X

Stanford unveils high-res 'micro-endoscope' thin as hair

The ultrathin, single-fiber endoscope boasts four times the resolution of existing designs and could result in minimally invasive surgeries for studying the brain, detecting cancer early, and more.

Elizabeth Armstrong Moore
Elizabeth Armstrong Moore is based in Portland, Oregon, and has written for Wired, The Christian Science Monitor, and public radio. Her semi-obscure hobbies include climbing, billiards, board games that take up a lot of space, and piano.
Elizabeth Armstrong Moore
3 min read
Professor Joseph Kahn (right) and graduate students Reza Nasiri Mahalati (left) and Ruo Yu Gu (center) with their prototype single-fiber micro-endoscope. John Todd

An electrical engineer at Stanford has led the effort to develop an endoscope that is not only as thin as a human hair but also boasts a resolution four times better than existing ultrathin models. The "micro-endoscope" could lead to far less invasive bio-imaging, making it easier and safer to peer inside living organs and tissue to, say, study the brain and detect cancer.

The researchers, led by Joseph Kahn at the Stanford School of Engineering, report in the journal Optics Express and the Optical Society of America's Spotlight in Optics that the prototype can resolve objects at just 2.5 microns in size. (A micron is one one-thousandth of a millimeter.) To put that into context, the naked eye can see objects as small as about 125 microns, and current high-res endoscopes can resolve objects down to about 10 microns, making the Stanford prototype a fourfold improvement.

As exciting as the development is, Kahn and his team still have their work cut out for them. While they've managed to achieve very good resolution using a single-fiber device, bending a multimode fiber scrambles the image. The current workaround is to house the fiber in a thin needle to maintain rigidity, but Kahn -- best known for his work in fiber-optic communications -- hopes to create a flexible single-fiber endoscope. (He admits that while his rigid single-fiber micro-endoscope could vastly improve our ability to image living organisms, as a physicist and engineer he is most driven by the technical challenges at play.)

Kahn first began working on endoscopes two years ago after talking about bio-photonics -- using light-based technologies to study biological systems -- with colleague Olav Solgaard, who wondered if it was possible to send light through a single fiber to create a bright spot within a living body and scan that to record images.

This is where multimode fibers come into play, where light travels via multiple paths, or modes. The engineers were well aware that light can convey complex information through multimode fibers, such as computer data, but that it can be scrambled as it travels.

So Kahn figured out how to essentially undo that scrambling of info with a spatial light modulator and an adaptive algorithm. (He'd used similar methods to set world records for transmission speeds by unscrambling computer data traveling through multimode fibers.) The team then looped in the work of another Stanford electrical engineer, John Pauly, who'd used random sampling to speed image recording in MRIs, and used random patterns of light to speed up imaging through a multimode fiber.

"And that was it," Kahn said in a school news release. "We were on our way. The record-setting micro-endoscope was born."

And yet the team encountered a puzzle. The members were able to light an object using the spatial light modulator to project random light patterns through the fiber into the body, and the light reflecting off the object traveled back through the fiber to a computer that measured the reflected power of the light and used algorithms to reconstruct an image. Their stunning, fourfold improvement in resolution, however, "meant that, somehow, we were capturing more information than the laws of physics told us could pass through the fiber," Kahn said. "It seemed impossible."

After several weeks, they realized what was happening. The answer was in the unusual algorithm they were using to reconstruct the image. "Previous research had overlooked the mixing," Kahn said. The random intensity patterns mix the modes, increasing the number of modes fourfold and thereby producing four times as much detail.

Kahn is already on to the next challenge: a flexible version. "No one knows if a flexible single-fiber endoscope is even possible, but we're going to try," Kahn said.