Contrast opacification on thoracic CT angiography: challenges and solutions

Inappropriate placement of ROI for bolus tracking scan is a common cause of non-diagnostic CT scan. Human error is a common source of inappropriate placement. Selection of the wrong target vessel, especially in the setting of complex vascular anatomy, and/or selection of an ROI which is too big or too small are common operator dependent errors. Patient movement between localizer slice selection/ROI placement and the start of contrast administration/imaging can also affect ROI placement (i.e. large respiratory effort, cardiac motion, and/or the patient physically shifting on the table), leading to premature, delayed or even no bolus triggering. In addition, intraluminal abnormalities, including dissection and embolus, may not be readily apparent on the precontrast localizer images and placement of the ROI overlying one of these structures may result in delayed or absent bolus triggering. For example, if the ROI is placed in the false lumen of a type B aortic dissection (Fig. 2), the contrast enhancement may or may not rise as quickly as expected (2). To avoid this, the indication of the scan should be well known to the operating technologist. A test bolus is preferable to bolus tracking in patients with post-surgical repair of complex congenital heart diseases.

Imaging pearl: In patients with known aortic dissection, test bolus can be more useful in identifying time to peak enhancement in true and false lumens. Optimal time for acquisition would be when both lumens are opacified. If false lumen dose not opacify at all on the bolus timing scan, a limited Z axis 60 s delayed image can be obtained to confirm slow flow/ thrombus or for follow-up, contrast-enhanced MRA may be obtained.

Extravasation of contrast material, in which contrast medium is injected outside the intended vessel, is an infrequent, but well known complication of CTA (Fig. 4a). Contrast extravasation rates during CT imaging range between 0.1 and 0.9 %, with an average rate of 0.4 % [13, 14]. Patients at risk for contrast extravasation include infants and small children, elderly, uncooperative, and unconscious patients, as they may not be able to communicate or complain of pain reliably during injection. Patients receiving chemotherapy also have an increased risk due to fragile, damaged, and often small caliber vessels. In addition, use of distal access sites (i.e. hand or foot), use of power injection, use of a vessel with multiple puncture attempts, and use of a peripheral IV that has been in place >24 h can also increase the risk of extravasation [13‐15]. Contrast extravasation should be considered if the power injector demonstrates unexpected rapid drop in pressure or exceeds the pressure limit with sudden decrease in flow rate before the full volume of contrast is administered to the patient. Radiograph or CT topogram imaging of the affected limb following an extravasation event may be useful to determine the magnitude of infiltration and verify if compartmentation is present (Fig. 4b) [15]. If some contrast has gone into the patient, the study may still be salvageable. This should be reviewed by the radiologist. If contrast is suboptimal, sometimes it can be amplified by using virtual monoenergetic images from a dual energy scanner. By using low energy virtual monoenergetic images, the energy levels of which are closer to the K edge of iodine, the contrast signal is amplified which can potentially salvage some suboptimal studies. However, if the study is not salvageable or if no contrast went into the area of interest, the study will have to be repeated. This can be done immediately if there is another venous access or later after obtaining appropriate venous access.

ACR Manual on Contrast Media discusses the treatment of contrast extravasation [9]. Raising the affected limb above the level of the heart may reduce swelling and facilitate absorption of extravasated fluid. Some favor cold compresses to decrease pain at the extravasation site and others prefer warm compresses to improve blood flow to the extravasation site and increase absorption of the contrast from the tissues into the vasculature and lymphatics. In addition, attempts to remove the extravasated contrast via aspiration have not been shown to be consistently beneficial. Regardless of the post-extravasation treatment method, patients should be evaluated by the radiologist. If pain is the main symptom, we use cold compresses, and if the extravasation has occurred in a location where there is a higher likelihood of compartment syndrome, we use hot compresses. We observe the patient in the radiology department for at least 1 h to ensure that there are no new symptoms, such as pain or numbness to suggest development of compartment syndrome. Before discharge, a radiologist discusses the findings that would suggest a developing compartment syndrome with the patient. The patient is instructed to seek medical attention if new neurologic or circulatory symptoms or skin ulceration develop [9].

