서브메뉴
검색
From Beats to Biomarkers: Unlocking the Power of Cardiac Diagnostic Tools with Artificial Intelligence
From Beats to Biomarkers: Unlocking the Power of Cardiac Diagnostic Tools with Artificial Intelligence
상세정보
- 자료유형
- 학위논문 서양
- 최종처리일시
- 20250211152750
- ISBN
- 9798342108911
- DDC
- 616
- 서명/저자
- From Beats to Biomarkers: Unlocking the Power of Cardiac Diagnostic Tools with Artificial Intelligence
- 발행사항
- [Sl] : Stanford University, 2024
- 발행사항
- Ann Arbor : ProQuest Dissertations & Theses, 2024
- 형태사항
- 152 p
- 주기사항
- Source: Dissertations Abstracts International, Volume: 86-04, Section: B.
- 주기사항
- Advisor: Guestrin, Carlos;Zou, James;Kundaje, Anshul.
- 학위논문주기
- Thesis (Ph.D.)--Stanford University, 2024.
- 초록/해제
- 요약Heart disease is the number one cause of mortality in the US and globally even as preventative and interventional care have improved significantly in the past decades, in part because treatment isn't getting to the right people at the right time. Hospitals collect huge amounts of cardiovascular data, in the form of imaging, biosignals, text, and structured data. Could artificial intelligence (AI) help use this data to improve outcomes? This thesis builds on work applying AI to echocardiogram and electrocardiogram data with the goal of improving the set of tasks AI used for, how models are evaluated and interpreted, and our understanding of what diagnostic modalities are capable of. We start by describing a new computer vision tool for detecting abnormal laboratory values from the echocardiogram. Next, we present a novel AI risk score for cardiovascular mortality and disease from the ECG, which in combination with current standard-of-care cardiovascular risk scores can improve risk stratification for decisions like statin prescription. Third, we demonstrate via case study a new way of thinking about deep learning for electrocardiography by exploring a range of risk scores for left ventricular systolic dysfunction (LVSD) ranging from single measurements taken from the ECG to deep learning models with hundreds of thousands of parameters. Finally, we present a novel deep learning phenome-wide association study, analyzing the connection between the ECG and over 500 different diseases.
- 일반주제명
- Cardiovascular disease
- 일반주제명
- Biomarkers
- 일반주제명
- Electrocardiography
- 일반주제명
- Medicine
- 일반주제명
- Public health
- 기타저자
- Stanford University.
- 기본자료저록
- Dissertations Abstracts International. 86-04B.
- 전자적 위치 및 접속
- 로그인 후 원문을 볼 수 있습니다.
MARC
008250123s2024 us c eng d■001000017163763
■00520250211152750
■006m o d
■007cr#unu||||||||
■020 ▼a9798342108911
■035 ▼a(MiAaPQ)AAI31520346
■035 ▼a(MiAaPQ)Stanfordzt172kh2421
■040 ▼aMiAaPQ▼cMiAaPQ
■0820 ▼a616
■1001 ▼aHughes, John Weston.
■24510▼aFrom Beats to Biomarkers: Unlocking the Power of Cardiac Diagnostic Tools with Artificial Intelligence
■260 ▼a[Sl]▼bStanford University▼c2024
■260 1▼aAnn Arbor▼bProQuest Dissertations & Theses▼c2024
■300 ▼a152 p
■500 ▼aSource: Dissertations Abstracts International, Volume: 86-04, Section: B.
■500 ▼aAdvisor: Guestrin, Carlos;Zou, James;Kundaje, Anshul.
■5021 ▼aThesis (Ph.D.)--Stanford University, 2024.
■520 ▼aHeart disease is the number one cause of mortality in the US and globally even as preventative and interventional care have improved significantly in the past decades, in part because treatment isn't getting to the right people at the right time. Hospitals collect huge amounts of cardiovascular data, in the form of imaging, biosignals, text, and structured data. Could artificial intelligence (AI) help use this data to improve outcomes? This thesis builds on work applying AI to echocardiogram and electrocardiogram data with the goal of improving the set of tasks AI used for, how models are evaluated and interpreted, and our understanding of what diagnostic modalities are capable of. We start by describing a new computer vision tool for detecting abnormal laboratory values from the echocardiogram. Next, we present a novel AI risk score for cardiovascular mortality and disease from the ECG, which in combination with current standard-of-care cardiovascular risk scores can improve risk stratification for decisions like statin prescription. Third, we demonstrate via case study a new way of thinking about deep learning for electrocardiography by exploring a range of risk scores for left ventricular systolic dysfunction (LVSD) ranging from single measurements taken from the ECG to deep learning models with hundreds of thousands of parameters. Finally, we present a novel deep learning phenome-wide association study, analyzing the connection between the ECG and over 500 different diseases.
■590 ▼aSchool code: 0212.
■650 4▼aCardiovascular disease
■650 4▼aBiomarkers
■650 4▼aElectrocardiography
■650 4▼aMedicine
■650 4▼aPublic health
■690 ▼a0800
■690 ▼a0564
■690 ▼a0573
■71020▼aStanford University.
■7730 ▼tDissertations Abstracts International▼g86-04B.
■790 ▼a0212
■791 ▼aPh.D.
■792 ▼a2024
■793 ▼aEnglish
■85640▼uhttp://www.riss.kr/pdu/ddodLink.do?id=T17163763▼nKERIS▼z이 자료의 원문은 한국교육학술정보원에서 제공합니다.
