Coronary artery disease has long been diagnosed through invasive procedures—catheters snaking through arteries, X-ray dye, and procedures that carried risk.
But a revolution is quietly unfolding in cardiology.
Artificial intelligence now allows doctors to measure coronary plaque with stunning precision using nothing more than a CT scan.
Dr. Campbell Rogers, an interventional cardiologist who ran the cardiac catheterization lab at Brigham and Women’s Hospital and taught at Harvard, recently explained how this technology is fundamentally changing how we detect, measure, and treat heart disease before symptoms ever appear.
Jump to:
- The Old Way: Invasive Tests for Too Many People
- Enter AI-Powered Coronary Plaque Analysis
- What Makes This Different From Standard CT Scans
- Why Non-Calcified Plaque Matters Most
- From Risk Factors to Actual Disease
- A Real-World Example: Tracking Plaque Regression
- Who Should Consider This Test
- Measuring Treatment Success
- The Future: From Surrogate Markers to Hard Outcomes
- Calcium Score Zero Doesn’t Mean No Disease
- Understanding Plaque Stabilization
- Practical Considerations and Limitations
- The Bottom Line
The Old Way: Invasive Tests for Too Many People
For decades, confirming coronary disease meant an invasive cardiac catheterization. Patients would undergo a procedure with real risks, only to discover—two-thirds of the time—that they didn’t actually need invasive treatment.
The proportion of people who present with a heart attack who have had symptoms before that heart attack is very small. It’s under 50%.
Rogers explained that waiting for symptoms isn’t enough. Many heart attacks strike without warning.
If you want something simple that actually works, this is one of the most effective tools I’ve used to build strength, conditioning, and endurance without needing a full gym setup.
- Full-body training with one weight using swings, squats, and presses
- Solid cast iron build that feels stable and lasts for years
- Comfortable grip that makes high-rep workouts easier to handle
Traditional stress tests and nuclear imaging offered some insight, but they couldn’t visualize plaque directly. Calcium scoring helped identify some disease, but missed non-calcified plaque entirely—often the most dangerous kind.
Enter AI-Powered Coronary Plaque Analysis
HeartFlow, founded on research from Stanford University, initially focused on analyzing blood flow through coronary arteries using CT scans and artificial intelligence. This helped doctors determine who actually needed stents or bypass surgery.
But the real breakthrough came when they turned their AI toward measuring plaque itself.
One person who’s ApoB may have diffuse severe coronary plaque. Another person may have whistle clean coronaries. And now we can know the answer to that.
This isn’t guesswork. The AI analyzes CT angiography images and provides precise, quantitative measurements of plaque volume, composition, and location—all validated against intravascular ultrasound, the gold standard.
What Makes This Different From Standard CT Scans
Most people who get coronary CT angiography receive a basic report: mild, moderate, or severe disease. Maybe a calcium score.
HeartFlow’s AI plaque analysis goes several steps further.
If you want something simple that actually works, this is one of the most effective tools I’ve used to build strength, conditioning, and endurance without needing a full gym setup.
- Full-body training with one weight using swings, squats, and presses
- Solid cast iron build that feels stable and lasts for years
- Comfortable grip that makes high-rep workouts easier to handle
Within about an hour, it returns detailed information: total plaque volume measured in cubic millimeters, plaque composition (calcified versus non-calcified), and percentile ranking compared to others of the same age and sex.
Rogers emphasized the revolutionary nature of this precision.
It’s hard to overstate the revolutionary nature of this for coronary disease. It’s like X-rays over 100 years ago.
Before X-rays, doctors could only guess whether a bone was broken. Now, with AI plaque analysis, doctors can see exactly how much coronary disease someone has—not estimate based on risk factors.
Why Non-Calcified Plaque Matters Most
Calcium scores only detect calcified plaque. But non-calcified plaque—soft, cholesterol-rich deposits—poses significant risk.
Heart attacks typically occur when non-calcified plaque ruptures, triggering blood clots that block arteries.
Rogers noted that more non-calcified plaque correlates with worse outcomes: higher likelihood of heart attack, chest pain, needing stents, or dying from heart disease.
If you want something simple that actually works, this is one of the most effective tools I’ve used to build strength, conditioning, and endurance without needing a full gym setup.
- Full-body training with one weight using swings, squats, and presses
- Solid cast iron build that feels stable and lasts for years
- Comfortable grip that makes high-rep workouts easier to handle
The AI analysis measures both types separately, providing crucial information calcium scoring alone cannot offer.
From Risk Factors to Actual Disease
Traditional cardiology relies heavily on biomarkers like LDL cholesterol or ApoB levels. These predict risk but don’t tell the whole story.
Two people with identical ApoB levels can have vastly different plaque burdens.
Is that where we’re moving where it’s treating the actual plaque seen on imaging versus just the biomarker? 100%.
Rogers believes cardiology is shifting toward treating disease rather than just numbers on blood tests. Seeing actual plaque helps patients understand their risk and improves adherence to medications and lifestyle changes.
A Real-World Example: Tracking Plaque Regression
During the discussion, Rogers analyzed two CT scans taken 16 months apart from a 38-year-old man with mild plaque.
At baseline, the analysis showed 116 cubic millimeters of total plaque, placing him in the 89th percentile for men his age—more plaque than 89% of peers.
If you want something simple that actually works, this is one of the most effective tools I’ve used to build strength, conditioning, and endurance without needing a full gym setup.
