DIGITAL SUBTRACTION ANGIOGRAPHY
Digital subtraction angiography (DSA) using an X-ray fluoroscopy system is the mainstay for visualizing and examining blood vessels in the interventional suite. Using a digital detector such as an X-ray image intensifier or, on modern systems, a flat-panel detector, two-dimensional (2D) DSA acquisitions are composed of two phases. In the first phase, a 2–4-second sequence of static “mask” images is acquired to register the anatomic structures in the field of view, such as bones and soft tissue. In the second phase, X-ray images are acquired as contrast is administered and the mask is automatically subtracted from each frame so that only the contrast opacified structures, such as blood vessels, are visible. Overlapping anatomic structures that may obstruct visualization of the opacified region of interest—such as bones, soft tissue, bowel gas, and calcifications—are removed from the images. The contrast agent most often used is an iodinated agent. For DSA, the iodinated contrast agent is selectively administered through a catheter inserted in the vessel of interest by either using an automated power injector or manually using a syringe. Gadolinium chelates should be used with caution in patients with GFR of less than 60 mL/min to avoid the risk of nephrogenic systemic fibrosis. Carbon dioxide (CO2) is another alternative for patients with allergies to iodinated contrast or renal failure. Its use is limited to subdiaphragmatic vessels to avoid risks to the coronary, cerebral, and spinal vasculature; it also requires imaging at a higher frame rate, with great care taken to avoid contamination with room air.
The digital subtraction technique can be used to visualize a variety of tubular structures, including arteries, veins, and ducts. Due to faster blood flow, arteriograms often use faster frame rates than venograms, though numerous considerations including radiation dose and contrast bolus size are important factors in choosing DSA imaging and contrast injection parameters.
Fig. 1.1: With contrast administered from a replaced right hepatic artery, digital subtraction angiography image in the anteroposterior projection shows overlapping, corkscrew arteries that are typical in a cirrhotic liver, that course toward a hypervascular hepatocellular carcinoma (arrow). The specific tumor-supplying branches are difficult to identify on a single projection, as shown in this image.
A single DSA imaging run overlays all contrast-enhanced vessel branches along one 2D X-ray projection, making it difficult to precisely determine the spatial relationships between overlaid and foreshortened vessels from a single acquisition (Fig. 1.1). This limitation can be overcome with two or more DSA acquisitions, often in orthogonal planes, at the expense of increased 4contrast and radiation doses. To reduce contrast dose, two simultaneous acquisitions can be acquired with equipment capable of biplane imaging. These machines have two X-ray sources and two detectors, which are arranged in orthogonal planes, allowing simultaneous acquisitions with a single contrast bolus. However, even biplane images cannot provide true three-dimensional (3D) spatial orientation.
Hence for complex procedures such as subselective microcatheter embolization, for problem solving and for applications that require 3D spatial resolution as well as soft tissue resolution, there is a need for more advanced technology such as cone-beam computed tomography (CBCT) that can be used in the interventional suite.
C-ARM MOUNTED CBCT
The ability to acquire contrast-enhanced volumetric images in the interventional suite complements DSA by addressing some of its shortcomings. Three-dimensional imaging using an angiographic C-arm system is often called C-arm CT, flat detector CT, or simply CBCT. Several angiographic C-arm systems are capable of reconstructing 3D images. The names vary by vendor: syngo DynaCT (Siemens AG, Forchheim, Germany), XperCT (Philips Medical Systems, Eindhoven, The Netherlands), Innova CT (GE Medical Systems, Waukesha, Wisconsin), Infinix (Toshiba Corporation, Minato/Tokyo, Japan), and Trinias or Bransist (Shimadzu Medical Systems, Kyoto, Japan). All systems are capable of conventional 2D DSA imaging, but also permit the acquisition of CT-like 3D images directly in the interventional room during any procedure.
Cone-beam CT has several strengths that can be useful in a range of clinical scenarios. The biggest benefit of CBCT over conventional 2D DSA is detailed 3D anatomic information. The CBCT data can be viewed in axial, coronal, and sagittal slices as well as volume renderings and planar reformats at any angle. Thus, the operator can elucidate detailed 3D vascular relationships and provide a roadmap that simplifies complex anatomy.1,2
Another strength of CBCT is its high spatial resolution in conjunction with soft tissue details. These provide visualization of small blood vessels and soft tissue structures that may not be apparent on DSA images.1,2 Superior soft tissue and contrast resolution improve visibility of the region of interest, such as tumor or stent structure that may not be readily apparent on subtracted images.3,4 Finally, since CBCT images are acquired in the interventional suite, the above information is made available to the interventional radiologist at the time of the procedure, and provides live feedback that allows modification of the plan during the procedure as opposed to after the procedure, which is often the case with conventional imaging.4
However, the nature of CBCT also presents some shortcomings relative to other cross-sectional imaging modalities. The primary limitation of CBCT relative to DSA and multidetector CT (MDCT) is its longer acquisition time. Because a CBCT rotation takes a minimum of 3 seconds while an MDCT rotation can be as fast as 0.3 seconds, motion artifacts are far more pronounced in CBCT. Unlike DSA, where motion can be corrected for by remasking the images, CBCT has no such recourse. Additionally, the physical gantry rotation required for CBCT may limit the field-of-view (FOV) coverage for some large organs or in large patients. Without correction, the latter results in a truncation artifact—a bright ring at the border of the FOV. In general, 2D DSA provides a larger FOV than a CBCT reconstruction using the same imaging system. Magnetic resonance imaging (MRI) and MDCT also have better soft-tissue contrast resolution, allowing better differentiation of anatomical structures. The larger detector pixels used in MDCT also typically permit better post-patient collimation, resulting in less noise from scattered radiation. While CBCT cannot supplant MDCT or MRI for diagnostic imaging, CBCT is a valuable tool for complex interventional applications.
ACQUISITION AND RECONSTRUCTION
Cone-beam CT acquisitions collect numerous high-resolution 2D projection X-ray images during a rotation of the C-arm gantry (both the X-ray source and detector) around the patient. These projection images are then reconstructed into a 3D volume.