Thoracic outlet syndrome (TOS) refers to the effects of dynamic compression of the nerve, artery, and/or vein as these structures cross the thoracic outlet due to changes in arm position, typically induced by elevation of the arms [16]. Dynamic CTA, with the arm in neutral position and then in elevated positions (130° of hyperabduction with external rotation), can be used to evaluate TOS [17, 18]. Contrast injection should be administered into the vein of the asymptomatic extremity to reduce beam hardening artifact [17, 18].

As the majority of thoracic CTAs are performed with the patient’s arms raised, compression of the subclavian vein (asymptomatic or symptomatic) can lead to compromises in IV contrast delivery to the central vascular structures, affecting bolus timing and leading to suboptimal opacification due to reductions in flow rate (Fig. 5a). If suboptimal contrast opacification of the target vessel is present, reimaging the patient with the arm in the neutral or adducted position should relieve the dynamic narrowing of the thoracic outlet, thereby improving opacification of the vessel (Fig. 5b). Secondary signs of venous stenosis include dynamic collateral vessel filling and distal venous thrombus in chronic cases (Fig. 6). Thoracic venous outlet obstruction should be considered when extensive collateral vessel filling is seen on the side of contrast administration when the patient’s arms are raised. Alternatively, new access from the contralateral extremity vein can be obtained.

Considered a physiologic artifact, transient attenuation or interruption of the contrast bolus refers to disruption of the normal opacified contrast column secondary to return of unopacified venous blood via the inferior vena cava (IVC) in the setting of deep inspiration (Fig. 7a and b). There are two significant imaging consequences of this artifact: missing a true pulmonary embolus due to decreased opacification of the pulmonary artery or misinterpreting the decreased vessel attenuation as an embolus when it is not present. The pathophysiologic mechanism of this artifact is secondary to the normal variable inflow of blood to the right heart during inspiration. Referred to as the “abdominal-thoracic pump”, initial deep inspiration decreases intrathoracic pressure and increases intraabdominal pressure, acutely increasing venous return, favoring flow from the IVC over the superior vena cava (SVC), resulting in a bolus of nonopacified blood entering the right heart from the abdomen [19, 20].

This artifact should be considered when there is decreased opacification of multiple bilateral pulmonary arteries at the same level without vessel lumen distention: true pulmonary emboli typically present at various levels and normally expand the vessel lumen acutely [19]. In patients unable to hold breath, alternatively a free breathing high pitch flash CTA may be obtained [21] (Fig. 7c).

Imaging pearl: Techniques to overcome this artifact often rely on patient respiratory coaching, as the command “take a breath in and hold it” can lead some patients to take a rapid deep inspiratory breath, increasing the risk of transient attenuation of the contrast bolus. Alternate breathing instructions include requesting the patient to stop breathing or to take a slow gentle breath [20, 22].

Sequential contrast opacification of central veins and cardiac chambers can be observed when bolus timing technique is used to identify contrast arrival. Normal sequence of enhancement follows right atrium, right ventricle, pulmonary artery, pulmonary vein, left atrium, left ventricle, and aorta. Earlier opacification of a distal chamber may be an indicator of intra or extra-cardiac shunt. When using bolus tracker technique, failure to adequately opacify the target vessels to reach the threshold needed for triggering the scan may also be an indication of decreased cardiac pump function.

Although the real incidence of cardiac arrest at the time of CT is not known, it is probably not rare [23]. In these patients, the contrast is distributed almost entirely in the venous system with no opacification of the right ventricle, pulmonary artery or aorta and indicates circulatory dysfunction (Fig. 8a and b). There may be retrograde opacification of IVC, hepatic veins, and even portal vein with dependent pooling of the contrast forming a blood-contrast level (Movie 1) [24]. The distribution of contrast medium is now being determined by the push from the power injector and the viscosity of the contrast medium. These patients are likely hemodynamically unstable at the time of presentation and may be on cardiopulmonary monitoring which should be evaluated by the attending radiologist. If the patient is not being monitored, and when such a finding is seen on a nondiagnostic CTA, it is imperative to call the code team and immediately begin cardiopulmonary resuscitation rather than planning for a reinjection.

