December 2016

Virtual Radiology: New Viewing Platform Creates Interactive Medical Imaging Holograms
By Beth W. Orenstein
Radiology Today
Vol. 17 No. 12 P. 14

What if physicians could interact with medical images the same way they do with patients lying on the operating room (OR) table? Thanks to advances in imaging technology, physicians are able to piece together multiple 2D images from CT or MRI and imagine patients' anatomy in three dimensions. "But they're forced to make assumptions about what the patient's anatomy truly looks like," says Sergio Aguirre, MSc, chief technology officer and founder of EchoPixel of Mountain View, California. Aguirre's company has developed software that enables radiologists, interventional and pediatric cardiologists, and other surgical specialists to actually see their patients' anatomy in open 3D space.

EchoPixel True 3D, powered by HP, works by using four cameras that track the user's head movements. Special lightweight glasses (rather than virtual reality headsets) allow users to turn images into 3D holographic visuals and, with a stylus, move and interact with the organs, tissues, and vessels they see on their workstations in real time. The detail is lifelike, and the images they see appear to float in space, Aguirre says. A start-up, EchoPixel developed the software in 2012; the True 3D viewer received FDA clearance in March 2015. An annual subscription for the EchoPixel software is approximately $22,000 to $25,000 a year.

"With traditional methods, you're using a mouse and viewing images on a flat monitor," says Judy Yee, MD, a professor and vice chair of radiology and biomedical imaging at UC San Francisco and chief of radiology at the San Francisco VA Medical Center. "With the EchoPixel True 3D platform, you're basically lifting that image out into a space where you can interact with it."

Gut Instinct
One of Yee's areas of interest is virtual colonoscopy, also known as CT colonography. She has used the EchoPixel True 3D viewer for about three years and finds that she prefers it for most of her cases. "It depends a bit on the case itself," she says, "but you could use it for every case"; she expects she might in the near future. Yee says the timing of its being brought to market is ideal for virtual colonoscopy.

In June 2016, the US Preventive Services Task Force (USPSTF) gave CT colonography an "A" rating as a screening tool; an "A" rating means that the benefit is substantial. As a result of the rating, "all private-payers who participate in the Affordable Care Act are mandated to include screening CT colonography as being reimbursable," Yee says. "That was a big step forward." The rating is also likely to mean that Medicare will expand its coverage policies to CT colonography, not only as a diagnostic test, but also as a screening test. "We're working on Medicare reimbursement," says Yee, who chairs the Colon Cancer Committee for the ACR.

In today's world, where everyone is used to touchscreens on their smartphones and tablets, this type of technology seems natural and more intuitive, Yee says. Using the EchoPixel True 3D platform requires some training, she adds, but once familiar with its features and operation, it's not difficult. In fact, Yee finds that because she's able to see large portions of the colon at one time with the True 3D viewer, it can reduce the time it takes to interpret the data sets.

Yee also believes that being able to manipulate the data set in this way makes it easier to find and better diagnose flat lesions in the patient's colon. Flat lesions can be more difficult to detect with other visualization platforms and can be a problematic area for those who perform virtual colonoscopies, Yee says. A study published in the Journal of the American Medical Association in March 2008 found that flat lesions within the lining of the colon and rectum may be more likely to be cancerous than polyps.

"But because we are able to combine 2D and 3D views with the EchoPixel platform, it allows you to more quickly assess the internal density of potential lesions and to do so more accurately," Yee says. "With the EchoPixel display, if you see any lesion, you can virtually dissect it out and look at it more closely."

Some sites performing virtual colonoscopy have their patients ingest oral contrast to tag residual material in the colon. "Then, using software, you can electronically subtract out the tagged material," Yee explains. "However, using contrast can produce artifacts. It has been found that even if you don't electronically subtract the tagged material, there is value from the contrast, especially on 2D views." The tagged material will highlight certain lesions, such as flat lesions, for the radiologist reading the images. "I think the use of tagging and electrical subtraction would be of additional value in the 2D and 3D images that you can get from the EchoPixel display," Yee says.

Less Trial and Error
Richard Kovach, MD, division director of interventional cardiology at Deborah Heart and Lung Center in Browns Mills, New Jersey, admits he is a "geek" when it comes to technology. When he saw images from the EchoPixel platform, he had to be one of the first to try it. "Because it's a relatively new technology, we're actively learning on it ourselves," he says.

An interventional cardiologist, Kovach places many balloons and stents. "But I fix everything artery-wise from head to toe," he says. He has found the EchoPixel to be helpful in planning and executing many of the interventional procedures he performs, especially structural heart procedures such as transcatheter aortic valve replacement (TAVR) and left atrial appendage (LAA) closure, as well as endovascular abdominal aortic aneurysm repair (EVAR).

Patients don't know that their physicians are rendering their images in 3D open space with the EchoPixel. The patients go for the same imaging procedures they would if their physicians were using standard imaging technology, Kovach notes. It's the post processing that's different, and that's done at the workstations, which have up to eight discrete processor cores, up to 128 GB of RAM, and multiple storage and peripheral component interchange configurations.

