Angiogram Images: How to Spot Normal vs. Abnormal Results

Angiogram images show blood vessels as dark or bright branching lines against a contrasting background, and telling normal from abnormal comes down to vessel shape, size, and how freely contrast dye flows through them. A healthy artery appears as a smooth, tapering tube with no abrupt narrowing, while an abnormal one may show focal pinch-points, bulging segments, or areas where dye stalls or leaks. But the eye alone is surprisingly unreliable at grading what it sees, and several common image artifacts can mimic real disease. Understanding both the visual landmarks and the known pitfalls gives you a much more realistic sense of what an angiogram can and cannot tell you.

What a Normal Angiogram Looks Like

During a catheter-based coronary angiogram, a thin tube is threaded into the heart’s arteries and contrast dye is injected while X-ray video records how the dye travels. On the resulting images, healthy coronary arteries look like a branching tree. The left main stem comes off the left side of the aorta and splits into the left anterior descending artery (LAD), which runs down the front of the heart, and the left circumflex (LCx), which wraps around the left side. The right coronary artery (RCA) comes off the right side and curves along the bottom of the heart. Each of these main trunks gives off smaller side branches at fairly regular intervals.

The key features of normal on any angiogram, whether coronary, cerebral, or peripheral, are smooth vessel walls, a gradual taper from larger to smaller branches, brisk and even flow of contrast dye, and no abrupt changes in diameter. The dye should fill each vessel fully and wash out at a consistent pace. When a cardiologist or radiologist reviews these images, they rotate the camera to multiple angles so no segment hides behind another branch. Standard projections are chosen specifically to separate overlapping vessels and give the clearest look at each segment.

For cerebral angiography, the normal picture is more complex because the brain’s arterial supply has a wider range of built-in variations. Fenestrations (where an artery briefly splits into two channels and then rejoins), duplications, and remnants of fetal circulation patterns are all seen in otherwise healthy people. A comprehensive review of three-dimensional reconstructions is usually needed to distinguish these harmless variants from genuine disease.

Anatomical Variants That Can Look Alarming

One of the most common sources of confusion for patients reviewing their own images is an anatomical variant, a blood vessel that takes an unusual route or originates from an unexpected spot. Coronary anomalies are found in roughly one to two percent of the general population, and most are completely benign. An anomalous circumflex artery that arises from the right side, for instance, simply takes a longer path to reach the left side of the heart. In a study using magnetic resonance coronary angiography to map these variants, the anomalous vessel was correctly identified in almost every case, though the exact course was occasionally hard to trace, especially when the vessel was small and non-dominant.¹

The rare variants that do matter are those where an anomalous artery passes between the aorta and the pulmonary artery. That particular squeeze can compress the vessel during exercise and, in unusual cases, trigger dangerous heart rhythms. On an angiogram, the artery itself may look perfectly smooth and healthy, so the concern is not narrowing or plaque but the route it takes. This is why imaging that shows the three-dimensional course of the vessel, rather than just whether it is open, is so important when an anomaly is suspected.

In the brain, similar logic applies. A comprehensive CT examination including three-dimensional and maximum-intensity-projection images can identify most normal variants. Recognizing features like fenestrations and persistent fetal arteries matters because they can mimic disease or change how a stroke is treated.²

What Stenosis Looks Like and Why the Eye Gets It Wrong

Stenosis, or narrowing of a vessel, is the single most common abnormality people look for on an angiogram. It shows up as a focal pinch in the contrast column: the smooth tube suddenly gets thinner at one point before widening again downstream. The more severe the narrowing, the thinner the column of dye at that spot, and a complete blockage shows the dye column stopping abruptly with no flow beyond it.

The trouble is that the human eye is not good at estimating exactly how severe a narrowing is. In a landmark study, when a known fifty-percent narrowing was shown to experienced angiographers, their estimates ranged from thirty percent all the way to ninety-five percent. Even the correlation between visual readings and careful quantitative measurements was only moderate, with a standard deviation of about fifteen percentage points. Averaging the readings of multiple experienced reviewers improved accuracy, but a single observer’s eyeball estimate was unreliable enough that the researchers concluded interventional decisions should not rest on it alone.³

A more recent comparison from the PROMISE trial reinforced this. Of patients whose site cardiologists reported obstructive disease based on visual assessment, a substantial number did not meet the same threshold when an independent lab measured the narrowing with quantitative software. The disagreement rate was about one in five patients, and in almost all of those discrepant cases the visual read had overcalled the severity, not undercalled it.⁴ Another study found low to moderate agreement between visual estimation and quantitative measurements, and a statistically large gap between diameter-based and area-based measurements of the same lesion.⁵

The practical upshot is that when your doctor tells you a narrowing is “about sixty percent” from looking at the angiogram, there is a real margin of error. Whether the narrowing is truly flow-limiting often requires additional testing beyond what the image alone can show.