Geometric sampling constraints require a rotation angle around the patient of 180° plus the fan angle (typically about 20°) in order to produce accurate reconstructions. During a single rotation, between 100 and 600 individual projections are acquired at an evenly spaced angular increment around the patient, with the total acquisition time varying from 3 to 20 seconds. The wide variation in these parameters is a result of the versatile applications of CBCT; high-contrast applications such as contrast-enhanced angiography require fewer projections and shorter scans than applications in which soft-tissue characterization is important. Cone-beam CT systems have numerous preprogrammed protocols optimized for different applications so that the operator does not need to 5select individual parameters during a procedure. These protocols include factors such as tube voltage, beam filtering, system dose, pixel binning, rotation trajectory, frame-rate, acquisition time, and more.2
The X-ray sources used in CBCT are primarily designed for angiography. They are typically operated at lower tube voltages (kV) than MDCT sources. However, unlike MDCT in which mAs is typically set by the protocol, CBCT systems use automatic exposure control to maintain a constant detector entrance dose, which ensures a constant signal-to-noise ratio throughout the scan by modulating the exposure parameters such as tube current, exposure time, and, if necessary, tube voltage from frame to frame.5
Modern CBCT systems utilize flat detectors (FDs), which contain a crystalline scintillator material to convert X-rays into visible light that is subsequently captured by a digital silicon photodiode array. Typical FDs have a measurement field of up to 41 × 41 cm, an isotropic pixel size as low as 154 µm, and a dynamic range as high as 14 bits. In order to achieve sufficient angular sampling with a reasonable acquisition time, CBCT projection images are typically acquired at 30–60 frames per second. Image data often cannot be read from each individual pixel quickly enough to enable such high frame-rates, so the intensity values measured at neighboring detector pixels are combined into one output pixel. This “pixel binning” technique typically combines either 2 × 2 or 4 × 4 adjacent pixels, reducing the effective pixel resolution of the detector isotropically while enabling frame rates of 30 fps and 60 fps, respectively. The 2 × 2 binning mode is most commonly used for head or neck applications, where a higher spatial resolution requirement and less involuntary patient motion justify the slower rotation speeds associated with this mode. Controlling motion artifacts in abdominal and thoracic imaging is more challenging, so the 4 × 4 binning mode with its corresponding faster scan times is preferable for these applications.
Contrast-enhanced CBCT acquisitions require injection of iodinated contrast, similar to DSA. Since iodinated contrast is a positive contrast agent, meaning it has a higher density than blood, it results in greater X-ray attenuation. Consequently, contrast-enhanced areas appear dark on 2D fluoroscopic images and bright in 3D CBCT images. Carbon dioxide, on the other hand, results in negative contrast enhancement: due to its low density, appears bright on 2D fluoroscopic images and appears dark on the CBCT image. Wong et al.6 explored the feasibility of using CO2 as an intra-arterial contrast agent for CBCT for arterial mapping and tumor visualization during transcatheter arterial chemoembolizations. Twelve patients with unresectable hepatocellular carcinoma (HCC) and at high risk of nephrotoxicity underwent 15 CBCT procedures. The results were compared to gadolinium contrast and showed good correlation. One shortcoming of the use of CO2 is that CO2 does not mix with blood and could result in bolus fragmentation and underestimation of the vessel lumen (Figs. 1.2A and B).
The raw data in a contrast-enhanced CBCT acquisition is composed of 2D DSA images that are reconstructed into a 3D image on a separate workstation. A few important steps can help improve image quality and ensure the C-arm can rotate around the patient. First, for thoracic and abdominal applications, it is critical that the patient can hold his or her breath for the duration of the scan to prevent excessive motion artifacts. To reduce beam-hardening artifact, the arms should be temporarily positioned out of the FOV, typically over the patient's head, similar to the position for MDCT. Beam hardening monitoring devices such as the tubing from a blood pressure cuff or a pulse oximeter, need to be temporarily disconnected or positioned on the lower extremities. Finally, oxygen tubing, tubing for intravenous medications, arm-boards, and other devices that can potentially get caught in the C-arm during its rotation need to be kept clear of the C-arm arc.
Some operators prefer to use diluted contrast to decrease the beam hardening from the iodinated contrast itself, though this is not universal. For most vascular applications, the acquisition of CBCT images is often triggered following a brief delay after contrast injection so that the vessel of interest and the parenchyma, if so desired, are optimally enhanced. The X-ray delay required for such opacification depends on the vessel of interest and transit time from the catheter to the organ of interest. Thanks to continuous technological improvements, the transfer of these 2D images to a workstation for reconstruction into 3D CBCT images that are available for review is now just a matter of a few seconds following acquisitions.7,8
CBCT APPLICATIONS
Over the last decade, CBCT has found a rapidly growing number of applications for procedures outside the brain. Although primarily used for liver-directed therapies, a number of applications also exist for procedures outside the liver. Below is a description of some of the more common applications.6
Figs. 1.2A and B: Carbon dioxide (CO2) can be used as an alternative contrast medium for both (A) digital subtraction angiography (DSA) and (B) cone-beam computed tomography (CBCT). (A) In this DSA image, the contrast has been digitally inverted; thus, the CO2 administered from the right hepatic artery appears dark. The tumor in the superior right hepatic lobe is well visualized. (B) Similarly, 3D reconstruction of minimum intensity projection CBCT images with CO2 administered from the proper hepatic artery shows a superior right hepatic lobe tumor and its supplying arterial branches.
A particular indication for which CBCT guidance is critical is prostate artery embolization for benign prostatic hypertrophy. This is covered in detail in Chapter 69. Detailed description of applications for neurointerventional procedures is outside the scope of this chapter, but can be found in the literature.
