How to read a dental CBCT scan: the three planes, slices and viewer tools

Simplified Orthodontics Academy · Published 28 October 2026

Key points

A CBCT scan is a volume: a block of small boxes called voxels, each holding one grey value. The voxel size, in millimetres, sets the scale of every measurement.
The volume is shown in three planes at right angles: axial (horizontal), coronal (seen from the front) and sagittal (seen from the side). With linked views, one click moves all three to the same point.
Brightness and contrast, the window, choose which grey values show as black, grey and white. They change how the image looks, not the data.
A panoramic reconstruction lays the arch out flat along a curve you draw. Cross-sections cut the arch at right angles to that curve.
A length is measured in the plane of the view. A structure tilted out of that plane measures shorter than it is.
CBCT grey values are not reliable as absolute density values. The whole volume is read, every slice in every plane, not only the area of interest.

What you open: the volume and the voxel

Scarfe and Farman describe how a cone-beam CT scanner acquires its image: a cone-shaped X-ray beam and a detector turn around the head, and the many projection images taken during the turn are reconstructed into one three-dimensional volume [1]. They also describe how image quality depends on geometric and software parameters [1]. So two scans can look different even before anyone reads them.

The volume is a block of voxels: small boxes, like the pixels of a photo, but with depth. Each voxel holds one number, its grey value, and the viewer shows that number as a shade of grey. The voxel size is the distance from one voxel to the next, in millimetres. It is fixed when the scan is made. No detail smaller than one voxel can show on its own, and the viewer uses the voxel size to turn every distance on the screen into millimetres. Voxels need not be cubes: the size can differ in the three directions.

In Simplified CBCT, the panel Scan information shows the size of the volume in voxels, the voxel size in each of the three directions, and the field of view in millimetres.

Figure 1. The scan volume is a block of voxels. Each voxel has a size in x, y and z, in millimetres, and holds one grey value. Schematic, not to scale.

The three planes

A screen is flat, so the volume is shown as slices: flat cuts through the block. Three slice directions at right angles are standard:

Axial: horizontal slices, as if looking at the head from above or below. Moving through them goes from the lower edge of the scan to the upper edge.
Coronal: upright slices from side to side, as if looking at the face from the front. Moving through them goes from the front of the face to the back.
Sagittal: upright slices from front to back, as if looking at the head from the side. Moving through them goes from one side to the other.

Each plane is at right angles to the other two. So every voxel lies on one slice of each plane, and the three slices through a point cross at that point.

Figure 2. The three planes through one point of the volume. Axial: horizontal. Coronal: parallel to the face. Sagittal: parallel to the side of the head. The colours are the ones Simplified CBCT uses for each plane. Schematic.

Linked views and the crosshair

Viewers usually show the three planes side by side, linked by a crosshair. In each view, two lines mark where the other two planes cut it. Click a point in one view, and the other two move to the slices through that point. This is how you follow a structure from one plane to the next: find it in one view, click it, and look at it in the other two.

The letters at the edges of each view give the direction: A anterior, P posterior, R right, L left, S superior, I inferior. Right and left are the patient’s.

Figure 3. Linked views. In each view, the two lines show the other two planes, in their colours: amber for axial, cyan for coronal, magenta for sagittal. The edge letters are those of Simplified CBCT. The outlines are simple shapes, not anatomy.

In Simplified CBCT, the axial view has the front at the top and the patient’s right on the left of the screen. The coronal view has the patient’s right on the left. The sagittal view has the front on the left. Click or drag with the Cross tool to move the crosshair. The views can be shown 2 × 2, 1 + 3, or one at a time.

Orienting the volume

The head is rarely perfectly straight in the scanner. A slight turn or tilt means the axial slices are not parallel to the reference planes you are used to, and matching structures on the left and right are not on the same slice. Some viewers let you turn the planes so that the slices line up with a reference you choose; slices at such an angle are called oblique.

Whatever the viewer, check the orientation first: the face at the front, the top of the head up, and the patient’s right where the letters say. A scan can be exported with one direction the wrong way round.

In Simplified CBCT, the three planes follow the scan’s own voxel grid, as the scan was exported. They cannot be turned, and there is no oblique plane. If the scan was exported the other way round, Orientation fixes flips it left–right, front–back or up–down. Because the planes cannot follow a tilted head, the tilt stays in every view: compare the two sides on more than one slice.

Scrolling through slices and slice thickness

A slice shows only what lies in that cut. To understand a structure, scroll through it: watch it appear, change and disappear, slice by slice, in at least two planes. Something that looks separate on one slice may join something else on the next.

A single slice is usually one voxel thick. A slab combines several neighbouring slices into one image, in one of three ways:

Brightest (maximum intensity projection, MIP): each point shows the brightest voxel in the slab. A bright structure that crosses several slices shows along its whole path.
Average: each point shows the mean of the voxels in the slab. The image looks smoother and less grainy.
Darkest (minimum intensity projection, MinIP): each point shows the darkest voxel. Dark spaces, such as the airway, stand out.

