CT Protocol Mastery — Day 3

Master the CTA brain and carotids protocol with 7 evidence-based steps covering bolus tracking, contrast injection parameters, venous contamination avoidance, and radiation dose optimisation for acute stroke imaging.

7 Critical CTA Brain & Carotids Protocol Steps Every Radiographer Must Master

📅 June 14, 2026 ⏱ 38-minute read 🗂 Neuroradiology | CT Protocols Medically Reviewed

At a glance — CTA brain & carotids

kVp

100

Low kV for better iodine CNR

Pitch

0.8

Balanced speed & resolution

300–400

Auto-mA preferred

Contrast

80 mL

@ 4.5 mL/s + 100 mL saline

Trigger

120 HU

Bolus tracking — aortic arch ROI

Key HU target

>250 HU

Endoluminal arterial opacification

Rotation time

0.5 s

Sub-second for temporal resolution

Coverage

Arch → Vertex

Caudal-to-cranial acquisition

⚠ Primary scanning pitfall: Venous contamination — if the bolus-tracking trigger threshold is set too low or the injection flow rate is too slow, contrast fills the jugular veins before the arterial phase is complete, obscuring carotid visualisation.

Introduction — why CTA brain & carotids is the cornerstone of stroke imaging

In the hyperacute management of ischaemic stroke, every minute of delayed arterial imaging translates into approximately 1.9 million neurons lost per minute.^[1] CTA brain and carotids — computed tomography angiography covering both the extracranial cervical vessels from the aortic arch and the intracranial circulation to the vertex — has become the single most time-critical imaging step in modern stroke care. It identifies large vessel occlusion (LVO), grades carotid stenosis, detects dissection and aneurysm, and directly determines eligibility for mechanical thrombectomy.

This is Day 3 in the 30-Day CT Protocol Mastery Series. Having covered the non-contrast brain CT (Day 1) and the contrast-enhanced brain CT (Day 2), this article addresses the technically more demanding CTA examination, which combines a fast injection bolus, sub-second rotation, low-kVp acquisition, and precise bolus-tracking timing into a single, seamless workflow.

For radiographers, the challenge is coordinating a 4.5 mL/s injection with an automated aortic arch trigger at 120 HU while minimising venous contamination. For radiologists, the challenge is interpreting heavily calcified carotid bifurcations, tortuous vessels, and subtle intimal flaps under time pressure. For clinicians, the risk lies in over- or under-estimating stenosis severity when calcium blooming or contrast artefact is present.

🩺 Clinical context

CTA brain and carotids is recommended as the first-line investigation in all patients with a National Institutes of Health Stroke Scale (NIHSS) score ≥6 presenting within 24 hours of symptom onset, and in any patient where mechanical thrombectomy is clinically considered. The American Heart Association / American Stroke Association (AHA/ASA) 2019 guidelines designate CTA as the imaging standard of care for LVO screening, superseding conventional catheter angiography as the initial diagnostic tool.^[2]

Unlike the non-contrast brain CT, which relies on density differences between brain structures, CTA brain and carotids is fundamentally a vascular contrast study — the diagnostic value depends entirely on achieving sufficiently opacified arteries (endoluminal HU >250) at precisely the right moment. Timing failure, low flow rates, or a poorly placed bolus-tracking region of interest (ROI) are the most common reasons for non-diagnostic examinations, wasted contrast, and delayed patient management.

This article provides a fully referenced, protocol-level breakdown of every parameter required for a diagnostic CTA brain and carotids study, aligned with the current ESO, AHA/ASA, ACR, and ESNR guidelines. It addresses all three professional audiences — radiographers who acquire the study, radiologists who interpret it, and clinicians who act on its results.

Anatomy & HU values — the vascular roadmap from arch to vertex

Understanding the anatomical territories covered by a CTA brain and carotids study is essential for protocol design, ROI placement, and systematic interpretation. The acquisition spans from the aortic arch superiorly through the cervical carotid and vertebral arteries, across the skull base, and terminates at the vertex of the intracranial compartment.

Full HU reference table for CTA brain & carotids

Structure	Normal HU range	Abnormal / pathological HU	Clinical significance
Opacified internal carotid artery (ICA)	250–450 HU	<200 HU = under-opacified (flow/timing error)	Diagnostic target; must exceed 250 HU
Opacified middle cerebral artery (MCA) M1	250–400 HU	Filling defect = thrombus / occlusion	LVO detection; abrupt cut-off = thrombectomy candidate
Basilar artery	250–400 HU	Hyperdense on NCCT >50 HU = acute thrombosis	Posterior circulation stroke
Aortic arch (bolus-tracking ROI)	120 HU trigger point	Below threshold = premature scan trigger	Sets automated scan start
Carotid bifurcation (soft plaque)	60–100 HU	Vulnerable plaque; risk of ulceration and embolism	Stroke aetiology assessment
Carotid bifurcation (calcified plaque)	400–1,200 HU	Blooming artefact overestimates stenosis	Requires MIP / curved-MPR post-processing
Carotid dissection — intimal flap	40–80 HU (mural haematoma)	Bright flap on T1 fat-sat MRI but CT may show crescentic wall thickening	Anticoagulation decision
Saccular aneurysm (opacified)	250–400 HU	Partially thrombosed aneurysm: 40–80 HU in thrombus	Rupture risk; surgical/endovascular planning
Venous structures (target-free)	<150 HU in pure arterial phase	>200 HU in jugular vein = venous contamination	Scan failure; rescan or adjust protocol
Grey matter (non-enhanced)	35–40 HU	<30 HU = cytotoxic oedema (early infarct)	Background brain assessment on unsubtracted images
Normal white matter	25–30 HU	Loss of grey–white differentiation = acute ischaemia	Assess ASPECTS score on CTA unsubtracted images
Moyamoya collateral vessels	200–350 HU	Multiple tiny enhanced vessels in basal ganglia region	Puff of smoke appearance on MIP
Dural venous sinus (opacified, delayed)	200–350 HU	Central filling defect = dural sinus thrombosis	CT venography timing required; separate protocol
AVM nidus	200–350 HU	Tangle of vessels with early draining vein on arterial phase	Neurosurgical planning; Spetzler–Martin grading

Extracranial vascular anatomy

The aortic arch gives rise to three principal branches in most individuals: the brachiocephalic (innominate) trunk, the left common carotid artery (CCA), and the left subclavian artery. The brachiocephalic trunk divides into the right CCA and right subclavian artery. Each CCA bifurcates at approximately the C3–C4 level into the external carotid artery (ECA) and internal carotid artery (ICA). The ICA has no branches in the neck and enters the skull base through the carotid canal, traversing the petrous and cavernous portions before entering the intracranial compartment as the supraclinoid ICA.

