What a CBCT scan shows, and why one is taken

If you have already had a lower wisdom tooth removed and the feeling in your lip, chin or tongue has changed, telephone the practice today. Nothing else on this page is time-critical. That is.

Everything below concerns a decision taken before an operation rather than a recovery from one. A cone-beam scan is one input to that decision, and it is an input with a cost attached: a dose of radiation, an appointment, and a volume of images that can be over-read as easily as under-read.

A panoramic radiograph is the standard first study for third molars, and for a substantial share of patients it is the only one. The X-ray source and the detector rotate around the head together on opposite sides, and the machine assembles one long image of a curved layer running through both jaws. Structures inside that layer — the focal trough — come out reasonably sharp. Structures outside it blur, magnify, or appear a second time on the opposite side as a ghost.

What it establishes is a great deal. Which third molars are present and how many roots each appears to have. The angle each one sits at, and how deep it is relative to the second molar. The general shape of the roots. The outline of the maxillary sinus floor above the upper teeth. The line of the inferior alveolar nerve canal running beneath the lower ones. Decay at the contact point with the second molar, bone loss behind it, and a follicular space around an unerupted crown that looks wider than it should.

What it cannot do is give you the third dimension. A panoramic image records height and length and collapses depth: everything between the cheek surface and the tongue surface is superimposed into a single shadow. So a lower root that appears to cross the canal may be sitting to the cheek side of it, to the tongue side of it, or genuinely against it, and the film looks broadly the same in all three cases. It is also magnified, variably and not uniformly across the image, which means a millimetre measured off a panoramic film is not a millimetre in the jaw.

The panoramic signs that mean look further

Rood and Shehab described the appearances that suggest a true relationship between a lower third molar root and the canal in 1990, and the list is still the one looked for.

• Darkening of the root where it crosses the line of the canal.
• Interruption of one or both of the white cortical lines that mark the canal walls.
• Diversion of the canal from its expected course.
• Narrowing of the canal where it passes the root.
• Narrowing of the root itself at the crossing point.
• Deflection of the root as it reaches the canal.
• A dark, blunted or bifid root apex.

These are signs of proximity, not of injury, and their presence is not a prediction about you. Several of them appear on films of teeth that are removed with nothing eventful happening at all. What they mean is that the flat image has reached the limit of what it can settle, and that a question now exists which a second study can answer and this one cannot. Where none of them is present and the roots sit clear of the canal, that question does not arise, and no further imaging is taken.

The machine looks much like a panoramic machine and the appointment feels much like one. You sit or stand upright, chin on a rest, head steadied by side supports or a strap, biting on a positioning block. The arm makes a single rotation around your head — a partial or a full circle depending on the machine — taking somewhere between roughly ten and forty seconds. You breathe normally throughout. The only instruction that matters is not to move.

During that rotation the machine fires a cone-shaped beam through the jaw onto a flat detector and captures several hundred images from different angles. Software reconstructs those into a block of data made of cubic elements called voxels. Because the voxels are cubes, the block can be resliced in any direction without distorting the measurements, which is the property everything else depends on. The images are then read as slices: across the jaw, along it, and — the view that answers most third molar questions — as a series of thin cross-sections cut perpendicular to the dental arch, one every fraction of a millimetre through the tooth.

It is not a medical CT scanner. The geometry is different, the dose is lower, and it carries considerably less soft-tissue information. It shows hard tissue well and soft tissue poorly, which has a specific consequence: it does not show the nerve. It shows the bony canal that carries the nerve. There is no injection, no contrast agent, no tunnel and nothing enclosed. The grey values it produces are also not calibrated the way Hounsfield units are on a medical scanner, so density readings taken off a cone-beam study are not a dependable measure of bone quality, whatever the software displays.

Field of view, and why it is kept small

The machine can be set to capture a small volume — one quadrant, or the back of the lower jaw — or a large one covering both jaws and the surrounding facial skeleton. A smaller field irradiates less tissue, generally resolves finer detail for the same exposure, and produces less anatomy that has to be examined and reported on. A field covering both jaws and the sinuses in order to examine one lower third molar irradiates tissue that no clinical question was attached to, which is why the field is matched to the question rather than set to whatever the machine will capture.

What degrades a scan

Movement is the main one, and the usual reason a scan has to be repeated. A swallow, a shift of the shoulders, or a jaw that relaxes off the bite block during the rotation smears the reconstruction. Metal is the other: crowns, amalgam fillings, posts, orthodontic bands and retainer wires scatter the beam and throw dark and bright streaks across the slices. Earrings, facial and tongue piercings, hair clips, spectacles, hearing aids and removable appliances come out beforehand for the same reason. A streak crossing exactly the region of interest can make an intact cortical line look interrupted, or hide one that genuinely is. That is one reason the scan is read alongside the panoramic film and the clinical findings rather than in place of them.

