no evidence of significant publication bias. In conclusion, we found that COPD is not associated with AAA growth. Further investigations would be required to elucidate why COPD is not associated with AAA growth despite its positive association with AAA presence.

A Meta-Analysis of the Association of Chronic Obstructive Pulmonary Disease with Abdominal Aortic Aneurysm Presence.

BACKGROUND: Several case-control and population-based abdominal aortic aneurysm (AAA) screening studies have reported inconclusive results of the association of chronic obstructive pulmonary disease (COPD) with AAA presence. To determine whether COPD is associated with AAA presence, we performed a meta-analysis of contemporary clinical studies. METHODS: To identify all contemporary case-control and population-based AAA screening studies evaluating the association of COPD with AAA presence, databases including MEDLINE and EMBASE were searched from January 2000 to May 2015 using Web-based search engines (PubMed and OVID). An adjusted odds ratio (OR) and 95% confidence intervals (CI) for COPD or AAA presence (using multivariable logistic regression) were abstracted from each individual study. We took an OR for AAA presence to be representative of an OR for COPD presence. RESULTS: Of 159 potentially relevant articles screened initially, there were 7 case-control and 4 population-based AAA screening studies that met eligibility requirements and were included. Pooled analysis of all the 11 studies (14 estimates, 155,731 participants), 7 case-control studies (4171 participants), and 4 population-based AAA screening studies (7 estimates, 151,560 participants) respectively demonstrated a statistically significant 1.78-fold (OR 1.78, 95% CI 1.38-2.30, P < 0.00001), 3.05-fold (OR 3.05, 95% CI 1.44-6.49, P = 0.004), and 1.24-fold (OR 1.24, 95% CI 1.04-1.48, P = 0.02) increased prevalence/incidence of COPD in patients with AAA relative to subjects without AAA (i.e., a statistically significant 1.78-, 3.05-, and 1.24-fold increased prevalence/incidence of AAA in patients with COPD relative to subjects without COPD) (P for subgroup differences = 0.02). CONCLUSION: The present meta-analysis demonstrated 1.8-fold increased prevalence/incidence of COPD in patients with AAA relative to subjects without AAA (i.e., 1.8-fold increased prevalence/incidence of AAA in patients with COPD relative to subjects without COPD), which suggests that COPD is associated with AAA presence.

A Meta-Analysis of Sutureless or Rapid-Deployment Aortic Valve Replacement.

Objective To summarize the safety of sutureless or rapid-deployment aortic valve replacement (AVR), we performed a systematic review and meta-analysis of single-arm studies. Methods MEDLINE and EMBASE were searched through December 2014. Studies considered for inclusion met the following criteria: the design was a single-arm study enrolling ≥50 participants; the study population consisted of patients undergoing sutureless/rapid-deployment AVR; and main outcomes included early (in-hospital or 30-day) mortality and/or overall survival. Results Of 250 potentially relevant articles screened initially, 11 eligible studies enrolling a total of 2,066 patients were identified and included. The Enable, Intuity, and Perceval bioprostheses were used in three, two, and six studies, respectively. Mean age of patients was 77.6 years, and 56.9% of patients were women. Mean logistic European System for Cardiac Operative Risk Evaluation I and II were 10.5 and 7.4%, respectively. Aortic cross-clamp times in overall patients, patients undergoing isolated AVR, those undergoing AVR with any concomitant procedures, and those undergoing AVR with coronary artery bypass grafting were 44.7, 41.9, 56.2, and 51.3 minutes, respectively. Arithmetic mean of early mortality was 2.6%, and fixed-effects combined early mortality was 3.2% (95% confidence interval, 2.5-4.2%). Arithmetic mean of 1-year survival was 89.7%, and fixed-effects combined 1-year mortality was 10.4% (9.0-12.1%). Conclusion Sutureless/rapid-deployment AVR is feasible and safe with approximate 3 and 10% of early and 1-year mortality, respectively. Large-size randomized controlled trials, however, are needed to determine whether sutureless/rapid-deployment AVR improves mortality compared with conventional AVR.

