1 Non-Contrast Scan 10+ Actionable Reports
Cardiovascular
Health
CardioMetabolic
Health
Lung Health
Bone Health
All-in-one AI-CVD

Comprehensive Bundle
AutoFracture
(coming soon)


New peer-reviewed study introduces Agatston-2.0, led by Morteza Naghavi, MD, with Arthur S. Agatston, MD, the originator of the original Agatston score, and leading cardiovascular imaging and prevention experts.
HOUSTON, TX, June 2026 — HeartLung Corporation today announced the publication of a landmark peer-reviewed study introducing Agatston-2.0, a next-generation artificial intelligence-based coronary artery calcium scoring method designed to modernize one of the most influential tools in preventive cardiology.
The original Agatston score, introduced in 1990 by Arthur S. Agatston, MD, and colleagues, transformed cardiovascular prevention by giving physicians a noninvasive way to quantify coronary artery calcium and identify silent coronary atherosclerosis before symptoms occur. Over the past three decades, CAC scoring has become one of the most validated and widely used imaging biomarkers in cardiovascular medicine.
Now, 36 years later, a multi-institutional team of leading physicians and researchers, led by Morteza Naghavi, MD, Founder and President of HeartLung Corporation, and Arthur S. Agatston, MD, of The Agatston Center in Miami Beach and originator of the Agatston score, has introduced the first major AI-based modernization of this historic method.
"The original score was designed to be simple, reproducible, and clinically useful," said Arthur S. Agatston, MD, The Agatston Center, Miami Beach. "Agatston-2.0 keeps that strength while using AI to measure what older CT methods could miss."
The study, titled "Agatston-2.0: A Next-Generation AI-Based Coronary Calcium Quantification Approach to Improve Risk Stratification Among Individuals with Zero Agatston Scores - Part I," was published in the American Journal of Preventive Cardiology.
"This is a defining moment for coronary calcium scoring," said Morteza Naghavi, MD, Founder and President of HeartLung Corporation. "Agatston-2.0 brings CAC scoring into the era of artificial intelligence, modern CT resolution, and voxel-level quantitative imaging."
Why Agatston-2.0 Matters
Traditional coronary artery calcium scoring, now referred to as Agatston-1.0, relies on technical assumptions established in the late 1980s and early 1990s, including a fixed 130-Hounsfield-unit threshold and conventional 2.5-3 mm CT slice thickness. These parameters made the method practical and reproducible, but they can also miss very small, low-density, fragmented, or partially calcified coronary plaques.
Agatston-2.0 was developed to address these limitations. Instead of relying only on rigid thresholds, it uses AI-based coronary segmentation and continuous voxel-wise calcium quantification. The result is a more sensitive and biologically faithful way to quantify coronary calcium, especially early or low-burden disease that may be invisible to conventional scoring.
"This is quantitative CT doing what it should do," said David F. Yankelevitz, MD, Department of Diagnostic, Molecular, and Interventional Radiology, Icahn School of Medicine at Mount Sinai. "Agatston-2.0 extracts more useful signal from the scan."
The advance is important because CAC scanning is already widely used in clinical prevention. Agatston-2.0 does not require an invasive test or a new imaging philosophy. It builds directly on the established value of CAC scoring and updates the measurement to match the capabilities of contemporary CT and AI.
Answering the "Soft Plaque" Criticism of CAC
One of the most persistent criticisms of CAC scanning is that non-contrast calcium scans may miss so-called soft or non-calcified plaque. That criticism is partly true but often oversimplified. A CAC scan is not a coronary CT angiogram and is not intended to characterize the coronary lumen, stenosis, or purely non-calcified plaque morphology. However, coronary atherosclerosis exists along a continuum, and many plaques described as soft, non-calcified, or CAC-negative may contain subtle, low-density, fragmented, or semi-calcified components that fall below the detection rules of Agatston-1.0.
