このページは自動翻訳されたものであり、翻訳の正確性は保証されていません。を参照してください。 英語版 ソーステキスト用。

Artificial Intelligence-enabled Management of Atrial Fibrillation(ABLE-AF Trial)

2026年11月9日 更新者:Jong-Il Choi、Korea University Anam Hospital

Artificial Intelligence-enabled Management of Atrial Fibrillation: a Randomized Clinical Trial (ABLE-AF Trial)

Atrial fibrillation (AF) is the most common type of irregular heartbeat in adults. People with AF may also have other heart problems, including reduced heart pumping function or diseases affecting the heart muscle or heart valves. Identifying these problems early is important because the condition of the heart can affect treatment decisions and long-term care.

An ultrasound examination of the heart, called transthoracic echocardiography (TTE) or an echocardiogram, is commonly used to check how well the heart pumps blood and to look for structural problems, such as heart valve disease. However, an echocardiogram requires special equipment and trained healthcare professionals to perform and interpret the examination. Because of these requirements, it may not always be available when needed. This can be particularly challenging in situations where heart function needs to be evaluated quickly, such as when a patient with AF has a very fast heart rate. Delays may also occur in outpatient settings because of limited availability of echocardiography.

Artificial intelligence (AI) technology may provide an additional way to identify patients who are more likely to have heart problems and who may benefit from further evaluation. A standard 12-lead electrocardiogram (ECG) is a simple and widely available test that records the electrical activity of the heart. AI-based computer programs can analyze information contained in an ECG that may not be easily recognized through routine interpretation.

This study will use two AI-based ECG programs, called AiTiALVSD and AiTiALVDD. AiTiALVSD analyzes a standard 12-lead ECG to identify patients who may have reduced heart pumping function. Previous studies have evaluated this program in different clinical settings, including patients with AF and a fast heart rate.

AiTiALVDD analyzes the ECG for signs that may be associated with difficulty in the heart relaxing and filling with blood. Problems with heart relaxation and filling may occur in several types of heart disease, including heart valve disease, abnormal thickening of the heart muscle, heart disease related to high blood pressure, diseases affecting the heart muscle, and coronary artery disease. Importantly, some patients may have these conditions even when the heart's overall pumping function is normal. Therefore, evaluating only pumping function may not identify all patients who have important underlying heart disease.

By using AiTiALVSD and AiTiALVDD together, this study aims to evaluate a broader AI-based approach to identifying heart problems in patients with AF. AiTiALVSD may help identify patients who are more likely to have reduced heart pumping function, while AiTiALVDD may provide additional information about possible problems with heart relaxation and filling. A high-risk result from AiTiALVDD may serve as a signal that further heart evaluation should be considered, including evaluation for significant heart valve disease or other underlying structural heart conditions.

The purpose of this study is to evaluate whether using these two AI-based ECG programs together can help identify patients with AF who may have previously unrecognized problems with heart function or underlying heart disease. Patients identified by the AI analysis as being at higher risk may receive further evaluation as appropriate to determine whether additional heart disease is present.

The study will also evaluate the clinical usefulness of this AI-based ECG approach in routine care. The goal is not to replace an echocardiogram or a physician's clinical assessment. Instead, AI-based ECG analysis may serve as an additional screening tool to help healthcare professionals identify patients who may benefit from further cardiac testing.

By using information from a routinely performed ECG, this approach may help support earlier and more efficient evaluation of heart function and underlying heart disease in patients with AF. The results of this study will help determine whether combining AiTiALVSD and AiTiALVDD can provide useful information for deciding which patients may need additional cardiac evaluation and whether this approach can improve the process of identifying clinically important heart conditions.

