Time of Flight Magnetic Resonance Angiography: A Trap for the Unwary

Introduction 

Magnetic resonance scanning is being increasing used in vascular imaging. However, liked any other modality, it does have some pitfalls. This case illustrates one such pitfall when time-of-flight magnetic resonance imaging is used to assess high flow vessels such as the carotid arteries.

Report 

A six-year-old girl was attacked by a Staffordshire terrier which locked its jaw on the child's neck and vigorously flung her around. When the dog released the child she was found to be bleeding profusely from multiple deep puncture wounds on the left neck, lower face and around the left ear. There was marked surgical emphysema of the neck. She underwent wound exploration under anaesthesia. The trachea was found to be partially transected and was repaired. The left eleventh cranial nerve was thought to be damaged but was grossly intact. No major vascular injury was found. She was ventilated and subsequently required a tracheostomy. She was then noted to have a left Horner's syndrome. A time-of-flight magnetic resonance angiogram (MRA) was arranged to image any carotid injury (Philips Achieva 1.5T scanner with a Sensor head coil: Tr = 15, Te = 3.3 with a flip angle of 20°). This demonstrated a focal area of expansion of the distal common carotid artery in combination with a defined area of vessel narrowing. This was reported as suspicious for a local dissection although no associated area of high signal wall haematoma or intimal flap could be clearly identified. She was fully anticoagulated with low molecular weight heparin on the basis of a presumed diagnosis of dissection.

Fig. 1 shows the area of abnormality but also demonstrates a catch for the unwary. In addition to the localised abnormality there is a laminar flow artefact visible throughout the left internal carotid artery. This might be interpreted as a dissection by physicians without sufficient MRA experience.

View Large Image Download to PowerPoint

Figure 1 Time of flight MRA showing left common carotid injury (A) and laminar flow void in left internal carotid artery (B). The flow void is also visible in the right carotid arteries (C).

Seven days later a contrast-enhanced (Magnevist 5ml) MRA was performed that confirmed the localised abnormality in the distal common carotid artery (Fig. 2). The left internal carotid is, however, shown to be pristine. Duplex scanning at 6 months again demonstrated a dissection flap localised within the common carotid artery causing an 85% stenosis, although the patient remained clinically well with no neurological sequelae.

View Large Image Download to PowerPoint

Figure 2 Gadolinium-enhanced MRA showing left common carotid injury (arrow) but no laminar flow void.

Discussion 

MRA has been reported as having 100% sensitivity and specificity for carotid dissection.1 The key features of dissection are vessel expansion, stenosis and intramural haematoma, with some studies describing the presence of intramural haematoma in all cases.2 However, a choice of imaging modalities for identifying dissection (CT angiography, duplex, conventional angiography and MRA) is often available and the clinical team and radiologist have to weigh up the potential complications of each test with their respective accuracies. In this case the patient was young so MRA has advantages over CTA as it is non-invasive and there is no ionising radiation. However, a CT angiogram, with its improved spatial resolution can be diagnostic.3 This case demonstrates that whatever imaging is used, correct interpretation of the scan is essential. MRA interpretation in particular can be difficult. The flow voids on time-of-flight (TOF) imaging can mimic dissection. This occurs because TOF depends on saturation differences between molecules flowing into an imaging slice and static molecules within the slice. These molecules have a different level of radiofrequency excitation which leads to the angiographic differences. This occurs because the flowing molecules in a particular slice must experience both the excitatory and refocusing radiofrequency pulses to produce a signal. At high velocities, such as in the carotid arteries, most of the flowing molecules have left the slice before the refocusing pulse is applied and thus a flow void is seen.4 As velocities in the centre of a vessel are greater than those at the periphery, this difference is particularly noticeable in the middle of the vessel. In effect, the rapid flow in the centre of the vessel escapes the image slice between excitation and signal measurement. Clinicians reviewing scans of their patients must have knowledge of such interpretation traps and liaison with an experienced radiologist is essential.

Conflict of Interest/Funding 

None.