- Full-body training with one weight using swings, squats, and presses
- Solid cast iron build that feels stable and lasts for years
- Comfortable grip that makes high-rep workouts easier to handle
Sixteen months later, total plaque volume dropped to 60 cubic millimeters. His percentile fell to the 80th.
Plaque in the left main artery had disappeared entirely. Non-calcified plaque in the left anterior descending artery decreased significantly.
Rogers acknowledged that some might question whether plaque regression can happen this quickly, but he emphasized the need for humility.
We’ve never had tools to look at this. So, how do we know that this hasn’t happened before? We don’t know.
World-class imaging combined with validated AI analysis revealed changes previously invisible to medicine.
Who Should Consider This Test
Rogers outlined three groups who could benefit most from AI coronary plaque analysis:
- People with symptoms like chest pain or shortness of breath—CT angiography is now the preferred first test in guidelines
- Those with positive calcium scores between 1 and 300, where treatment decisions remain ambiguous
- High-risk individuals without symptoms, including those in safety-critical occupations or with strong family histories of early heart disease
Currently, insurance typically covers CT angiography only for symptomatic patients. Those without symptoms often pay out of pocket.
If you want something simple that actually works, this is one of the most effective tools I’ve used to build strength, conditioning, and endurance without needing a full gym setup.
- Full-body training with one weight using swings, squats, and presses
- Solid cast iron build that feels stable and lasts for years
- Comfortable grip that makes high-rep workouts easier to handle
Measuring Treatment Success
One exciting application involves tracking how plaque responds to therapy.
Rogers cited a randomized trial published in JAMA Cardiology studying men with prostate cancer. Researchers used HeartFlow’s plaque analysis to compare two drugs, finding one caused significantly more plaque progression after just one year.
The study required only 65 patients total—far fewer than typical cardiovascular trials—because the plaque measurements were so precise.
Another study called BESIDE showed that using AI plaque analysis led to meaningful reductions in LDL cholesterol. Lower LDL levels have been proven across decades of research to reduce heart attacks and cardiovascular death.
The Future: From Surrogate Markers to Hard Outcomes
While the technology demonstrates clear benefits for risk assessment and treatment decisions, Rogers acknowledged that proving it reduces actual heart attacks and deaths remains crucial.
Long-term follow-up studies are underway. Results from the DECIDE trial, tracking patients over time, are expected throughout this year.
If you want something simple that actually works, this is one of the most effective tools I’ve used to build strength, conditioning, and endurance without needing a full gym setup.
- Full-body training with one weight using swings, squats, and presses
- Solid cast iron build that feels stable and lasts for years
- Comfortable grip that makes high-rep workouts easier to handle
NIH-funded research called PREEMPT is examining young people with small amounts of plaque to understand what drives regression.
We need to gather the clinical evidence that these are the right choices so we can compel patients, physicians, guideline writers, etc. to move in this direction.
Calcium Score Zero Doesn’t Mean No Disease
Many people receive calcium scores of zero and assume they’re safe from coronary disease.
Rogers clarified an important distinction: calcium score zero indicates very low risk but doesn’t guarantee absence of disease.
Non-calcified plaque—invisible to calcium scoring—can still exist and pose risk.
For patients with zero calcium scores but very high risk factors, CT angiography with AI plaque analysis may provide valuable additional information.
Understanding Plaque Stabilization
Some patients on statin therapy notice their calcium scores increase over time and worry their disease is worsening.
If you want something simple that actually works, this is one of the most effective tools I’ve used to build strength, conditioning, and endurance without needing a full gym setup.
- Full-body training with one weight using swings, squats, and presses
- Solid cast iron build that feels stable and lasts for years
- Comfortable grip that makes high-rep workouts easier to handle
Rogers explained this often represents plaque stabilization—soft plaque converting to calcified plaque—which most experts consider beneficial vascular healing.
AI plaque analysis can track these compositional changes, showing whether non-calcified plaque decreases even as calcium increases.
Practical Considerations and Limitations
CT angiography requires intravenous contrast dye, though volumes used are small and generally well-tolerated except in people with allergies.
Radiation exposure is modest but not zero—a consideration when deciding whether to scan.
Image quality matters significantly. About 97% of scans provide sufficient quality for AI analysis. The remaining 3% are too fuzzy, often due to inadequate heart rate control during scanning.
For serial scans tracking plaque changes, using the same scanner and protocol reduces variability and improves accuracy.
If you want something simple that actually works, this is one of the most effective tools I’ve used to build strength, conditioning, and endurance without needing a full gym setup.
- Full-body training with one weight using swings, squats, and presses
- Solid cast iron build that feels stable and lasts for years
- Comfortable grip that makes high-rep workouts easier to handle
The Bottom Line
AI-powered coronary plaque analysis represents a fundamental shift in how medicine approaches heart disease.
For the first time, doctors can see and measure coronary disease with precision comparable to invasive procedures, but using only non-invasive CT scans.
This technology identifies disease earlier, guides treatment decisions more accurately, and tracks whether therapies are working—all potentially before symptoms develop or heart attacks occur.
As Rogers emphasized, we’re witnessing the cardiology equivalent of X-rays arriving over a century ago: moving from educated guesses to definitive answers about what’s actually happening inside arteries.
The science continues evolving, clinical trials are ongoing, and guidelines are gradually shifting. But the trajectory is clear: treating coronary disease based on actual plaque burden rather than biomarkers alone represents the future of preventive cardiology.