The vertebral arteries arise from the subclavian arteries bilaterally, ascending through the transverse foramina of C6–C1 before entering the skull through the foramen magnum, ultimately uniting to form the basilar artery at the pontomedullary junction. Together, the carotid and vertebrobasilar systems form the Circle of Willis — the critical anastomotic ring whose completeness determines collateral flow potential in the setting of acute occlusion.^[3]

Intracranial vascular anatomy

The supraclinoid ICA divides into the anterior cerebral artery (ACA) and the middle cerebral artery (MCA). The MCA M1 segment runs horizontally in the Sylvian fissure before dividing into superior and inferior M2 divisions, then M3 and M4 opercular branches. Large vessel occlusion most commonly involves the M1 segment, producing the maximal ischaemic territory eligible for thrombectomy. The ACA A1 and A2 segments supply the medial frontal and parietal cortex. The posterior communicating arteries connect the ICA to the posterior cerebral arteries (PCA), which arise from the basilar tip and supply the occipital lobes, thalami, and upper brainstem.

Key anatomical variants affecting the CTA protocol

Several common anatomical variants have direct implications for CTA acquisition and interpretation. A bovine arch — where the left CCA originates from the brachiocephalic trunk rather than the aortic arch directly — is present in approximately 13% of patients and may alter contrast bolus dynamics.^[4] A hypoplastic vertebral artery (diameter <2 mm) may appear occluded on CTA and must be correlated with the contralateral side before a pathological diagnosis is made. An incomplete Circle of Willis — present in 20–25% of the population — predicts poorer collateral territory and heightened stroke risk in the setting of proximal ICA or MCA occlusion.

⚠ Anatomical trap

A fenestrated basilar artery — a congenital duplication where the basilar artery splits into two separate channels before reuniting — can mimic a basilar dissection or thrombus on axial CTA images. Recognition of this normal variant prevents inappropriate anticoagulation decisions.

🩻

Upgrade your CT vascular imaging workflow

SATMED Health’s pressure-rated line sets and SATLine injector accessories are engineered for high-flow CTA protocols requiring consistent 4.5 mL/s injection rates across all patient access types.

Explore SATLine for Vascular CTA →

Scanning technique — 7 essential steps for diagnostic CTA brain & carotids

Patient positioning and preparation: Supine position, arms by the sides (upper extremities should not be raised for head and neck CTA — this risks venous obstruction artefact in the thoracic inlet). Head in the head holder with midline alignment to minimise asymmetric streaking. Instruct the patient explicitly: “Do not swallow during the scan, and hold your breath gently when I tell you.” A single shallow respiratory hold during the neck acquisition dramatically reduces swallowing artefact across the larynx and carotid bifurcation. For agitated or confused stroke patients, time the acquisition during a natural respiratory pause.
IV access and pressure-rated equipment: Establish 18-gauge IV access in the right antecubital fossa as the preferred site. The left arm may be used, but contrast injection from the left may pass through a persistent left superior vena cava in up to 0.3% of patients, delaying arterial opacification. Connect a high-pressure-rated line set (rated to ≥300 PSI) and pre-fill with normal saline. Confirm no extravasation risk at 1 mL/s test injection. The contrast injector must be set to 4.5 mL/s for the 80 mL contrast bolus and 3.0 mL/s for the 100 mL saline chaser, which maintains the contrast column velocity and reduces venous contamination.^[5]
Scout and scan range: Acquire a lateral scout projection first. Define the scan range from the carina (to include the aortic arch for bolus-tracking ROI placement) superiorly through to the vertex of the skull. On a 64-slice scanner, a typical field of view of 400–450 mm (neck) transitioning to 250 mm (brain) with a 512 reconstruction matrix yields adequate voxel size. Confirm coverage includes the posterior fossa, as basilar artery and posterior inferior cerebellar artery (PICA) territory infarcts are frequently missed if superior extent is insufficient.
Bolus-tracking ROI placement and trigger threshold: Place the circular bolus-tracking ROI (minimum 50 mm²) in the centre of the descending aorta at the aortic arch level — not in the ICA, ascending aorta, or left ventricle. The 120 HU trigger threshold reflects the established consensus for adequate arch opacification preceding carotid peak enhancement by approximately 4–6 seconds at 4.5 mL/s injection rates.^[6] Monitor the dynamic tracking curve live; in patients with low cardiac output, the bolus may arrive 5–10 seconds later than average — allow the threshold to be reached naturally before initiation.
Acquisition parameters and scan direction: Acquire caudal-to-cranial (feet-to-head direction), starting at the aortic arch and progressing to the vertex. This prevents the scan from “outrunning” the contrast bolus in the intracranial vessels while the neck vessels are still filling. At pitch 0.8, rotation time 0.5 s, and 100 kVp, the acquisition proceeds at approximately 40–60 mm/s on a 64-slice scanner. At 320-slice, total acquisition time is reduced to under 3 seconds. Reconstruction interval: 0.625 mm with 50% overlap for MPR/MIP post-processing.
Post-processing and reconstructions: Automated workstation reconstructions should include: (a) axial slices at 1.25 mm and 2.5 mm in soft-tissue and bone windows; (b) coronal and sagittal multiplanar reconstructions (MPR) at 3 mm for cervical vessels; (c) curved-MPR of the ICA bilaterally from origin to terminus; (d) Maximum Intensity Projection (MIP) in coronal, sagittal, and oblique planes for aneurysm screening and collateral assessment; and (e) Volume Rendering Technique (VRT) for neurosurgical planning and carotid stenosis measurement. Send all series to PACS simultaneously to avoid reporting delays.
Quality check before releasing patient: Before disconnecting the patient, review the preliminary images on the scanner console for: (1) adequate endoluminal arterial HU >250 in the ICA and MCA M1; (2) absence of jugular or sigmoid sinus contamination obscuring carotid bifurcation; (3) complete cranio-caudal coverage including the full cervical ICA and basilar artery; and (4) absence of motion artefact crossing the skull base. If any criterion is not met, discuss with the supervising radiologist whether an immediate re-scan (with contrast reduction) is warranted or whether post-processing can salvage the images.

Scanner comparison table — CTA brain & carotids

Scanner type	Slices	Rotation time	kVp	mA range	Pitch	Acquisition time	Key consideration
16-slice	16	0.5 s	120	280–350	0.75	~20–25 s	Slow; consider test bolus over tracking; use triphasic if venous contamination risk
64-slice	64	0.5 s	100	300–400	0.8	~10–14 s	Workhorse; bolus tracking standard; 100 kVp feasible in most patients
128-slice dual source	128×2	0.28 s	100/Sn140	300–400	0.8	~5–7 s	Dual-energy; virtual non-contrast for aneurysm coil artefact reduction
256-slice	256	0.27 s	100	280–380	0.8	~4–5 s	Wide coverage; single rotation for skull base through vertex
320-slice	320 (16 cm)	0.275 s	100	280–380	N/A (volumetric)	~2–3 s	Full brain in single rotation; 4D-CTA time-resolved option; preferred for complex AVM/aneurysm
Photon-counting CT (PCCT)	144×2	0.25 s	90–100	220–350	0.8	~4–6 s	Spectral data at all kVp; superior plaque composition; 0.2 mm spatial resolution; K-edge contrast

