Data presented at TCT 2026 demonstrate highest sensitivity and specificity of any FFR-CT validation study
BOSTON & SAN DIEGO--(BUSINESS WIRE)--#CCTA--Elucid, an innovator of AI-powered cardiovascular imaging software, today announced results from ELUCIDATE, presented at the annual Transcatheter Cardiovascular Therapeutics Conference (TCT 2026). The ELUCIDATE study, which supported Elucid’s recent U.S. Food and Drug Administration (FDA) 510(k) clearance of the FFR-IQTM software, demonstrated 92.9 percent vessel-level diagnostic accuracy for Elucid’s FFR-IQ technology versus invasive fractional flow reserve (FFR). ELUCIDATE also demonstrated high accuracy in the “gray zone,” FFR values between 0.75 and 0.85, which represent a physiological area of uncertainty when evaluating moderate stenosis.
ELUCIDATE (NCT06018194) is an international, multicenter retrospective validation study using prospectively collected and retrospectively identified datasets, including 491 coronary vessels from 347 patients at 11 U.S. and European institutions. It met both pre-specified primary endpoints, with 92.2 percent sensitivity and 93.1 percent specificity, the highest of any reported FFR-CT validation study, and demonstrated vessel-level diagnostic accuracy of 92.9 percent.1 The invasive measurements and their locations were independently reviewed and identified by the Cardiovascular Research Foundation core laboratory.
“The ability to transform coronary CT angiography from a purely anatomical assessment into one that integrates both anatomy and physiology is critical to the broader adoption of CT for coronary artery disease assessment,” said Ziad A. Ali, MD, DPhil, principal investigator of the study. “ELUCIDATE demonstrated sensitivity and specificity exceeding 92 percent for predicting physiological significance as defined by invasive FFR, with independent core-laboratory adjudication and a model locked before validation.”
FFR helps clinicians assess whether coronary artery disease is limiting blood supply to the heart. FFR-IQ applies Elucid’s BioIntegrated FFR-CT model, which incorporates coronary anatomy and plaque composition into the estimation of coronary physiology, to an existing coronary CT angiography (CCTA) scan to generate numerical FFR-CT values along the coronary vessels. The deep learning model integrates coronary lumen and vessel geometry with quantitative plaque characteristics, including calcified plaque, non-calcified matrix, and lipid-rich necrotic core. Trained against invasive FFR, it estimates coronary physiology directly from these features without performing computational fluid-dynamics equations.
“A coronary CT scan contains valuable information about both the artery and the disease within its wall,” said Jagat Narula, MD, PhD, co-first author of the ELUCIDATE abstract. “BioIntegrated FFR-CT brings anatomy and plaque biology together to estimate coronary physiology. ELUCIDATE shows what that approach can achieve, with high diagnostic accuracy and close agreement with invasive measurements.”
Performance near the 0.80 clinical decision threshold, or gray zone, was a particular focus of the study. In 145 vessels with invasive FFR from 0.75 to 0.85, the mean absolute error was 0.034. Diagnostic accuracy in this gray zone was 84.1 percent, with an AUC of 0.879.
Adjacent to the decision threshold, the negative predictive value was 95.2 percent across FFR-CT values of 0.81 to 0.90, and the positive predictive value was 79.3 percent for FFR-CT values of 0.71-0.8, providing complementary information for interpreting results on either side of the cutoff.
“What excites me about the ELUCIDATE results is the accuracy of FFR-CT, especially in the gray zone, where small differences matter,” said Amir Ahmadi, MD, of the Icahn School of Medicine at Mount Sinai, senior author of the ELUCIDATE abstract and lead scientific advisor to Elucid. “Pair that with Plaque-IQ’s validated estimation of stenosis, plaque burden and composition, and its ability to identify lipid rich plaque (LRP), and we bring anatomy, plaque biology and physiology into one assessment. This is the foundation for a one-stop approach to coronary risk assessment, one lesion at a time, with the whole patient in view.”
Elucid’s AI-based software combines plaque quantification, stenosis assessment, and CT-derived coronary physiology. Its FDA-cleared Plaque-IQTM and FFR-IQ technologies support clinicians evaluating coronary artery disease. The ELUCIDATE cohort had no overlap with the data used for model development or internal testing of Elucid’s FFR-IQ technology.
References
- Elucid FFR-IQ (B.1). FDA 510(k) Summary, K262015. September 29, 2026. https://www.accessdata.fda.gov/cdrh_docs/pdf26/K262015.pdf. Highest reported performance of FDA cleared FFR-CT validation studies
About Elucid
Elucid is a Boston-based medical technology company developing AI that sees beyond what the human eye can see in coronary and carotid CT angiography scans to advance personalized cardiovascular care. Elucid’s CT-Virtual Histology technology platform is based on ground-truth histology, the gold standard for vascular plaque analysis. Its FDA-cleared Plaque-IQ™ software precisely quantifies and characterizes atherosclerotic plaque, uniquely including high risk lipid-rich plaque, providing physicians with deeper insights into the underlying disease; while its FFR-IQ™ software, also FDA cleared, delivers noninvasive assessment of coronary physiology (BioIntegrated FFR-CT) derived from plaque analysis. Together, these technologies integrate biology, anatomy, and physiology to provide a more comprehensive understanding of cardiovascular disease and support personalized treatment decisions aimed at preventing heart attack and stroke.
Contacts
Media Contact
Sam Choinski
Pazanga Health Communications
schoinski@pazangahealth.com
(860) 301-5058