Transarterial Chemoembolization
Randomized control studies have demonstrated the survival benefit of transhepatic arterial chemoembolization (TACE) for patients with unresectable HCC.9,10 Based on these randomized studies as well as other prospective and retrospective data, the Barcelona Clinic Liver Cancer treatment algorithm at the 2000 European Association of the Study of Liver disease recommends TACE for patients unsuitable for resection and/or to serve as a bridge to transplantation.11
Early and frequent screening is advised for patients at risk for developing HCC.12 With the better detection of early-stage, small HCCs,13 the importance of superselective catheterization that maximizes drug delivery to the tumor, while minimizing nontarget delivery, cannot be overemphasized. Several studies have demonstrated better treatment efficacy, local response, and lower toxicity with superselective TACE.14-17 Although DSA remains the default modality for guidance and navigation, the underlying cirrhosis increases the technical challenge by rendering the hepatic arteries tortuous, with a corkscrew configuration and crowding and overlapping of vessels on 2D imaging. Prior to the advent of angiographic units capable of CBCT, operators often performed CT hepatic arteriogram (CTHA) to identify and target small tumors; however, this entailed laborious and potentially unsafe transfer of patients from the angiographic suites to CT.18 The availability of hybrid angiographic units overcomes this limitation, meeting the demands for both excellent temporal resolution provided by DSA and high spatial resolution, soft tissue contrast and 3D roadmaps provided by the CBCT acquisition.1,2,19
The use of CBCT during TACE is twofold; as a problem-solving tool, CBCT is used for identification and better characterization of the tumor itself,3,4 particularly for small (<2 cm) HCCs.20,21 Cone-beam CT allows for identification of all potential segmental and subsegmental arteries that supply the tumor(s)3,4,22 (Figs. 1.3A to D). Cone-beam CT has been successfully used in identifying the presence of dual supply in watershed tumors as well as identification of small arteries supplying the caudate lobe,23 replacing the multitude of DSA acquisitions that would be needed to identify these small arteries often arising from the common hepatic artery. The identification of extrahepatic supply to a tumor24 (Figs. 1.4A to F) and the pivotal role provided by CBCT for navigational guidance of the tortuous hepatic arteries3,4,25 is well documented (Fig. 1.5).7
Figs. 1.3A to D: Contrast-enhanced cone-beam computed tomography (CBCT) allows for identification of potential tumor-supplying arteries in transarterial oncologic interventions. (A) Contrast-enhanced CBCT axial image shows three arterial branches (arrow segment 5, dotted arrow cystic artery and arrowhead segment 8) supplying a hepatocellular carcinoma (HCC) in a watershed area. (B) The arteries can also be visualized in the coronal reconstruction of the CBCT data set and confirmed arterial supply from segment 8 (arrow), segment 5 (not shown, out of plane), and the cystic artery (arrowhead). (C) The presence of multiple segmental arteries supplying this HCC is further confirmed when, after superselective transhepatic arterial chemoembolization is performed from both the cystic artery (shown) and segment 5 artery, only partial “staining” (dark) of the tumor is visualized. The superolateral aspect of the tumor (dotted arrow) shows no uptake of the treatment agents. (D) Digital subtraction angiography image with contrast administered from a tumor-supplying segment 8 branch shows supply to the superolateral aspect of the HCC. Chemoembolization from this branch completed treatment of this tumor.
Finally, as a confirmatory tool, CBCT is used postembolization to confirm circumferential and complete uptake of the chemoembolic emulsion in the tumor (Fig. 1.6).3,4,26-28
In a study published by Tognolini et al., the utility of CBCT in 100 consecutive TACEs in 84 patients was analyzed. In this study, CBCT provided additional information not discernible on DSA in 36% of the patients and resulted in a change in diagnosis, treatment planning and/or delivery in 28%.4 Of these, CBCT identified additional or angiographically occult tumors in 17% of patients, identified additional segmental arterial supply in 7% of patients, and helped lay out the arteries for easier navigation in 9.5% of patients.48
Figs. 1.4A to F: Contrast-enhanced axial magnetic resonance imagings (MRIs) demonstrate a hepatocellular carcinoma in the inferior tip of the right hepatic lobe (arrows) with (A) enhancement in the arterial phase and (B) contrast washout in the portal venous phase. (C) On cone-beam computed tomography (CBCT) axial image with contrast administered from the right hepatic artery, the tumor should appear similar to the arterial phase MRI image. Here, there is only enhancement of the medial aspect of the tumor. The absence of partially unenhanced, but known viable tumor (arrow), indicates supply from parasitized extrahepatic arteries and serves as a trigger for interrogation of adjacent extrahepatic artery, in this case an omental artery. With contrast injected from the gastroduodenal artery, (D) axial CBCT, (E) coronal CBCT, and (F) digital subtraction angiography anteroposterior plane images show omental branches supplying the lateral (arrow) and inferior aspects (arrowhead) of the tumor.
Fig. 1.5: An advantage of cone-beam computed tomography (CBCT) is the rotational three-dimensional (3D) vascular roadmap (MIP images) that can be created from the raw CBCT data. Here, a single contrast-enhanced acquisition from the common hepatic artery not only provided the operator with multiplanar reconstruction (shown in Figs. 1.3A to D), but also processed to create a 3D rotation view of the arterial anatomy. The 3D image can be rotated to identify the specific tumor-supplying branches (arrows) and map out the intravascular course required to selectively catheterize those branches. The images can be manually rotated to provide the best angle to lay out the segmental arteries of interest to allow for easy catheterization.
Fig. 1.6: Following superselective transhepatic arterial chemoembolization using chemoembolic emulsion containing ethiodized oil, axial noncontrast CBCT image shows dense circumferential and complete internal uptake of the therapeutic emulsion in this hepatocellular carcinoma, indicating a technically complete treatment. Complete, circumferential uptake of the chemoembolic emulsion has been correlated with excellent response and is used as a surrogate for technical and clinical success.
Wallace et al. demonstrated similar findings, where the addition of CBCT changed management during a TACE procedure in 19.2% of patients.3 Similarly, the identification of extrahepatic supply to a tumor is critical for a durable response.24 Tumors that are in Couinaud segments at risk for extrahepatic supply, such as those in segment VII abutting the diaphragm and hence in close proximity to the phrenic artery, may be treated incompletely if the extrahepatic supply goes unrecognized. Along those lines, tumors on the anterior surface of the liver may receive additional supply from the internal mammary artery and tumors in segment VI may have additional supply from the omental artery. In these scenarios, a contrast-enhanced proper hepatic artery CBCT can demonstrate paucity of perfusion in a part of the tumor, indicating extrahepatic supply. Likewise, a completion unenhanced CBCT can demonstrate paucity of chemoembolic emulsion deposition in parts of the tumor, triggering a search for additional vascular supply to that portion of the tumor. Finally, the importance of assessing the deposition of the chemoembolic emulsion and its correlation to tumor response cannot be overstated. Studies have demonstrated the CBCT can be used in lieu of helical CT for assessing ethiodized oil uptake in HCC28,29 and further, the use of CBCT has been positively correlated to improved local progression-free and overall survival.30,31 The recent use of drug-eluting embolics due to widespread shortage of lyophilized doxorubicin and iodized oil does limit the ability to evaluate the uptake of the chemoembolic emulsion. However, Suk Oh et al. demonstrated that contrast retention in the tumor observed on an unenhanced postembolization acquisition can be correlated to local response (Fig. 1.7).32
The addition of CBCT to DSA does come at the expense of operator and patient radiation as discussed below. However, Kothary et al. demonstrated that a single volumetric acquisition can replace several DSA acquisitions for guidance, decreasing overall exposure.33 More importantly, in patients with HCC, studies have correlated the use of CBCT to improved local control31 and as an independent factor associated with longer progression-free (hazard ratio, 0.25, P = 0.003) and overall survival (hazard ratio, 0.40, P = 0.03) with a 3-year survival of 65% compared to 44% in patients where CBCT was not utilized.30 The acceptance of CBCT for TACE has led to further development of the technology.10
Fig. 1.7: Following transhepatic arterial chemoembolization using drug-eluting embolics delivered with iodinated contrast, axial noncontrast CBCT image shows circumferential and complete contrast retention in this hepatocellular carcinoma. Similar to circumferential uptake of ethiodized oil, circumferential retention of the iodinated contrast used to suspend the drug-eluting embolics, has been correlated with excellent response.
