Occlusal caries

Occlusal caries is a carious lesion on the pit-and-fissure surface of a posterior tooth. On a bitewing radiograph it stays largely invisible until it reaches dentine: pooled sensitivity for any occlusal lesion is 0.35 in clinical studies, rising to 0.56 once the lesion is into dentine [1, 2]. This surface is examined visually first.

Synonyms: occlusal decay; hidden caries; occult caries; fissure caries

What is occlusal caries?

It's caries that starts in the pits and fissures of the chewing surface of a molar or premolar. Most of the evidence below concerns first and second permanent molars.

The reason it gets its own entry is that the radiograph plays a different role here than it does on the proximal surfaces. Proximally the radiograph is the primary instrument, and how early a proximal lesion shows up is a question the radiograph itself answers. Occlusally it is an adjunct: enamel-only lesions are outside what it reports [3], and the ADA and AAOMR recommendations put the clinical examination first [12].

How accurate is a bitewing on the occlusal surface?

Start with the accuracy figures, and read them carefully — they don't say quite what people expect.

In a 2015 meta-analysis of radiographic caries detection — 117 studies, 13,375 teeth, 19,108 surfaces — diagnostic odds ratios were higher for proximal lesions than for occlusal ones [1, 2]. Pooled sensitivity for detecting any occlusal carious lesion was 0.35 in clinical studies and 0.41 in vitro, with pooled specificity of 0.78 and 0.70 respectively [2]. So roughly two thirds of occlusal lesions present at the moment of the radiograph produced no radiographic sign. The specificity figures are the other half of the picture, and at this threshold they don't rescue it: 0.78 and 0.70 mean roughly a fifth to a third of sound occlusal surfaces get flagged positive [2]. A positive finding on an any-lesion basis is not self-verifying.

Note what the meta-analysis could not report. No included study covered cavitated occlusal lesions at all [1]. That absence is worth carrying: the evidence base thins out exactly where you'd most want a number.

The geometry is worth holding in your head while you read. A bitewing sends the beam through the crowns roughly parallel to the occlusal plane, a projection built around the contact areas — which is what the technique exists to show.

At what stage does it become visible?

Depth is the whole story, and it's the same axis as the E1–D3 depth classification you already use proximally.

Enamel only. In a 1998 review of caries diagnosis, Wenzel put it flatly: radiography "is of no value for the detection of initial (enamel) occlusal lesions" [3]. That review reflects analogue and first-generation digital systems, but nothing in the later pooled data overturns it, and the any-lesion pool that produces the 0.35 above includes these lesions by construction.

Into dentine. This is where the radiograph starts working. Pooled sensitivity for occlusal dentine lesions is 0.56 (95% CI 0.53 to 0.59); across dentine lesions generally — proximal and occlusal, clinical and in vitro — specificities ran from 0.87 to 0.95 [1]. At the dentine threshold, then, a positive finding is usually real. It's the negative that carries little information. Wenzel's earlier review reported 0.6 to 0.8 for occlusal lesions into dentine, with false-positive fractions of 5% to 10% [3]. Take the range rather than a single number — it moves with lesion depth and with whether the study was done in vitro.

One in vivo study in general practice reported bitewing sensitivity of 0.70 and specificity of 0.96 for occlusal dentine caries, but in a sample where 224 of 248 molars turned out to have dentine caries on operative exposure [4]. That's a population selected for diagnostic doubt, and the figures don't transfer to routine screening.

What you're looking for once it does appear is a radiolucency in dentine, beneath the enamel of the fissure. A dentine radiolucency is a good predictor of demineralisation [3].

What is hidden (occult) caries?

Hidden caries is occlusal dentine caries that is missed on visual examination but is large enough and demineralised enough to be detected radiographically [5]. Seow describes the same entity: occlusal caries not diagnosed clinically because the surface "appears ostensibly intact," with radiographs showing radiolucencies in dentine [6].

So it's the inverse of the usual problem. The surface looks fine. The dentine underneath doesn't.

Reported prevalence ranges from 2.2% to over 50% of permanent molars [6]. That's an order of magnitude, and the spread is structural rather than sloppy. Detection rate depends on how much caries there is in the population and how often bitewings are actually taken [5]. Two anchor points from screened populations:

• In 1,833 permanent molars scored as clinically sound in children aged 5 to 12, 72 (4%) showed a dentine radiolucency on bitewings [7].
• Among clinically sound occlusal surfaces in 15-year-olds, radiolucencies into dentine appeared in 16.9% in a fluoridated town and 24.6% in a non-fluoridated one [8]. The bitewings in that study were taken in 1968–69.

