Cone cut

A cone cut is the part of a radiograph the X-ray beam never reached, because the beam and the receptor were not lined up. The region is blank and sharply bordered, and nothing in it has been imaged. A surface it covers was never examined, so “no caries seen” there is not a negative finding.

Synonyms: cone cutting; collimator cut; conecut; partial image; cone-cut artifact

What is a cone cut?

The beam leaving the tubehead is collimated down to roughly the size of the receptor. UK guidance caps the beam at the end of the spacer cone at 40 mm by 50 mm for a size 2 receptor, and prefers it under 35 mm by 45 mm.

Line the two up and the whole receptor is exposed. Miss, and the part that falls outside the beam records nothing at all.

That is a cone cut, and it belongs to the family of radiographic artifacts — appearances produced by the imaging, not by the patient.

The important word is nothing. A cone cut is not a shadow, not a faint area, not a region you can squint at. It is an absence of image.

What it looks like

A blank, pale region at an edge of the radiograph, with a hard border.

Two things about it are worth fixing in your head.

1. The border is geometric.

   It runs straight through whatever happens to be there — enamel, dentine, restoration, bone — without bending, branching or changing character. Anatomy does not do that. Every real structure on a radiograph has a border that means something: a lamina dura follows a root, a caries lesion spreads within enamel. A cone cut's border follows the edge of the beam, which knows nothing about the anatomy underneath it. The shape of that edge follows the shape of the collimator aperture.

2. It is not a

   radiopacity. This is the part that catches people out. The region is pale, and pale usually means something dense attenuated the beam. Here nothing attenuated anything. The beam never arrived, so the receptor recorded no signal, and no signal renders pale. Reading it as a dense structure is reading a hole as an object.

It can be at any edge. In the largest series retrieved for this entry — a dental school in Amsterdam, 53,684 intraoral radiographs assessed and 44,685 classified, all taken with rectangular collimation — the 9337 classified radiographs carrying a cone cut had it above in 22%, below in 31%, distal in 21% and mesial in 27%. There is no characteristic corner to check first.

Where it turns up by view is worth knowing, because it runs against what you may have been told. In that same series, posterior periapicals carried the most cone cut, at 26%, and bitewings the least, at 16%. The common claim that cone cut is above all a bitewing problem is not what that data shows.

What it's confused with

All four are listed in the UK guidance's table of common image quality faults, alongside cone cut itself.

One look-alike deliberately has no row: a genuine radiolucency. It points the other way. A real radiolucency is dark because the beam got through easily; a cone cut is pale because the beam never came. If you're deciding between those two, the density direction settles it before anything else does.

Why it happens

The guidance names two causes, and they are the same failure seen from two ends.

A misorientated collimator — the rectangular aperture turned the wrong way relative to the receptor, so the beam's long axis and the receptor's long axis disagree. And a mispositioned spacer cone — the beam aimed off-centre, so part of the receptor sits outside it. Either one leaves the region of interest only partially imaged.

Neither takes much.

Two things make it more or less likely, and neither is about how hard the operator is trying.

The equipment can be the cause. A 2014 technical report found that where cone cuts persist in the same position, the problem may originate in the X-ray machine rather than in the aim, and can be resolved by tapping the collimator or repositioning the position indicating device. That is a small study — two machines, four radiographs — so hold it as a possibility to check rather than a frequent explanation. But if you are cutting the same corner off every image, checking the set is cheaper than blaming yourself.

The collimator and holder change the rate. Comparing rectangular collimator designs on training manikins, one study found a magnetic collimator that testers preferred also generated more cone cut errors; another found collimator centering errors occurred three times more often with a universal device than with an enhanced one. Both are manikin studies, so the numbers are not chairside performance — but they locate the variable in the hardware rather than in operator effort.

Then there is collimation shape, which is where the literature actually argues.

Rectangular collimation is the recommended default — the UK guidance says it should always be used for periapical, bitewing and occlusal radiography unless there are sound clinical reasons not to, because it substantially reduces dose. The long-standing objection is that it causes more cone cuts. The evidence is mixed on how much.