Automated vessel tracking systems have been reported to have higher sensitivity of detecting small subsegmental tumor-feeders than DSA; this does come at the expense of increased false positives and needs further refinement.34,35 Finally, perfusion imaging including parenchymal blood volume and dual phase imaging are other software algorithms that are being explored as prognostic tools.36,37
Transarterial Radioembolization
The benefits of CBCT in TACE not only have similar applications for transarterial radioembolization (TARE) but in some instances, may have an even greater impact. In TARE, resin or glass microspheres containing Yttrium-90 (Y-90), a pure-β-emitting radioisotope, are administered to the tumor(s). Transarterial radioembolization is often used for bulky or multifocal tumors, including HCC and other metastatic disease to the liver. The administration of Y-90 particles could be segmental, but more commonly, to a whole lobe or the whole liver, depending on the extent of the disease. Prior to administration of these radioactive particles, meticulous mapping of vascular anatomy is needed for several reasons. First, the particles themselves are radiolucent and their distribution can only be inferred from the preadministration imaging. Second, unlike conventional chemoembolic emulsion or even drug-eluting embolics, the Y-90 microspheres are severalfold smaller, ranging in size from 20 to 60 µm, and can traverse small, hair-thin vessels that would be too small for the chemoembolic agents. Further, for tumors that are supplied by extrahepatic arteries or by accessory hepatic arteries, incomplete tumor perfusion and hence administration would result in incomplete coverage. And finally, although the risk for nontarget embolization exists for both TACE and TARE, the added risk of persistent radiation-induced damage from TARE adds to the morbidity of nontarget embolization. Since TARE is often lobar or whole liver, a common hepatic CBCT acquisition can identify small arteries arising from the hepatic artery that supply extrahepatic structures such as the right gastric or supraduodenal arteries38,39 (Figs. 1.8A and B). Even if the culprit vessel is not obviously visible on the rotational images, enhancement of extrahepatic organs can be identified on the multiplanar soft tissue imaging, leading to further interrogation.38,39 Hence the incorporation of CBCT into TARE planning can increase operator confidence and reduce the risk of life-threatening complications. Similar to its application in TACE, CBCT can identify and address tumors with parasitized extrahepatic arteries.38 In a study by Louie et al., the authors reported that CBCT identified extrahepatic enhancement or incomplete tumor perfusion in 52% of TARE patients, of which 33% was evident exclusively on CBCT.38 Finally, in an application that is unique to TARE, Abdelmaksoud et al. reported a technique of consolidating hepatic arterial inflow by embolizing variant hepatic arteries. In this study, the authors noted that patients with multifocal hepatic neoplasia with tumor supply from both standard and variant hepatic arteries would require Y-90 administration from two separate sites, each requiring separate catheterization and hence increasing delivery set up time, risk of nontarget embolization, and increased radioactive waste. Following embolization of the variant hepatic arteries, the authors reported redistribution or cross-perfusion of the tumor from the standard hepatic arteries via intrahepatic collaterals as confirmed on a contrast-enhanced CBCT acquisition.40 Finally, although not commonly used, the 3D information provided by CBCT can also be used to calculate the treatment volume required for calculating the dose for TARE.8
TIPS and Other Portal Vein Applications
The first reported use of CBCT for a liver-directed intervention was by Sze et al., in which the authors described the use of CBCT to guide the creation of a transjugular intrahepatic portosystemic shunt (TIPS) in a patient with polycystic disease involving the liver.41 Since then, operators have attempted to use CBCT to successfully direct the needle from the hepatic vein to the portal vein.11
Figs. 1.8A and B: In transarterial radioembolization, it is critically important to identify potential variant hepatic arterial branches that supply extrahepatic structures such as the stomach. (A) With contrast injection of the common hepatic artery, digital subtraction angiography image in the anteroposterior projection shows a branch of the segment 3 artery that courses beyond the liver margin and toward the stomach (arrow). (B) Axial cone-beam computed tomography image with contrast administered from the same artery confirms supply to the anterior body of the stomach (arrow). This variant segment 3 hepatic artery branch supplying the stomach was embolized with coils to prevent nontarget delivery of radioactive material to the stomach.
This process has not been widely adopted as it is time-consuming and technically challenging compared to triangulation using orthogonal CO2 portograms or, more recently, the use of intravascular ultrasound. Wallace et al. suggested the possibility of obtaining a CBCT generated indirect map of the portal vein, created by manually injecting CO2 into a wedge hepatic venous catheter throughout a C-arm rotation. The resulting 3D map of the portal vein is then used as an overlay with live fluoroscopy to provide a better roadmap of the portal vein.1 This technique, while not commonly practiced, can be a valuable problem-solving tool for difficult TIPS procedures.41
While TIPS remains the preferred method for portal decompression for variceal bleeding in United States and Europe, many Asian countries prefer to obliterate varices using sclerotherapeutic approaches such as balloon-occluded retrograde transvenous obliteration (BRTO).43 Balloon-occluded retrograde transvenous obliteration not only obliterates the culprit varix, but can theoretically improve hepatopedal portal venous flow and liver function. However the advantage of BRTO is often weighed against possible severe complications from spillage of the liquid sclerosant into the systemic circulation, including hemolysis and acute respiratory distress syndrome.44 To decrease the rate of complications, Koizumi et al. described a technique of foam sclerotherapy (gaseous mixture) in which delivery into the varix was visualized using either CT or, more conveniently, using CBCT.43 The same group found that BRTO using foam sclerotherapy under CBCT guidance allowed significant reduction of the sclerosant dose and resulted in a negligible complication rate while maintaining a high success rate.45
The use of CBCT for portal vein embolization is similar to that of TACE and TARE, in that CBCT can provide a 3D vascular road map of all the portal branches along with multiplanar reformats that can provide information on the liver segments and their associated portal branches. Of particular benefit is depiction of segment IV and its branches, which may need to be embolized or spared, depending on the extent of the surgery. A single contrast-enhanced acquisition can depict the size and the location of the segment IV portal branch, increasing operator confidence and safety.1,41
Other Vascular Applications
The availability of multiplanar imaging that displays soft tissues as well as the vascular roadmap provided by CBCT has found other applications, including those for venous and aortic interventions.12
For the treatment of challenging abdominal aortic aneurysms treated with fenestrated grafts, CBCT has been used to identify visceral branches for catheterization following placement of the fenestrated graft. In a study published by Dijkstra et al., the authors used a technique in which images from a preprocedure planning MDCT were fused with a procedural CBCT for placement of the fenestrated graft followed by a postdeployment CBCT.46 The authors reported that the use of CBCT resulted in significantly lower volume of iodinated contrast required, a clear benefit in this subgroup of patients who often have impaired renal function. Additionally, postdeployment CBCT identified endoleaks not identifiable on DSA that were addressed at the time of the primary intervention.46 The advantage of CBCT over DSA in demonstrating postdeployment endoleaks has been reported in other studies, allowing for treatment at the time of the primary procedure and hence decreasing the need for subsequent interventions.47
The use of fusion imaging, which combines and overlays preprocedural imaging with a procedural CBCT and live fluoroscopy, has found applications in direct puncture of the aneurysmal sac46 followed by a percutaneous and/or endovascular approach to obliterate type II endoleaks.