The inferior alveolar nerve runs forward through the lower jaw in a bony canal beneath the molar roots and emerges near the corner of the mouth to supply feeling to the lower lip, the chin and the gum in front of it on that side. It is purely sensory. Injury to it alters sensation; it does not weaken the lip, change the smile, or affect how the face moves. Lower third molar roots complete their formation towards that canal, and in some patients around it.

A cone-beam scan resolves the relationship into four separate facts, none of which a panoramic film can supply. Whether the cortical outline of the canal is intact along the length of the roots, or lost where they meet. Which side of the canal the roots actually lie on — cheek side, tongue side, directly above it, or straddling it. Whether a root is merely touching the canal, grooved by it, notched around it, or perforated by it. And how much bone, if any, separates the two.

Those are anatomically different situations with different implications, and on a flat film they can look identical. A root sitting a comfortable distance to the cheek side of the canal and a root wrapped around it produce the same overlapping shadow. Loss of the cortical outline where root meets canal is the finding that most changes the conversation afterwards, because it is the one that establishes there is no bone in between.

The lingual nerve is the limitation. It is the second nerve at risk in a lower third molar operation, running in soft tissue on the tongue side of the jaw and supplying sensation and taste to that side of the tongue, and it is not visible on a cone-beam scan or on any other radiograph. Its position varies between people and it sometimes lies at or above the level of the bone crest. Nothing in the imaging says anything about it. A scan addresses one of the two nerves, not both, and surgical technique on the tongue side is what addresses the other.

The nerve canal is the usual reason a scan is taken, and it is not the only thing the images are read for.

What the roots are shaped like

How many roots there are, whether they are separate or fused into a single mass, how widely they splay, and whether any of them curves. A sharply bent or hooked root — dilaceration — is the finding that most often decides how a tooth is divided, because a tooth with a hook cannot be delivered along the path a straight-rooted one takes. A bulbous apex from hypercementosis has the same effect. Ankylosis, where the periodontal ligament space has been lost and the root is fused directly to bone, means the tooth will not move under normal elevation and more bone has to be removed instead of more force applied. Each of those is worth knowing before the flap is raised rather than discovering it underneath.

The upper jaw and the maxillary sinus

Upper third molars have no nerve canal beneath them. The structure that constrains them is the maxillary sinus above, and the relevant question is how far the sinus floor has expanded down between and around the roots and how much bone is left. A root projecting into the sinus, or separated from it by a floor thin enough to be barely visible, makes an opening between mouth and sinus a realistic outcome of the removal rather than a remote one. It also raises the possibility of a root being displaced upward into the sinus if it is pushed in the wrong direction, and of the bone behind the tooth fracturing with it. None of those is common. All of them change what is said before consent is taken, and what is planned for.

Everything else in the volume

A follicular space around an unerupted crown that is wider than normal, or that has widened since an earlier film, raises the question of a cyst — usually a finding that produces no symptoms, picked up because a film was taken for another reason. External resorption of the second molar root, where a horizontal third molar has been pressing against it, is similarly silent and only visible on imaging. Decay at the contact point between the two teeth, bone loss behind the second molar, retained root fragments from an old extraction, and an unexpected fourth molar all appear in the same volume.

The volume also contains structures that have nothing to do with teeth: bone, sinuses, part of the airway, and depending on field size, the upper cervical spine and the region where the carotid arteries run. Whoever takes a scan is responsible for reviewing all of it, not only the part the question was about. This practice treats third molars and nothing else. Anything found outside that scope is explained to you at the time and referred to the appropriate service rather than managed here, and you are told what was found either way.

The dose from a dental cone-beam scan is small, it is not nothing, and it varies between machines and settings by more than most patients expect — by enough that quoting a single number would mislead. Doses are described in microsieverts (µSv), a unit weighted for how sensitive the particular tissues in the beam are. A thousand microsieverts is one millisievert.

Two things that table does not say. Effective dose is a population-level construct built for comparing exposures, not a prediction about what will happen to one person. And the ranges overlap heavily and deliberately. A small-field scan on low-dose settings can land beneath a full-mouth series of intraoral films that most patients accept without a second thought. A large-field scan on high-resolution settings sits well above a panoramic film. Where a particular scan falls is decided by field size, resolution, the arc the machine rotates through and the exposure factors. The background-equivalence column is arithmetic from an assumed 3,000 µSv per year, which is roughly 8 µSv per day; it is a way of picturing the size of the number, not a measurement.

Age matters here in a way that is easy to skip past. Risk from a given dose is higher the younger the patient, and third molar surgery is concentrated in the late teens and twenties. That is an argument for the smallest field that will answer the question and for justifying each exposure individually. It is not an argument against imaging a young patient who has a question that genuinely needs answering, because the alternative to a scan is not an absence of risk — it is operating with less information near a nerve.

The governing principle has two halves. Every exposure must be justified by a specific clinical question whose answer changes what is done. Then it must be optimised: the smallest field, the shortest arc and the lowest exposure that will still answer it. A scan that cannot be tied to a stated question fails the first half, and no amount of optimisation rescues it.

