BASICS OF COMPUTED TOMOGRAPHY
The basic single-detector computed tomography (CT) scanner operates with a single row of detectors mounted opposite an x-ray source on a rotating gantry through which a patient moves during image acquisition. In 1992, the first dual-section helical CT scanner was introduced, which had two rows of detectors.1 In subsequent years, technology has continued to improve, with additional rows of detectors added to the rotating gantry. Currently, multidetector scanners with a single x-ray source are available with up to 128 rows of detectors. Scanners with two x-ray sources have 256 or more detectors.
A multidetector CT (MDCT) scanner houses parallel rows of detectors aligned along the long axis of the patient on a rotating gantry. These detectors can be of equal width, matrix detectors, or unequal width, adaptive array detectors. Multiple rows of detectors allow a larger length of the patient to be covered per gantry rotation, which means that CT scans can be performed faster and using higher x-ray tube current. This improves temporal resolution, decreases motion artifacts, and decreases noise. The more the detectors, the faster the scan can be performed. A conventional single detector CT scanner makes a complete 360-degree rotation in approximately 1 second. Multidetector scanners have faster gantry rotation speeds and obtain multiple sections per revolution. For example, a scanner with four rows of detectors is actually eight times faster than its single detector counterpart.
In addition, thinner sections can be acquired with multiple detectors, improving spatial resolution. Thinner sections also allow for the creation of multiplanar reformatted (MPR) images, maximum intensity projection (MIP) images, and three-dimensional reformatted images. MPR images are created by fusing data collected from the multiple rows of detectors. This data can be reformatted into coronal, sagittal, and oblique planes. Interpolation algorithms can also be used to average overlapping data points and further decrease noise in the resultant images.2
Intravenous (IV) contrast administration is at the cornerstone of imaging the venous and lymphatic systems. Faster scan times allow for improved vascular concentration of contrast and better separation of arterial and venous phases of imaging. The timing of the CT scan depends on the vascular system or organ of interest. There are three techniques for determining the appropriate delay to maximize contrast opacification. The first is bolus tracking, which involves low-dose (low tube current) monitoring scans through the level of interest. The actual scan begins when a predetermined Hounsfield unit threshold is reached. The second technique involves use of a test bolus. A small quantity of contrast, the test bolus, is administered, and low dose monitoring scans are again obtained through a predetermined region of interest. The time it takes for the test bolus to reach the region of interest is used to determine the scan delay after the actual, full dose of contrast. The third and final technique is to simply use an approximation based on experience with prior scans in similar patients.
In 2006, approximately 62 million CT examinations were performed in the United States.3 As the applications for CT increase, it is important to remain cognizant that this modality uses ionizing radiation. The radiation dose from a study depends on the number of photons and the individual energies of those photons. The number of photons is proportional to the x-ray tube current and the time that the x-ray tube is “on.” The approximate effective dose from a CT of the chest is 5 to 7 millisieverts (mSv) and from an abdomen or pelvis CT is 8 to 14 mSv.3,4 For comparison, in the United States, the naturally occurring annual background radiation is 3 mSv.4 Although the dose for a single CT scan is small, it is not insignificant. The risks of cumulative radiation exposure to a patient from multiple CT studies should be weighed against the benefits of the information provided by the study.
Pulmonary Angiography. Although the pulmonary arteries are not technically part of the venous system, pulmonary embolism (PE) is included in the spectrum of venous thromboembolic disease because it is a low-velocity system. Deep venous thrombosis (DVT) and PE result in 300,000 to 600,000 hospitalizations per year in the United States, and as many as 50,000 patients die every year from PE.5 In the past, conventional pulmonary angiography was considered the gold standard, and ventilation/perfusion scanning was the primary means of noninvasive diagnosis of suspected PE. The advent of MDCT and the development of timed contrast injection protocols for imaging the pulmonary arteries have revolutionized noninvasive diagnosis of PE. Because MDCTs are now widely available and the test can be performed rapidly, it has emerged as the diagnostic test of choice in many clinical settings for PE. The reported specificity and sensitivity of MDCT for diagnosing PE is variable but has been shown to be at least equivalent to ventilation/perfusion scanning.6 Another source found that MDCT may actually detect more subsegmental emboli than ventilation/perfusion scanning.7 However, the clinical benefit of this possible increased detection rate has yet to be determined.