When a Blockage Looks Tight but Blood Flow Is Fine (and Vice Versa)

An angiogram is essentially a shadow picture of the inside of the vessel. It tells you the lumen diameter but says almost nothing about the vessel wall itself or about whether the narrowing is actually restricting blood flow in a meaningful way. Fractional flow reserve (FFR) measurement, which uses a pressure-sensing wire to directly measure the drop in blood pressure across a narrowing, has revealed a striking disconnect between what looks bad on the image and what actually starves the heart of blood.

In one study comparing visual assessment by experienced interventional cardiologists to FFR, visual reads had decent sensitivity (about eighty percent) at catching genuinely significant narrowings but poor specificity: only about half the lesions flagged as significant by eye actually were. The positive predictive value was just twenty-five percent, meaning three out of four visually concerning lesions turned out not to be flow-limiting.⁶

The mismatch runs in both directions. A large registry found that among non-left-main lesions that looked at least fifty percent narrowed on the angiogram, more than half had an FFR above the treatment threshold, meaning they were not actually restricting flow enough to warrant a stent. Conversely, among left-main lesions that looked less than fifty percent narrowed, about forty percent were flow-limiting by FFR.⁷ This visual-functional mismatch is one of the most important concepts in modern cardiology and a major reason why angiogram images alone are no longer considered sufficient for deciding whether to place a stent in many situations.

Plaque Shape and What Angiograms Miss

Atherosclerotic plaque does not always bulge inward in a neat ring. It frequently grows in an eccentric pattern, building up more on one wall of the artery than the other. On an angiogram, the lumen may look only mildly narrowed because the camera catches the wider profile, while the opposite view might reveal a much tighter pinch. A systematic study of over 1,400 lesions compared how angiography classified plaque shape against intravascular ultrasound, which images the vessel wall directly. Angiography called about fifty-five percent of lesions eccentric, but the ultrasound-based classification agreed with that call in fewer than half of cases.⁸

Beyond eccentricity, angiography has a well-known blind spot for calcified plaque. Heavy calcium deposits show up as bright white spots on CT angiography and can obscure the lumen behind them, a phenomenon called “blooming.” Research into the root cause has found that partial volume averaging, where the scanner blends the brightness of calcium with the surrounding tissue, is the main culprit. Motion artifacts can worsen blooming, though modern scanners with advanced motion correction handle this better than older machines. Beam hardening, once considered a major contributor, turns out to have a relatively small effect with current technology.⁹ If you have heavy coronary calcium and your CT angiogram report mentions limited assessment of certain segments, this blooming effect is usually why.

Aneurysms and Dissections on the Angiogram

While narrowing is the most common abnormality, arteries can also be too wide. A coronary artery aneurysm is a focal ballooning to at least one and a half times the normal vessel diameter. On the angiogram, it appears as a round or oval bulge along the otherwise tubular vessel. If the dilation is long and diffuse rather than focal, it is called ectasia. Giant aneurysms, defined as a fourfold enlargement of the vessel, are exceedingly rare.¹⁰ These dilated segments matter because blood tends to swirl and stagnate inside them, raising the risk of clot formation.

Spontaneous coronary artery dissection (SCAD) has a very different look. In SCAD, the inner lining of the artery tears and blood tracks into the vessel wall, creating a false channel that compresses the true lumen. On angiography, a single-center series identified five characteristic features: the absence of atherosclerotic plaque in other arteries, a radiolucent flap creating two visible channels, contrast dye staining within the arterial wall, the abnormality starting or ending at a side branch, and a long, smoothly tapered narrowing that can mimic a “stick insect” appearance. Multiple flaps were seen in about forty percent of cases, and dye staining of the wall was considered essentially diagnostic when present.¹¹ SCAD is an important diagnosis to recognize because it predominantly affects younger women and its treatment is very different from the standard approach to atherosclerotic blockages.

Collateral Vessels and Slow Flow

When a coronary artery has been completely blocked for a long time, the heart sometimes builds its own detour. These collateral vessels are tiny channels that reroute blood from a healthy artery to the territory downstream of the blockage. On an angiogram, they appear as faint, winding threads connecting two vascular beds that would not normally be linked. Cardiologists grade collateral development using the Rentrop scoring system, which runs from 0 (no visible collaterals) to 3 (the collateral vessels fully fill the blocked artery’s territory).¹²

The grade matters. A study using cardiac MRI found that higher Rentrop scores were significantly associated with viable heart muscle in the territory supplied by the blocked artery. Among patients with no visible collaterals, more than half had non-viable tissue, while those with well-developed collaterals (Rentrop 2) had viable muscle in nearly ninety percent of cases. Each step up in collateral grade roughly doubled the odds that the downstream heart muscle was still alive.¹³ This is directly relevant to treatment decisions: if muscle is viable, reopening the blockage may restore function, but if the tissue is already dead, intervention offers less benefit.

