Diagnostics & Imaging Week National Editor

Looking for, and finding, the proverbial needle in the equally proverbial haystack is the metaphorically most difficult search a human can undertake.

But certainly just ais difficult, and with assuredly much greater risk/reward, is finding a metal object – or any other unwanted bit of material – in living tissue.

Hoping to make that task highly accurate, and extremely simple, is the thrust of new work by bioengineering researchers at the Pratt School of Engineering at Duke University (Durham, North Carolina). They have developed a laboratory robotic system that can successfully locate tiny pieces of metal within flesh and guide a needle to its exact location. Importantly, the system enables this to be done automatically – meaning, without human assistance.

In proof-of-feasibility experiments, the researchers combined three technologies, 3-D ultrasound, a robotic probe system and advanced image processing, to do pinpoint location of a piece of metal in simulated human flesh. This could make it possible for someone other than a surgeon or highly trained technician – such as generic Homo sapiens on the battlefield to do the extraction.

Ned Light, PhD, a member of the Pratt research team, told Diagnostics & Imaging Week that the development of advanced ultrasound was the primary contribution to this tech threesome by the Duke University Ultrasound Transducer Group, in the lab of Stephen Smith, PhD.

Advances in ultrasound technology have made these latest experiments possible, the researchers said, by generating detailed, 3-D moving images in real-time.

Light noted that the Duke team has a robust track record of modifying traditional 2-D ultrasound, like that used to image babies in utero, into advanced 3-D scans. Since inventing the technique in 1991, the team has shown its utility by developing specialized catheters and endoscopes for real-time imaging of blood vessels in the heart and brain.

Now, demonstration of the bit-of-metal location, Light said, could ultimately make it possible for a soldier to use a robot to find and treat shrapnel injuries on the battlefield.

Other medical applications, the researchers say, could be in the placement and removal of radioactive seeds used to treat prostate cancer, or other cancers, or the removal of metal bits from the even more delicate tissues of an eyeball.

Light told MDD that the team first used a basic tabletop robot and provided it with 3-D ultrasound technology developed by the lab, thus providing the robot with "eyes." Next, the system was given a "brain," in the form of an artificial intelligence (AI) program that translated the real-time 3-D ultrasound images into the necessary instructions to the robot.

That instruction was for the robot's probe to find the piece of metal, accomplished by adding to the probe an electromagnet, the magnetism causing the metal bit to vibrate slightly.

Light supervised the work of bioengineering undergrad student A.J. Rogers – who has now moved on to medical school – and Rogers explained: "Once the shrapnel's coordinates were established by the computer, it successfully guided a needle to the site of the shrapnel.

"The movement caused by the electromagnet on the shrapnel was not visible to the human eye," he said. "However, on the 3-D color Doppler system, the moving shrapnel stood out plainly as bright red."

To extract a piece of metal, the needle probe could simply be replaced with a tiny tool, such as a grabber, the researchers say, and the AI system could make it possible for anyone to use such a system to perform the extraction.

The initial work with the system was done to locate tiny bits of metal in a water bath, approximating the watery composition of the human body.

To prove that this could be done within the body – in the dark, you might say – the researchers then demonstrated the same ability in what Light described as an opaque agar/polymeric "slurry," a basic material used to mimic human tissue.

The initial feasibility demonstrations were done some weeks ago using a robot probe that could move in the three basic directions. More recently the team has improved the flexibility of the system by using a probe that can be manipulated along seven axes, "the way a human arm would move," according to Light.

Like much basic research, there is a good long way to go before the system is small enough and flexible enough to reach the clinic, let alone the battlefield, he acknowledged.

He said that ultrasound systems are already often found in handheld format, and the artificial intelligence system required could be incorporated into that modality. But while he noted the existence of many "where's Waldo?" surgical robots, he said the robotic portion of the threesome will have to be much smaller to be sufficiently portable.

"We showed that in principle, the system works," Smith said. "It can be very difficult using conventional means to detect small pieces of shrapnel, especially in the field. The military has an extensive program of exploring the use of surgical robots in the field, and this advance could play a role."

"That's a big project, but a lot of things are pushing our healthcare in that direction," Light said. And he noted that such a system also could be developed that would be usable and cheaper "for people and places in the world where it's hard to find doctors."

A paper by Light, Smith and John Angle, describing the laboratory's work in guiding interventional devices with real-time 3-D ultrasound, can be accessed at www.duke.edu/~edl/.