Why is cervical burnout so easily mistaken for caries?

Two separate things go wrong at the cervical region of a posterior tooth, and most teaching names only one. The first is in the radiograph: there really is less tooth there to absorb the beam, so that strip really is darker. The second is in you: your visual system exaggerates the edge of that dark strip before you have consciously decided anything about it. Those are different kinds of problem, and only one of them can be fixed by looking again.

That matters because the standard advice — recognise the pattern, look more carefully — is aimed squarely at the half of the problem that isn't perceptual. If you don't know the second mechanism exists, the experience of being fooled feels like carelessness. It isn't.

What the question actually is

Cervical burnout is the radiolucent zone that appears near the cementoenamel junction on intraoral radiographs and can look like decay. The Mach band effect is the perceptual illusion that exaggerates contrast wherever two densities meet. Both have their own entry in the library, and this article isn't a second description of what they look like. The question here is narrower: why the two of them together are so hard to defeat, and what check survives both.

Worth separating out at the start: cervical burnout is a loss of attenuation — a real difference in how much of the beam got through. The Mach band is not. It is a difference in how your retina reports a boundary that already exists. One is about the tooth, one is about the observer, and they stack.

What the evidence shows

What physically produces the radiolucent zone

The 2025 Oral Radiology systematic review by Dioguardi and colleagues states the mechanism directly: cervical burnout "occurs when the X-ray beam passes tangentially across the curved proximal surface of the tooth, encountering less dental structure compared to other areas, thereby resulting in a localized apparent radiolucency," and it "commonly appears between the CEJ and the alveolar crest, where the dentin is thinner, leading to reduced absorption."¹ Berry's 1983 account in JADA — the origin description of both phenomena, and a film-era one — puts the same thing as overpenetration of the thinner cervical tooth areas, producing radiolucent areas with "rounded, diffuse inner borders."²

So the zone is not an error of technique in the ordinary sense. It's the expected consequence of shooting a beam through the narrowest part of the tooth.

Anatomy modulates how strong it looks. Dioguardi and colleagues note that burnout is more pronounced in multirooted teeth when the buccal and lingual roots don't perfectly overlap in the mesiodistal projection, and that a shallow root bifurcation accentuates it further.¹ Crown shape contributes on its own account too. Kühnisch and colleagues attributed the non-carious triangular-shaped radiolucencies seen on the mesial of upper molars to the crown's receding mesial surface running from mesiobuccal toward distopalatal, combined with an often prominent palatal cusp and a smaller mesiodistal diameter at the cervical neck.³ In their sample of 113 children aged 11 to 12, those radiolucencies were present on 60.3% of upper second primary molars, 35.5% of upper first primary molars and 24.8% of upper first permanent molars — and on none of the lower molars, primary or permanent.³

That last detail is the useful one. A radiolucency whose distribution is governed by which arch you're in is being produced by shape, not by disease.

What the visual system adds

Berry's description of the second mechanism is precise: Mach bands are "produced by lateral inhibition of neural receptors in the eye," appearing in dentine along the proximal dentinoenamel junction, or in dentinal peaks bounded by occlusal and proximal enamel.² A 2025 review of visual perceptual artefacts in cardiothoracic imaging describes the same physiology from the radiology side — Mach bands are "caused by lateral inhibition in the retina" and "create false contrasts at object boundaries," mimicking pathologies such as fractures.⁴

The reason this defeats effort is in how that review frames the whole category. Perceptual artefacts arise "from the human visual system rather than hardware or imaging acquisition processes," and they emerge across stages from the retina through the visual pathways and cortex to cognitive interpretation.⁴ The exaggerated edge is generated upstream of the deliberate, careful look. By the time you are examining the boundary attentively, the boundary you are examining has already been altered.

It also cuts both ways. Dioguardi and colleagues report that Mach bands can raise the detection threshold for small features nearby, which may account for false negatives as well as false positives — a true early lesion sitting next to a strong band can be harder to see, not easier.¹

And expectation moves the call. Nielsen asked 33 fourth-year dental students and 40 dentists to read the same radiograph under two scenarios. Students were more susceptible than dentists to misreading a Mach band as a horizontal root fracture; but when the scenario was recent trauma, both groups became more likely to call it a fracture.⁵ This was a simulated task with no pathological gold standard, and the systematic review that later assessed it rated it high risk of bias on the index test — so read it for direction, not for magnitude.¹ The direction is that what you were told before you looked shapes what you see.

How often these effects are called as disease

This is where the honest answer is narrower than the question. Dioguardi and colleagues pooled five studies and found that radiographic optical effects — "mainly categorized as Mach bands, TSR, and, to a lesser extent, cervical burnout" — were present in about 26% of maxillary molars, and that roughly 13% of observations of these effects were falsely interpreted as caries or fracture, with individual study proportions ranging from about 5% to 26%.¹

Two caveats have to travel with those figures. First, the measured phenomena are triangular-shaped radiolucencies and the Mach band; no included study measured misdiagnosis of cervical burnout on its own. Second, heterogeneity was high (I² > 90%) and the review's authors graded the certainty of the evidence as low to very low.¹ So the family of effects is now quantified, weakly; cervical burnout specifically is not.

