NIH grant: TARGET-TTR: Targeted Artificial Intelligence-Driven Screening of Transthyretin Amyloid Cardiomyopathy in the Electronic Health Record (1K23HL181318-01A1)

retained text

NIH grant 1K23HL181318-01A1: TARGET-TTR: Targeted Artificial Intelligence-Driven Screening of Transthyretin Amyloid Cardiomyopathy in the Electronic Health Record
Institute: {'code': 'HL', 'abbreviation': 'NHLBI', 'name': 'National Heart Lung and Blood Institute', 'admin_org_id': '12500068', 'admin_funding_url': ''}. PI: Evangelos  Oikonomou. Award: $170964.
Start: 2026-08-15T00:00:00.
PROJECT SUMMARY
Transthyretin amyloid cardiomyopathy (ATTR-CM) is associated with significant morbidity and mortality but
remains largely underrecognized. Novel therapies can delay disease progression, improve quality of life, and
prolong survival, but their real-world benefits are limited by the extent of ATTR-CM underdiagnosis. To date,
disease detection has relied on advanced multimodality imaging, which is poorly suited for scalable screening of
at-risk individuals during the earliest stages of the disease. Artificial intelligence (AI) methods applied to routine
cardiac diagnostic tests, such as electrocardiography (ECG), may help discriminate ATTR-CM, offering an
opportunity for timely risk stratification and diagnosis. However, broad AI-enabled screening in the electronic
health record (EHR) is limited by the condition’s low prevalence, high false positive rates, and the challenge of
relying on opportunistic healthcare interactions for deployment. In this patient-oriented mentored career
development award proposal, Dr. Evangelos Oikonomou proposes a multimodal informatics-guided approach
for the precision AI-assisted screening of ATTR-CM in the EHR of large health systems. First, he introduces an
innovative computational strategy that leverages the depth of clinical phenotypes encoded in the EHR as time
series of structured and unstructured data to identify key intersections between encounters and patient profiles
associated with high precision and sensitivity for AI-ECG screening (Aim 1). Second, he will develop and test an
EHR-based digital twinning paradigm to identify computable drivers of variability in access to cardiac testing that
may be amplified by opportunistic AI deployment (Aim 2). Throughout both aims, Dr. Oikonomou will collaborate
with clinical experts to enhance natural language processing methods using consensus-based textual
representations. Finally, he will design and deploy a computational simulation framework to gather quantitative
and qualitative feedback from clinicians and patients in order to define the features of an optimal clinical decision
support (CDS) system for AI-ECG-guided screening of ATTR-CM in the EHR (Aim 3). The resulting algorithms
will be made open source, enabling institutions to apply them to their individual care settings. Notably, the
proposed approach may serve as a model for the AI-assisted screening of other rare conditions in the EHR. The
applicant will benefit from access to a broad pool of mentors and extensive computational resources at the Yale
School of Medicine. The mentored research period will support Dr. Oikonomou’s training in applied clinical
informatics, natural language processing, and mixed methods, through direct mentorship from experts in each
of these domains. The experience gained during this time will support Dr. Oikonomou’s transition to
independence, during which he plans to lead multi-center collaborations to evaluate the efficient deployment of
AI-enabled screening tools for underdiagnosed cardiomyopathies. The proposed career development plan will
position Dr. Oikonomou to become a national leader in the implementation science of AI-assisted diagnostics in
cardiovascular disease.

document fetched 2026-08-23 06:34