The protocols for MDCT pulmonary angiography involve varying parameters, depending on the number of detectors and the institution. In general, the scans are performed during breath hold or shallow breathing if the patient cannot tolerate breath hold. Approximately 100 to 150 mL of nonionic iodinated contrast material is injected with a power injector usually in an antecubital vein at a rate of 3 to 5 mL/s. Timing of the scan can be determined via bolus tracking or a test bolus. The scan is usually started 10 to 20 seconds after contrast injection to optimize opacification of the pulmonary arteries. Scanning is performed from the aortic arch to below the inferior pulmonary veins. Images are acquired as thin sections of between 1 and 3 mm with a pitch of between 1.0 and 1.75.
Interpretation of PE protocol CT in most instances is straightforward. A study is considered positive for acute PE if there is an intraluminal filling defect outlined by contrast material within a pulmonary artery (Figure 12-1) or if an artery is completely occluded by low-density material. Absence of a filling defect is considered negative for PE. Factors that can make interpretation difficult include motion, inadequate contrast opacification, and image noise. Motion artifact can be caused by an inability of the patient to remain still or from involuntary patient breathing or cardiac motion. Inadequate artery opacification occurs as a result of parameters affecting the timing of the bolus. For instance, use of smaller gauge peripheral IV lines or altered hemodynamics can adversely affect the timing of the contrast bolus and result in insufficient contrast material within the pulmonary arteries at the time of the scan. Patient body habitus or inadequate current or kilovoltage used for the scan may lead to a decrease of the image signal-to-noise ratio and limit the diagnostic utility of the CT scan.
Axial (A) and coronal (B) computed tomography (CT) images demonstrating a low-density, intraluminal filling defect with the main, right, and left pulmonary arteries consistent with a large pulmonary embolus. C. Axial CT image demonstrating small filling defects within segmental pulmonary arterial branches bilaterally (blue arrows).
Pulmonary Venography. Atrial fibrillation is the most common cardiac arrhythmia and is a leading cause of stroke. More than 90% of cases of paroxysmal atrial fibrillation are caused by spontaneous activity originating in the pulmonary veins.8,9 The goal of endovascular or surgical therapy is to disconnect the electrical connection between the left atrium and the pulmonary veins and CT of the pulmonary veins is a useful adjuvant to these treatments, particularly endovascular ablation.
The protocol for performing CT of the pulmonary veins is similar to that for CT coronary angiography. Images are obtained with thin collimation through the chest in a caudal to cranial direction during breath hold. Electrocardiogram gating improves image quality. One hundred to 150 mL of IV nonionic is administered via an antecubital vein with a power injector at a rate of approximately 4 mL/s, and scanning is started after a delay of 30 seconds.
On a CT of the pulmonary veins, it is important to identify the number and location of the pulmonary veins (Figures 12-2A and B). Normal pulmonary vein anatomy is two superior pulmonary veins (one right and one left) and two inferior pulmonary veins (one right and one left). The superior right pulmonary vein drains the right upper lobe and right middle lobe. The superior left pulmonary vein drains the left upper lobe and lingula. The inferior pulmonary veins drain the respective lower lobes. However, the pulmonary vein anatomy is often variable. Up to a quarter of patients have a common left or right pulmonary vein.10 Another common anomaly is a separate vein draining the right middle lobe (Figure 12-2C). There could also be an anomalous pulmonary venous–systemic connection with an aberrant pulmonary vein draining into the superior vena cava (SVC), coronary sinus, portal vein, or other systemic vein.
A. Three-dimensional reformatted image of normal pulmonary venous anatomy. Two right (blue arrows) and two left pulmonary veins (black arrows) are draining into the left atrium. B. Oblique, coronal maximum intensity projection (MIP) image of normal pulmonary venous anatomy, two left pulmonary veins (black arrows), and two right pulmonary arteries (blue arrows). C. Three-dimensional reformatted computed tomography image of the pulmonary veins and left atrium. There are three right pulmonary veins entering the left atrium, a superior (S) and inferior vein (I), and an accessory vein draining the right middle lobe (M). Coronal MIP images of the left atrium and pulmonary veins before (D) and after (E) pulmonary vein ablation for atrial fibrillation. The superior left pulmonary vein is widely patent on the pre-ablation image (arrow) (D) but is significantly stenotic on the postablation image (arrow) (E).