A separate phenomenon to watch for is coronary slow flow, where the contrast dye moves through the arteries at a noticeably sluggish pace despite no visible blockage. On the angiogram, it looks like the dye is wading through molasses in one or more arteries while other vessels clear normally. Slow flow has been attributed to problems in the tiny vessels that angiography cannot see, a condition broadly called microvascular dysfunction.¹⁴ However, recent research has challenged the assumption that slow flow on the angiogram reliably indicates microvascular disease. When formal invasive testing of the small vessels was performed, the correlation with angiographic slow flow was weak, suggesting that slow flow should not be used as a stand-in for a proper microvascular diagnosis.¹⁵

Artifacts That Mimic Real Disease

Not everything abnormal-looking on an angiogram is real pathology. A review of 246 trauma angiograms found that about twelve percent contained artifacts mimicking arterial injury. The most common were stationary waves, where the dye column takes on a wavy contour that looks like a series of small tears, and streaming effects, where contrast separates into layers within the vessel, creating the illusion of a flap or dissection.¹⁶ Mach bands, an optical illusion in which the boundaries between areas of different brightness appear exaggerated, accounted for another handful of false findings.

In CT angiography, the blooming artifact discussed earlier is the most clinically significant pitfall. Stents also produce their own artifacts because the metal struts scatter the X-ray beam. This can make it difficult to see whether the inside of a stent has re-narrowed, a limitation that has driven development of newer scanner technology with higher resolution.¹⁷ Despite these limitations, 64-slice and newer CT scanners can still diagnose stent occlusion with high reliability. In an early evaluation using 64-slice CT, all completely occluded stents were correctly identified, and the overall sensitivity and specificity for detecting in-stent restenosis and occlusion were both near ninety percent.¹⁸

Evaluating Bypass Grafts and Stents After Treatment

If you have had a coronary stent or bypass surgery, angiogram images take on an additional layer of complexity. A patent stent looks like a short metallic cage with contrast dye flowing smoothly through its center and out both ends. In-stent restenosis, where scar tissue grows inside the stent and re-narrows the vessel, appears as a focal dimming or thinning of the dye column within the stent’s boundaries. Complete stent thrombosis shows up the same way as any total blockage: the dye stops at the stent’s entrance with no flow beyond.

Bypass grafts, whether made from a leg vein or an internal chest artery, are checked the same way as native arteries. A healthy graft fills briskly with contrast from its origin at the aorta (or from the chest wall, for mammary artery grafts) to where it connects to the native artery downstream of the original blockage. Graft disease tends to appear as diffuse narrowing along the length of a vein graft or as a focal narrowing at the connection points. Mammary artery grafts generally hold up better over time and tend to look healthier at follow-up angiography, which is one reason surgeons prefer them when possible.

Where Artificial Intelligence Fits In

Given how unreliable the human eye is at grading narrowings, there has been intense interest in training AI systems to read angiograms. Deep-learning models have now been tested on both coronary artery segmentation, meaning tracing the vessel outlines automatically, and stenosis detection. One recent model achieved a per-lesion detection rate of about sixty-seven percent, which matched the agreement rate between human experts reading the same images.¹⁹ Another integrated model, tested on over 1,600 angiographic images, reached about eighty-nine percent accuracy for detecting and classifying stenosis severity.²⁰

These numbers are encouraging but deserve some context. A detection rate matching expert agreement does not mean AI has solved the problem; it means AI is roughly as inconsistent as the experts are, which, as the studies above show, is fairly inconsistent. The real promise of AI may be less about replacing human readers and more about providing a standardized second opinion, flagging lesions a reader might miss and reducing the variation between observers that plagues visual assessment. For now, no AI tool has replaced the combination of experienced human judgment and functional testing like FFR in making actual treatment decisions.

How Heart Imaging Has Shifted Away from Pictures Alone

The broader trend in cardiology is a move away from making treatment decisions based solely on what the angiogram looks like. The visual-functional mismatch described earlier was a wake-up call. Large randomized trials over the past two decades have shown that guiding stent placement by FFR measurements, rather than by how bad a narrowing looks, leads to better patient outcomes. As a result, guidelines now recommend functional assessment for most intermediate-severity lesions rather than relying on the angiographic image alone.

Intravascular ultrasound and optical coherence tomography add another dimension that the angiogram cannot provide. These catheter-based tools sit inside the artery and image the vessel wall in cross-section, revealing plaque composition, calcium distribution, and the true lumen size behind an ambiguous shadow on the angiogram. Their use has grown steadily, particularly in complex cases where the stakes of misjudging a lesion are highest. The angiogram remains the first and most widely used imaging tool during catheterization, but it is increasingly understood as a starting point rather than the final word.