Cone-Beam Computed Tomography for Interventional Oncology: an Interview With Nishita Kothary, MD

At the 2015 Annual Scientific Meeting of the Society of Interventional Radiology (SIR), SIR member Nishita Kothary, MD, FSIR, an interventional radiologist and associate professor of general radiology at Stanford University Medical Center, presented a hands-on workshop and categorical course on the use of cone-beam computed tomography (CBCT) in interventional oncology applications with Michael J. Wallace, MD, FSIR, interim chair, interventional radiology, the University of Texas MD Anderson Cancer Center, Houston. Stanford University was one of the initial sites in the United States to begin using CBCT, and Dr. Kothary has been using CBCT in her clinical practice since 2007 at Stanford. Interventional Oncology 360 asked Dr. Kothary to share some details about CBCT in the IO setting.

IO360: Could you describe some challenges that you face when you’re trying to get good cone-beam CT images?

Kothary: One of the shortcomings or limitations of cone-beam CT is respiratory artifact. For liver applications, most vendors have an approximately 10-second sweep during which patients have to hold their breath, and sometimes that is difficult for patients who have underlying respiratory conditions, speak a different language, or can’t hold their breath for other reasons. So in about 5% to 10% of cases, patients can’t hold their breath, and unlike digital subtraction angiography (DSA), those images presently cannot be corrected. 

IO360: What are some ways that you can try to get around those limitations?

Kothary: One of the first things that we do is practice the breath-hold with the patient before actually acquiring the CBCT acquisition. I’ll talk to the patients and explain that the whole machine is going to go around them and it won’t touch them, so that they won’t be surprised when the machine does do its sweep. And, because during contrast-enhanced acquisitions there are a few initial seconds (2-4 seconds) during which contrast is being injected but images are not being acquired, I will let the patients breathe through that phase, limiting the breath hold to approximately 6 seconds of the 10-second acquisition. But this requires well-coordinated effort between the technologist and the physician, to time it perfectly. With our frequent use of CBCT, our technologists are very comfortable with this, but there was a learning curve, as with any new technology. 

The other thing that we do in the interest of saving time is set the room up so that we can do a CBCT at the drop of a hat. We have what I call the “arm rests,” which are pads mounted on a C-shaped contraption. The patients rest their forearms on this, so that the arms are out of the way, similar to helical CT. The blood pressure cuff is on the lower extremity when possible, and the IV pole is at the head of the table, so that the tubing does not get caught in the sweep. Finally, all EKG leads are translucent to avoid beam hardening from the metal. A little preparation prior to starting the case goes a long way in efficiency. With this, if I do want to do a cone-beam CT, all we have to do is get the arms up from the sides and we’ll be ready to go and nothing gets tangled or pulled. So that’s how we make life a little easier on patients. 

IO360: And what comprises the CBCT device?

Kothary: Cone-beam CT is mounted on your C-arm, so it’s a software program essentially. It reconstructs 3-dimensional images from a 2-dimensional angiogram — from a 2-dimensional rotation. So it’s not really a separate device, but one that is made available or integrated onto your fluoroscopy/angio suite. Most major vendors that are selling high-end angiography units will have this already integrated into the machine. The reconstruction algorithm is software that is loaded on the machine. Some of the newer technologies, such as perfusion imaging (blood volume) can be purchased as an additional software license.  

IO360: Are there areas of the body where this works best and other areas where it does not work well?

Kothary: For the most part it can work at any part of the body. It has been used in interventional neuroradiology a lot because imaging the brain is not impacted by respiratory or cardiac motion and as long as the patient doesn’t swallow or move, the images are of good quality. As interventional radiologists, liver intervention is where we, as an IO community, have good experience with CBCT. Cone-beam CT in the liver is often used for problem solving, providing a 3D image of the hepatic and extrahepatic vasculature, identifying small tumors, and confirming precise targeting. Outside of the liver, there are good data on the use of CBCT for arterial interventions such as endovascular repair of aneurysms, as well as in venous interventions and nonvascular applications such as biopsies. I’ve occasionally used it for biliary work but by far the most widely used non-IR application is locoregional therapy in the liver. 

IO360: Are there methods that you have learned over the years for how to reduce contrast dose?

Kothary: For a common or proper hepatic artery CBCT acquisition, the contrast volume required is about 20mL of iodinated contrast. At Stanford we use half strength contrast, i.e. diluted down to 50% strength by adding equal volume of saline, however this is not practiced universally or at other centers. However, even with full strength contrast, one CBCT acquisition can replace orthogonal views required by standard DSA imaging. 

The workflow in our practice is to start with a CBCT acquisition, which gives us a rotational image of all the hepatic vessels as well as identification of the tumors. Once we have the lay of the land, we can map out the arteries feeding the tumors and lay the origins of these arteries, which is often a challenge in cirrhotic livers, where the arteries are corkscrew, the liver is shrunken, and all the vessels are overlapping each other. If needed, we will do a single DSA angiogram, in the best projection determined by the CBCT rotation, which serves as a roadmap. After that, I prefer to do hand injections with “fluoro-saves” to limit contrast and radiation dose. Once I get to my super-selective position, I do one more injection, a DSA angiogram to essentially verify my position, and that’s it. 