A less dramatic, but equally important observation may be seen in patients with congestive heart failure with resultant poor or no opacification of left cardiac chambers and aorta during a CT pulmonary angiogram (Fig. 9). The likely explanation for these findings can be increased pulmonary transit time. When using a scanner with shorter acquisition time, non target vessel enhancement may be less than expected, and these vessels should be interpreted with caution. Contrast-blood mixing artifacts are often seen in the right atrium, right ventricle and pulmonary artery during a pulmonary artery CTA due to unopacified blood returning from the IVC. These are, however, not commonly seen in left atrium or left ventricle, and whenever seen should be considered abnormal (Fig. 10).

Decreased systolic function of left ventricle can result in dependent contrast pooling and layering in the aorta [25]. This is likely due to decreased stroke volume with resultant contrast blood pooling with dependent layering of the higher viscosity contrast. Such dependent contrast pooling in descending aorta can also be seen in patients with acute cardiac tamponade, likely due to decreased stroke volume (Fig. 11).

Imaging Pearl: In patients with known heart failure, test bolus can be more useful in identifying time to peak enhancement, which can be delayed compared to contrast arrival time. Use of delayed images after 30 s can help differentiate soft plaque/thrombus from slow flow when dependent pooling is seen.

Differential enhancement of pulmonary arteries during a pulmonary artery CTA can be seen in patients with Fontan circulation (Fig. 12), extra-cardiac shunts such as patent ductus arteriosus, bronchial artery, or coronary artery fistulas (Fig. 13), and when using prospective ECG triggered CTA (Fig. 14). Cavopulmonary shunts that connect the caval and pulmonary circulation are performed in patients with single ventricle physiology. Glenn shunt is performed as the second stage of surgical repair and involves anastomosis between the SVC and the right pulmonary artery, which can either be unidirectional or bidirectional. Fontan shunt is performed as the third stage of ventricular repair and involves anastomosis between the IVC and the left pulmonary artery. In the lateral tunnel Fontan, the right atrial wall is used to create a baffle, whereas in an extra-cardiac Fontan, a conduit is used to connect IVC blood to the pulmonary artery. In classic Fontan, the right atrium and the pulmonary artery are anastomosed. Total cavopulmonary connection involves a Glenn shunt connecting SVC to the right PA and Fontan shunt connecting IVC to left PA. CT angiography in these patients to visualize the pulmonary arteries or the conduits themselves is challenging since the SVC flow is directed to the right lung and the IVC flow is directed to the lung (Fig. 13a). Hence, injecting contrast only through the arm will not result in opacification of the left pulmonary arteries and injection through the lower extremity will not result in opacification of right pulmonary arteries, resulting in non-diagnostic studies [26]. This is important since there is a higher risk of pulmonary thromboembolism (3–19 %) in these patients [27]. Furthermore, due to the absence of pumping action of right ventricle, there is passive laminar flow of Fontan circulation, which causes inhomogeneous enhancement, particularly within the conduit [28]. It has been shown that 13 % of these patients have mural thrombus in the extracardiac conduit [27], even without symptoms, which may be missed with suboptimal studies [27] Solutions for this are (1) Simultaneous upper and lower extremity (femoral vein/foot vein) injections at 4–5 mL/s, so that both the SVC and IVC are opacified simultaneously [27]. However, the arrival of contrast media may not always be simultaneous due to different resistance, collaterals, and flow velocities. (2) Two-phase CT angiography, with both arterial and delayed venous phases (Fig. 12b, Movie 2) [26]. Delayed phase scan only. Delayed phase scan at 3 min has been shown to be good in visualizing entire vasculature during recirculation, regardless of the intravenous route or surgical technique [28]. Some authors use a 1-min delay provided the injection is antecubital due to shorter distance to pulmonary artery and in patients with cavopulmonary connections than atriopulmonary connections [28].

