MDCT is the gold standard imaging investigation for AAS with a sensitivity of 100 % and specificity of 98–99 % (Hiratzka et al. 2010). Although MRI and transoesophageal echo have similar sensitivity and specificity, MDCT is generally preferred in an acute setting because it is near-universally available and provides speedy acquisition and high accuracy (Salvolini et al. 2008). The ability to reconstruct in three planes and to assess aortic branches including aberrant anatomy makes CT invaluable for surgical and endovascular planning (Maddu et al. 2013; Agarwal et al. 2009). Both pre- and post-contrast CT should be obtained to differentiate between the three components of AAS (Maddu et al. 2013). Table 2 summarises the CT findings in all three entities of AAS (see later text for further explanations).

At least 64 detector rows are required to achieve multiplanar isotropic spatial resolution. ECG gating is required to reduce cardiac motion artefacts – this is particularly important in accurately diagnosing dissection at the aortic root and avoiding false positives due to cardiac motion artefact mimicking an acute dissection flap. ECG gating also improves coronary artery visualisation. The field of view for pre-contrast ECG-gated CT should extend from above the aortic arch to the inferior diaphragm. This should be acquired prospectively in mid diastole. Premedication with beta-blockers or other medications should be avoided (31 Chin in Abbas). On the non-contrast CT, high-density IMH can be identified, and local pleural or pericardial rupture can also be demonstrated (Abbas et al. 2014; Castaner et al. 2003). A bolus tracking CT angiogram achieving opacification of the aorta of >250 HU follows, covering from the thoracic inlet to the aortoiliac bifurcation or common femoral arteries, with precise timing of image acquisition and imaging parameters varying slightly with machines (Castaner et al. 2003). The usual volume of iodinated 370 mg/ml contrast medium of 120 ml is decreased in elderly patients to 80–100 ml in view of decreased cardiac output. The contrast should be injected via the right arm to avoid streak artefact from the left brachiocephalic vein near head and neck vessels. A saline flush of 20 ml further minimises streak artefact (Weininger et al. 2011). The patient is able to free breathe during abdominal aortic imaging, but chest imaging is acquired in breath-hold.

Multiplanar reformat assessment is advisable since it helps in the evaluation of the aortic root. Moreover, clinicians appreciate 2D and 3D reconstructions as they are useful for planning surgical or endovascular procedures (Agarwal et al. 2009). Maximum intensity projection reconstructions are valuable for visualisation of branch vessel involvement (Abbas et al. 2014).

Alternative techniques to MDCT in patients in whom iodinated contrast is contraindicated include transoesophageal echocardiography and MRI, but both of these techniques are time-consuming and not widely available in all hospital settings. Contrast-enhanced MR angiography has excellent spatial and contrast resolution and allows multiple vascular imaging phases and post-processing. Its limitations include the inability to demonstrate arterial wall or intimal calcification, lack of suitability for unstable patients or those with implanted electronic devices and the problem of artefacts due to stent placement (McMahon and Squirrell 2010). Initial unenhanced CT followed by MRI (where feasible) or transoesophageal echocardiography may be a reasonable compromise in patients who cannot receive iodinated contrast or in whom repeated examinations are required, as CT will demonstrate IMH, intimal calcification displacement, rupture into pleural or pericardial space and aortic size (Abbas et al. 2014).

Classification of AAS is based on the location and extension of the dissection, with the Stanford system preferred over the DeBakey system in the emergency setting because it dictates immediate clinical management (broadly speaking, surgical = type A, medical = type B) (Maddu et al. 2013). Stanford type A affects the ascending aorta or aortic arch and accounts for 75 % of aortic dissection (Castaner et al. 2003). The dissection flap may extend to the descending aorta (McMahon and Squirrell 2010). Stanford type B begins distal to the left subclavian artery. For all three entities of AAS (dissection, IMH, PAU), the diagnosis of type A involvement requires immediate referral to a cardiothoracic surgical centre.

There are 2000 new cases of aortic dissection each year in the USA and 3000 in Europe (McMahon and Squirrell 2010). Dissection is the most common aortic emergency, being more prevalent than thoracoabdominal aortic aneurysm rupture (Castaner et al. 2003). Acute thoracic dissection is life-threatening and requires immediate diagnosis and treatment (Castaner et al. 2003): 75 % of deaths from aortic dissection occur within 2 weeks of clinical presentation.

The normal aortic wall has three layers (from external to internal: adventitia, media and intima) (Fig. 1). The pathophysiology of aortic dissection includes disruption of the aortic intima and inner layer of the media such that blood can track along the media (Macura et al. 2003) (Fig. 2). This propagation can be both retrograde and anterograde (Maddu et al. 2013) and results in formation of a true and a false lumen with the false lumen having pressures greater than or equal to the true lumen (Williams et al. 1997). Since the intervening ‘intimal flap’ between the two lumens consists of intima and media, it should more precisely be termed an ‘intimomedial’ flap. It is theorised that reduced elastic recoil of the dissection flap enables the false lumen to dilate. This resulting collapse of the true lumen is highly variable based on the number of fenestrations and tears between the true and false lumen, the chronicity of the flap and the degree of control of the patient’s blood pressure. In our clinical experience, the patient’s response to the compressed true lumen and their overall clinical status dictates further management (McMahon and Squirrell 2010; Williams et al. 1997). The overall degree of dilation of the false lumen depends on blood pressure, residual wall thickness and percentage of wall circumference involved in the dissection. The false lumen may subsequently compress or obstruct the true lumen. The dissection may remain patent as a false lumen, may thrombose or may recommunicate with the true lumen through fenestrations (Macura et al. 2003). Alternatively, it may rupture into a pericardial or pleural space (McMahon and Squirrell 2010). The greater the proportion of the media involved in the flap, the thinner the external wall of the false lumen, and therefore, the higher the risk of rupture (Roberts 1981). The initial intimal tear occurs at sites of greatest hydraulic pressure – usually the right lateral wall of the ascending aorta or the proximal segment of the descending aorta (Maddu et al. 2013).