This is one of the key problems Agatston-2.0 addresses. By replacing a binary threshold with continuous AI-derived calcium probability, spatial neighborhood analysis, and automated coronary segmentation, Agatston-2.0 can detect subthreshold calcium signal that older scoring rules may miss. In practical terms, some risk previously attributed to "CAC-negative soft plaque" may actually reflect semi-calcified disease that was present but not counted by the original scoring method.
"The soft-plaque argument against CAC has often been overstated," said Dr. Naghavi. "Agatston-2.0 shows that part of the blind spot may be outdated scoring, not CAC imaging itself."
This distinction is clinically important. Agatston-2.0 does not replace CCTA when the clinical question is plaque morphology or stenosis. But it strengthens the scientific foundation of CAC scanning by narrowing a commonly cited gap: the inability of conventional scoring to capture subtle, early, or semi-calcified coronary plaque signal.
​
Refining the “Power of Zero”
A CAC score of zero has long been one of the most reassuring findings in preventive cardiology. This concept, widely known as the "Power of Zero," reflects the very low near-term risk seen in many patients with no detectable coronary calcium by conventional scoring.
However, a small but clinically important number of patients with CAC=0 still experience coronary heart disease events. The newly published study directly addresses this unresolved question: can modern AI detect hidden coronary calcium signal in some patients whose traditional Agatston score is zero?
The answer appears to be yes.
In the study, investigators pooled 3,965 participants with CAC=0 from two major prospective cohorts: the Multi-Ethnic Study of Atherosclerosis and the Framingham Heart Study. Agatston-2.0 detected AI-derived coronary calcium in 862 participants, representing 21.7% of individuals who were classified as CAC=0 by conventional scoring.
Those with AI-CAC greater than zero had significantly higher long-term coronary heart disease risk than those with AI-CAC of zero. Over 20 years, coronary heart disease incidence was 7.7% in participants with AI-CAC greater than zero compared with 3.8% in those with AI-CAC of zero. After adjustment for traditional cardiovascular risk factors, AI-CAC greater than zero remained independently associated with incident coronary heart disease.
Agatston-2.0 also predicted future conversion from CAC=0 to a positive traditional CAC score on repeat CT imaging, suggesting that the AI-detected signal represents clinically meaningful early coronary calcification rather than random image noise.
"The Power of Zero remains powerful and is even more powerful with Agatston-2.0," said Nathan D. Wong, PhD, Heart Disease Prevention Program, Mary and Steve Wen Cardiovascular Division, University of California Irvine. "Agatston-2.0 helps identify which zero is truly zero as well as those with possible earlier, semi-calcified disease associated with greater future risk than those previously thought to be 0 based on traditional Agatston scoring."
"A more precise zero can help clinicians refine prevention," said Robert A. Kloner, MD, PhD, Huntington Medical Research Institutes and Keck School of Medicine of the University of Southern California. "That is the clinical value of this advance."
A New Chapter in Preventive Cardiology
The publication of Agatston-2.0 is especially significant because CAC scoring is already embedded in cardiovascular prevention and supported by decades of outcomes research. Unlike many emerging imaging tests, Agatston-2.0 does not require invasive procedures or a new biological theory of risk. It builds directly on the most established CT-based biomarker in preventive cardiology and updates it for the AI era.
The study suggests that Agatston-2.0 may help physicians better personalize prevention by identifying a higher-risk subgroup within the conventional CAC=0 population while also defining an even lower-risk subgroup with AI-CAC=0. This could have practical implications for shared decision-making, intensity of risk-factor treatment, and timing of repeat scanning.
The study authors emphasize that additional validation in other cohorts and clinical settings will be important. However, the results establish a new framework for modern calcium scoring: one that preserves the clinical simplicity of the original Agatston score while using AI to improve sensitivity, reproducibility, and biological precision.
"This is not a replacement of CAC scoring. It is its modernization," said Dr. Naghavi. "Agatston-2.0 brings the world's most established CT-based prevention test into the AI era."
