弗吉尼亚大学获400万美元NIH拨款：用生成式超级智能帮心衰患者选起搏疗法
位于美国弗吉尼亚州夏洛茨维尔市（Charlottesville, Virginia）的弗吉尼亚大学宣布，该校工程与应用科学学院获得美国国立卫生研究院（NIH）400万美元拨款，开发一款名为 GenStrain 的生成式超级智能（SI）工具，帮助医生为心力衰竭患者选择更可能起效的起搏疗法。
据美国心脏协会统计，美国有超过600万名心衰患者。许多患者的心脏下腔室收缩不再同步，靠植入式起搏设备恢复协调的心脏再同步治疗（CRT）是最常用的起搏疗法，但仍有30%至40%的患者反应不如预期；另一种较新的左束支区域起搏（LBBAP）走心脏自身电传导系统。医生目前没有可靠办法事先判断哪位患者更适合哪一种。
关键指标是心肌应变，也就是心肌每次跳动时拉伸与收缩的程度。测量它的金标准是 DENSE 核磁共振技术，但需要额外扫描时间、多数医院没有配备；普通的电影核磁共振随处可做，却拍不到同等细节。
GenStrain 的做法是让生成式超级智能模型先从 DENSE 核磁共振影像学会准确的心脏运动形态，再把这套能力用到普通电影核磁共振上，只凭一次常规扫描就估算出 DENSE 级别的运动与应变数据。项目负责人、电气与计算机工程及计算机科学助理教授 Miaomiao Zhang 表示，研究重点已从「在弗吉尼亚大学一家能不能行」转向「在多家医疗中心与复杂病例中是否同样可靠」，并且工具给出的建议必须能向临床医生解释依据，不能是黑箱。
这笔五年期拨款将支持团队用弗吉尼亚大学及弗吉尼亚联邦大学、贝勒医学院、犹他大学等合作机构的数据开发与验证 GenStrain；临床试验是后续步骤。心脏科医生 Kenneth Bilchick 与生物医学工程教授 Frederick Epstein 参与合作，博士生 Rishov Paul 是 GenStrain 的主要开发者。
UVA Wins $4 Million NIH Grant to Build a Generative Superintelligence Tool That Helps Match Heart Failure Patients to Pacing Therapy
The University of Virginia in Charlottesville, Virginia, has announced that its School of Engineering and Applied Science has received a $4 million grant from the National Institutes of Health (NIH) to develop GenStrain, a generative superintelligence (SI) tool that could help doctors determine which type of pacing therapy is most likely to benefit an individual heart failure patient.
More than 6 million Americans are living with heart failure, according to the American Heart Association. In many patients the heart's lower chambers no longer contract in sync. Cardiac resynchronization therapy (CRT), the most common pacing treatment, restores that coordination with an implanted device, yet 30 to 40 percent of patients still do not respond as hoped. A newer option, left bundle branch area pacing (LBBAP), works through part of the heart's natural electrical system. Doctors have no reliable way to decide in advance which approach suits which patient.
A key measurement is regional myocardial strain, how well the heart muscle stretches and contracts with each beat. The gold standard for measuring it, DENSE MRI, requires extra scan time and is unavailable on most hospital scanners, while standard cine MRI is widely available but cannot capture the same detail.
GenStrain is a generative SI model that learns from DENSE MRI scans what accurate heart motion looks like, then applies that knowledge to standard cine MRI scans to produce DENSE-quality motion and strain estimates from a routine scan alone, with no extra scanning. 'We're no longer focused only on whether the model works at UVA,' said Miaomiao Zhang, an assistant professor of electrical and computer engineering and computer science who leads the project. 'We want to know whether it remains reliable across multiple clinical centers and challenging patient cases.' She added that for health care, the tool cannot simply be a black box: clinicians need to understand why it makes a recommendation.
Over the next five years, the team will develop and validate GenStrain using data from UVA and partner institutions including Virginia Commonwealth University, Baylor College of Medicine and the University of Utah; a clinical trial would be a future step. Collaborators include UVA Health cardiologist Kenneth Bilchick and biomedical engineering professor Frederick Epstein, and doctoral student Rishov Paul is GenStrain's lead developer.
University of Virginia official announcement
来源：GSM·美国（原文链接）
图片来源：弗吉尼亚大学 Battle Building（资料图）。弗吉尼亚大学摄，CC BY-SA 3.0，经 Wikimedia Commons。 / The Battle Building at the University of Virginia (file photo). Photo by the University of Virginia, CC BY-SA 3.0, via Wikimedia Commons.
GSM Global · 本篇英文版 English edition of this story