The use of CBCT for problem solving during other peripheral arterial interventions has been documented in the literature, albeit predominantly in the form of case-reports or small series. Most of these involve the use of CBCT to provide a vascular roadmap, especially for visceral aneurysms/pseudoaneurysms where inflow and outflow vessels need to be identified.8 Similarly, CBCT is a useful adjunct in patients undergoing nephron-sparing embolization of large renal angiomyolipomas (AML). In these cases, not only does CBCT identify every artery supplying the AML but also serves to verify adequate and complete devascularization of the AML at completion (Figs. 1.9A to E).
Interestingly, CBCT has found significant applications in venous interventions. The most notable one is for adrenal vein sampling. Adrenal vein sampling requires catheterization of both the right and the left adrenal vein. However, other small veins, especially on the right, may mimic an adrenal vein, leading to false positive catheterizations. Digital subtraction angiography imaging is often equivocal in these situations, and confirmation of correct catheterization of the adrenal vein can be performed by CBCT48 (Fig. 1.10). Other ad hoc applications reported in the literature include guidance during difficult inferior vena cava filter removal and direct puncture of venous lakes associated with venous malformations. In the former, CBCT has been used to identify the position of the apex of the filter, allowing the operator to define a view best suited to snare the apex as well as providing additional information to the operator of wall penetration by the legs of the offending filter.49 For venous malformations, CBCT has been used to directly puncture a venous lake that is too deep to be identified by ultrasound, akin to direct percutaneous access to arterial aneurysmal sacs.50
Nonvascular Applications
Biliary Interventions
Simple decompression of a dilated biliary system can be done efficiently with ultrasound imaging and fluoroscopy and rarely requires the use of CBCT. Cone-beam CT can provide anatomic detail that is useful for surgical planning. A single 3D rotational acquisition of the opacified biliary system can provide detailed information on the segments drained by a particular duct, its insertion into the main hepatic ducts, and any variant anatomy if present. Similar to other applications of CBCT for navigational guidance, a single 3D acquisition can often replace multiple views typical of a biliary mapping that results in overall less radiation and iodinated contrast dose.42,51
Percutaneous Biopsies
The effectiveness of CBCT-guided lung biopsies is comparable to alternative techniques, such as fluoroscopy-guided and conventional CT-guided. Reported diagnostic accuracy of CBCT guided lung biopsies of nodules <3 cm is upward of 95%, paralleling that of conventional CT-guided lung biopsy.52,53
Many newer fluoroscopy systems have features that allow safe path planning based on the CBCT data, permitting the operator to avoid critical structures such as vessels and bullae. These systems, including XperGuide on Philips C-arm systems and iGuide on Siemens C-arm systems allow incorporation of preprocedural 3D imaging data (such as MDCT or MRI) into the approach planning process as well as overlay of the volumetric imaging data on live 2D fluoroscopy. Motion artifact due to breathing is still a rate-limiting factor for widespread use of this technology for percutaneous biopsies, but it remains a viable option when resources are limited and alternative methods such as ultrasound and conventional CT are not available.13
Figs. 1.9A to E: Cone-beam computed tomography (CBCT) is similarly useful in transarterial embolization of angiomyolipomas (AMLs). (A) Left renal artery digital subtraction angiography image shows an exophytic mass (arrow) with tortuous internal vascularity consistent with known AML. Contrast-enhanced CBCT in (B) axial and (C) coronal planes better identify the individual supplying branches to the AML, allowing nephron-sparing therapy. Following embolization with a mixture of absolute alcohol and ethiodized oil administered via a microcatheter after superselective catheterization of individual tumor supplying arteries, a noncontrast CBCT in (D) axial and (E) coronal images verify adequate treatment as confirmed by the hyperdense therapeutic material within the vascular portions of the mass.
Fig. 1.10: In adrenal vein sampling, it can be difficult to determine whether the adrenal vein or an adjacent venous structure is selected. In this axial cone-beam computed tomography image, successful catheterization of the right adrenal vein is confirmed as contrast injection via the catheter opacifies the right adrenal gland (arrow).
Ablation
The success of CBCT in needle guidance for biopsies has led to interest in using CBCT to place the electrodes for radiofrequency ablation (RFA). Cazzato et al.54 compared 40 RFA-treated lung tumors using either MDCT or CBCT guidance and discovered that CBCT guidance is faster regardless of tumor size, or to require electrode repositioning, or to cause pneumothoraces, and results in fewer recurrences. The same group established feasibility of using 3D CBCT hepatic angiography to guide percutaneous RFA of liver tumors under 1.5 cm.55
Cone-beam CT can also be used to assess RFA endpoints and success. Abdel-Rehim et al. published a series of 23 cases in which contrast-enhanced CBCT was shown to provide adequate assessment of the ablation zone.56 Similarly, Ierardi et al. showed in a seven-case series that a CBCT-based treatment planning software could accurately predict the ablation volume for both microwave and RFA of lung tumors.57
CBCT LIMITATIONS AND ARTIFACTS
As with any imaging modality, the nature of CBCT acquisitions results in some notable image artifacts that are important to understand when acquiring and interpreting CBCT data.5
Fig. 1.11: A limitation of cone-beam computed tomography (CBCT) is image degradation due to motion artifact from a poor breath-hold. As seen here, the contrast-enhanced axial CBCT image is limited by respiratory motion as the vessels are blurred and not well-defined. The respiratory motion and movement of the catheter, further adds to the beam hardening artifacts due to contrast that is too dense.
Motion
Due to the slower rotational speed of the C-arm system compared to MDCT, the image reconstruction quality is highly sensitive to motion during the acquisition.