The association between body mass index and abdominal aortic aneurysm growth: a systematic review.

Diabetes, a state of relative insulin resistance, is negatively associated with both the presence and growth abdominal aortic aneurysms (AAA), which could suggest a protective role of obesity against AAA presence or growth. A recent meta-analysis demonstrated a trend toward a positive, though statistically non-significant, association between body mass index (BMI) and the presence of AAA. With respect to the association between obesity and AAA growth, however, the evidence had been very limited. To determine whether obesity (or BMI) is associated with AAA growth, we reviewed currently available studies with a systematic literature search. Our comprehensive search identified seven eligible studies reporting the association of BMI and AAA growth rates, which included data on a total of 3,768 AAA patients. All seven identified studies demonstrated no association between BMI and AAA growth. Despite a trend toward a positive association between BMI and AAA presence, the reason why BMI is not associated with AAA growth (suggested in the present review) is unclear. A discrepancy between associated comorbidities (coronary artery disease, peripheral artery disease, and chronic obstructive pulmonary disease) and AAA presence and between the same comorbidities and AAA growth, however, could be identified. Further investigations are required to elucidate why BMI is not associated with AAA growth despite the trend for a positive association with AAA presence.

Effects of rosuvastatin versus atorvastatin on small dense low-density lipoprotein: a meta-analysis of randomized trials.

In addition to their high-intensity effects on the reduction in low-density lipoprotein (LDL) levels, rosuvastatin and atorvastatin would be expected to also reduce small dense LDL (sdLDL) levels. To determine which reduces sdLDL levels more, we performed the first meta-analysis and meta-regression of randomized head-to-head trials of rosuvastatin versus atorvastatin therapy. MEDLINE, EMBASE, and the Cochrane Central Register of Controlled Trials were searched through April 2012. Eligible studies were prospective, randomized controlled trials of rosuvastatin versus atorvastatin therapy reporting final sdLDL (directly measured or calculated) levels as an outcome. For each study, data regarding final sdLDL levels in both the rosuvastatin and atorvastatin groups were used to generate mean differences (MD) and 95 % confidence intervals (CI). Meta-regression analysis was performed to determine whether the effects of rosuvastatin therapy were modulated by the prespecified factors. Of 159 potentially relevant articles screened initially, 28 reports of randomized trials enrolling a total of 7802 patients were included. Pooled analysis suggested a significant reduction in final sdLDL levels among patients randomized to rosuvastatin versus atorvastatin therapy (MD, -1.56 mg/dl; 95 % CI, -2.30 to -0.83 mg/dl; P < 0.0001). The meta-regression coefficients were statistically significant for the baseline LDL/sdLDL level and the difference in LDL changes between the two groups. In conclusion, rosuvastatin rather than atorvastatin therapy is likely more effective in reduction of sdLDL levels. It should be further investigated whether the reduction in sdLDL levels implies overt clinical benefits of rosuvastatin over atorvastatin.

Circulating interleukin-6 levels are associated with abdominal aortic aneurysm presence: a meta-analysis and meta-regression of case-control studies.

