google ads

Everything you need to know about Computed Tomography (CT) & CT Scanning

Chest: Spiral CT of the Chest Part II: Pulmonary embolism, aneurysm and dissection

Leo P. Lawler, MD, FRCR.
Assistant Professor


In this lecture we are mainly concerned with the larger vascular structures of the thorax. We shall address the pulmonary arteries delivering deoxygenated blood to the secondary pulmonary lobules and the thoracic aorta delivering oxygenated blood to the systemic circulation.

Optimal imaging of the pulmonary artery and aortic circulations requires a clear understanding of thoracic CT angiography and clinical conditions it attempts to address.

Thoracic CT angiography

CT angiography of the thoracic vasculature is performed after the administration of iodinated contrast material. Our goal is to match the image acquisition and duration to the plateau of intravascular contrast peak density. We aim for density values over 200HU and this is achieved by using an empiric 15s scan delay for the pulmonary arteries and a 30s scan delay for the aorta. Non-ionic contrast is administered through an antecubital vein at a rate of 3-4cc per second by power injector. We decrease the rate for more peripheral injections and patients with perceived diminished cardiac output. Hand injections are used for central line injections with an added 10s delay. Bolus tracking and test bolus techniques will improve consistency of HU measurements but are time consuming for patient through-put and may not significantly affect image interpretation. Saline chasers can increase the contrast plateau by adding the pooled contrast in peripheral veins to the central pool but in the absence of dual power injectors this protocol requires hand administration. Z-axis 26second breath hold coverage can be obtained with single detector helical scanning and 3mm collimation or high resolution 1mm MDCT detectors. For faster or greater volume coverage (~10s) 2.5mm detectors are utilized producing less noisy images. As scan duration shortens with multidetector row scanning the margin for error deceases but the ability to capture the peak of contrast enhancement improves. The scan direction for the pulmonary arteries if from caudad to cephalad and starts just above the diaphragm and extends to above the aortic arch. For poor breath holders this ensures the basal vessels are best seen which is the commonest site for clot. Aorta evaluation must include the great vessel origins (carotids and subclavian) and must extend to at least the celiac axis. Many aortic pathologies extend into the abdominal aorta and MDCT will allow coverage to the femoral vessels without loss of bolus capture.

In angiography more than most other areas of CT imaging 3D imaging has become of paramount importance. CTA is expected not only to simulate angiography but to improve on it with greater sensitivity for disease, greater measurement accuracy an more extra-luminal detail.

Pulmonary embolism

This common condition accounts for 15% of hospital mortality. Patients with poor mobility or procoagulant states are at risk. Clinical assessment is imprecise and 1% are clinically unsuspected. Initiation of therapy does not remove pulmonary emboli (PE) but prevents further clot formation and though it does decrease the risk of mortality but it carries significant risks of morbidity(7%) so it can only be used when there is clear indication. Ventilation/perfusion scintigraphy studies when normal, low likelihood or high likelihood have high predictive value for PE evaluation. However a large number of patients who have this test receive an intermediate result which basically does not change the pre-test suspicion of clot. Though pulmonary angiogram can improve sensitivity and specificity it is an invasive test with associated morbidity and it is no longer considered the gold standard.

CT has proven highly sensitive and specific for the detection of pulmonary embolism (over 90% for both). It is easy to perform and is non-invasive. It relies on high contrast in the main and branch vessel pulmonary arteries. Visualization of the 4th order branches and higher requires good bolus and small collimation. Image windows must be appropriately set for film printing and cine scrolling on soft copy monitors is proven more sensitive for detection.

The CT features of PE are


Pulmonary artery filling defect

Enlargement of the pulmonary artery


Wedge shaped lung parenchyma infarct

Mosaic perfusion abnormality

Pleural effusion

Chronic clot

Serpiginous filling defect


Laminated filling of branch vessels

Vessel wall distortion or reduction in caliber

CT is not yet a perfect test for PE. 4% of cases are indeterminate (similar to pulmonary angiography) and false positive and negatives exist. Small or incidental pulmonary embolism is treated at present but its true clinical implication is unclear. There is, however, greater interobserver agreement for clot detection compared with V/Q scanning.

Most pulmonary emboli are thought to arise in deep venous thrombosis. If you think of these two conditions as a continuum it makes sense that 66% of those with PE will have DVT. 17% of those DVTs will be in the pelvis or areas above which are not routinely imaged by Doppler sonogram. Delayed imaging (3-4mins) from the proximal tibia to the IVC has proven efficacious in detecting DVT. However adding this to a routine PE protocol significantly increases body dose and in particular to radiation sensitive areas such as the genitals. It is also important to review with the clinicians at your institution whether finding DVT in the presence of already diagnosed PE will change management.