Pulmonary vein CT is performed before and after ablation. There are two reasons for performing preprocedure pulmonary vein CT. The first is to localize the pulmonary vein ostia, the target of endovascular ablation therapy. This is difficult to accomplish and time consuming by conventional angiography. The second is to evaluate for thrombus in the left atrium or left atrial appendage, a contraindication for endovascular ablation therapy. Thrombus appears as a low attenuation filling defect. Pulmonary vein CT is also performed following ablation to evaluate for pulmonary vein stenosis, the most common complication. Stenosis occurs to a variable degree and has been reported in up to 40% to 100% of cases after endovascular ablation (Figures 12-2D and 12-2E).8
The SVC is formed in the superior mediastinum by the convergence of the right and left brachiocephalic veins. The azygos vein empties into the posterior aspect of the SVC, which then drains into the right atrium. Occlusion of the SVC may result in a variety of clinical presentations depending on the acuity. SVC occlusion usually occurs gradually, and there is time for collateral circulation to develop. SVC occlusion may be caused by external compression, which is termed SVC syndrome or internal thrombus. SVC syndrome may be caused by compression; by lymphadenopathy; fibrosing mediastinitis; hematoma; or direct compression by a tumor (Figure 12-3A), usually lung cancer. Thrombus within the SVC is not as uncommon as it once was because of the increased prevalence of central venous catheters.11,12 Determining the underlying cause of SVC occlusion is important in terms of treatment and identifying possible malignancy. The diagnosis of SVC syndrome or occlusion can often be made clinically or with ultrasonography; however, MDCT is important for evaluating the underlying cause and determining the degree of obstruction and the extent of disease.13 MDCT can also play a role in diagnosis of SVC occlusion or stenosis when the diagnosis is uncertain or symptoms are not clinically apparent.
A. Axial computed tomography (CT) image through the superior mediastinum demonstrating malignant superior vena cava (SVC) syndrome. Because of external compression from a right upper lobe mass (M), the SVC (blue arrow) is narrowed to just slightly larger than the caliber of the central venous catheter (black arrow). B. Axial CT image through the mediastium demonstrating SVC stenosis (red arrow) with prominent contrast opacification of an enlarged azygos vein (blue arrow) and several anterior chest wall collateral veins (black arrows).
There are multiple findings on CT scans through the chest, with and without injection of IV contrast material, which can aid in the diagnosis of SVC syndrome or occlusion. A dedicated protocol for evaluation of the SVC is not necessarily indicated; however, a scan delay of 25 to 30 seconds after contrast injection is usually adequate for opacification of the SVC and collaterals.14 After contrast administration, low-density filling defect within the SVC indicates thrombus. Contrast enhancement of the SVC allows for measurement of the diameter and evaluating the degree and extent of stenosis (Figure 12-3B). A third benefit of contrast enhancement is identification of collateral vessels, an indication of SVC obstruction (see Figure 12-3B). MDCT with or without contrast enhancement is a useful tool for evaluating the mediastinum and lungs for lymphadenopathy, hematoma, mass, fibrosis, or other possible source of SVC compression.
Limitations of MDCT evaluation of the SVC are primarily because of contrast-related artifacts or motion. Mixing of contrast opacified blood and non-opacified blood within the SVC may result in flow artifacts that can simulate thrombus.14 Respiratory motion or cardiac motion can also create artifacts that limit evaluation of the SVC on CT imaging.
Thoracic outlet syndrome refers to dynamically induced compression of the neural and vascular structures within the confined space of the thoracic outlet. This space extends from the cervical spine and mediastinum to the pectoralis minor muscle and includes the subclavian artery and vein and nerves of the brachial plexus.15 The cause of thoracic outlet syndrome is congenital or an acquired abnormality of bordering bone or soft tissue (e.g., a cervical rib or postoperative scarring). The diagnosis can be difficult to make on a solely clinical basis; positive clinical findings with provocative maneuvers can be seen in normal patients.16,17 Therefore, the use of an imaging examination is a useful adjuvant to confirm the suspected diagnosis of thoracic outlet syndrome. Imaging examinations used in this situation include ultrasonography, CT, and magnetic resonance imaging. Here we will discuss the role for CT in evaluating compression of the neurovascular structures within the thoracic outlet.
When evaluating for compression of thoracic outlet structures with CT, the patient is imaged in two positions—the first with the arms at the patient’s side and the second with the arms elevated to provoke compression. Contrast is injected via a vein in the asymptomatic arm to obtain opacification of only the arteries in the symptomatic side, at a rate of approximately 3 to 4 mL/s.15,18 Images are obtained with narrow collimation to allow for creation of alternate plane, three-dimensional, and MIP reformatted images.
