Routine CT head (non-contrast brain)

The most-requested CT examination: fast, widely available and the first test for head injury, stroke and sudden neurological change. Get the head straight, the range tight and both a brain and a bone reconstruction out of one acquisition, and every other head protocol is a variation on it.

By the ScanFlo Editorial Team

CT protocols/Head

When this study is used

• Head injury: suspected intracranial bleed, skull fracture or raised intracranial pressure.
• Suspected stroke, before thrombolysis, to exclude haemorrhage (often followed by CT angiography and perfusion).
• Sudden severe headache, new seizure, acute confusion or reduced consciousness.
• Follow-up of known haemorrhage, hydrocephalus or a shunt.
• When MRI is unavailable, contraindicated or the patient cannot lie still for it.

Patient preparation

• Confirm identity, the request and the clinical question; check pregnancy status where relevant.
• Remove hairpins, earrings, glasses, hearing aids and removable dentures: metal in the field causes streak artefact.
• Explain the scan: it is quick, the table moves, and the patient must keep the head completely still.
• Non-contrast: no fasting or cannula needed for a routine head.

Positioning

• Patient head first and supine, with the head in the dedicated head holder.
• Tuck the chin slightly so the orbitomeatal line (OML) is perpendicular to the table top. This lines the axial slices up with the skull base and keeps both sides symmetric.
• Keep the head straight: no rotation and no tilt, the interpupillary line horizontal.
• Immobilise with foam pads beside the head and a forehead strap.
• Switch on the lasers: the sagittal laser on the midline, the lateral laser at the external auditory meatus (EAM); adjust the table height until the lateral laser meets the EAM.
• Move the table into the gantry with the head at the isocentre.

Topogram

Take a lateral topogram (Tube Position: Lateral) covering the whole head and upper neck. It is a low-dose scout: Load, hold the table move key, then press Start. The range, field of view and centring are all planned on it.

Range planning

• Select the head sequence; the yellow box on the topogram is the scan range.
• Start at the skull base (the foramen magnum) and end at the vertex, so the posterior fossa and the whole calvarium are covered.
• Do not extend far above the vertex or below the foramen magnum: every extra centimetre adds dose (DLP rises in direct proportion to length) and adds nothing.
• This case: −295 to −445 mm, a 15 cm range, craniocaudal direction.
• Check CTDIvol and DLP on the Scan tab before you scan, not after.

Acquisition parameters

Values from the ScanFlo head case the video scans. Scanners and departments differ. What must not change is understanding why each sits where it does.

Reconstructions

One acquisition, several reconstructions: a soft kernel for the brain and a sharp kernel for bone, thick for reading and thin for reformats and 3D. More in Reconstruction, kernels and windows.

3D, MPR and Batch Loop

• Load the 1 mm brain series in the 3D card: three MPR panes (axial, coronal, sagittal) and one volume-rendered (VR) pane.
• VR renders the skull from a bone-windowed copy; drag to rotate, and use the A, L and S buttons for straight anterior, lateral and superior views. In trauma, VR shows fractures and depressed bone at a glance.
• Batch Loop builds reformatted series for filming: choose the plane (here coronal), Set Start on the first slice and Set End on the last, choose the number of images (20) or the spacing, name the series (Brain COR) and Send to Filming.
• Reformat from the 1 mm reconstruction. Reformats of 5 mm slices are coarse and stepped.

Filming

• Batch series arrive in Filming already laid out; choose the layout (here 4 by 5) so the series fits one page.
• Right-click a cell for Properties: layout, text and apply-to-all-pages.
• Film or send both brain and bone windows. A head CT is read on both.
• On a real scanner, print or send to PACS from here (printing is switched off in ScanFlo for now).

Dose

• CTDIvol describes the protocol's output in a 16 cm head phantom; DLP is CTDIvol × scan length. Neither is the patient's dose.
• The range is the lever you control most: keep it from the foramen magnum to the vertex.
• Tucking the chin keeps the orbits out of the direct beam for most of the scan; never scan the lenses for convenience.
• Compare the displayed CTDIvol and DLP with your department's diagnostic reference level before you press Start.

How CTDIvol and DLP are defined and compared: CTDIvol and DLP.

Review checklist

• Head straight: the falx vertical and both sides symmetric on the first axial images.
• Coverage complete: foramen magnum to vertex, no missing calvarium.
• No motion: sharp bone edges, no doubled contours.
• Both brain and bone reconstructions present, each in its own window.
• 1 mm series available for MPR and 3D; reformats checked in all three planes.
• CTDIvol and DLP recorded and within the reference level.

Mistakes that cost repeats

• Head tilted or rotated: asymmetric ventricles and petrous bones mimic or hide pathology.
• Range extended above the vertex or into the neck: extra dose, no extra information.
• Earrings, hairpins or dentures left in: streaks across the posterior fossa and skull base.
• Only a 5 mm series reconstructed: no usable reformats or 3D for fractures.
• Reading a fracture on the brain window: fine fractures disappear; always review the bone window.

Common questions

Why tuck the chin so the OML is perpendicular to the table?

It aligns the axial slices with the skull base so both sides are symmetric and comparable, reduces beam-hardening streaks from the petrous bones, and keeps the orbits out of the direct beam for most of the scan.

Where should a CT head range start and end?

From the foramen magnum (skull base) to the vertex. That covers the posterior fossa and the whole calvarium; anything beyond adds dose without information.

Why reconstruct both a soft and a sharp kernel?

A soft kernel such as H31s keeps noise low so grey and white matter, bleeds and infarcts are visible; a sharp kernel such as B70s preserves edges so fractures show. One acquisition gives both; there is no second scan.

What window settings are used for a head CT?

Brain around W80 / L40 for the brain itself, and bone around W2000 / L500 for the skull. A head CT is always read on both.

Is contrast needed for a routine CT head?

No. A routine head is non-contrast: acute blood is bright on its own. Contrast is added for specific questions such as tumour, infection or a CT angiogram.