The one check that survives both mechanisms

Berry's discriminator is structural rather than tonal: burnout produces radiolucent areas that "show intact tooth edges."² That is the check. Not how dark the zone is, not how sharp its border looks — both of those are exactly what the two mechanisms are corrupting — but whether the outline of the tooth runs unbroken through it.

The logic holds because neither mechanism can break an outline. Reduced attenuation makes a region darker; it does not remove surface. Lateral inhibition redistributes apparent contrast at a boundary; it does not delete the boundary. Caries does remove structure, and a surface that has lost structure has a broken edge.

Treat that as the classic discriminator rather than a validated test. None of the sources retrieved for this piece measured its sensitivity or specificity.

For the Mach band specifically there's a second manoeuvre, and its logic is just as clean. In the masking test, an opaque card is placed over the enamel margin: if the radiolucent area disappears when the adjacent enamel is covered, it was likely a Mach band; if it remains visible, it indicates true demineralisation.¹ Manual adjustment of contrast can do the same job on a digital image.¹ Berry described a masking technique for the same purpose in 1983.²

Technique and projection

Projection changes what you see, and the direction is worth knowing. Rahmatulla and Wyne, reviewing 25 full-mouth intraoral series, recorded cervical burnout more often in maxillary than mandibular teeth, and reported that in molars and premolars, burnout present on a periapical radiograph may disappear on a bitewing view.⁶ With a sample that small the percentages aren't worth carrying, but the qualitative point — that this is projection-dependent — is consistent with the tangential-beam mechanism.

Horizontal angulation is less clear-cut than it's often assumed to be. Shokri and colleagues imaged 150 extracted premolars at 0°, 5°, 10° and 15° horizontal angles and found no significant effect on detection of proximal enamel caries; specificity was 92% at 0° and 76% at 15°, but the difference in overall accuracy was not significant (p = 0.846).⁷ It's an in vitro study on artificially demineralised teeth. The specificity numbers trend the way you'd expect, and the study still does not license the conclusion.

Where it's unsettled

Three gaps are worth stating plainly.

No burnout-specific error rate exists. The 13% figure covers Mach bands and triangular-shaped radiolucencies at low to very low certainty.¹ Anyone quoting it as the rate at which burnout is called as decay is quoting it wrongly.

The outline check is conventional, not validated. It is stated in the origin literature² and it follows from the mechanism, but no retrieved study puts an accuracy figure on it.

Exposure settings are unstudied here. Berry's description invokes overpenetration,² but no retrieved study measures burnout severity against kilovoltage, exposure time, sensor type or collimation. If you have been taught a specific exposure protocol for reducing burnout, the evidence for it is not in this literature.

What this means at the chairside

Check the outline first, before you decide how dark the zone is. Density is the property both mechanisms distort; structural continuity is the one neither can touch.

If the tooth edge is intact and the radiolucency is diffuse-bordered and sits between the CEJ and the alveolar crest, you're looking at an artefact of attenuation, not a lesion. If the edge is broken, you have a genuine proximal surface finding to assess — and at that point the depth of the lesion starts to matter for what happens next, along with how early a lesion becomes visible at all.

When it's genuinely equivocal, the literature's answer is not to stare harder at the same image.

• Change the image, not the effort. Cover the adjacent enamel, or adjust the contrast on a digital image, and see whether the finding survives.¹² A second projection is reasonable on the same logic, given that burnout on a periapical may not persist on a bitewing view.⁶
• Go back to the mouth. The review's own conclusion is that clinicians should interpret radiographic findings with caution and always correlate them with a thorough clinical examination.¹ For confirming secondary caries specifically, the EFCD–ORCA–DGZ consensus statement is more specific: avoiding false positive detection and over-treatment is a priority, and bitewing radiographs should be combined with visual-tactile assessment.⁸
• Let the threshold for intervening do some work. That same consensus holds that for proximal lesions, only cavitated lesions — confirmed by visual-tactile examination, or radiographically extending into the middle or inner dentine third — should be treated invasively, and that non-cavitated lesions may be arrested non-invasively in low-risk individuals or where they are radiographically confined to enamel.⁸ An equivocal cervical shadow does not clear that bar.
• Sequence the whole examination correctly. The 2026 ADA and AAOMR recommendations state that a thorough evaluation of patient history and clinical findings should precede radiographic examination, and that imaging decisions should rest on history, clinical findings, disease risk assessment and the specific clinical situation.⁹ The panel notes that limitations in the available evidence meant consensus recommendations rather than formal guidelines were developed.⁹

The cost of getting this wrong is not abstract. Misinterpreting these effects may lead to unnecessary conservative, endodontic or even extraction treatment if the effect is read as a fracture rather than as caries.¹

None of this is an argument that bitewings are unreliable. It is an argument that a specific, well-described, mechanically predictable pattern sits on a specific part of every posterior tooth, and that the reader's job is to know it is there. Burnout is simply the mimic with two mechanisms behind it instead of one — and it is one stop on a longer reading sequence that also has to account for overlapped contacts, radiolucent restorative materials, occlusal surfaces, restoration margins and what remains under a restoration.

Knowing the mechanism isn't the same as calling it under time pressure, with a chart open and a patient in the chair. Try three real bitewings and see whether the outline check holds up when you have to commit.

Start the 2-minute reading test →