Dual-energy and photon-counting CT protocol for CTA brain & carotids

Technique	kVp pairs / setting	Key CTA application	Post-processing output	Clinical gain
Dual-source DECT	100 kVp / Sn 140 kVp	Calcified carotid stenosis grading	Virtual non-calcium (VNCa) images, monoenergetic series at 40–70 keV	Removes blooming; true lumen diameter assessment
Single-source rapid kV-switching	80 / 140 kVp rapid alternation	Iodine map; small aneurysm detection	Material decomposition: iodine map, water map	Iodine quantification; reduces contrast volume in CKD patients
Low-keV monoenergetic (40 keV)	From any DECT acquisition	Enhances iodine signal in under-opacified vessels	Monoenergetic image series	Salvage of suboptimal contrast bolus; improved SNR in bariatric patients
Photon-counting CT (PCCT)	120 kVp polychromatic or spectral	Plaque characterisation; ultra-high-res CTA	0.2 mm reconstructions; spectral maps	Identifies vulnerable plaques; superior calcification margin definition

Deep learning reconstruction (DLR) in CTA brain & carotids

Vendor-deployed deep learning reconstruction (DLR) algorithms — including GE’s TrueFidelity, Siemens’ AI Rad Companion (Neuro), and Canon’s AiCE — have demonstrated 30–45% noise reduction over hybrid-iterative reconstruction at equivalent dose levels in CTA head and neck studies.^[7] For CTA brain and carotids, DLR specifically benefits the skull base region, where conventional iterative reconstruction tends to amplify beam-hardening noise from the petrous bone — precisely the zone where the carotid siphon and vertebral arteries must be clearly delineated. Incorporating DLR into the standard reconstruction chain (typically at a “medium-high” strength setting to preserve edge sharpness without texture smoothing) is recommended in all institutions where the technology is available.

✅ Best practice

Always reconstruct CTA brain and carotids with at least two kernel series: (1) a smooth/standard kernel for brain parenchymal windows and vessel lumen assessment, and (2) a sharp/edge kernel for skull base, temporal bone, and dense calcification evaluation. A single kernel reconstruction is insufficient for diagnostic quality across all anatomical zones.

Contrast media protocol — achieving peak arterial opacification without venous contamination

The contrast media protocol for CTA brain and carotids is one of the most technically demanding in routine CT practice. The goal is to deliver a sharp, high-concentration iodine bolus that achieves endoluminal arterial opacification exceeding 250 HU at the moment the acquisition traverses each vessel segment — from the aortic arch through to the pericallosal arteries. Achieving this requires precise coordination of contrast concentration, injection volume, flow rate, saline chaser, and scan timing.

Full injection protocol specification

Parameter	Standard value	Range / adjustment	Rationale
Contrast concentration	350–370 mgI/mL	300 mgI/mL in CKD / cardiac risk	Higher iodine concentration improves arterial CNR at 100 kVp
Contrast volume	80 mL	60 mL (DECT 40-keV salvage); up to 90 mL (16-slice, slow scan)	Sufficient bolus length for 10–14 s acquisition at 4.5 mL/s
Flow rate	4.5 mL/s	4.0 mL/s (22G access); 5.0 mL/s (16G access, bariatric)	High flow rate ensures sharp bolus and peak arterial HU within scan window
Saline chaser volume	100 mL	80–120 mL	Flushes contrast from peripheral IV into systemic circulation, maintaining bolus peak and reducing venous contamination from axillary/subclavian stasis
Saline chaser flow rate	3.0–3.5 mL/s	Match or slightly below contrast flow rate	Maintains upstream pressure without disrupting arterial bolus geometry
Trigger method	Bolus tracking — aortic arch	Test bolus (10 mL at 4 mL/s) as alternative for cardiac low-output patients	Automated tracking eliminates variable scan delays; adapts to individual cardiac output
Trigger threshold	120 HU	100 HU (paediatric/low output); 140 HU (high output / large patients)	Arch enhancement at 120 HU predicts peak ICA enhancement within 4–6 s in standard physiology
Diagnostic delay after trigger	4–6 seconds	6–8 s in low cardiac output; 3–4 s in high output	Allows bolus to propagate from arch to proximal ICA before scan initiation
Injection temperature	37°C (body temperature warming)	Room temperature acceptable; warming reduces viscosity	Reduces vascular resistance and bolus dispersion; lowers patient discomfort
Injection site preference	Right antecubital fossa, 18G	Left arm acceptable; avoid wrist/hand in high-flow protocols	Prevents left-sided injection artefact from persistent left SVC
Pre-medication	Per institutional allergy history protocol	IV hydrocortisone 200 mg + antihistamine for prior moderate reaction history	Risk stratification per ACR Manual on Contrast Media v10.3^[8]

Renal function and contrast volume reduction

In patients with eGFR 30–60 mL/min/1.73m², contrast volume should be targeted at ≤80 mL with adequate pre-scan hydration (500 mL IV normal saline over 1 hour pre-scan if time permits in the stroke setting). For patients with eGFR <30 or on dialysis, the clinical urgency of acute stroke imaging almost always outweighs contrast nephropathy risk in the acute setting — the ESO and ACR guidance both support proceeding with CTA in the setting of LVO stroke without delaying for creatinine results.^[9]

Contrast media safety protocol

🚨 Safety critical — extravasation risk

At 4.5 mL/s, extravasation of contrast media into surrounding soft tissue can cause severe compartment syndrome, particularly in elderly or immunocompromised patients with fragile veins. Always perform a saline test flush at 1–2 mL/s (10 mL) before injecting contrast, observe the injection site throughout the bolus, and ensure the power injector’s pressure limit is set appropriately for the cannula gauge. An 18G cannula should not exceed 325 PSI; a 20G cannula should not exceed 250 PSI.

Iodinated contrast agents are classified into iso-osmolar (iotrolan, iodixanol) and low-osmolar (iohexol, iomeprol, iopamidol) agents. For CTA brain and carotids, low-osmolar agents at 350–370 mgI/mL concentration are the standard of care, offering an optimal balance between arterial conspicuity and safety profile. Iso-osmolar agents (iodixanol) may be preferred in patients with prior severe contrast reactions or renal impairment, as evidence suggests marginally lower nephrotoxicity, though this remains debated in the literature.^[10]

💉

Precision contrast delivery starts with the right injector accessories

SATJect pressure-rated syringes and SATMix mixing chambers are engineered for consistent 4.5 mL/s bolus delivery in stroke CTA protocols — reducing venous contamination events across your department.

Explore SATJect Injector Solutions →

Radiation dose — DRL benchmarks and 5 dose reduction strategies for CTA brain & carotids

CTA brain and carotids delivers a moderate radiation dose, significantly higher than a standard non-contrast brain CT due to the extended field of view (skull base to aortic arch) combined with increased tube output needed for vascular CNR. Accurate benchmarking against national and international diagnostic reference levels (DRLs) is essential for justifying the dose and demonstrating ALARA compliance.

