突然心臓死の隠れた兆候を可視化する新技術を開発 (Doctors May Now See the Hidden Signs of Sudden Cardiac Death)

ad

2026-08-17 コロンビア大学

コロンビア大学工学部(Columbia Engineering)の研究チームは、突然心臓死(Sudden Cardiac Death: SCD)のリスクを従来より高精度に予測できる人工知能(AI)技術を開発した。突然心臓死は多くの場合、事前の症状が乏しく、既存の臨床指標だけでは高リスク患者の特定が困難であった。本研究では、心電図(ECG)データと患者の医療情報を統合して解析する深層学習モデルを構築し、人間の医師には認識が難しい微細な電気生理学的パターンを検出できることを示した。その結果、従来のリスク評価法を上回る予測性能を達成し、突然心臓死の危険性が高い患者をより正確に識別できた。これにより、植込み型除細動器(ICD)の適応判断や予防的治療の最適化が期待される。研究成果は、循環器疾患の予防医療や個別化医療の発展に貢献するものであり、AIを活用した次世代診断支援技術の有効性を示している。

突然心臓死の隠れた兆候を可視化する新技術を開発 (Doctors May Now See the Hidden Signs of Sudden Cardiac Death)
EWI imaging of a 22-year-old female patient with mitral valve prolapse and an abnormal heart rhythm.

<関連情報>

小児僧帽弁疾患の特徴付けのための電気機械的活性化および回復波イメージング Electromechanical activation and recovery wave imaging for pediatric mitral valve disease characterization

Melina Tourni, Christina Proestaki, Seungyeon Julia Han, +12 , and Elisa E. Konofagou
Proceedings of the National Academy of Sciences  Published:August 17, 2026
DOI:https://doi.org/10.1073/pnas.2503559123

Abstract

Mitral valve (MV) disease, particularly MV prolapse (MVP) and mitral regurgitation (MR), affects 2 to 5% of the population and poses a substantial arrhythmogenic risk, with 43% of MVP patients developing arrhythmias. Although less common in children, the absence of comorbidities can uniquely isolate early electromechanical alterations that may elucidate mechanisms later contributing to arrhythmic risk and devastating effects such as sudden cardiac death. Conventional echocardiography lacks sensitivity for MV electromechanics, motivating advanced approaches. We introduce Electromechanical Wave Imaging (EWI), a high-frame-rate echocardiography modality, to map MV-complex activation and diastolic recovery in N = 21 MVP, MR, and control pediatric subjects (13.10 ± 4.51 y old, 43% male). A preclinical canine study (n = 3) established EWI’s ability to observe temporally coupled electromechanical wave propagation across the atrioventricular junction through the closed MV-following atrial and preceding ventricular activation (73.0 ms, 60 BPM). MVP patients exhibited significantly delayed left ventricular (LV) activation (76.04 ± 12.51 ms vs. 47.64 ± 2.57 ms in controls, P = 0.0013), primarily in papillary muscles. Both MVP and MR-only subjects exhibited prolonged LV recovery, with MR-only patients showing significantly longer recovery intervals (MR-only: 277.0 ± 27.61 ms, Control: 248.7 ± 10.43 ms, MVP vs. Control: MR-only vs. Control: P = 0.0395). In two arrhythmogenic MVP cases, EWI localized arrhythmic exit sites adjacent to LV papillary muscles, coinciding with regional delayed sinus activation, aligning with invasive electrophysiology. This study demonstrates that atrioventricular valve function dictates cardiac electromechanical function and establishes full-cycle EWI as a transformative tool for diagnosing MV disease electromechanical effects, assessing arrhythmic risk, guiding interventions, and advancing noninvasive cardiac imaging.

医療・健康
ad
ad
Follow
ad
タイトルとURLをコピーしました