Three major sources of motion commonly corrupt CBCT images: cardiac motion, involuntary patient movement, and respiration. In order to reduce motion and respiratory artifacts, CBCT images for thoracic and abdominal applications are obtained during a breath-hold.7 Most patients undergoing liver-directed therapy are capable of a 5–12 seconds breath-hold; however, respiratory motion can degrade images substantially in up to 10% of the patients4 (Fig. 1.11). For cardiac vasculature applications, the heart motion is another source of motion artifacts. A clinically available product can perform multiple electrocardiogram-triggered sweeps around the patient, allowing for correction of cardiac motion.58
Truncation
The flat-detectors of current C-arm systems are not always large enough to capture the entire patient width. Consequently, some or all 2D projections are truncated; in other words, they do not show the whole patient during a 3D acquisition. These truncated projections cause a bright ring at the border of the field of view and incorrect density values. Modern systems utilize truncation correction 15algorithms, so the most practical technique is to align the region of interest with the C-arm system's isocenter before the 3D acquisition. For large patients, this can mean positioning the patient slightly off-center on the table to allow isocentering with the C-arm.4
Image Noise
Noise in CBCT images manifests as random imprecision in the measured brightness level of projection image pixels as well as in the Hounsfield unit (HU) value of reconstructed voxels. There are two main sources of image noise in CBCT images: X-ray fluence variations and scatter. The statistical variation in the number of X-ray photons reaching the detector results in variations in the brightness of reconstructed voxels. Another source of noise in X-ray images is scattered radiation, which arises from interactions between incident radiation and the patient. A scattered photon erroneously adds to the brightness detected at whichever pixel it is incident upon, reducing the contrast of the overall image. Modern CBCT systems incorporate antiscatter grids to reduce the amount of scattered radiation that reaches the detector, but these are less effective than MDCT antiscatter grids due to the smaller size of CBCT detector pixels. To reduce scatter in CBCT images, it is important to minimize the amount of material irradiated. Many CBCT systems also permit reduced-FOV imaging (such as slab collimation) that reduces the area of the patient irradiated, reducing both scatter and radiation dose to patient and operator.
Metal Artifact
Metal objects, such as implanted devices or coil masses, can strongly attenuate X-rays, resulting in few photons, if any, passing through to the detector. When this occurs, it becomes impossible to calculate the attenuation of the tissue beyond the metal object. This results in black bands appearing in the image behind the object. Using a higher energy beam can improve X-ray penetration at the cost of lower contrast in the image. For larger metal objects, however, an unacceptable amount of radiation may be required to penetrate the object, so the only way to circumvent the artifact is to coordinate the patient position and scan trajectory so that the artifact appears in an area of diagnostic unimportance. Many CBCT systems offer multiple scan trajectories, such as anterior-to-posterior and lateral-to-lateral position. Choosing a scan that minimizes projection of important areas over the metal object can reduce the clinical impact of the artifact.
Beam-Hardening
The polyenergetic spectrum of an X-ray source “hardens” while passing through an object as low-energy photons are more likely to be absorbed by the object than the high-energetic photons. Thus, the mean energy of the beam increases as it traverses the object, resulting in two types of artifacts: the so called “cupping” artifact and dark bands or streaks around edges of dense objects. After traveling through a dense object such as bone, the hardened beam is attenuated less by soft tissue, resulting in artificially reduced CT numbers. This typically manifests as a dark cup around bony objects or densely opacified vessels. Beam hardening can also manifest as streak artifacts between multiple highly attenuating components, e.g. between bony structures or structures densely opacified with contrast media.
Ring Artifact
Each pixel of the flat panel detector behaves slightly differently when exposed to X-rays. Due to inhomogeneities in the scintillator coating as well as in the semiconductor material, the same incident photon energy can result in varying signal output at different pixels. To correct for this, flat panel detectors undergo pixel-by-pixel intensity calibration, and modern systems continuously collect additional calibration data for ongoing corrections. However, if a defective pixel behaves differently from the calibration, the intensity measured at that point will be inaccurate for each projection, which will appear as a ring in an axial slice of the back-projected volume. Over time, as numerous pixels begin to stray from the calibration, the reconstructions will exhibit ring artifacts, which are concentric rings most clearly visible in axial slices. Ensuring that the system undergoes regular maintenance, including calibration, will remove this artifact.
PATIENT SAFETY
As with an X-ray-based imaging modality, the amount of ionizing radiation exposure to the patient is a major concern applying CBCT imaging. There must always be a balance between the utility of the information provided by CBCT and the risks inherent to radiation. There is a deterministic and stochastic risk inherent with ionizing 16radiation. The deterministic effect is characterized by a threshold dose; in other words, the deterministic effect is not observed until the dose exceeds a certain threshold. If the absorbed dose of an individual is over the specific threshold, the risk and severity of the injury increases. Examples of injuries due to deterministic effect include skin burns and hair loss. The probability of a stochastic effect increases with an increase in radiation dose; however, the severity of the effect is independent of the dose. Radiation-induced cancers are stochastic injuries.
The absorbed patient radiation dose is measured by the dose-area-product (DAP) and the cumulative dose (CD), which provide a good estimate of stochastic and deterministic risks, respectively.33,59
Kothary et al. presented a prospective study with 87 patients with HCC who underwent TACE to evaluate the impact of CBCT on radiation exposure.33 They compared a group who only underwent conventional 2D DSA (control group) and a second group who underwent a combined CBCT/DSA imaging (test group) with respect to the DAP and the CD. A marginal increase in DAP in the test group was offset by a substantial decrease (50%) in CD compared to the control group. In general, the routine use of CBCT can increase DAP, but can decrease the deterministic risk (CD) from DSA. The increase in DAP is operator-dependent and can thus be reduced to under 10%. Further, a single CBCT acquisition can often replace multiple orthogonal DSA acquisitions, thereby reducing overall DAP. The authors also state that CBCT provides additional information in 33% of the patients compared to DSA, while decreasing the amount of iodinated contrast required.
In a comparison of rotational angiography (RA) and conventional angiography (CA) of the coronary arteries, Loomba et al.60 conducted a meta-analysis of 11 studies including 1,916 patients, with approximately half of the patients in each group. Their analysis showed a statistically significant decrease in total radiation dose in the RA group as measured by air kerma. Based on the range of mean air kerma for the RA groups of the included studies, the meta-analysis concluded that the use of RA reduces total radiation exposure by nearly half. Rotational angiography also required less contrast, with a difference of 19.4 mL between the groups. While CBCT reconstruction of rotational coronary angiograms remains challenging due to cardiac motion, the results of this study support the use of rotational acquisitions over conventional angiograms from the perspective of radiation and contrast exposure. Overall, all ionizing radiation should be used with caution, with image acquisition choices tailored to the patient's disease state and age.
SUMMARY
The use of CBCT continues to steadily increase as more interventional applications take advantage of the modality's benefits. The ability to obtain 3D imaging along with soft tissue resolution, complements the role of DSA in the interventional suite. Its problem-solving role for liver-directed therapies and for other applications makes it an invaluable tool in the interventional radiologist's arsenal and should be exploited to its full capacity, but with the understanding of its additional radiation dose.