Diagnostic reference level table — CTA brain & carotids

Parameter	European DRL (EC RP 185)	AAPM TG-204 reference	Typical departmental target	Units
CTDIvol (head component)	60 mGy	50–70 mGy	45–55 mGy	mGy
CTDIvol (neck component)	15 mGy	12–18 mGy	10–14 mGy	mGy
DLP (combined head + neck)	1,200–1,500 mGy·cm	1,100–1,600 mGy·cm	900–1,200 mGy·cm	mGy·cm
Effective dose	~4–6 mSv	3–7 mSv	3–5 mSv	mSv
SSDE (size-specific dose estimate)	—	Apply water-equivalent diameter correction	Calculated per patient at scanner console	mGy
Thyroid dose (neck acquisition)	≤20 mGy	≤25 mGy	≤15 mGy	mGy
Lens of eye dose	≤25 mGy (deterministic threshold >500 mGy)	≤30 mGy	≤20 mGy	mGy
Benchmark standard	EC RP 185 (2019)	AAPM Report 204 (2022)	ICRP Publication 135^[11]	—

5 dose reduction strategies for CTA brain & carotids

1. Reduce kVp to 100 (or 80 in patients <70 kg). Lowering tube voltage from 120 kVp to 100 kVp for CTA brain and carotids reduces dose by approximately 40% while simultaneously increasing iodine contrast-to-noise ratio (CNR) due to the closer alignment of the beam spectrum to the K-edge of iodine at 33.2 keV. In patients weighing under 70 kg, 80 kVp with compensatory mA increase remains dose-neutral to 100 kVp while improving vessel CNR by a further 20–30%.^[12]

2. Apply automatic tube current modulation (ATCM). Angular and longitudinal ATCM systems (Siemens CARE Dose4D, GE SmartmA, Philips DoseRight) reduce tube output in low-attenuation anatomical zones (frontal sinuses, posterior fossa air cells) while maintaining adequate photon fluence in high-attenuation zones (skull base, petrous bone). This typically reduces the effective dose by 20–35% without impacting arterial vessel conspicuity.

3. Optimise scan length to clinical indication. If intracranial aneurysm screening is the primary indication and carotid stenosis is not clinically indicated, restrict the inferior extent of the scan range to the skull base (C1 level) rather than the aortic arch. This eliminates the neck contribution to total DLP (approximately 300–400 mGy·cm reduction). A separate, low-dose CT neck with contrast can be added if carotid assessment becomes clinically necessary following intracranial CTA.

4. Implement deep learning reconstruction (DLR) to enable dose reduction. By replacing conventional filtered back projection (FBP) or first-generation iterative reconstruction with vendor-approved DLR, tube current can be reduced by 30–40% while maintaining equivalent image noise. Multiple prospective studies have validated DLR’s ability to preserve spatial resolution and vessel edge sharpness in CTA neuro protocols at significantly lower dose.^[13]

5. Eliminate unnecessary delayed acquisitions. For the standard stroke CTA brain and carotids protocol, a single arterial-phase acquisition is sufficient for LVO detection and carotid stenosis grading. Delayed-phase acquisitions (venous phase or 5-minute delay) should not be performed routinely — they are only indicated when dural venous sinus thrombosis (requiring a CT venography phase) or arteriovenous malformation (requiring time-resolved 4D-CTA) is specifically suspected. Unjustified multi-phase acquisitions are the single largest avoidable dose contributor in CTA neuro practice.

ℹ ICRP & ALARA reminder

CTA brain and carotids falls under ICRP Publication 105 guidelines for diagnostic and interventional radiology. The dose to the thyroid gland during CTA neck acquisitions warrants particular attention in patients who may require multiple follow-up examinations. Document all CTDIvol, DLP, and SSDE values in the radiology report as per the 2024 ACR dose index registry recommendations.

⚡

Streamline dose documentation and protocol compliance

SATMED Health’s SATPro platform integrates dose tracking, protocol optimisation alerts, and DRL benchmarking directly into your CT workflow — supporting ALARA compliance across your entire department.

Start Your Dose Compliance Journey →

Top 10 pathologies detected on CTA brain & carotids

The following pathologies represent the primary diagnostic targets of CTA brain and carotids, each requiring specific knowledge of its CTA appearance, HU characteristics, and protocol implications. Sharp endoluminal contrast opacification above 250 HU is the baseline requirement for confident filling defect detection; complete filling defects indicate thromboembolic occlusion or dissection. Where relevant, vessel-specific HU thresholds are noted.

1 Large vessel occlusion (LVO)

0 HU (abrupt cut-off)

A sudden complete filling defect in the MCA M1 or ICA terminus, surrounded by opacified branches above and below. The hyperdense MCA sign on the preceding non-contrast scan (60–90 HU) confirms acute thrombus composition. Protocol impact: precise bolus timing is critical — under-opacified MCA due to early scanning mimics LVO. Always confirm arterial HU >250 HU before calling occlusion.^[14]

2 Intracranial aneurysm

250–400 HU (opacified sac)

A saccular, fusiform, or blister aneurysm appears as an outpouching of the arterial wall, opacifying at the same attenuation as the parent artery. Partially thrombosed aneurysms demonstrate a peripheral ring of enhancement surrounding central thrombus (40–80 HU). MIP reconstructions are essential for neck-to-parent-artery relationship. Detection sensitivity of CTA for aneurysms ≥3 mm approaches 98% on modern 64-slice and above scanners.^[15]

3 Carotid artery dissection (intimal flap)

40–80 HU (mural haematoma)

A crescent-shaped mural thickening reducing the true lumen, with or without a visible intimal flap on axial CTA. The false lumen may be partially opacified or completely thrombosed. Look for an eccentric narrowing of the ICA proximal to the skull base on curved MPR. Dissection is the most common cause of stroke in patients under 45 years. Protocol: 0.625 mm reconstructions with curved MPR are mandatory for accurate lumen diameter measurement.^[16]

4 Carotid stenosis (atherosclerotic)

Mixed: soft plaque 40–100 HU; calcium 400–1,200 HU

CTA grades stenosis at the carotid bifurcation using NASCET criteria (residual lumen diameter vs distal ICA diameter). CTA has been validated at 94% concordance with catheter angiography for ≥50% stenosis. Mixed soft and calcified plaques require both windowed axial images and virtual non-calcium dual-energy reconstructions to avoid blooming-related overestimation. The primary scanning pitfall — venous contamination — can simulate or mask stenosis at the carotid bifurcation.^[17]

5 Atherosclerotic plaque (soft vs calcified)

Soft: 40–100 HU; Calcified: >400 HU

CTA distinguishes soft (lipid-rich, vulnerable) plaque from calcified plaque based on HU attenuation on native and contrast images. Soft plaques with intraplaque haemorrhage demonstrate areas of higher density (80–100 HU) within a generally low-attenuation region. Photon-counting CT improves characterisation further with reduced partial-volume averaging. Protocol impact: DLR and dual-energy virtual non-calcium series are the reconstruction chain of choice for plaque characterisation studies.

6 Fibromuscular dysplasia (FMD)

Normal arterial density; alternating segments

FMD produces the classic string of beads appearance on CTA — alternating areas of stenosis and dilatation in the mid-ICA, typically at the C2–C3 level. It is predominantly medial-fibroplasia type and must be distinguished from dissection (which is proximal) and atherosclerosis (which is at the bifurcation). CTA is the diagnostic investigation of choice, with sensitivities approaching 89% for the beaded pattern when sub-millimetre reconstructions are employed.^[18]

7 Vertebrobasilar insufficiency

Variable HU depending on stenosis severity

CTA of the posterior circulation demonstrates atheromatous stenoses at the vertebral artery origin (V1 segment), the intracranial V4 segment, and the basilar artery. Hypoplastic vertebral arteries (diameter <2 mm) are common and must be distinguished from stenosis by calibre consistency along the vessel course. The V1 origin is best assessed on coronal MIP or curved MPR, as the axial view at the subclavian artery level frequently underestimates origin stenosis severity.

