Overlapped contacts

Overlapped contacts are the superimposition of two adjacent proximal surfaces on a bitewing radiograph, caused by horizontal angulation error — the central ray passes obliquely through the contact instead of parallel to it. Where the surfaces overlap, neither one can be assessed, and a small lesion can sit inside the superimposition unseen.

Synonyms: overlapping contacts; horizontal angulation error; contact overlap; closed contacts

What are overlapped contacts?

A bitewing is taken to answer one question: is there a lesion on the proximal surface. Answering it requires that you can actually see the surface — a single enamel outline, with nothing else projected on top of it.

When the beam is angled wrong horizontally, you don't get that. You get the distal surface of one tooth and the mesial surface of its neighbour printed over each other. What you're looking at is not a surface. It's two surfaces added together, and the sum has no diagnostic meaning.

Reading a radiolucency inside an overlap is reading an artifact of geometry.

What it looks like

The contact between the two crowns is closed. Instead of a thin dark interproximal space with a crisp enamel margin on each side, there's a band of increased density where the two enamel layers add up, and the outer margin of the proximal surface can't be traced.

That last point is the one the literature actually operationalises. In the validation study by Cascante-Sequeira and colleagues, a surface was scored as not overlapped only when the observer could recognise the proximal surface's outer margin; any recognisable addition of the two surfaces counted as overlapped. Two oral radiologists with ten years' experience, reading in vitro, applied that criterion and agreed perfectly (κ = 1.0).

So: can you trace the outer margin of the surface you're about to call? If not, you aren't looking at that surface.

What it's confused with

Two other look-alikes have their own entries rather than a row here, because the sources retrieved for this entry don't support a discriminator against overlap specifically: cervical burnout and the Mach band effect.

Why it happens

The mechanism is straightforward geometry. As Farman puts it in the American Dental Assistants Association's technique course, if the central beam isn't perpendicular to the interproximal spaces, the teeth overlap and the spaces close on the image — a flashlight on a picket fence, where the gaps stay open only while the beam stays square to the fence. The degree of overlap grows as the beam departs from perpendicular.

How steeply it grows is worth knowing. McDonald radiographed eleven child skulls across a 50° range of horizontal angulation and found a linear relationship between beam angle and overlap width in every skull. For each 0.1 mm increase in overlap width, 95% of the beam deviations were under 2.5°. That's a deciduous dentition on dry skulls in the film era, so don't carry the millimetres to your own chair — carry the shape of the finding. Small angular errors produce visible overlap.

McDonald's own conclusion was blunt: positioning the beam so that no overlapping occurs is "at best difficult, and at worst impossible."

The data agree. With a conventional XCP holder on dry skulls, 71.2% of proximal surfaces came out overlapped. With standard technique and an operator of more than five years' experience imaging teaching mannequins, median overlap was 13% of the surface — but the interquartile range was 40% and the maximum was 100%. Those two figures measure overlap differently, so they don't compare to each other; together they say the same thing, which is that overlap is normal, not exceptional.

How much overlap is too much?

No threshold appears in the literature retrieved for this entry. It's worth being direct about that, because the figure you may have been taught — that overlap up to about half the enamel thickness is acceptable — isn't in there either. Treat it as teaching convention rather than evidence.

What the literature has instead is three incompatible ways of defining the same word. Some studies score overlap as binary: any recognisable superimposition counts. Some measure it as a percentage of the surface using image-analysis software. Olegário and colleagues graded it by how far it encroaches — no overlap, overlap into enamel, or overlap into dentine — which is the closest thing to a severity scale among the sources retrieved here, and the one that maps best onto how you'd actually think about it. In their 550 primary molars on vertical bitewings, 13.27% showed overlap into enamel and 2.18% into dentine. Paediatric, primary teeth, vertical technique: the grading travels better than the percentages do.

In the absence of a number, margin traceability is the usable standard — bearing in mind it was defined to classify overlap on in vitro radiographs, not validated as a way to tell an overlap shadow from caries. If the outer margin of the surface is traceable, you can assess it. If it isn't, you can't — and no proportion of enamel makes that judgement for you.

What it means clinically

Overlap measurably degrades detection. Schropp and colleagues had 74 dental students read bitewings against a reference standard for enamel-only proximal caries. Thirty-eight per cent of the surfaces were overlapped, and both positive and negative agreement were significantly lower on those surfaces than on clear ones (p < .001). Overlap doesn't just hide lesions; it also produces calls on surfaces that are sound.

There's a second cost that gets less attention. Overlap is visible evidence that projection geometry changed — which means a lesion can't be validly compared across a series. McDonald made the point that monitoring progression requires identical projections, because small changes cause overlaps. Benn tested whether clinicians act on it, showing radiographic tests to 28 general dental practitioners and 21 teachers of conservative dentistry. In the test using approximal overlaps as evidence of irradiation geometry change, 54% of practitioners and 71% of teachers ignored that evidence. In a separate test, 43% of practitioners and 29% of teachers recommended restorations on radiolucencies they had been explicitly told were affected by geometry change.

That is a film-era study of a small, non-randomised UK sample. But it describes a failure that's easy to repeat: treating a depth classification as stable when the projection under it moved.

Retakes have a cost too. In a review of 26 studies covering all dentomaxillofacial radiograph types, the pooled reject rate for bitewings was 11.25%, with positioning error one of the two commonest reasons — though that figure covers all causes of rejection, not overlap alone.

How to be sure

Correcting horizontal angulation means putting the central ray back through the contact, parallel to the proximal surfaces rather than across them.

Which way the tubehead was off is conventionally read from the buccal object rule: buccal structures shift opposite to tubehead movement, lingual structures shift with it. The dentalcare CE course states the rule that way and goes on only to say it tells you whether the tubehead sat too far anteriorly or too far posteriorly. It does not state which way to move to correct it, and no source retrieved for this entry does — so this entry doesn't either. Hold the rule as standard technique teaching and an orientation aid, not as an evidence-backed correction protocol.

Equipment helps more than technique alone. A wedge-modified receptor holder cut overlapped surfaces from 71.2% with a conventional XCP to 19.2%, and stationary intraoral tomosynthesis reduced median overlap from 13% to 1%. Both results are in vitro and both devices are prototypes, so neither is a solution you can order this afternoon — but they show the variable is holder and receptor geometry, not operator effort.

One thing that isn't the lever: collimation shape. Comparing round with rectangular collimation across 176 round-collimation and 106 rectangular-collimation paediatric bitewings, Campbell and colleagues found differences in horizontal positioning errors and in the size of overlapped contacts that were statistically significant but not clinically important.

Common errors

• Calling a lesion inside an overlap. The commonest and the costliest. Run the margin test before the depth call.
• Comparing a lesion across a series without checking geometry. If overlap differs between two radiographs, the projection changed and the comparison is invalid — which is precisely what most of Benn's practitioners and teachers overlooked.
• Treating overlap as a binary pass or fail. Overlap into enamel and overlap into dentine are different problems; a surface can be partly assessable.
• Assuming an experienced operator produces clear contacts. Median overlap was 13% for an operator with more than five years' experience, with a maximum of 100%.
• Accepting the radiograph because retaking feels like a failure. Positioning error is ordinary rather than exceptional. In one series of 150 vertical bitewings in children, the receptor was placed too far posteriorly in 48.67% of cases and coning off occurred in 38.67%. Those are receptor-placement and collimation errors rather than beam angulation — that study reports no figure for beam angle — but they make the same point about how routine a degraded radiograph is.