BACKGROUND: In a number of relatively small-size studies, investigators have measured and compared circulating interleukin-6 (IL-6) levels in cases with abdominal aortic aneurysm (AAA) and controls without AAA to assess its possible role in the pathogenesis or progression of AAA. To summarize the present evidence for an association between circulating IL-6 levels and AAA presence, we performed a meta-analysis of case-control studies that compared circulating IL-6 levels between patients with AAA and subjects without AAA. METHODS: MEDLINE and EMBASE were searched through December 2013 using Web-based search engines (PubMed and OVID). Eligible studies were case-control studies of patients with AAA and subjects without AAA reporting circulating IL-6 levels. For each study, data regarding plasma or serum IL-6 levels in both the AAA and control groups were used to generate standardized mean differences (SMDs) and 95% confidence intervals (CIs). Mixed-effects (unrestricted maximum likelihood) meta-regression analyses were performed to determine whether the differences in circulating IL-6 levels were modulated by the prespecified factors. RESULTS: Our search identified 13 eligible studies enrolling a total of 1,029 cases with AAA and 924 controls without AAA. Pooled analysis of the 13 studies demonstrated significantly greater circulating IL-6 levels in the AAA group than those in the control group (random-effects SMD 0.59; 95% CI 0.37-0.80; P for effect < 0.00001; P for heterogeneity < 0.0000). The meta-regression coefficient for the mean age (P = 0.10196) and the proportion of current smokers (P = 0.29893) was not statistically significant. That for the mean AAA diameter, however, was significantly positive (coefficient 0.02789; 95% CI 0.00778-0.04800; P = 0.00657), and that for the proportion of men was significantly negative (coefficient -0.01823; 95% CI -0.03202 to -0.00445; P = 0.00952). CONCLUSIONS: Circulating IL-6 levels are greater in patients with AAA than those in subjects without AAA, which suggest that greater circulating IL-6 levels are associated with AAA presence. As the mean AAA diameter and the proportion of men increase, circulating IL-6 levels in patients with AAA are respectively more and less greater than those in controls without AAA.

Identifying patients with acute aortic dissection using an electrocardiogram with convolutional neural network.

Background: The potential of utilizing artificial intelligence with electrocardiography (ECG) for initial screening of aortic dissection (AD) is promising. However, achieving a high positive predictive rate (PPR) remains challenging. Methods and results: This retrospective analysis of a single-center, prospective cohort study (Shinken Database 2010-2017, N = 19,170) used digital 12-lead ECGs from initial patient visits. We assessed a convolutional neural network (CNN) model's performance for AD detection with eight-lead (I, II, and V1-6), single-lead, and double-lead (I, II) ECGs via five-fold cross-validation. The mean age was 63.5 ± 12.5 years for the AD group (n = 147) and 58.1 ± 15.7 years for the non-AD group (n = 19,023). The CNN model achieved an area under the curve (AUC) of 0.936 (standard deviation [SD]: 0.023) for AD detection with eight-lead ECGs. In the entire cohort, the PPR was 7 %, with 126 out of 147 AD cases correctly diagnosed (sensitivity 86 %). When applied to patients with D-dimer levels ≥1 µg/dL and a history of hypertension, the PPR increased to 35 %, with 113 AD cases correctly identified (sensitivity 86 %). The single V1 lead displayed the highest diagnostic performance (AUC: 0.933, SD: 0.03), with PPR improvement from 8 % to 38 % within the same population. Conclusions: Our CNN model using ECG data for AD detection achieved an over 30% PPR when applied to patients with elevated D-dimer levels and hypertension history while maintaining sensitivity. A similar level of performance was observed with a single-lead V1 ECG in the CNN model.

Lead-Specific Performance for Atrial Fibrillation Detection in Convolutional Neural Network Models Using Sinus Rhythm Electrocardiography.

Background: We developed a convolutional neural network (CNN) model to detect atrial fibrillation (AF) using the sinus rhythm ECG (SR-ECG). However, the diagnostic performance of the CNN model based on different ECG leads remains unclear. Methods and Results: In this retrospective analysis of a single-center, prospective cohort study, we identified 616 AF cases and 3,412 SR cases for the modeling dataset among new patients (n=19,170). The modeling dataset included SR-ECGs obtained within 31 days from AF-ECGs in AF cases and SR cases with follow-up ≥1,095 days. We evaluated the CNN model's performance for AF detection using 8-lead (I, II, and V1-6), single-lead, and double-lead ECGs through 5-fold cross-validation. The CNN model achieved an area under the curve (AUC) of 0.872 (95% confidence interval (CI): 0.856-0.888) and an odds ratio of 15.24 (95% CI: 12.42-18.72) for AF detection using the eight-lead ECG. Among the single-lead and double-lead ECGs, the double-lead ECG using leads I and V1 yielded an AUC of 0.871 (95% CI: 0.856-0.886) with an odds ratio of 14.34 (95% CI: 11.64-17.67). Conclusions: We assessed the performance of a CNN model for detecting AF using eight-lead, single-lead, and double-lead SR-ECGs. The model's performance with a double-lead (I, V1) ECG was comparable to that of the 8-lead ECG, suggesting its potential as an alternative for AF screening using SR-ECG.