8 Moyamoya disease

200–350 HU (collateral vessels)

Progressive stenosis and occlusion of the supraclinoid ICA and proximal MCA/ACA with development of extensive basal perforating collateral vessels produces the pathognomonic puff of smoke appearance on MIP reconstructions. CTA allows comprehensive assessment of bilateral disease, including posterior circulation collaterals. Protocol impact: a pure arterial phase CTA acquisition is required — venous contamination or inadequate opacification renders collateral vessel assessment unreliable.^[19]

9 Arteriovenous malformation (AVM)

200–400 HU (nidus and feeding arteries)

An AVM presents on CTA as a cluster of abnormal, directly connected arterial and venous structures with no intervening capillary bed. The nidus appears as a tangle of opacified vessels on MIP, with early draining cortical veins visible in the arterial phase. Feeding arteries are enlarged; draining veins are opacified within the first 1–2 seconds after arterial peak. CTA-based Spetzler–Martin grading requires accurate delineation of nidus size, eloquent cortex proximity, and deep venous drainage on 3D reconstructions.

10 Dural venous sinus thrombosis (DVST)

Filling defect within opacified sinus

While the standard CTA brain and carotids targets the arterial phase, a delayed venous-phase acquisition (90–120 seconds) — converting the protocol to CTA + CTV — reveals non-opacified filling defects within the superior sagittal, transverse, or sigmoid sinus. The empty delta sign — a central hypodense thrombus surrounded by enhanced dural leaves — is pathognomonic for superior sagittal sinus thrombosis. If clinical suspicion exists pre-scan, request a combined CTA/CTV protocol from the outset to avoid repeat contrast administration.^[20]

🧠

Enhance LVO detection with AI-assisted CTA triage

Connect with SATMED Health’s platform to integrate FDA and CE-cleared AI LVO detection tools that flag stroke CTA studies for immediate clinical escalation — reducing door-to-needle times across your stroke network.

Explore Stroke AI Solutions →

Pitfalls for radiographers — avoiding the 7 most dangerous CTA brain & carotids scanning errors

🟡 Primary scanning pitfall from protocol data

Venous contamination: If the smart-prep (bolus-tracking) trigger is set too low or the injection flow rate is too slow, contrast fills the jugular veins and internal jugular bulb during the acquisition, producing dense venous enhancement that directly overlies and obscures the carotid bifurcation and ICA. This creates diagnostic uncertainty over carotid stenosis, dissection, and thrombus. Mitigation: Set the trigger threshold to 120 HU at the aortic arch ROI, maintain flow rate at ≥4.5 mL/s, and scan caudal-to-cranial so that the neck vessels are acquired during the pure arterial phase before venous return occurs.

Category	Description	Mechanism	Mitigation strategy
Bolus timing (primary)	Venous contamination at carotid bifurcation	Trigger threshold too low (<100 HU); flow rate <3.5 mL/s; scan direction cranial-to-caudal	Set ROI in descending aortic arch; 120 HU trigger; ≥4.5 mL/s; caudal-to-cranial direction
IV access failure	Extravasation or flow-rate limitation with 20G or 22G cannula	Peripheral vein calibre insufficient for 4.5 mL/s; pressure-rated tubing not used	18G antecubital preferred; confirm test flush at 2 mL/s (10 mL saline); use pressure-rated line set rated ≥300 PSI
Scan direction error	Cranial-to-caudal acquisition outrunning contrast in neck	Contrast not yet reached cervical ICA when scan passes through neck	Acquire caudal-to-cranial (arch to vertex) as standard; document in protocol parameters
Trigger ROI misplacement	ROI placed in ascending aorta, left ventricle, or pulmonary trunk	These sites reach 120 HU earlier than the descending arch, triggering premature scan start	Place ROI only in descending aortic arch; verify position on scout before injection
Insufficient scan length	Inferior margin excludes aortic arch; superior margin excludes vertex	Scout prescribed from insufficient coverage range by inexperienced operators	Protocol should mandate: inferior = carina; superior = top of skull vault (not just circle of Willis)
Patient motion (stroke patients)	Head movement during acquisition produces ghost artefact at skull base	Agitation, confusion, involuntary tremor; long acquisition time on older scanners	Use head restraint; brief instruction pre-scan; 320-slice single-rotation eliminates motion on modern systems
Saline chaser omission	Contrast stasis in antecubital vein prolongs venous contamination	Operator forgets or reduces saline chaser; bolus tail enters systemic circulation too slowly	Programme saline chaser (100 mL at 3.0 mL/s) as a mandatory injector phase; audit omission rate
Wrong kVp selection	Defaulting to 120 kVp reduces iodine CNR and increases dose	Technologist overrides protocol to 120 kVp for larger patients without adjusting mA	Implement ATCM with hard-coded 100 kVp default for all CTA brain/carotids unless BMI >35 (where 120 kVp may be appropriate)

Pitfalls for radiologists — avoiding the 8 most dangerous CTA brain & carotids interpretation errors

🔴 Primary interpretation pitfall from protocol data

Calcium blooming overestimation of stenosis: Heavy circumferential calcification at the carotid bifurcation causes blooming artefacts — beam-hardening-related widening of calcified plaque beyond its true boundary — leading to severe overestimation of stenosis percentage when measured on standard axial or MIP images. A vessel that is 50–60% stenosed on catheter DSA may appear 80–90% stenosed or even occluded on standard CTA MIP, leading to inappropriate surgical referral or incorrect exclusion from thrombolysis. Mitigation: Always measure stenosis using curved-MPR reconstructions of the carotid, dual-energy virtual non-calcium series (where available), or classify as “heavily calcified, stenosis percentage unreliable — DSA correlation recommended” rather than committing to a specific NASCET percentage.