If validated in additional cohorts and implemented at scale, Agatston-2.0 could become a new standard for coronary calcium scoring and risk stratification.
About the Study
The study was conducted by a broad multi-institutional team including investigators and collaborators from HeartLung.AI, Cornell University, The Agatston Center, UCLA / Harbor-UCLA and The Lundquist Institute, University of California Irvine, Icahn School of Medicine at Mount Sinai, Stanford University, Houston Methodist, Cedars-Sinai, Kaiser Permanente, University Medical Center Groningen, University Hospital Basel, and other leading cardiovascular imaging and preventive cardiology institutions.
The Part I analysis focused on individuals with conventional CAC=0 in MESA and FHS and tested whether Agatston-2.0 could identify occult coronary calcium and improve long-term risk stratification. The findings support a more nuanced interpretation of CAC=0: conventional zero remains clinically valuable, but AI-confirmed zero may represent an even lower-risk phenotype.
Key Coauthor Affiliations Quoted in This Release
• Arthur S. Agatston, MD - The Agatston Center, Miami Beach, FL; originator of the original Agatston score.
• Morteza Naghavi, MD - Founder and President, HeartLung Corporation / HeartLung.AI, Houston, TX.
• Robert A. Kloner, MD, PhD - Huntington Medical Research Institutes, Pasadena, CA, and Keck School of Medicine of the University of Southern California, Los Angeles, CA.
• David F. Yankelevitz, MD - Department of Diagnostic, Molecular, and Interventional Radiology, Icahn School of Medicine at Mount Sinai, New York, NY.
• Nathan D. Wong, PhD - Heart Disease Prevention Program, Mary and Steve Wen Cardiovascular Division, University of California Irvine, Irvine, CA.
Publication details:
Article: Agatston-2.0: A Next-Generation AI-Based Coronary Calcium Quantification Approach to Improve Risk Stratification Among Individuals with Zero Agatston Scores – Part I
Journal: American Journal of Preventive Cardiology
DOI: 10.1016/j.ajpc.2026.101698
Publication Link:
https://www.sciencedirect.com/science/article/pii/S2666667726002734#fig0002
About HeartLung Corporation
HeartLung Corporation is a pioneer in AI-driven preventive imaging focused on early detection of cardiovascular disease, lung cancer, COPD, osteoporosis, fatty liver disease, and other conditions detectable on CT scans. Its flagship platform, AI-CVD, transforms routine CT imaging into a scalable preventive health assessment by quantifying cardiovascular and systemic biomarkers, including coronary artery calcium, cardiac chamber size, aortic and valvular calcification, epicardial fat, lung density, liver fat, bone mineral density, and muscle-fat composition.
About AI-CVD
AI-CVD is an FDA-cleared, AI-powered platform that performs fully automated analysis of CT scans to detect and quantify cardiovascular disease and related cardiometabolic imaging biomarkers. By enabling opportunistic screening across large populations, AI-CVD supports earlier intervention, improved patient engagement, and more scalable prevention pathways.
AI-CVD delivers a comprehensive automated assessment from a single CT scan, including coronary artery calcium, aortic and valvular calcification, cardiac chamber volumetry, aorta and pulmonary artery size, epicardial and visceral fat, liver density, lung density, bone mineral density, and muscle-fat composition.
About AI-CAC and AutoCAC
HeartLung AI’s AutoCAC represents automated Agatston-1.0, delivering fast, reproducible coronary artery calcium scoring aligned with the traditional CAC method at scale.
HeartLung AI’s AI-CAC represents Agatston-2.0, an advanced AI-native approach designed to quantify coronary calcium more sensitively and detect early, subtle, or lower-burden disease that may be underestimated by conventional scoring.
Company
Resources
2450 Holcombe Blvd
TMC Innovations
Houston, TX 77021








.avif)