A thick slab shows a structure that a thin slice would cut. But it lays everything in the slab on top of everything else, as a radiograph does: structures overlap, and depth within the slab is lost.

Figure 4. One slice and a slab. The slanted bar leaves a thin slice quickly, so one slice shows only a short piece of it. A slab combined as brightest shows the part of the bar inside the whole slab, flattened into one image. Schematic.

In Simplified CBCT, you scroll with the mouse wheel or the up and down arrow keys, or with the slice bar under each view on a phone or tablet. Each view shows its slice number, its position in millimetres and the grey value at the crosshair. Slice thickness 0 shows one slice. A slab up to 40 mm thick combines slices as brightest (MIP), average or darkest (MinIP). Slab views are in the Plus and Pro plans.

Brightness and contrast (windowing)

A scan holds far more grey values than a screen can show as separate shades. The window chooses which part of that range you see. It has two numbers:

Width: how many grey values are spread from black to white. A narrow window gives strong contrast within a small range. A wide window shows a larger range with less contrast.
Centre, also called level: the grey value in the middle of the window. Values below the window show black, and values above it show white.

Figure 5. The window maps a range of grey values to the shades of the screen. A narrower width gives more contrast; moving the centre makes the whole image darker or brighter.

The window changes only how the image is shown, never the data. But it changes what you see. A narrow window can push a thin structure into black or white, and it moves where an edge seems to be. When you measure, or compare two scans, use the same setting each time.

In Simplified CBCT, the presets Bone, Teeth, Soft tissue and Airway, with Auto and Full range, set the window from the spread of grey values in the scan you opened, not from fixed numbers. So the same preset can give different numbers on two scans. The Contrast tool sets the window by dragging: sideways for contrast, up and down for brightness. The panel also takes the centre and width by hand, and Invert swaps black and white for a film look. The centre (C) and width (W) in use are shown under each view. All of this is in the Free plan.

Panoramic reconstruction and cross-sections

A panoramic reconstruction starts from a curve drawn along the dental arch on an axial slice. The viewer samples the volume along the curve, through the height of the scan, and lays the curve out straight: the arch unrolled into one flat image. A cross-section is a slice at right angles to the curve at one point. It shows that part of the arch from the cheek or lip side to the tongue or palate side.

The panorama depends on the curve. A curve placed off the middle of the arch shows the teeth thinner, or leaves part of them out of the image. Distances across the panorama are measured along the curve, not in straight lines through the head.

Figure 6. Building a panorama. Points clicked along the arch on an axial slice give a curve (blue); the shaded band is the slab sampled around it. The panorama lays the curve out straight. Each short line across the arch is a possible cross-section; the middle one is the cross-section shown, marked on the panorama by the vertical line. Schematic shapes, not anatomy.

In Simplified CBCT, choose the Arch tool and click 5 to 9 points along the arch on one axial slice, from the right molar to the left molar, then Build panorama. The tool draws a smooth curve through the points. The panorama combines a slab around the curve as brightest values (MIP); the slab is 20 mm thick unless you change it (4 to 40 mm). The curve is the same at every height of the scan, so the surface it cuts is upright: it does not lean with the teeth. A slider moves the cross-section along the arch and gives its position in millimetres along the arch. The cross-section is one slice thick and 30 mm wide unless you change it (10 to 60 mm), with the outer surface on the left and the tongue or palate side on the right. Both images can be saved as PNG files. Lengths and angles are measured in the axial, coronal and sagittal views, not on the panorama or the cross-section. The panorama and cross-sections are in the Plus and Pro plans.

Making a measurement

A length needs two points. An angle needs three, with the vertex in the middle. The viewer turns the distance on the screen into millimetres with the voxel size. Three things limit the result:

Where the points go. The edge of a structure is a change in grey over a few voxels, and the window you use moves where the edge seems to be.
The voxel size. A structure only one or two voxels across is hard to measure, because its edges are blurred over the same distance.
The plane. A length is measured in the plane of the view. A structure tilted out of that plane is shorter on the screen than it really is.

Worked example: a tilted structure

This example is a drawing, not a scan: a straight structure 20.0 mm long, tilted 30° to the axial plane, with the tilt from front to back.

On the sagittal slice through it, the whole structure lies in the view. The length reads 20.0 mm, and the angle to the axial plane 30.0°.
On one axial slice, only a short piece shows: the structure climbs through 20.0 × sin 30° = 10.0 mm of height, many slices.
On an axial slab thick enough to hold all of it, its whole outline shows. But the length you can draw is its shadow on the axial plane: 20.0 × cos 30° = 17.3 mm, 2.7 mm short.

The true length needs a plane that contains the structure.

Figure 7. The worked example. In the sagittal view, which contains the structure, it measures its full 20.0 mm. On the axial slab, it is flattened onto the axial plane and measures 17.3 mm. Drawn to scale.