Aprons, collars and pregnancy

Lead aprons and thyroid collars were standard for decades. The current position of the oral and maxillofacial radiology bodies is that routine abdominal and thyroid shielding is no longer recommended for dental radiography, on the reasoning that modern collimation confines the beam tightly enough that shielding outside it achieves very little, while a collar sitting in the wrong place can obscure the area being examined and force a repeat exposure. If you would prefer shielding, say so before the exposure and it is discussed with you rather than dismissed.

Tell the office if you are pregnant or think you may be. The dose reaching the uterus from a dental exposure is extremely small, but small and necessary are different tests, and a decision about an asymptomatic third molar will usually keep. Elective imaging and elective surgery are generally deferred. Where there is infection or a problem that will not wait, the calculation changes and it is made explicitly rather than by default.

A cone-beam scan is not part of a routine third molar assessment and is not taken as a matter of course. The panoramic film is the default study, and for many patients it is the last one.

• An erupted or soft-tissue impacted tooth whose roots sit clearly away from the canal. There is no question for the scan to answer.
• A lower tooth where the roots do not approach the canal on the panoramic film and none of the Rood and Shehab signs is present.
• Most upper third molars. The relationship to the sinus floor is usually adequate on the panoramic film, and a scan is taken where that film suggests roots sitting in or through the floor and a difficult removal is expected — not for every upper tooth.
• A decision to monitor rather than operate. If the tooth is not coming out, there is no surgical plan for the scan to change, and radiographic review at intervals is the plan.
• A recent cone-beam scan already taken elsewhere covering the same region for the same question.

That last one is worth being practical about. What is useful is the dataset — DICOM files on a disc or through a transfer link — because the dataset can be resliced in any plane. A single exported picture answers only the question it was exported to answer, and a scan whose field excludes the region in question is not usable however recent it is. Age of the study matters in a young patient too, because roots that were incomplete two years ago are not incomplete now.

A scan also does not decide a borderline case for you. If the appropriate plan for a particular tooth is monitoring with radiographs at intervals, imaging it in three dimensions does not convert that into an operation. The scan tells you how a removal would go, not whether it should happen. Those are separate questions and they are answered in that order — the second one first.

If a scan is recommended, ask what question it is expected to answer and what would change depending on the answer.

The justification for any scan is that something downstream depends on it. In practice the changes fall into four groups: how the gum flap is designed and which side the bone is removed from, how the tooth is divided and in what sequence, whether the roots are approached at all, and what the consent conversation actually says. Sometimes it changes the date as well.

The row that gets least attention is the first. A scan showing intact bone between the roots and the canal has changed nothing about the operation, and it is not a wasted scan. It has removed a question that was genuinely open, closed off the coronectomy discussion, and allowed the risk to be described to you specifically rather than in the widest terms that would otherwise have to be used. A study whose result is reassuring answered its question as completely as one whose result is not.

The caveat belongs with all of it. A cone-beam scan changes what is planned in a meaningful share of the cases it is taken for. Whether it lowers the rate of permanent nerve injury is a separate question, and the trial evidence has not established that it does. The scan improves the accuracy of the plan and the precision of the consent conversation. It does not by itself make the operation safe, it does not remove the risk, and it should not be presented as though it does.

Where the scan confirms that roots are genuinely bound to the canal, a coronectomy becomes a real option rather than a theoretical one. The crown is removed, what remains of the roots is reduced below the level of the surrounding bone, and the gum is closed over them. No attempt is made to retrieve the roots. That is the purpose of the operation, not a shortcoming in it. The crown is the part associated with pericoronitis and with decay on the back of the second molar, so removing it addresses the problems the crown was causing while leaving the part that sits against the nerve where it is.

The indication is narrow, and a good number of teeth that look like candidates on a panoramic film turn out not to be once the scan and the clinical findings are read together. A tooth that is already mobile is ruled out, because the technique depends on the roots staying still. Active infection involving the roots themselves rules it out. So does root structure that is diseased in its own right. It applies to lower third molars and to very little else — there is no upper equivalent, because the constraint up there is a sinus rather than a nerve canal, and a retained upper root sitting against a sinus floor solves nothing.

Consent for a coronectomy is taken for two operations rather than one. If a root becomes mobile during sectioning it has lost the stability and the blood supply the technique relies on, and it is removed there and then; leaving it would store up trouble. So the plan is described as a coronectomy with complete removal as the alternative outcome, and not as a certainty.

The exchange is this. The intention is a lower chance of a permanent change in feeling in the lip and chin — an intention rather than a promise, since neither operation carries a certain result — in return for retained root structure, an obligation to be followed with radiographs over years, and the possibility of a further operation if a root later migrates towards the surface or the site becomes symptomatic. It is not a smaller experience than a surgical removal: the same flap is raised and a comparable amount of bone is taken. What is avoided is the part of the operation that moves roots sitting against the nerve. Which side of that exchange suits you is a legitimate preference rather than a clinical fact. It is discussed against your own scan, on screen, with the anatomy pointed at rather than described.