Differential enhancement of ascending and descending aorta during a thoracic aortic CTA can be seen by using a prospectively triggered acquisition, coarctation, large aneurysms, and dissections. When prospective ECG gating is used (Fig. 14), there may be a delay between consecutive axial acquisitions which is exaggerated in the presence of irregular heart rate. This can lead to differential enhancement in different segments of the aorta, which merely indicates different contrast density at different time points. There is no solution to this artifact once acquired, but this can be avoided by using spiral instead of axial acquisitions.

Differential aortic enhancement can also be seen in patients with coarctation of aorta (Fig. 15). This is due to dilution of contrast within the blood pool of the post stenotic dilated aortic lumen. Unless sagittal images are also reviewed, this subtle sign may be the only significant clue seen on axial CTA images.

Mixing artifacts can be seen in large aortic aneurysms and should not be confused with a thrombus (Fig. 16a). Delayed images can help in opacification of the lumen (Fig. 16b).

Differential enhancement of false lumen of an aortic dissection can also be due to delayed opacification due to higher inherent luminal pressures. This should not be confused with a thrombus. In type B dissection, identification of false lumen thrombus can be overestimated by first pass CTA/MRA. True estimation of this false lumen thrombosis after aortic dissection is important as this can be important for prognosis [29]. Delayed phase acquisition is recommended for a more accurate estimation of true extent of false lumen thrombus vs. slow flow. Aortic dissection with partial thrombosis of the false lumen has a significantly higher annual aortic growth rate when compared with those patients with complete thrombosis of the false lumen [30].

Imaging pearl: In patients with known aortic aneurysm, ROI for test bolus or bolus tracking should be placed in that portion of the aorta which has the largest diameter.

Extracorporeal membrane oxygenation or ECMO is increasingly being used in adults for pulmonary or cardiopulmonary support in not just pediatric, but also adult patients with severe respiratory failure or following failure to wean from cardiopulmonary bypass after cardiac surgery [31]. The hemodynamics of flow in these patients, especially those on a venoarterial ECMO, are altered, with retrograde flow occurring in the access artery and in case of femoral artery access, in the aorta [32]. This can lead to variable enhancement pattern (Fig. 17a–f) of aorta, poor opacification of cardiac chambers, and suboptimal enhancement of the pulmonary vessels.

Imaging Pearl: Different approaches have been suggested to perform contrast-enhanced CTA in patients on ECMO: injection into the arterial cannula of the ECMO after the membrane oxygenator or into the venous line distal to the membrane oxygenator [33]. In our experience, slowing the flow of the circuit to the minimal flow rate that would prevent thrombus formation for the duration of the scan (15–20 s) has worked well in cases of suspected pulmonary embolism (Fig. 18).

Central venous catheters are often used for contrast injection. Although there are safety issues related to this such as the risk of catheter rupture, fragmentation, or thromboembolism, these devices can be safely used if appropriate precautions including manufacturer specifications are followed [34]. Studies have shown that vascular enhancement is superior with central venous catheter injections compared to peripheral route injections due to the short time to peak enhancement facilitated by shorter travel distance for contrast bolus. There is also reduced individual patient variability [8]. However, the contrast injection is typically performed slower than peripheral routes due to safety concerns. Since the contrast media will directly opacify the lower SVC or the right atrium and the subsequent cardiovascular structures, the upper SVC and other veins will not be adequately opacified in the first pass as with a peripheral route injection. Hence, if venous visualization is the primary clinical objective, a delayed phase should be obtained in addition to the arterial phase.

Imaging Pearl: Manufacturer recommendations for the central venous catheter that is being used should be adhered to for peak flow rate. A delayed phase, 40 s acquisition can help in identifying any thrombus, vegetation, or fibrin sheath attached to the catheter.