Evaluation of subclavian artery compression is most easily done with sagittal reformatted images, with the artery in cross-section. Evaluation of the subclavian vein is more difficult because the vein is more compressible and can appear very narrow even in normal patients with arm elevation.15,19 If there is contrast enhancement of the vein, a low-density filling defect can be observed in cases of venous thrombosis. Opacification of the collateral vessels can be easily observed on CT as a secondary sign of late-stage venous occlusion.15 CT imaging is also a good modality for evaluating the bony structures of the thoracic outlet. Postural changes of the clavicle, differences in the distance between the first rib and clavicle, or the presence of a cervical rib may be indications of thoracic outlet compression.18
Hepatic Veins. The liver has a dual blood supply via the hepatic arteries and portal venous system, but the hepatic veins are the only vascular drainage from the liver. The leading cause of obstruction of the major hepatic veins is thrombosis. Hepatic outflow obstruction causing abdominal pain, congestive hepatomegaly, and ascites is a clinically defined entity known as Budd-Chiari syndrome. Acute onset of hepatic venous thrombosis may induce severe portal hypertension and functional liver damage.20 In addition to thrombosis, other pathologies can adversely affect the hepatic veins, including venous congestion from heart failure and direct neoplastic invasion. Multiphase MDCT can provide important and detailed anatomic information about the hepatic venous structure and the cause of hepatic venous obstruction.
MDCT of the liver is generally performed in at least two phases, an arterial and a portal venous phase. IV contrast is injected via a power injector at a rate of 2 to 3 mL/s, and imaging is begun after 20 to 30 seconds for the arterial phase and again 60 to 90 seconds after injection for the portal venous phase. Occasionally, unenhanced images may be obtained initially or additional delayed images may provide further information.
Findings of hepatic venous occlusion on MDCT include filling defect within the involved hepatic vein(s) or inferior vena cava (IVC), a heterogeneous hepatic parenchymal perfusion pattern, as well as intrahepatic venovenous collaterals. Whereas intrahepatic veins that form collaterals with the systemic venous pathway enhance on the surface of the liver, the ascending lumbar, azygous, and hemiazygous veins are collaterals frequently seen with obstruction of the intrahepatic IVC.20,21 The left renal–hemiazygos pathway, inferior phrenic–pericardiophrenic collaterals, and superficial collaterals of the abdominal wall are other routes of venous blood diversion with intrahepatic IVC thrombosis. Noninvasive CT hepatic venography can better elucidate these collaterals as well provide anatomical information revealing the extent of hepatic congestion and the damage of the hepatic parenchyma.20
Hepatocellular carcinoma or hepatic metastases may directly invade, compress, or indirectly lead to thrombosis of the hepatic veins or IVC. This may lead to arterioportal, arteriosystemic, and portosystemic venovenous collateral formation. Findings on CT venography include filling defects in the portal or hepatic veins, enhancement of the malignant thrombus on arterial phase images, and expansion of the vein lumen.
MDCT is also useful for preoperative evaluation of living-donor liver transplantations documenting anatomic variations.22,23,24 Branching patterns and venous variations are crucial for determining a suitable line of resection and in devising strategies for venous reconstruction.22 Postoperatively, CT venography can display changes of hepatic venous congestion secondary to hepatic venous stenosis, reported in 5% of living donor liver transplantations.21 Nonvisualization of a preoperatively visualized vein, mosaic heterogeneous enhancement, and hepatomegaly on CT venography are indicative of postoperative Budd-Chiari syndrome.
Portal Venography. Slow-flow or obstruction of the portal venous system may occur because of a variety of causes, including infections, neoplasms, emboli, surgery, portal hypertension, venous stasis, and hypercoagulable states.25 Identification of the underlying cause of portal hypertension or thrombosis is important in terms of treatment. For example, malignant thrombi are an absolute contraindication for liver transplantation, resective surgery, and percutaneous ablation techniques, and they are also a relative contraindication for transarterial chemoembolization.25 Portal hypertension secondary to cirrhosis or thrombosis can lead to reversal of venous blood into the systemic venous circulation, leading to varices.26 Esophageal and gastric varices are the two most frequently visualized varices and have been widely studied using MDCT portal venography (Figures 12-4A and B).27
A. Perigasric varices (arrows) secondary to portal hypertension. B. Computed tomography (CT) scan of the abdomen demonstrating esophageal varices in a patient with portal hypertension (arrow). C. Three-dimensional reformatted image constructed from axial slices of a CT of the abdomen during the portal venous phase demonstrating a recannalized paraumbilical vein (arrow). D. Axial CT performed during the portal venous phase depicting a low-attenuation filling defect (arrow) within the portal vein, which represents a non-occlusive thrombus. E. Reformatted maximum intensity projection image demonstrating thrombus extending in the splenic and superior mesenteric veins (blue arrow).