Pitfall	Mechanism	Consequence if missed	Mitigation
Calcium blooming overestimation (primary)	Beam hardening expands calcified plaque HU range beyond true margins on MIP	Overgraded stenosis → unnecessary endarterectomy; or lumen appears occluded → inappropriate thrombectomy withholding	Curved-MPR on thin reconstructions; DECT virtual non-calcium; DSA if >70% on CTA with heavy calcium
Premature trigger — under-opacification mistaken for filling defect	ROI in wrong location; trigger fired before contrast reaches vessels	Pseudo-occlusion of MCA or ICA reported; unnecessary stroke code activation	Always confirm arterial HU >250 HU in ICA or MCA M1 before diagnosing filling defect
Hypoplastic vertebral artery called stenosis/occlusion	Congenitally small calibre V1–V4 segments misinterpreted as atheromatous disease	Incorrect posterior circulation stroke mechanism; anticoagulation commenced inappropriately	Symmetric calibre throughout course; no post-stenotic dilatation; correlate with contralateral side
Infundibular origin missed as aneurysm	A normal cone-shaped posterior communicating artery origin mimics a small (<3 mm) aneurysm on thin-slice CTA	Unnecessary neurosurgical referral; patient anxiety; repeat imaging radiation	Infundibulum: symmetrical cone <3 mm, arises from parent artery; report as “infundibular origin, not aneurysm”
Dural AVF missed on arterial-phase-only CTA	Small dAVF feeding arteries may not be visible without late-phase or 4D-CTA; early draining vein subtly opacified	Aggressive dural AVF type III–V missed → haemorrhage risk unreported	If pulsatile tinnitus or unexplained intracranial venous engorgement: request 4D-CTA or catheter DSA
FMD called dissection	Mid-ICA beaded appearance of FMD can appear similar to segmental narrowings of dissection on axial images	Anticoagulation commenced for dissection when antiplatelet therapy is appropriate for FMD	FMD is mid-vessel; affects C2–C3 ICA; beaded; no mural haematoma; dissection is proximal with crescentic thickening
Partial volume averaging at skull base masking basilar tip aneurysm	Thick-slice or obliquely reconstructed series at the sellar region merge the basilar tip with adjacent bone	Basilar tip aneurysm — highest rupture risk — missed on standard axial images	Review MIP oblique sagittal and coronal reconstructions at 1.25 mm; inspect basilar tip systematically on every CTA
Post-thrombectomy re-occlusion confused with successful recanalisation	Contrast staining of ischaemic tissue (haemorrhagic transformation) after thrombectomy mimics vessel opacification	Re-occlusion unreported; patient loses repeat thrombectomy window	On post-procedure CTA: always compare vessel calibre with pre-procedure baseline; non-contrast NCCT first eliminates haemorrhagic staining

Pitfalls for non-radiology physicians — understanding the clinical dangers of CTA brain & carotids misinterpretation

Non-radiology clinicians — including neurologists, emergency physicians, and stroke fellows — frequently review CTA brain and carotids images in the acute setting before formal radiology reporting is available. Several common clinical misinterpretations can lead to direct patient harm, including withheld or inappropriate thrombolysis and thrombectomy decisions.

Pitfall	What the clinician sees	What it actually is	Clinical danger	Recommended action
Calling “no LVO” on an under-opacified CTA	Absent MCA filling; no obvious filling defect	Technically inadequate scan — premature trigger or low flow rate; vessels not opacified to diagnostic threshold	LVO missed; thrombectomy window lost; devastating stroke outcome	Before reporting “negative,” confirm MCA M1 HU >250 HU; if in doubt, request radiologist review urgently
Overestimating carotid stenosis severity due to calcium blooming	ICA appears near-occluded at bifurcation on MIP	Circumferential calcium blooming artefact — true lumen 50–60% by NASCET	Unnecessary urgent surgical referral; anticoagulation commenced; patient anxiety	Verify with curved-MPR; document uncertainty; request formal radiology report before surgical action
Calling basilar artery thrombosis on non-diagnostic images	Basilar artery not clearly visualised on axial CT due to bone artefact	Posterior fossa beam-hardening artefact obscuring basilar; may be normal or partially visualised	Emergency anticoagulation commenced on the basis of artefact; haemorrhagic transformation risk increased	If posterior circulation stroke is suspected clinically, request formal CTA with dedicated skull base MPR or MRA if CTA is non-diagnostic
Attributing venous contamination to carotid pathology	Dense jugular vein overlying and obscuring the carotid bifurcation	Venous contamination from inadequate trigger threshold or slow bolus — normal carotid underneath	Normal carotid misclassified as thrombosed; unnecessary anticoagulation; intervention for a non-existent lesion	Observe that opacified jugular vein is the dominant hyperdense structure; request repeat CTA or formal radiology review
Ignoring ASPECTS score on unsubtracted CTA source images	Clinician views only MIP reconstructions; does not review source axial images	Early ischaemic changes (ASPECTS ≤5) visible on source images predict poor thrombectomy outcome	Thrombectomy performed in patient with established infarct core; outcome worse than medical management	Always review CTA source images (brain windows 80/40 HU) for ASPECTS scoring, not only MIP vascular images
Confusing tandem occlusion topology	MCA M1 filling defect identified; ICA not reviewed	Concomitant proximal ICA occlusion or critical stenosis (“tandem lesion”) present	Thrombectomy plan targets only distal occlusion; proximal lesion causes recurrent occlusion post-procedure; incomplete revascularisation	Systematically review the full CTA from arch to vertex; report both lesions if present; confirm tandem status in thrombectomy team communication

🚨 Critical clinical safety point

In the acute stroke setting, CTA brain and carotids reports should never be acted upon definitively by non-radiology clinicians without either radiologist verification or, at minimum, confirmation that the examination was technically adequate (arterial HU >250 HU confirmed in the MCA M1). A technically failed CTA — however its images appear — must be treated as a non-diagnostic study.

🩺

Reduce CTA protocol errors with standardised injector accessories

SATMED Health’s SATSyringe and SATLine systems are purpose-designed to ensure consistent high-flow bolus delivery in your stroke CTA pathway — reducing technically failed scans by up to 37% in clinical deployments.

Discover SATSyringe for Stroke CTA →

Pitfall comparison summary — three professional perspectives on CTA brain & carotids failure modes

The following three-column summary allows teams to identify the precise point of failure in a CTA brain and carotids examination and assign responsibility for mitigation. Each column addresses the same examination through a different professional lens.

🟡 Scanning pitfalls (radiographers)

Venous contamination — wrong trigger threshold or slow bolus flow rate
IV access <18G limiting injection rate
Cranial-to-caudal scan direction outrunning neck contrast
ROI placed in ascending aorta or left ventricle
Saline chaser omitted; contrast stasis in peripheral vein
Scan length too short — excludes posterior fossa or aortic arch
Head motion — inadequate immobilisation in agitated stroke patient
120 kVp used instead of 100 kVp — increased dose, reduced iodine CNR

🔴 Interpretation pitfalls (radiologists)

Calcium blooming overestimates stenosis — MIP artefact; requires curved-MPR or DECT
Under-opacified vessels called “filling defect” or “occlusion”
Hypoplastic vertebral artery mistaken for stenosis
Infundibular PCoA origin called small aneurysm
FMD mid-vessel beading misidentified as segmental dissection
Basilar tip aneurysm missed on thick-slice axials at skull base
dAVF early draining vein overlooked on arterial-only CTA
Post-thrombectomy contrast staining confused with vessel patency

🟣 Clinical pitfalls (physicians)

Acting on under-opacified CTA as “no LVO” — missed occlusion
Calling carotid occlusion based on calcium blooming on MIP
Not reviewing ASPECTS on source images — thrombectomy in established infarct
Tandem occlusion (ICA + MCA) not identified — incomplete thrombectomy plan
Venous contamination artefact attributed to carotid pathology
Basilar non-visualisation called “thrombosis” — inappropriate anticoagulation

AI & automation in CTA brain & carotids — the evidence and available tools

The integration of artificial intelligence into CTA brain and carotids workflows has accelerated dramatically since 2019, driven by the clinical urgency of stroke care and the evidence base supporting mechanical thrombectomy for LVO. AI tools now span every stage of the workflow, from automated acquisition optimisation at the scanner to real-time LVO triage, stenosis measurement, and aneurysm detection on the PACS workstation.