The EchoPixel display has taken a lot of the guesswork Kovach often has to make and "turned it into more precise measurements," he says; for example, stretched 2D imaging very often overestimates or underestimates distances in the anatomy that he will be intervening upon. Data from several studies demonstrate that 3D imaging or dynamic 3D imaging is much more accurate than stretched 2D imaging. "This takes it one step further," Kovach says. With EchoPixel, "when I'm planning a procedure, I not only get much more accurate measurements but also a [significantly] better understanding of what I'll be dealing with anatomically. It makes it so much easier to do my job."

Having a more precise picture of the patient's anatomy preprocedure also helps Kovach choose the proper devices for the surgical procedure before he begins. Most interventional devices are available in different lengths and diameters. "Anytime you can go into a case knowing exactly what equipment to prepare ahead of time is great," Kovach says. Currently, he might make his best guess based on standard imaging but find, when the patient is on the table, that X won't work and that he needs to use Y instead. "I might need to change during the procedure because of the different aspects of the patient's anatomy," he says.

With EchoPixel, Kovach feels that he has a much better idea of what he will encounter and can prepare accordingly. Reconstructing the patient's anatomy with EchoPixel "can save you time because you're using the right devices from the start and can cut down on equipment usage because you have less trial and error as you're doing the procedure." Choosing the right equipment at the start means less waste, which can translate into cost savings. For balloons and stents, the savings might not be as significant, but for some devices, such as TAVR valves or WATCHMAN, which is implanted to prevent harmful blood clots from escaping the heart's left atrial appendage, it can be, Kovach says.

It's also possible, Kovach says, that the EchoPixel display might allow him to see that the procedure he was planning won't work because of the angles he would have to navigate in the patient's anatomy. It's certainly better to know that than to prepare the patient for surgery and have to tell him, after leaving the OR, that the procedure couldn't be performed, Kovach says. However, he adds, the imaging system hasn't been in use long enough for anyone to show that it improves success rates. "We haven't done enough patients to show superiority," he says, but he believes users will do such studies and publish them in the future.

Modeling Potential
At Penn State College of Medicine, physicians are using the EchoPixel platform as a learning tool. Typically, first-year medical students learned anatomy in a four- to 10-week class they started as soon as they walked into the classroom, says Timothy J. Mosher, MD, a professor of radiology and orthopaedic surgery at the medical school in Hershey, Pennsylvania. Medical schools use cadavers to teach anatomy, even though cadavers don't last very long and are very expensive, he says.

"As a medical student, that was the last you saw of anatomy, unless you went into surgery or some other specialty that required an advanced knowledge of anatomy," Mosher says.

The professors at Penn State had long wanted to develop elective anatomy classes that could be offered to fourth-year students. About four years ago, when they were developing the curriculum for the electives, they learned about the EchoPixel and saw a perfect fit.

"One of the things that intrigued us about this technology is that we could show the students 2D images the way they would see it with traditional CT or MRI and then we could give them a projection using the EchoPixel that resembled what they would see if they were looking at a skeleton or a cadaver or the patient," Mosher says. "It would be very helpful in allowing students to make that relationship between imaging and a physical exam."

The Penn State professors are exploring the possibility of using the EchoPixel to develop a library of virtual cadavers showing different clinical diseases. "We have rich image archives, and they could be used to develop these virtual models that could be used to teach our students and trainees a variety of different anatomical conditions," Mosher says.

The students would be able to work on the virtual cadavers and learn how to handle cases. For example, Mosher says, the virtual cadavers could show complex fracture patterns that orthopedic residents might see and have to treat in the OR. He also sees the EchoPixel 3D models being helpful to teach residents stereotactic breast biopsies. "Now, it requires having 3D models in their heads while looking at 2D images," he says.

As an orthopedic radiologist, Mosher says the EchoPixel renderings could help him to examine the mechanics of the foot. "We have new technology where you can get a standing CT of the foot. If you take that anatomy and have a 3D model, you could use modeling software to determine [what would happen] if you applied forces to the foot and where the strains would be," he says. "You could look at the foot anatomy and have heat maps and develop a better understanding about the areas where you have to worry about fractures or rapid degenerative change."

Mosher thinks the EchoPixel platform could play a role in teaching functional neurologic imaging as well. "That could be a real interesting thing to explore with this," he says. The medical school has put the EchoPixel in its "technology sandbox" and has asked everyone to brainstorm: "Where do you see it adding value? There are a lot of options and different ways it could be applied," he says.

As with any new technology, Mosher says the EchoPixel has a learning curve. However, he believes radiologists can learn it quickly because it fits with their training. "As radiologists, we're more used to looking at 3D imaging," he says. "It does what you conceptually do in your head all the time after you've been in practice for awhile." It also could be, Mosher says, that radiologists will find fewer applications for it than some of their clinical colleagues "because [3D] tends to be the way we think, particularly those of us who deal with cross-sectional imaging."

— Beth W. Orenstein of Northampton, Pennsylvania, is a freelance medical writer and frequent contributor to Radiology Today.