調査の概要

研究の種類

介入

入学 (推定)

1006

段階

  • 適用できない

連絡先と場所

このセクションには、調査を実施する担当者の連絡先の詳細と、この調査が実施されている場所に関する情報が記載されています。

研究連絡先

研究連絡先のバックアップ

研究場所

      • Seoul、韓国、02841
        • Korea University Anam Hospital
        • コンタクト:
          • Jong-Il Choi
          • 電話番号:82-10-2122-5476

参加基準

研究者は、適格基準と呼ばれる特定の説明に適合する人を探します。これらの基準のいくつかの例は、人の一般的な健康状態または以前の治療です。

適格基準

就学可能な年齢

18年歳以上 (大人、高齢者)

健康ボランティアの受け入れ

いいえ

受講資格のある性別

全て

説明

Inclusion Criteria

  • Atrial fibrillation (AF) with rapid ventricular response confirmed on the day of screening or within 90 days before screening.
  • Rapid ventricular response is defined as (i) a ventricular rate ≥110 bpm or (ii) a ventricular rate ≥100 bpm with related symptoms.
  • AF can be confirmed by any guideline-accepted electrocardiographic (ECG) method, including 12-lead ECG, Holter ECG, or single-lead ECG (e.g., patch ECG or smartwatch).
  • AF must be documented for at least 10 seconds on a 12-lead ECG or at least 30 seconds on Holter or single-lead ECG.

Exclusion Criteria

  • Left ventricular ejection fraction (LV EF) assessed by transthoracic echocardiography (TTE) within the past 6 months.
  • If TTE was performed at another institution within the past 6 months, this exclusion criterion applies only when objective imaging data confirming LV EF are available.
  • Hemodynamically unstable patients requiring immediate direct current cardioversion.
  • Acute decompensated heart failure requiring immediate hospitalization.
  • Contraindication to AF pharmacotherapy (e.g., pregnancy).
  • Hospitalization for acute myocardial infarction (STEMI or NSTEMI) within the past 3 months.
  • 12-lead ECG not available in a format analyzable by the AI model.
  • Individuals aged 18 years or younger.
  • Diagnosis of severe valvular disease (e.g., severe aortic stenosis).

研究計画

このセクションでは、研究がどのように設計され、研究が何を測定しているかなど、研究計画の詳細を提供します。

研究はどのように設計されていますか?

デザインの詳細

  • 主な目的:診断
  • 割り当て:ランダム化
  • 介入モデル:並列代入
  • マスキング:ダブル

武器と介入

参加者グループ / アーム
介入・治療
アクティブコンパレータ:Conventional Treatment

Pharmacologic treatment will be initiated after TTE is performed, in accordance with guideline recommendations.

  • Rate control: In accordance with guideline recommendations, the choice of rate-control agent may be based on LV EF (≤40% or not). However, in patients requiring acute rate control, agents that can be used regardless of LV EF (beta-blocker, digoxin) may be used without prior TTE.1
  • Rhythm control: In accordance with guideline recommendations, antiarrhythmic drug therapy may be initiated after assessment of LV EF and structural heart disease. Because oral amiodarone is not recommended as first-line therapy when LV EF >40%, it may not be used preferentially without information on LV EF. 1,2
  • When underlying heart failure is suspected, guideline-recommended pharmacotherapy will also be administered; however, when acute or congestive heart failure is suspected, agents that can be used without LV EF information (e.g., diuretics) may be initiated.
  • Classification code (grade): E01020.01(2)
  • Mechanism of action: When biosignal data (12-lead ECG) from an ECG device are automatically transmitted, or uploaded by the user, this software applies an artificial intelligence algorithm to the ECG data to predict the likelihood of left ventricular systolic dysfunction and displays the analysis results on a web-based screen.
実験的:AI-Guided Treatment

Guideline-directed AF pharmacotherapy may be initiated based on the AI-predicted LV EF obtained on the day of enrollment. Because patients with AF undergo a 12-lead ECG for clinical purposes, in the investigational group the AI-predicted LV EF becomes available immediately once the AI-enabled 12-lead ECG is performed.

  • TTE may also be performed in the investigational group, in accordance with guideline recommendations, to assess structural heart disease and left atrial substrate, and treatment may be modified based on the TTE results.
  • Rate control: The choice of agent may be based on the AI-predicted LV EF.1
  • Rhythm control: If the AI-predicted LV EF is >40%, there is no history of coronary artery disease, and there is no suspicion of moderate-to-severe valvular disease (e.g., no murmur), a Class IC antiarrhythmic drug (propafenone, flecainide, etc.) may be used in accordance with guideline recommendations. If the AI-predicted LV EF is ≤40%, oral amiodarone may be administered.1,2
  • Classification code (grade): B2ABXA1(3)
  • Mechanism of action: When biosignal data (raw 12-lead ECG data) measured by an ECG device are automatically transmitted or uploaded, an artificial intelligence algorithm displays a score and risk level for the likelihood of left ventricular diastolic dysfunction, for use as a diagnostic aid. Analyzed ECGs can be stored in HL7 aECG XML, JPG, or DICOM format.