Automated LVO detection and alerting

The most clinically validated application of AI in CTA brain and carotids is large vessel occlusion (LVO) detection. Multiple FDA-cleared and CE-marked tools are now in clinical use, including Viz.ai LVO (CE mark 2018, FDA 510(k) cleared 2018), Rapid AI (iSchemaView, FDA cleared 2020), Brainomix e-Stroke (CE mark 2018), and Aidoc Brain AI (FDA cleared 2018). These tools apply convolutional neural network (CNN) architectures trained on thousands of CTA datasets to flag examinations with a suspected M1/ICA/basilar occlusion, typically generating a read and mobile clinician alert within 2–5 minutes of image upload to PACS.^[21]

A 2022 prospective multicentre study of Viz.ai LVO across 15 US stroke centres demonstrated a reduction in door-to-CTA-review time from a median of 24 minutes to 8 minutes when AI alerts were delivered directly to the interventionalist’s mobile device, with a sensitivity of 87% and specificity of 93% for M1 LVO detection.^[22]

AI-assisted carotid stenosis grading

Automated carotid stenosis measurement algorithms, including those within the Siemens Teamplay AI ecosystem and GE’s AIRx platform, apply NASCET criteria to curved-MPR reconstructions generated automatically from the CTA data, delivering lumen diameter measurements with inter-reader variability below 5%.^[23] This is particularly valuable in high-volume stroke centres where the CTA must be reviewed quickly by on-call clinicians, reducing the risk of human error in NASCET calculation under time pressure.

AI aneurysm detection

AI aneurysm detection tools — including Aidoc Intracranial Aneurysm Detector (FDA cleared) and HeadXNet (Stanford University, validated externally) — demonstrate sensitivities of 94–97% for aneurysms ≥3 mm on 64-slice and above CTA acquisitions, outperforming inexperienced observers and matching attending neuroradiologist performance.^[24] These tools are not yet validated for aneurysms <3 mm or for dissecting aneurysms, and require expert radiologist oversight and confirmation before clinical action.

Automated collateral scoring

Collateral vessel status — scored using the Tan collateral grading scale (0–3) or the regional Pial Arterial Supply score — is a key predictor of thrombectomy outcome and therapeutic window extension. Manual collateral scoring on CTA MIP images is highly variable (κ = 0.45–0.60 across readers). AI-based collateral quantification tools (Rapid AI ASPECTS and Collateral, CE-marked) automate this process with demonstrated agreement with expert consensus of κ >0.75, reducing one of the largest sources of inter-institutional variability in stroke imaging triage.^[25]

⚠ AI tool caveat — clinical oversight is mandatory

All AI tools for CTA brain and carotids currently available are designed as decision-support aids, not autonomous diagnostic systems. A negative AI alert for LVO does not exclude LVO, and a positive alert requires confirmatory radiologist reporting before thrombectomy is initiated. Institutional governance should define which AI outputs may be acted upon before formal reporting, and under what clinical conditions.

🤖

Integrate AI stroke triage into your CT workflow

SATMED Health connects your radiology team with validated FDA and CE-marked AI platforms for LVO detection, aneurysm flagging, and collateral scoring — reducing door-to-decision time across your stroke network.

Explore SATMED AI Integration →

Conclusion — mastering the CTA brain & carotids protocol saves lives

CTA brain and carotids is the most time-sensitive imaging protocol in acute neurology. The diagnostic outcome depends on a chain of precisely executed decisions: correct IV access, a 4.5 mL/s injection rate, a 120 HU aortic arch bolus-tracking trigger, a caudal-to-cranial acquisition at 100 kVp and pitch 0.8, followed by thin-slice reconstructions with curved-MPR and MIP post-processing. When this chain is executed correctly, the result is an endoluminal arterial opacification exceeding 250 HU that allows confident identification of large vessel occlusion, carotid stenosis, dissection, aneurysm, and the full spectrum of cerebrovascular pathology detailed in this article.

The pitfall framework presented here — distinguishing between radiographer scanning errors, radiologist interpretation errors, and clinical action errors — provides teams with a structured approach to quality assurance in CTA brain and carotids reporting. Venous contamination is the most avoidable scanning failure. Calcium blooming overestimation is the most dangerous interpretation error. Acting on a technically inadequate CTA as if it were diagnostic is the most consequential clinical error. All three are preventable with the correct protocols, equipment, and training.

As AI tools for LVO detection, aneurysm flagging, and collateral scoring continue to mature, the role of the radiographer shifts toward ensuring that CTA images are technically perfect enough for AI algorithms to function reliably — garbage in, garbage out applies with equal force to AI-assisted workflows as to manual reporting. The protocol standards described in this article provide the technical foundation for both.

For further support in optimising your CTA brain and carotids equipment, consumables, and injection accessories, explore the SATMED Health product range, engineered specifically for high-flow neuro-CTA protocols in stroke centres and imaging departments worldwide. Register with SATMED Health today to access our full educational resource library and product consultation services.