REFERENCES
- Wallace MJ, Kuo MD, Glaiberman C, et al. Three-dimensional C-arm cone-beam CT: applications in the interventional suite. J Vasc Interv Radiol. 2008;19(6):799–813.
- Orth RC, Wallace MJ, Kuo MD, Technology Assessment Committee of the Society of Interventional R. C-arm cone-beam CT: general principles and technical considerations for use in interventional radiology. J Vasc Interv Radiol. 2008;19(6):814–20.
- Wallace MJ, Murthy R, Kamat PP, et al. Impact of C-arm CT on hepatic arterial interventions for hepatic malignancies. J Vasc Interv Radiol. 2007;18(12):1500–7.
- Tognolini A, Louie JD, Hwang GL, et al. Utility of C-arm CT in patients with hepatocellular carcinoma undergoing transhepatic arterial chemoembolization. J Vasc Interv Radiol. 2010;21(3):339–47.
- Fahrig R, Dixon R, Payne T, et al. Dose and image quality for a cone-beam C-arm CT system. Med Phys. 2006;33(12): 4541–50.
- Wong AA, Charalel RA, Louie JD, et al. Carbon dioxide contrast enhancement for C-arm CT utility for treatment planning during hepatic embolization procedures. J Vasc Interv Radiol. 2013;24(7):975–80.
- Tognolini A, Louie J, Hwang G, et al. C-arm computed tomography for hepatic interventions: a practical guide. J Vasc Interv Radiol. 2010;21(12):1817–23.
- Angle JF. Cone-beam CT: vascular applications. Tech Vasc Interv Radiol. 2013;16(3):144–9.
- Llovet JM, Real MI, Montana X, et al. Arterial embolisation or chemoembolisation versus symptomatic treatment in patients with unresectable hepatocellular carcinoma: a randomised controlled trial. Lancet. 2002;359(9319):1734–9.
- Lo CM, Ngan H, Tso WK, et al. Randomized controlled trial of transarterial lipiodol chemoembolization for unresectable hepatocellular carcinoma. Hepatology. 2002;35(5): 1164–71.
- Bruix J, Sherman M, Llovet JM, et al. Clinical management of hepatocellular carcinoma. Conclusions of the Barcelona- 2000 EASL conference. European Association for the Study of the Liver. J Hepatol. 2001;35(3):421–30.
- Murakami R, Yoshimatsu S, Yamashita Y, et al. Transcatheter hepatic subsegmental arterial chemoembolization therapy using iodized oil for small hepatocellular carcinomas. Correlation between lipiodol accumulation pattern and local recurrence. Acta Radiologica. 1994;35(6):576–80.
- Antoch G, Roelle G, Ladd SC, et al. Selective and sequential transarterial chemoembolization: survival in patients with hepatocellular carcinoma. Eur J Radiol. 2012;81(9):2290–7.
- Bouvier A, Ozenne V, Aube C, et al. Transarterial chemoembolisation: effect of selectivity on tolerance, tumour response and survival. Eur Radiol. 2011;21:1719–26.
- Daniels JR, Wallman M. Subselective intra-arterial chemotherapy infusion in the treatment of hepatocellular carcinoma. Semin Oncol. 2010;37(2):83–8.
- Ji SK, Cho YK, Ahn YS, et al. Multivariate analysis of the predictors of survival for patients with hepatocellular carcinoma undergoing transarterial chemoembolization: focusing on superselective chemoembolization. Korean J Radiol. 2008;9(6):534–40.
- Sze DY, Razavi MK, So SK, et al. Impact of multidetector CT hepatic arteriography on the planning of chemoembolization treatment of hepatocellular carcinoma. Am J Roentgenol. 2001;177(6):1339–45.
- Wallace MJ. C-arm computed tomography for guiding hepatic vascular interventions. Tech Vasc Interv Radiol. 2007;10(1):79–86.
- Iwazawa J, Ohue S, Hashimoto N, et al. Detection of hepatocellular carcinoma: comparison of angiographic C-arm CT and MDCT. Am J Roentgenol. 2010;195(4):882–7.
- Miyayama S, Matsui O, Yamashiro M, et al. Ultraselective transcatheter arterial chemoembolization with a 2-f tip microcatheter for small hepatocellular carcinomas: relationship between local tumor recurrence and visualization of the portal vein with iodized oil. J Vasc Interv Radiol. 2007;18(3):365–76.
- Iwazawa J, Ohue S, Mitani T, et al. Identifying feeding arteries during TACE of hepatic tumors: comparison of C-arm CT and digital subtraction angiography. Am J Roentgenol. 2009 192(4):1057–63.
- Choi WS, Kim HC, Hur S, et al. Role of C-arm CT in identifying caudate arteries supplying hepatocellular carcinoma. J Vasc Interv Radiol 2014;25(9):1380–8.
- Kim HC, Chung JW, Park JH, et al. Transcatheter arterial chemoembolization for hepatocellular carcinoma: prospective assessment of the right inferior phrenic artery with C-arm CT. J Vasc Interv Radiol. 2009;20(7):888–95.
- Kakeda S, Korogi Y, Hatakeyama Y, et al. The usefulness of three-dimensional angiography with a flat panel detector of direct conversion type in a transcatheter arterial chemoembolization procedure for hepatocellular carcinoma: initial experience. Cardiovasc Interv Radiol. 2008;31(2):281–8.
- Kakeda S, Korogi Y, Ohnari N, et al. Usefulness of cone-beam volume CT with flat panel detectors in conjunction with catheter angiography for transcatheter arterial embolization. J Vasc Interv Radiol. 2007;18(12):1508–16.
- Jeon UB, Lee JW, Choo KS, et al. Iodized oil uptake assessment with cone-beam CT in chemoembolization of small hepatocellular carcinomas. World J Gastroenterol. 2009;15(46):5833–7.
- Chen R, Geschwind JF, Wang Z, et al. Quantitative assessment of lipiodol deposition after chemoembolization: comparison between cone-beam CT and multidetector CT. J Vasc Interv Radiol. 2013;24(12):1837–44.
- Iwazawa J, Ohue S, Kitayama T, et al. C-arm CT for assessing initial failure of iodized oil accumulation in chemoembolization of hepatocellular carcinoma. Am J Roentgenol. 2011;197(2):W337–42.
- Iwazawa J, Ohue S, Hashimoto N, et al. Survival after C-arm CT-assisted chemoembolization of unresectable hepatocellular carcinoma. Eur J Radiol. 2012;81(12):3985–92.
- Miyayama S, Yamashiro M, Hashimoto M, et al. Comparison of local control in transcatheter arterial chemoembolization of hepatocellular carcinoma ≤6 cm with or without intraprocedural monitoring of the embolized area using cone-beam computed tomography. Cardiovasc Interv Radiol. 2014;37(2):388–95.