この研究は何を測定していますか?

主要な結果の測定

結果測定
メジャーの説明
時間枠
Major adverse cardiovascular events (MACE), a composite outcome
時間枠:1year

Major adverse cardiovascular events (MACE) will be assessed as a composite outcome comprising all-cause death, cardiovascular death, stroke (ischemic or hemorrhagic), acute coronary syndrome, and cardiovascular hospitalization.

The outcome will be reported as the number of participants experiencing at least one of these events during the follow-up period. Participants experiencing multiple component events will be counted only once in the composite outcome.

1year

二次結果の測定

結果測定
メジャーの説明
時間枠
Number of Participants with All-Cause Death
時間枠:1year

The number of participants who die from any cause during the follow-up period will be assessed. Each participant will be counted only once.

Unit of Measure - Number of participants

1year
Number of Participants with Stroke (Ischemic or Hemorrhagic)
時間枠:1year

The number of participants experiencing at least one ischemic or hemorrhagic stroke during the follow-up period will be assessed. Each participant will be counted only once, regardless of the number of stroke events.

Unit of measure - number of participants

1year
Number of Participants with Acute Myocardial Infarction
時間枠:1year

The number of participants experiencing at least one acute myocardial infarction during the follow-up period will be assessed. Each participant will be counted only once, regardless of the number of events.

Unit of Measure- Number of participants

1year
Number of Participants Hospitalized for Heart Failure
時間枠:1year

The number of participants experiencing at least one hospitalization for heart failure during the follow-up period will be assessed. Each participant will be counted only once, regardless of the number of hospitalizations.

Unit of Measure- Number of participants

1year
Number of Participants with an Emergency Room Visit or Hospitalization for Cardiovascular Causes
時間枠:1year

The number of participants experiencing at least one emergency room (ER) visit or hospitalization for cardiovascular causes during the follow-up period will be assessed. Participants experiencing both an ER visit and hospitalization for cardiovascular causes will be counted only once.

Unit of Measure- Number of participants

1year
Percentage of Participants Achieving Atrial Fibrillation Rate Control
時間枠:1year

Atrial fibrillation (AF) rate control will be assessed as the percentage of participants achieving a heart rate of less than 110 beats per minute (bpm) during the specified assessment period.

Unit of Measure: Percentage of participants

1year
Time to Achievement of Atrial Fibrillation Rate Control
時間枠:1year

Time to AF rate control will be measured in days from the protocol-defined starting point to the first documented achievement of a heart rate of less than 110 bpm.

Unit of Measure: Days

1year
Time to Initiation of Oral Antiarrhythmic Drug Therapy
時間枠:1year

Time to initiation of oral antiarrhythmic drug therapy will be measured in days from the protocol-defined starting point to the first documented initiation of oral antiarrhythmic drug therapy.

Unit of Measure: Days

1year
Time to Atrial Fibrillation Catheter Ablation
時間枠:1year

Time to atrial fibrillation (AF) catheter ablation will be measured in days from the protocol-defined starting point to the date of the first AF catheter ablation procedure.

Unit of Measure: Days

1year
Time to Restoration of Sinus Rhythm
時間枠:1year

Time to restoration of sinus rhythm will be measured in days from the protocol-defined starting point to the first documented restoration of sinus rhythm.

Unit of Measure: Days

1year
AUC for LV Systolic Dysfunction
時間枠:1year

The predictive performance of the artificial intelligence (AI) model for identifying left ventricular (LV) systolic dysfunction will be assessed using the area under the receiver operating characteristic curve (AUC). AUC values range from 0 to 1, with higher values indicating better discrimination. The AUC of the AI model will be compared with that of N-terminal pro-B-type natriuretic peptide (NT-proBNP).