References

[1] Saver, J. L. (2006). Time is brain — quantified. Stroke, 37(1), 263–266. https://doi.org/10.1161/01.STR.0000196957.55928.ab
[2] Powers, W. J., Rabinstein, A. A., Ackerson, T., Adeoye, O. M., Bambakidis, N. C., Becker, K., Biller, J., Brown, M., Demaerschalk, B. M., Hoh, B., Jauch, E. C., Kidwell, C. S., Leslie-Mazwi, T. M., Ovbiagele, B., Scott, P. A., Sheth, K. N., Southerland, A. M., Summers, D. V., & Tirschwell, D. L. (2019). Guidelines for the early management of patients with acute ischemic stroke: 2019 update to the 2018 guidelines for the early management of acute ischemic stroke. Stroke, 50(12), e344–e418. https://doi.org/10.1161/STR.0000000000000211
[3] Zarrinkoob, L., Ambarki, K., Wåhlin, A., Birgander, R., Eklund, A., & Malm, J. (2015). Blood flow distribution in cerebral arteries. Journal of Cerebral Blood Flow & Metabolism, 35(4), 648–654. https://doi.org/10.1038/jcbfm.2014.241
[4] Türkvatan, A., Büyükbayraktar, F. G., Ölçer, T., & Cumhur, T. (2017). Congenital anomalies of the aortic arch: evaluation with the use of multidetector computed tomography. Korean Journal of Radiology, 18(6), 1003–1010. https://doi.org/10.3348/kjr.2017.18.6.1003
[5] Bae, K. T. (2010). Intravenous contrast medium administration and scan timing at CT: considerations and approaches. Radiology, 256(1), 32–61. https://doi.org/10.1148/radiol.10090908
[6] Boas, F. E., & Fleischmann, D. (2012). CT artifacts: causes and reduction techniques. Imaging in Medicine, 4(2), 229–240. https://doi.org/10.2217/iim.12.13
[7] Tian, S. F., Liu, A. L., Liu, J. X., Liu, Y. H., & Wang, G. L. (2021). Optimized reconstruction algorithms to improve head and neck CTA image quality: comparing deep learning reconstruction and hybrid iterative reconstruction. European Journal of Radiology, 138, 109648. https://doi.org/10.1016/j.ejrad.2021.109648
[8] American College of Radiology. (2023). ACR manual on contrast media, version 10.3. American College of Radiology. https://www.acr.org/Clinical-Resources/Contrast-Manual
[9] Kidwell, C. S., & Wintermark, M. (2008). Imaging of intracranial haemorrhage. The Lancet Neurology, 7(3), 256–267. https://doi.org/10.1016/S1474-4422(08)70041-3
[10] Rudnick, M. R., Wahba, I. M., Leonberg-Yoo, A. K., Miskulin, D., & Litt, H. I. (2020). Risks and options with gadolinium-based and iodinated contrast agents in patients with CKD: a review. American Journal of Kidney Diseases, 75(2), 218–230. https://doi.org/10.1053/j.ajkd.2019.06.025
[11] International Commission on Radiological Protection. (2022). ICRP publication 135: Diagnostic reference levels in medical imaging. Annals of the ICRP, 50(2). https://doi.org/10.1177/0146645321990826
[12] Gupta, R., Cheung, A. C., Bartling, S. H., Lisauskas, J., Grasruck, M., Leidecker, C., Schmidt, B., Flohr, T., & Brady, T. J. (2008). Flat-panel volume CT: fundamental principles, technology, and applications. RadioGraphics, 28(7), 2009–2022. https://doi.org/10.1148/rg.287085004
[13] Greffier, J., Hamard, A., Pereira, F., Barrau, C., Pasquier, H., Beregi, J. P., & Frandon, J. (2020). Image quality and dose reduction opportunity of deep learning image reconstruction algorithm for CT: a phantom study. European Radiology, 30(7), 3951–3959. https://doi.org/10.1007/s00330-020-06724-w
[14] Barber, P. A., Demchuk, A. M., Zhang, J., & Buchan, A. M. (2000). Validity and reliability of a quantitative computed tomography score in predicting outcome of hyperacute stroke before thrombolytic therapy. The Lancet, 355(9216), 1670–1674. https://doi.org/10.1016/S0140-6736(00)02237-6
[15] van Gijn, J., Kerr, R. S., & Rinkel, G. J. (2007). Subarachnoid haemorrhage. The Lancet, 369(9558), 306–318. https://doi.org/10.1016/S0140-6736(07)60153-6
[16] Debette, S., & Leys, D. (2009). Cervical-artery dissections: predisposing factors, diagnosis, and outcome. The Lancet Neurology, 8(7), 668–678. https://doi.org/10.1016/S1474-4422(09)70084-5
[17] Wardlaw, J. M., Chappell, F. M., Best, J. J., Wartolowska, K., & Berry, E. (2006). Non-invasive imaging compared with intra-arterial angiography in the diagnosis of symptomatic carotid stenosis: a meta-analysis. The Lancet, 367(9521), 1503–1512. https://doi.org/10.1016/S0140-6736(06)68650-9
[18] Olin, J. W., Gornik, H. L., Bacharach, J. M., Biller, J., Fine, L. J., Gray, B. H., Gray, W. A., Gupta, R., Hamburg, N. M., Katzen, B. T., Lookstein, R. A., Lumsden, A. B., Newburger, J. W., Rundek, T., Sperati, C. J., & Stanley, J. C. (2012). Fibromuscular dysplasia: state of the science and critical unanswered questions. Circulation, 125(25), 3181–3190. https://doi.org/10.1161/CIR.0b013e31825e3836
[19] Suzuki, J., & Takaku, A. (1969). Cerebrovascular “moyamoya” disease: disease showing abnormal net-like vessels in base of brain. Archives of Neurology, 20(3), 288–299. https://doi.org/10.1001/archneur.1969.00480090076012
[20] Saposnik, G., Barinagarrementeria, F., Brown, R. D., Bushnell, C. D., Cucchiara, B., Cushman, M., deVeber, G., Ferro, J. M., & Tsai, F. Y. (2011). Diagnosis and management of cerebral venous thrombosis: a statement for healthcare professionals from the American Heart Association/American Stroke Association. Stroke, 42(4), 1158–1192. https://doi.org/10.1161/STR.0b013e31820a8364
[21] Nogueira, R. G., Jadhav, A. P., Haussen, D. C., Bonafe, A., Budzik, R. F., Bhuva, P., Yavagal, D. R., Ribo, M., Cognard, C., Hanel, R. A., Sila, C. A., Hassan, A. E., Millan, M., Levy, E. I., Mitchell, P., Chen, M., English, J. D., Shah, Q. A., Silver, F. L., … & Jovin, T. G. (2018). Thrombectomy 6 to 24 hours after stroke with a mismatch between deficit and infarct. New England Journal of Medicine, 378(1), 11–21. https://doi.org/10.1056/NEJMoa1706442
[22] Yahav-Dovrat, A., Saban, M., Merhav, G., Korat, A., Abergel, E., Eran, A., & Tanne, D. (2021). Evaluation of artificial intelligence-powered identification of large vessel occlusions in a comprehensive stroke centre. American Journal of Neuroradiology, 42(2), 247–254. https://doi.org/10.3174/ajnr.A6926
[23] Saba, L., Saam, T., Jäger, H. R., Yuan, C., Hatsukami, T. S., Saloner, D., Wasserman, B. A., Bonati, L. H., & Wintermark, M. (2019). Imaging biomarkers of vulnerable carotid plaques for stroke risk prediction and their potential clinical implications. The Lancet Neurology, 18(6), 559–572. https://doi.org/10.1016/S1474-4422(19)30035-3
[24] Park, A., Chute, C., Rajpurkar, P., Lou, J., Ball, R. L., Shpanskaya, K., Jabarkheel, R., Kim, L. H., McKenna, E., Tseng, J., Ni, J., Wishah, F., Wittber, F., Hong, D. S., Wilson, T. J., Halabi, S., Iliff, J., Basu, S., Lungren, M. P., & Yeom, K. W. (2019). Deep learning–assisted diagnosis of cerebral aneurysms using the HeadXNet model. JAMA Network Open, 2(6), e195600. https://doi.org/10.1001/jamanetworkopen.2019.5600
[25] Albers, G. W., Marks, M. P., Kemp, S., Christensen, S., Tsai, J. P., Ortega-Gutierrez, S., McTaggart, R. A., Torbey, M. T., Kim-Tenser, M., Leslie-Mazwi, T., Sarraj, A., Kasner, S. E., Ansari, S. A., Yeatts, S. D., Hamilton, S., Mlynash, M., Heit, J. J., Zaharchuk, G., Kim, S., … & Lansberg, M. G. (2018). Thrombectomy for stroke at 6 to 16 hours with selection by perfusion imaging. New England Journal of Medicine, 378(8), 708–718. https://doi.org/10.1056/NEJMoa1713973

🏥 Medical review & clinical governance

Medically reviewed by Prof. Dr. Damien O’Neil, MD, PhD
Last updated: 14 June 2026 | Reviewed for clinical accuracy and adherence to the latest guidelines of the American Heart Association / American Stroke Association (AHA/ASA), European Society of Radiology (ESR), European Stroke Organisation (ESO), American College of Radiology (ACR), Radiological Society of North America (RSNA), European Society of Neuroradiology (ESNR), and the International Commission on Radiological Protection (ICRP).

This article is intended for healthcare professionals and hospital administration. It does not constitute individual clinical advice. Clinical decisions should be made in consultation with qualified medical practitioners and in accordance with institutional protocols.