How Does A Low-Cost 3D Ultrasound Come into Being

    

 

Many great inventions are invented by accident. Is  the 3D ultrasound the same too?

 

Joshua Broder was using a Wii handset to bat a ping-pong ball back and forth when the idea struck. An emergency physician at Duke University Medical Center, he uses ultrasound to understand what’s happening inside a patient’s body, and treat wounds and illnesses. But the picture he gets, while rapid enough to operate in real time, is two dimensional and hard to parse.

 

“The controller in my hand is really an inexpensive thing,” he thought. “Why is it that expensive medical devices don’t use that kind of low cost technology?”

 

With some help from engineers at Duke and Stanford, Broder 3D printed a body for an ultrasound wand that’s meant to house accelerometers and gyroscopes similar to the ones found in phones or Wiimotes. These small devices, which have become ubiquitous and cheap thanks to the smartphone revolution, work together to determine the angle, position and orientation of your phone, so you can play games, keep the screen upright and use gestures. Attached to the wand of the ultrasound, which emits and receives the ultrasound like radar, the same sensors track its precise position. Then, as the images are taken, software uses that information to stitch them all together into a three-dimensional file. The output, while not approaching the image quality of an MRI or CT scan, is much easier to understand than a 2D ultrasound image, which can appear grainy and confusing.

 

The ultrasound machines Broder is building upon are different from the ones doctors use to image unborn fetuses. While those cart-sized machines do provide 3D images, they cost hundreds of thousands of dollars, and are not extremely portable. What Broder describes is a small, 3D-printed attachment for a $25,000, laptop-sized 2D ultrasound machine.

 

Point-of-care ultrasound, in which doctors use ultrasound during a physical exam to inform further care, is becoming more common—a market which P&S Market Research expects to grow at 7 percent per year until 2025—but it still remains an underutilized resource, says Chris Fox, director of instructional ultrasound at the University of California-Irvine. He teaches ultrasound techniques to doctors across a wide variety of specialties, from the emergency room to internal medicine, how to capture and read ultrasound images. “The quality of care simply improves when you can look through the patient’s skin at the organs you’re concerned about, right there at the point of care, and not have to wait for another test to come back,” says Fox.

 

An ultrasound view into the abdomen can tell a physician whether the patient is experiencing a bowel obstruction, a gallstone or a blocked kidney, for example. Shortness of breath can be attributed to pneumonia, fluid in the chest or fluid around the heart. In these ways, doctors can use ultrasound to determine whether a patient needs to be sent for further imaging or not. And they frequently use ultrasound to guide needle placement in laparoscopic surgery and other procedures that require the precise placement of implements, because it can show a real-time image of the needle entering the tissue.

 

But that’s where 2D ultrasound gets tricky; you can’t see much of the tissue and it’s hard to differentiate vasculature, nerves, muscle and bone. “All we’re seeing is a slice, and we have to decide right now, are we going to look at this in a longitudinal plane, or a transverse plane? That is confusing to have to commit to one of those two planes,” says Fox. A transverse view would show the needle coming toward the viewer, and a longitudinal view would show the needle entering from the side, but in these two dimensional planes it’s very hard to determine depth, and therefore whether the needle is positioned properly. “Three-dimensional ultrasound is so much easier to interpret that it really would remove this layer of insecurity I think a lot of doctors have, when it comes to trying to learn ultrasound.”

 

More simply put, 2D ultrasound is hard to use. “It’s hard for people who’ve never done ultrasound before to learn how to take pictures and interpret them,” says Broder. “We want this to be such an intuitive technology that many different medical personnel could use it immediately with almost no training.”

 

Presenting at the American College of Emergency Physicians research forum, Broder described what he sees as a primary function of the technology: brain imaging in young children. Kids under two years old have soft skulls, and ultrasound can see right in, and help diagnose hydrocephalus, where cerebrospinal fluid causes pressure in the brain. He used it to record an image of the brain of a 7-month-old child, while the baby sat peacefully in his mother’s lap. It required no radiation, like a CT scan, and the child didn’t have to be motionless or sedated, like an MRI. They simply drew the wand across the boy’s head, in a painting motion. In ten seconds it was done.

 

Open-source software called 3D Slicer renders the result on-screen with three axes and a slider that allows physicians to open up the image and view a cross section. Technically, it’s a stack of 2D images—up to 1,000 of them—laid next to each other, but the software can also estimate the volume of features within them, which is especially useful in diagnosing tumors.

 

“It’s just a much more dynamic dataset than when you take a still picture,” says Broder. “Think of the analogy of a photograph on your camera. Once you’ve taken the picture, you can play around with it, but if you didn’t like the angle that you took the picture from, you can’t fix it … when you’ve got a three-dimensional dataset, you really have a lot of control over what questions you want to ask and how you answer them.”

 

Even the more expensive ultrasound machines don’t offer the accuracy of CT or MRI imaging, nor can they image an entire body, but that’s not the point, says Broder. “We want to bring the cost in line,” he says. “We suffer in western medicine by doing a lot of things to perhaps to a greater degree of accuracy or precision than we need, and that drives the cost high. So what we want to do is exactly what the patient needs—provide the level of detail required for their best care.”

 

As point-of-care ultrasound use surges, Broder’s team isn’t the only one trying to improve the machines. Clear Guide ONE, built by doctors from Johns Hopkins, also uses a wand attachment, but employs a visual system to track needle insertion, though it’s restricted to that application. And, while it only offers two-dimensional ultrasound, a device called Clarius pairs wirelessly to a smartphone to sidestep the computer altogether and drive the price down below $10,000.

 

The small size and low cost of Broder’s device makes it useful in areas around the globe where it’s impossible or not cost effective to use the bigger machines. GE agreed, awarding Broder $200,000 in its inaugural Point of Care Ultrasound Research Challenge. As it is, the device is currently undergoing clinical trials, and Broder and his collaborators hold an international patent on it. In the future, Broder imagines pairing the device with an EKG to get real time imaging of heartbeats. If the data from the EKG is matched to the individual images taken by the ultrasound, you can sort the pictures based on when they occurred within the cardiac cycle. This “4D” imaging could give better pictures of the heart, as it compensates for the motion of the heart itself, as well as breathing.

 

“We can do a lot of the same things that the expensive 3D machines can do, but at a much lower cost,” says Broder. “We’re just at this incredible time where computing technologies have really facilitated what we’ve done.”