Eliminating most of the DSA imaging that is often done while doing a chemoembolization, for example, not only saves contrast dose but in the overall picture reduces the radiation dose too. One of the biggest fears physicians and patients have is the “additional” radiation dose from CBCT. But by replacing multiple DSAs, the cumulative dose to the patients is actually reduced by CBCT. In fact, the risk of radiation injury from the cumulative dose, which is manifested as skin redness and thermal burns, is lower with CBCT because the dose is distributed to the body over 200 degrees, as opposed to DSA, which is concentrated on a small area of the body, making it more detrimental. 

The stochastic risk, i.e., the risk of genetic mutations and development of radiation-induced cancers that can happen a decade or two later, is independent of a threshold dose, and is hence harder to determine. However, the benefit offered by CBCT when treating a patient who already has a cancer, such as HCC, often outweighs the risk of stochastic injury. As with everything else, all use of radiation-emitting devices should be judicious.

IO360: In other facilities, are physicians hesitant to use cone-beam CT because of contrast dose?

Kothary: There are three big factors that I think limit the use of CBCT and I think these concerns are fast becoming obsolete. The first is the learning curve. Compared to the number of steps needed to get a CBCT in 2007, all vendors have made acquisitions and reconstructions simpler and faster, so learning a difficult process is not much of a hurdle now. We started using this in 2006 to 2007 for body interventions, and at that time, the set-up, image reconstruction, and the overall process added time and delay to the actual procedure time. At that time, we didn’t have an efficient workflow in place, nor were the technologists or the physicians particularly adept in using CBCT to their advantage. 

Most importantly, we were unaware of how much CBCT improves patient care and technical completeness of the administered treatments, such as chemoembolization and radioembolization. But with all technologies, there is a learning curve, and in 2008 we published a “how-to” paper, which really addressed the practicalities of getting a good CBCT acquisition. Since then, the process has been reduced to seconds. 

Once you get used to how to manipulate the images in the right way, you’ve got your entire roadmap in less time than it takes to do DSA acquisitions and work through them. Secondly, there are now robust data on the value of CBCT in liver interventions as well as a host of other interventions. What used to be an academic argument is pretty much mainstream. So there are data supporting the use of CBCT as a complementary tool. And finally, although all angiographic units are expensive, this technology is now considered standard for most major vendors, so availability is no longer an issue. I believe if a technology is made available, people will use it.  

IO360: What are some tips that you have for manipulating the images?

Kothary: Every vendor has a different software algorithm. I can speak for Siemens because that’s the system that we have. There’s a 3-D work tab and an in-space work tab, and once you get your images in the right grayscale, you can run your axial, sagittal, and coronal images in sync, see where the tumor is in all planes, and sort out the feeding arteries. With the newer software, most vendors also have an automatic planning tool – the physician circumscribes the tumor and marks a spot where his or her catheter is, and the automated software will highlight all the arteries supplying the tumor. It’s essentially a GPS. It tells you, you’re at point A and here’s how you get to point B. You really just have to follow the directions. I am old-school and will still scroll through the images and get a mental image of all the arteries I need to treat sub-selectively, but that’s an option. 

Based on the angulation that best lays out the culprit vessel(s), I will then either do a hand run or an image save, and use it as my roadmap. For a chemoembolization, following selective catheterization and treatment, we all do an unenhanced CBCT to determine if there is circumferential deposition of the chemoembolic emulsion (or the drug-eluting embolics suspended in iodinated contrast) in the tumor. If there is a “bite” of the tumor that does not show retained chemoembolic emulsion, it triggers us to interrogate other arteries that could be supplying that part of the tumor, ensuring technical completeness of treating the entire tumor.

IO360: Any other tips or information that you think is important for interventional oncology clinicians to know?

Kothary: I think the data are pretty compelling on the advantage of CBCT for liver-directed therapies. Further, although not as robust as its use in liver-directed therapies, the use of CBCT as a problem-solving tool for other interventions like endovascular aneurysm repair, is also established. For IO, I think we are now at the second stage of optimizing the use of CBCT and using it as a prognostic tool also. Newer applications like measuring the blood volume to a tumor before and after embolization, determining the dose of radioembolic particles to be administered based on segmentation using CBCT, and multiphase CBCT are all being explored. We are in the infantile stage of these applications, so it’s hard to predict what the data will show. But I think the value of CBCT as a tool in our arsenal is now established. CBCT is one more example of how interventional radiologists are continuing to pioneer minimally invasive medicine.

__________________________

Editor’s note: Disclosure: Dr. Kothary has completed and returned the ICMJE Form for Disclosure of Potential Conflicts of Interest. She reports unpaid consultancy to Siemens Healthcare. 

Suggested citation: Ford J. Cone-beam computed tomography for interventional oncology: an interview with Nishita Kothary, MD. Intervent Oncol 360. 2015;3(7):E78-E82.