Prediction of device-related complications after continuous-flow left ventricular assist device implantation using artificial intelligence trained on acoustic spectra

2022 Fiscal Year Final Research Report

• Project/Area Number: 20K17715
• Research Category: Grant-in-Aid for Early-Career Scientists
• Allocation Type: Multi-year Fund
• Review Section: Basic Section 55030:Cardiovascular surgery-related
• Research Institution: Osaka University (2020, 2022); National Hospital Organization Osaka National Hospital Institute for Clinical Reserch (2021)
• Principal Investigator: Misumi Yusuke 大阪大学, 大学院医学系研究科, 特任助教(常勤) (20631477)
• Project Period (FY): 2020-04-01 – 2023-03-31
• Keywords: 補助人工心臓 / 在宅管理 / 音響解析 / 人工知能

Outline of Final Research Achievements

Left ventricular assist devises (LVADs) has become an important treatment of choice for patients with end-stage heart failure. As the growing number of patients receiving implantable LVAD therapy, there is an increasing clinical need to sensitively detect or predict LVAD malfunction or associated complications in daily lives of patients at home as in the case of tele-monitoring medical care system in the near future. In patients receiving implantable LVAD therapy, diagnosing modalities for major device-related complications require hospital visits and not suitable for a long-term tele-monitoring at home. LVAD acoustic signals are easily obtainable with an electronic stethoscope from the chest wall and could be potentially a suitable modality for tele-monitoring care of LVAD patients at home. In this study, we investigated the relationships between the changes in LVAD sound and relevant clinical outcomes.

Free Research Field

心臓血管外科

Academic Significance and Societal Importance of the Research Achievements

本研究成果を元に、LVAD駆動音の音響解析を用いた早期異常検出システムの技術的基盤を構築し、LVAD在宅管理における遠隔診断への応用の足がかりとしたい。駆動音の音響解析によりLVAD駆動異常の早期診断や循環動態の把握が可能となれば、LVAD装着患者が在宅にて手軽に自身のLVAD駆動状況を把握可能なシステムに応用できると考えられる。重篤な合併症の兆候を早期に検出し治療へ繋げることで、今後増加すると見込まれる在宅LVAD植込み患者の生活の質を維持すると共に、医療機関の救急・外来対応の負担を軽減させることが見込まれるものである。