- Suk Oh J, Jong Chun H, Gil Choi B, et al. Transarterial chemoembolization with drug-eluting beads in hepatocellular carcinoma: usefulness of contrast saturation features on cone-beam computed tomography imaging for predicting short-term tumor response. J Vasc Interv Radiol. 2013;24(4):483–9.
- Kothary N, Abdelmaksoud MH, Tognolini A, et al. Imaging guidance with C-arm CT: prospective evaluation of its impact on patient radiation exposure during transhepatic arterial chemoembolization. J Vasc Interv Radiol. 2011;22(11):1535–43.
- Iwazawa J, Ohue S, Hashimoto N, et al. Accuracy of software-assisted detection of tumour feeders in transcatheter hepatic chemoembolization using three target definition protocols. Clin Radiol. 2014;69(2):145–50.
- Miyayama S, Yamashiro M, Ikuno M, et al. Ultraselective transcatheter arterial chemoembolization for small hepatocellular carcinoma guided by automated tumor-feeders detection software: technical success and short-term tumor response. Abdom Imaging. 2014;39(3):645–56.
- Zhuang ZG, Zhang XB, Han JF, et al. Hepatic blood volume imaging with the use of flat-detector CT perfusion in the angiography suite: comparison with results of conventional multislice CT perfusion. J Vasc Interv Radiol. 2014;25(5):739–46.
- Loffroy R, Lin M, Yenokyan G, et al. Intraprocedural C-arm dual-phase cone-beam CT: can it be used to predict short- term response to TACE with drug-eluting beads in patients with hepatocellular carcinoma? Radiology. 2013;266(2): 636–48.
- Louie JD, Kothary N, Kuo WT, et al. Incorporating cone-beam CT into the treatment planning for yttrium-90 radioembolization. J Vasc Interv Radiol. 2009;20(5):606–13.
- Abdelmaksoud MHK, Louie JD, Kothary N, et al. Consolidation of hepatic arterial inflow by embolization of variant hepatic arteries in preparation for yttrium-90 radioembolization. J Vasc Interv Radiol. 2011;22(10):1364–71 e1.
- Sze DY, Strobel N, Fahrig R, et al. Transjugular intrahepatic portosystemic shunt creation in a polycystic liver facilitated by hybrid cross-sectional/angiographic imaging. J Vasc Interv Radiol. 2006;17(4):711–5.
- Kapoor BS, Esparaz A, Levitin A, et al. Nonvascular and portal vein applications of cone-beam computed tomography: current status. Tech Vasc Interv Radiol. 2013;16(3): 150–60.
- Koizumi J, Hashimoto T, Myojin K, et al. C-arm CT-guided foam sclerotherapy for the treatment of gastric varices. J Vasc Interv Radiol. 2010;21(10):1583–7.
- Hirota S, Matsumoto S, Tomita M, et al. Retrograde transvenous obliteration of gastric varices. Radiology. 1999;211(2):349–56.
- Koizumi J, Hashimoto T, Myojin K, et al. Balloon-occluded retrograde transvenous obliteration of gastric varices: use of CT-guided foam sclerotherapy to optimize technique. Am J Roentgenol. 2012;199(1):200–7.
- Dijkstra ML, Eagleton MJ, Greenberg RK, et al. Intraoperative C-arm cone-beam computed tomography in fenestrated/branched aortic endografting. J Vasc Surg. 2011;53(3): 583–90.
- Biasi L, Ali T, Hinchliffe R, et al. Intraoperative DynaCT detection and immediate correction of a type Ia endoleak following endovascular repair of abdominal aortic aneurysm. Cardiovasc Interv Radiol. 2009;32(3):535–8.
- Plank C, Wolf F, Langenberger H, et al. Adrenal venous sampling using Dyna-CT—a practical guide. Eur J Radiol. 2012 81(9):2304–7.
- Bozlar U, Edmunds JS, Turba UC, et al. Three-dimensional rotational angiography of the inferior vena cava as an adjunct to inferior vena cava filter retrieval. Cardiovasc Interv Radiol. 2009;32(1):86–92.
- Aadland TD, Thielen KR, Kaufmann TJ, et al. 3D C-arm conebeam CT angiography as an adjunct in the precise anatomic characterization of spinal dural arteriovenous fistulas. Am J Neuroradiol. 2010;31(3):476–80.
- Nanashima A, Abo T, Sakamoto I, et al. Three-dimensional cholangiography applying C-arm computed tomography in bile duct carcinoma: a new radiological technique. Hepato-gastroenterology. 2009;56(91-92):615–8.
- Jin KN, Park CM, Goo JM, et al. Initial experience of percutaneous transthoracic needle biopsy of lung nodules using C-arm cone-beam CT systems. Eur Radiol. 2010;20(9): 2108–15.
- Choi MJ, Kim Y, Hong YS, et al. Transthoracic needle biopsy using a C-arm cone-beam CT system: diagnostic accuracy and safety. Br J Radiol. 2012;85(1014):e182–7.
- Cazzato RL, Battistuzzi JB, Catena V, et al. Cone-beam computed tomography (CBCT) versus CT in lung ablation procedure: which is faster? Cardiovasc Intervent Radiol. 2015;38:1231–6.
- Cazzato RL, Buy X, Alberti N, et al. Flat-panel cone-beam CT-guided radiofrequency ablation of very small (≤ 1.5 cm) liver tumors: technical note on a preliminary experience. Cardiovasc Interv Radiol. 2015;38(1):206–12.
- Abdel-Rehim M, Ronot M, Sibert A, et al. Assessment of liver ablation using cone beam computed tomography. World J Gastroenterol. 2015;21(2):517–24.
- Ierardi AM, Petrillo M, Xhepa G, et al. Cone beam computed tomography images fusion in predicting lung ablation volumes: a feasibility study. Acta Radio. 2016;57(2):188–96.
- Al-Ahmad A, Wigstrom L, Sandner-Porkristl D, et al. Time-resolved three-dimensional imaging of the left atrium and pulmonary veins in the interventional suite—a comparison between multisweep gated rotational three-dimensional reconstructed fluoroscopy and multislice computed tomography. Heart Rhythm. 2008;5(4):513–9.
- Miller DL, Balter S, Wagner LK, et al. Quality improvement guidelines for recording patient radiation dose in the medical record. J Vasc Interv Radiol. 2009;20(7 Suppl): S200–7.
- Loomba RS, Rios R, Buelow M, et al. Comparison of contrast volume, radiation dose, fluoroscopy time, and procedure time in previously published studies of rotational versus conventional coronary angiography. Am J Cardiol. 2015 116(1):43–9.