Antiarrhythmic effect of cardiac resynchronization therapy with triple-site biventricular stimulation.

AIMS: The antiarrhythmic effect of triple-site biventricular stimulation (Tri-V) is poorly understood. This study aims to evaluate the effect of cardiac resynchronization therapy (CRT) on ventricular arrhythmia (VA) with Tri-V using a single right ventricular (RV) and double left ventricular (LV) lead. METHODS AND RESULTS: Over a period of 3.5 years, 58 consecutive patients with New York Heart Association class II-IV heart failure, an LV ejection fraction of ≤ 0.35, and a QRS interval of ≥ 120 ms were enrolled. Acute haemodynamic responses to dual-site biventricular stimulation (Bi-V) and Tri-V were evaluated by assigning patients to a Bi-V or Tri-V group. Electrocardiogram parameters [QT interval, JT interval, and transmural dispersion of repolarization (TDR)] were measured over time after CRT. Spontaneous VA detected by telemetry was reviewed and confirmed. During a mean follow-up of 481 days after implantation, VA occurred in 2 of 22 patients in the Tri-V group and 14 of 36 patients in Bi-V group. Triple-site biventricular stimulation was thus associated with a decreased VA risk (P = 0.044). Multivariate Cox analysis showed that Tri-V pacing prevented arrhythmic events as compared with Bi-V pacing (hazard ratio, 0.13; 95% confidence interval, 0.029-0.610; P = 0.009). Ventricular repolarization indices at 6 months were significantly shortened in Tri-V compared with Bi-V (QTc, -23.6 vs. -14.1%, P = 0.008; JTc, -21.4 vs. -7.7%, P = 0.005; TDRc, -39.9 vs. -17.0%, P < 0.001). CONCLUSION: Compared with Bi-V, Tri-V reduced VA during long-term follow-up. Improvements in repolarization parameters may result in antiarrhythmic effects.

Cardiovascular events and artificial intelligence-predicted age using 12-lead electrocardiograms.

Background: There is increasing evidence that 12-lead electrocardiograms (ECG) can be used to predict biological age, which is associated with cardiovascular events. However, the utility of artificial intelligence (AI)-predicted age using ECGs remains unclear. Methods: Using a single-center database, we developed an AI-enabled ECG using 17 042 sinus rhythm ECGs (SR-ECG) to predict chronological age (CA) with a convolutional neural network that yields AI-predicted age. Using the 5-fold cross validation method, AI-predicted age deriving from the test dataset was yielded for all ECGs. The incidence by AgeDiff and the areas under the curve by receiver operating characteristic curve with AI-predicted age for cardiovascular events were analyzed. Results: During the mean follow-up period of 460.1 days, there were 543 cardiovascular events. The annualized incidence of cardiovascular events was 2.24 %, 2.44 %, and 3.01 %/year for patients with AgeDiff < -6, -6 to ≤6, and >6 years, respectively. The areas under the curve for cardiovascular events with CA and AI-predicted age, respectively, were 0.673 and 0.679 (Delong's test, P = 0.388) for all patients; 0.642 and 0.700 (P = 0.003) for younger patients (CA < 60 years); and 0.584 and 0.570 (P = 0.268) for older patients (CA ≥ 60 years). Conclusions: AI-predicted age using 12-lead ECGs showed superiority in predicting cardiovascular events compared with CA in younger patients, but not in older patients.