In Simplified CBCT, Length takes two clicks and Angle three, the middle one the vertex. The values use the voxel size of the scan in each direction. Each measurement belongs to the view and the slice where it was drawn, and shows when you come back to that slice; the list in the panel takes you back to it. Because the planes cannot be turned, a structure tilted in two directions at once lies flat in none of the three views, and any length drawn on it is shorter than its true length. Length and angle are in the Free plan. Plus and Pro can copy every measurement as text.

Common artefacts

An artefact is something in the image that is not in the patient. Three are common, and they are worth recognising before you look at anything else:

Metal streaks. Dense metal, such as fillings, crowns, brackets and implants, sends bright and dark streaks across the slice, outward from the metal. They can hide what lies next to the metal, or between two pieces of metal. Scrolling through the nearby slices, and changing the window, shows how far they reach.
Motion. If the patient moves during the scan, edges look doubled or blurred. The whole head moves, so the effect is not limited to one small area.
Noise. A grainy, speckled look, strongest in thin slices and with a narrow window. An average slab smooths it, at the cost of fine detail.

Figure 8. Three common artefacts, drawn as simple shapes. They are illustrations of what to look for, not scans.

Grey values are not Hounsfield units

In medical CT, grey values are calibrated to the Hounsfield scale, which is anchored to water and air, so a value can be read as a measure of density. The Simplified CBCT landing page puts the difference this way: “CBCT gray values are not calibrated like hospital CT. Use them for shape and position, not for bone density.”

Dental CBCT grey values are not reliable as absolute density values. Pauwels and colleagues found large errors when dental CBCT grey values were used quantitatively, for density estimation [2]. Read a grey value as a shade within one scan, not as a number to compare with a table or with another scan.

In Simplified CBCT, the Value under each view is the number stored in the scan file for the voxel at the crosshair, with the scaling written in the file. It is a grey value, not a density.

Read the whole volume

The volume holds everything inside the field of view, not only the region the scan was taken for. A complete read goes through every slice in each of the three planes, from one edge of the volume to the other, with a window for bone and another for the soft tissues and the airway, before you look closely at the area of interest. A fixed order, used every time, makes it less likely that a part of the volume is skipped.

In Simplified CBCT, the arrow keys and the ‹ and › buttons of the slice bar move one slice at a time, and the presets change the window in one click.

Published guidelines

Guidelines on dental CBCT address justification, optimisation and referral criteria [3]. A position statement on CBCT in orthodontics found no clear evidence supporting its routine use in orthodontics, and says that its use should be decided case by case [4]. A review states that its use should be justified on the basis of its merits relative to its risks [5]. This article covers how to read a scan, not when one should be taken.

Limits

Schematic figures. The figures are drawings made of simple shapes. A real scan has far more detail, and more noise.
Viewers differ. Names, presets, the layout on screen and the tools on offer change from one viewer to the next. This article gives the general ideas, and what Simplified CBCT does.
The screen matters. A dim, small or reflective screen hides contrast that the scan holds.
Navigation, not interpretation. This article explains how to move through the image and measure it. It does not explain what any appearance means, and it does not replace a radiology report.

Do this in Simplified CBCT

Simplified CBCT opens a CBCT scan exported as uncompressed DICOM: a folder of files, a multi-frame file or a ZIP of the folder. The Free plan has the three linked views, the brightness presets and length and angle measurements. Plus adds slab views, the panorama with cross-sections, and screenshots. Pro adds synthetic cephalograms and the 3D view with STL export. The landing page says: “The scan is opened and viewed in your browser, and it is not uploaded, even when you are signed in. When you are signed in, the Studio sends only a random number that reveals nothing about the scan, so that each scan is counted once for your plan.” To practise without a scan, choose Try the demo scan: a computer-generated test object made of simple shapes, not a patient scan.

Open Simplified CBCT About Simplified CBCT

Educational content for clinicians and students. This page explains how to move through and measure a CBCT image. It does not cover diagnosis, or when a scan should be taken, and it does not replace clinical judgement.

References

Scarfe WC, Farman AG. What is cone-beam CT and how does it work? Dent Clin North Am. 2008;52(4):707-730. doi: 10.1016/j.cden.2008.05.005
Pauwels R, Nackaerts O, Bellaiche N, et al. Variability of dental cone beam CT grey values for density estimations. Br J Radiol. 2013;86(1021):20120135. doi: 10.1259/bjr.20120135. PMID: 23255537
European Commission. Radiation Protection No 172: Cone beam CT for dental and maxillofacial radiology. Evidence-based guidelines. Luxembourg: European Commission; 2012.
American Academy of Oral and Maxillofacial Radiology. Clinical recommendations regarding use of cone beam computed tomography in orthodontics. Position statement. Oral Surg Oral Med Oral Pathol Oral Radiol. 2013;116(2):238-257.
Kapila SD, Nervina JM. CBCT in orthodontics: assessment of treatment outcomes and indications for its use. Dentomaxillofac Radiol. 2015;44(1):20140282. doi: 10.1259/dmfr.20140282. PMID: 25358833