Whether fluoride exposure actually changes hidden caries prevalence is unsettled. The fluoridation study above found a proportional reduction [8]; Seow's review concluded prevalence "does not appear to be affected by fluoride exposure" [6]; and the children's study found no significant association with fluoride [7]. Most of the retrieved evidence leans against a fluoride effect, but it isn't settled.

Whether hidden caries is a distinct entity at all is also open. Ricketts and colleagues note it may simply be an artefact of inadequate looking, and that a better visual examination — careful cleaning and drying — might make the category disappear. They also say plainly that this has not been tested clinically [5].

What it's confused with

What it means clinically

A dentinal radiolucency under an apparently intact occlusal surface is not a stable finding.

In a six-year longitudinal study of 705 occlusal surfaces in Dutch adolescents followed from 17 to 23, about one third of sound occlusal surfaces developed a new dentinal radiolucency, and over 70% of existing radiolucencies progressed. Both held irrespective of whether a sealant was present [11].

The same group found clinical prevalence of occlusal caries in first and second molars was highly underestimated compared with the radiographs [10]. That is the argument for taking the occlusal surface seriously on a bitewing you took for another reason — not because the radiograph is sensitive there, but because a dentinal radiolucency, once visible, is usually real and usually progressing.

How to be sure

The examination comes first, not the radiograph. The ADA and AAOMR recommendations updated in January 2026 — consensus recommendations rather than formal guidelines, because of the limits of the available evidence [12] — state that radiographic imaging "serves as an adjunct to a comprehensive clinical examination," and that a thorough evaluation of history and clinical findings "should precede radiographic examinations" [12]. Which radiograph you take — bitewing or periapical — should follow from the lesion surface, the anatomic considerations and clinical judgment [12].

Occlusally, the eye outperforms the radiograph. A systematic review drawing on 240 studies found visual inspection better than radiographic examination on occlusal surfaces at all caries lesion thresholds — a head-to-head comparison drawn largely from laboratory studies — and concluded that visual detection alone is adequate for most patients in daily clinical practice regardless of tooth type or surface [13].

Adjuncts are adjuncts. Laser fluorescence and electrical conductance methods perform around 70% to 80% for sensitivity and specificity on occlusal dentine lesions, and the review treats them as supplements to visual and radiographic examination rather than replacements, on evidence graded as limited [14]. An in vivo comparison reached the same order: assess occlusal surfaces by meticulous visual inspection first, with fluorescence and bitewings as second-choice tools [4].

Not CBCT. The 2026 ADA and AAOMR statement records a lack of evidence supporting cone-beam CT for caries detection [12].

Common errors

Reading the occlusal surface with proximal confidence. The asymmetry between the two surfaces is real, but it doesn't run the way most people assume. Diagnostic odds ratios in the pooled data are higher for proximal than for occlusal lesions, and occlusal specificity is the weaker of the two [1, 2] — yet pooled sensitivity at the dentine threshold is actually higher occlusally than proximally, 0.56 against 0.36 [1]. The occlusal problem isn't that the radiograph finds less here than it does proximally. It's that the radiograph discriminates less well, and that a better instrument for this surface — your own eyes, on a clean, dry tooth — is already in the room [13].

Treating "no radiolucency" as "no lesion." Enamel-only occlusal lesions are essentially outside what the modality reports [3]. The absence of a finding here carries much less information than its presence.

The reverse error: chasing a stained fissure onto a radiograph. Staining is a visual finding. It cannot produce a radiolucency, so a radiograph cannot confirm or exclude it. If the question is whether a discoloured fissure is sound, the radiograph is not the instrument that answers it — and where it does answer, it answers late, at dentine.

Skipping the occlusal surfaces altogether. They're easy to slide past on a radiograph taken for the contacts. Give them a deliberate stop in the order you read a bitewing.

Forgetting sealed surfaces. Sealant does not stop the surface being readable, and it does not stop progression — radiolucencies progressed irrespective of whether a sealant was present in the longitudinal data [11].

Averaging the sensitivity figures into one number. They're threshold-dependent by design. Quoting a single headline sensitivity for "occlusal caries on bitewings" throws away the only part of the evidence that's clinically useful — the step from enamel to dentine.