Chronic PE is a distinct condition thought to occur due to the slow deposition of emboli and thrombus in the pulmonary arteries with subsequent endothelialization. This ultimately leads to vessel distortion and raised right heart pressures. It is one of the few treatable causes of right heart failure but effective therapy involves removing a cast of clot from the pulmonary artery branches. Surgical planning in this case is well served by 3D mapping of the pulmonary arteries and the clot present.


Thoracic aneuryms are seen in about 500 per 100,000 and 20% are asymptomatic. 3/100,000 are associated with dissection and 1/100,000 with rupture. They account for 17,000 deaths in the US annually and there is a 21% 5 year survival quoted for those not surgically treated. Those with associated dissection tend to rupture at a smaller size. The ascending aorta may be up to 2.6-2.9cm in size and 4cm is considered aneurysmal. The descending thoracic aorta is normally about 1.6cm and 3cm is considered aneurysmal. It is important that consistent sites of measurement are given in a report to reduce inter-observer error in aneurysmal assessment over time. A change of less than 1cm may dictate surgical intervention and there is no substitute for direct comparison with prior studies. The site and shape of an aneurysm may give some insight into its etiology. Diffuse ascending aneurysms may be seen in cystic medial necrosis (e.g. Marfan’s) whereas tubular aortic aneurysms are often the sequela of aortitis. Sinus aneurysms raise the suspicion of congenital anomaly or infection. Saccular aneurysms at the aortic root, the arch or near the ductus remnant raise the question of pseudoaneurysms at points of fixation. Focal supravalvular aneurysms are associated with aortic stenosis and distal arch aneurysms are seen proximal to a coarctation with a distal diminutive descending thoracic aorta. Atherosclerotic aneurysms tend to affect the proximal descending thoracic aorta. The interpretation should reflect size and configuration as well as any branch vessel involvement and mural thrombus. Most aneurysmal aortas are also quite tortuous and measurements taken from axial imaging are rarely a true of the actual aneurysm size. 3D interpretation is imperative for true orthogonal sizes taken perpendicular to the lumen.

Thoracic aortic dissection

Dissections result from a tear in the intima and a separation of the media creating a flap and two lumens. 70% of the flaps are seen on CT. It is the most common catastrophe of the aorta and is 2-3 times more common than aneurysmal rupture. The most commonly used classification scheme is the Stanford scheme,which calls a dissection involving the ascending aorta type A regardless of the descending extent. The remainder are type B. 60% are type A, 40% are type B. Debakey is the other major classification (I=ascending and beyond, II=ascending, III=beyond subclavian). There is also an etiological classification, which also attempts to place dissection in a group with other disease states;

Classic dissection, 2 lumen, +/-communication of lumens

Intramural hematoma

Subtle or discrete wall bulging

Ulceration of plaque

Iatrogenic or traumatic

The transverse tear is variable in size but usually is about half the aortic circumference. This condition is not usually related to atherosclerosis and in the majority the tear is in the ascending aorta about 2cm above the sinotubular junction. About 10% occur in the arch and 25% in the descending thoracic aorta. Most ascending dissections spread distally and not vice versa. Proximal extension into the coronary arteries and pericardium is always a concern but is in fact unusual. The extent of the tear is thought to be dictated by media scarring and vessel branch points. On CT we can rarely accurately identify either the entry or re-entry tears. The false channel is usually the outer one and it extends posterolaterally. On CT it is important to assess the flap extension into or coverage of any branch vessels and to assess visceral perfusion. Renal involvement is morel likely than mesenteric. Delayed flow in the false lumen may require additional delayed imaging-usually immediately after the first spiral. Close attention must be paid to the false lumen as this enlarges in 20% of patients. Signs that help differentiate the false lumen;

The true lumen is not a blind sac but follows to continuity with the non-diseased aorta

The false lumen pressure tends to compress the true lumen

If the flap wraps on itself in the arch the center lumen is usually the false one

There is beak sign of hematoma within the false lumen

There are signs of cobwebs of sheared intima in the false lumen

Luminal thrombus a good sign for the false lumen but the true lumen may thrombose.

© 1999-2019 Elliot K. Fishman, MD, FACR. All rights reserved.