1year
Sensitivity for LV Systolic Dysfunction
時間枠:1year

Sensitivity of the AI model for identifying LV systolic dysfunction will be calculated as the percentage of participants with reference-standard-confirmed LV systolic dysfunction who are correctly identified by the AI model. Values range from 0% to 100%, with higher values indicating better sensitivity.

1year
Specificity for LV Systolic Dysfunction
時間枠:1year

Specificity of the AI model for identifying LV systolic dysfunction will be calculated as the percentage of participants without reference-standard-confirmed LV systolic dysfunction who are correctly identified as negative by the AI model. Values range from 0% to 100%, with higher values indicating better specificity.

1year
Positive Predictive Value
時間枠:1year

Positive predictive value (PPV) will be calculated as the percentage of participants with a positive AI prediction for LV systolic dysfunction who have reference-standard-confirmed LV systolic dysfunction. Values range from 0% to 100%, with higher values indicating greater positive predictive value.

1year
Negative Predictive Value
時間枠:1year

Negative predictive value (NPV) will be calculated as the percentage of participants with a negative AI prediction for LV systolic dysfunction who do not have reference-standard-confirmed LV systolic dysfunction. Values range from 0% to 100%, with higher values indicating greater negative predictive value.

1year
Atrial Fibrillation Burden at 1 Year Assessed by 3-Day Holter Electrocardiography
時間枠:1year

Atrial fibrillation (AF) burden at 1 year will be assessed using 3-day Holter electrocardiography (ECG) monitoring. AF burden will be calculated as the total duration of AF episodes divided by the total analyzable ECG monitoring time, multiplied by 100. AF burden will be expressed as a percentage, ranging from 0% to 100%, with higher percentages indicating greater AF burden.

1year
Change From Baseline to 1 Year in Atrial Fibrillation-Related Quality of Life Assessed by the Atrial Fibrillation Effect on QualiTy-of-Life Questionnaire (AFEQT)
時間枠:Baseline and 1 year

Atrial fibrillation (AF)-related quality of life will be assessed using the Atrial Fibrillation Effect on QualiTy-of-Life Questionnaire (AFEQT) at baseline and 1 year. The AFEQT overall score ranges from 0 to 100, with higher scores indicating better AF-related quality of life. Change from baseline will be calculated as the score at 1 year minus the baseline score. A positive change indicates an improvement in AF-related quality of life.

Baseline and 1 year

協力者と研究者

ここでは、この調査に関係する人々や組織を見つけることができます。

出版物と役立つリンク

研究に関する情報を入力する責任者は、自発的にこれらの出版物を提供します。これらは、研究に関連するあらゆるものに関するものである可能性があります。

一般刊行物

研究記録日

これらの日付は、ClinicalTrials.gov への研究記録と要約結果の提出の進捗状況を追跡します。研究記録と報告された結果は、国立医学図書館 (NLM) によって審査され、公開 Web サイトに掲載される前に、特定の品質管理基準を満たしていることが確認されます。

主要日程の研究

研究開始 (推定)

2026年12月1日

一次修了 (推定)

2029年12月1日

研究の完了 (推定)

2030年6月1日

試験登録日

最初に提出

2026年10月7日

QC基準を満たした最初の提出物

2026年11月9日

最初の投稿 (実際)

2026年11月16日

学習記録の更新

投稿された最後の更新 (実際)

2026年11月16日

QC基準を満たした最後の更新が送信されました

2026年11月9日

最終確認日

2026年11月1日

詳しくは

本研究に関する用語

個々の参加者データ (IPD) の計画

個々の参加者データ (IPD) を共有する予定はありますか?

未定

医薬品およびデバイス情報、研究文書

米国FDA規制医薬品の研究

いいえ

米国FDA規制機器製品の研究

いいえ

この情報は、Web サイト clinicaltrials.gov から変更なしで直接取得したものです。研究の詳細を変更、削除、または更新するリクエストがある場合は、register@clinicaltrials.gov。 までご連絡ください。 clinicaltrials.gov に変更が加えられるとすぐに、ウェブサイトでも自動的に更新されます。

購読する