A 2026 retrospective study of dental students' radiographs found cone cut in 20.9% of images with rectangular collimation against 1.8% with circular — a large difference. A dental school series using rectangular collimation throughout found cone cut in 21%, consistent with that figure but with no comparator to check it against. Pointing the other way, a 2025 practice audit of 360 radiographs found three cone cuts in periapicals taken without rectangular collimation and none with it.

Those three don't reconcile, and the reasons are visible: two are teaching institutions and one is general practice, the samples run from 360 images to 53,684. Take the direction — rectangular collimation makes cone cut more likely — and leave the magnitude open.

When does a cone cut mean a retake?

Not automatically, and the standard is narrower than "is there a cone cut on it".

UK guidance rates every image as diagnostically acceptable or not acceptable, and the test for acceptable is that it has no errors, or minimal ones, and is "of sufficient image quality to answer the clinical question". An image is rated not acceptable when errors render it diagnostically unacceptable. Nothing in that turns on whether an error is visible. It turns on whether the error took away the thing you took the radiograph for.

So the question to ask is not is there a cone cut but is the surface I need inside it.

The numbers say most cone cuts survive that test. In the Amsterdam series of 44,685 classified radiographs, 79% had no cone cut, 18% had a cone cut that left diagnostic value unaffected, and 3% were judged diagnostically unusable because of it. The 2026 student study put cone cuts requiring a retake at 5.7% with rectangular collimation and 0.7% with circular. Those two figures measure different things — one is images judged unusable, the other is images actually retaken, in different populations — so don't merge them. Together they say the same thing: a cone cut usually costs you a corner you didn't need.

When it does cost you something, retaking is the answer and it has a price. The retake doubles that view's exposure for the patient. That is the whole reason the criterion is what it is: a retake has to buy back diagnostic information worth the dose, and a blank corner over a region nobody asked about doesn't qualify.

One more thing follows from rating an image not acceptable. The reason has to be recorded — the guidance requires it, and reject records are what turn a run of cone cuts into a fixable problem instead of a habit.

How to be sure

Use a holder and a beam-aiming device. The guidance is direct: beam-aiming devices and receptor holders should be used for periapical and bitewing radiography wherever clinically possible, to ensure the X-ray set is positioned correctly relative to the receptor, and this is essential where rectangular collimation is used. It states the effect plainly — better image quality, and fewer reject images requiring retakes. The 2026 study reaches the same practical conclusion from the other direction, recommending rectangular collimation for clinicians with proper training used together with a positioning device.

If it repeats in the same place, look at the machine rather than at your technique.

When you report, say what you couldn't see. This is the part that belongs to reading rather than to taking. The blank region contains no image of anything, so the surfaces under it have not been assessed — not assessed and found sound, not assessed at all. Write down which surfaces those were. A search pattern that runs over a cone cut and records nothing has quietly converted "not imaged" into "nothing found", and those are not the same claim.

Common errors

• Reading the blank region as a dense structure. It is not a radiopacity. Nothing attenuated the beam there — the beam never arrived.
• Calling a surface sound when the cone cut covered it. The commonest and the costliest. An unimaged surface has no finding, positive or negative.
• Assuming cone cut is mainly a bitewing problem. In the largest series here it was least common on bitewings and most common on posterior periapicals.
• Retaking every image with a cone cut on it. The standard is whether the image still answers the clinical question, and a retake costs the patient a second exposure of that view.
• Blaming the operator for a cone cut that repeats in the same position. That pattern may originate in the X-ray set rather than in the aim.
• Abandoning rectangular collimation because of cone cuts. Cone cuts do get more frequent with it, but the 2026 study found the overall retake rate was not significantly different between rectangular and circular collimation — 16.9% against 16.7% — while the dose reduction is the reason to use it in the first place.

Cone cut sits alongside overlapped contacts and elongation and foreshortening as errors that don't hide the whole radiograph — they take away one specific thing, and the risk is reading past them.