BIOMEDICAL IMAGING: OCT scanner requires no moving parts

Researchers at the University of Texas, Austin, have developed a method of optical coherence tomography (OCT) for two- and three-dimensional subsurface scanning that requires no moving parts.

Dec 1st, 1999
Th Acf3f2

Researchers at the University of Texas, Austin, have developed a method of optical coherence tomography (OCT) for two- and three-dimensional subsurface scanning that requires no moving parts. Prospective applications include medical imaging and semiconductor inspection.

The traditional tomographic process used in medical x-ray tomography takes a series of x-ray exposures (tomograms) of a stationary patient while varying the depth of focus of the x-ray machine, thereby obtaining images of a desired layer of the body while blurring out structures in other layers. Computed tomography (CT) scanning takes the process one step further by moving the x-ray camera in an arc around the patient and using a computer to reconstruct the x-ray data into a cross-sectional anatomical image.

The OCT technology—currently under development as an optical alternative for medical imaging and non-destructive-testing applications performed by CT, magnetic resonance, and ultrasound imaging—also relies on a tomographic process (see related story, p. 47). This effect is achieved in most OCT schemes by varying the optical path length of the reference arm in a Michelson interferometer structure and observing interference fringes when that path length falls within one coherence length of a reflective location within the material under study. The sample arm of the interferometer illuminates the material under study using a broadband light source. The detection mechanism is based on cross correlating the reference- and sample-arm signals.

Because OCT is a fiberoptic-based technology intended to deliver 2 to 10 µm of resolution from the distal ends of catheters, endoscopes, and laparoscopes, creative solutions have been developed to provide the scanning motion needed for tomography, according to Andrés Zuluaga at the university. These solutions include the use of piezoelectric transducers, rotational or translational fiber displacement, and radially firing probes. Zuluaga and coworker Rebecca Richards-Kortum have gotten around the need for mechanical displacement, however.1

Click here to enlarge image

FIGURE 1. Optical coherence tomography can be performed in a device with no moving parts using spectral inteferometry. In this scheme, a line on the sample surface is optically imaged onto the entrance slit of an imaging spectrograph coupled to an array detector. Depth information (z-axis) at the sample is encoded in the spectrograph output wavelength (l). Three-dimensional data can be obtained by imaging the sample through a two-dimensional fiber bundle.

The researchers accomplished this by passing the interferometer signal through an imaging spectrograph and recording the resulting fringe pattern on a charge-coupled-device (CCD) camera—effectively moving the process from the time domain to the frequency domain and taking the Fourier transform of the spectrum of the interferometer output. They also extended the technique from one to two dimensions by optically imaging a line on the sample, instead of just one point. Configuring a fiber bundle to image the sample surface in two dimensions could further extend the motionless OCT technique from two to three dimensions, according to the researchers (see Fig. 1).

Their experimental system provided about 40 µm of depth resolution and 13 µm of surface resolution in a sample with depth and transverse dimensions of 2.4 mm and 460 µm, respectively. The experimental device was illuminated by a 500-µW superluminescent diode source with a center wavelength and bandwidth of 855 nm and 25 nm full width at half maximum (FWHM), respectively.

Click here to enlarge image

FIGURE 2. Concept of OCT based on spectral interferometry was demonstrated on a sample consisting of microscope cover slips and a Ronchi grating.

To date the researchers have used their methods to "look at easy things" such as glass slides and semiconductors, Zuluaga said (see Fig. 2). Plans include increasing the brightness of the light source by a factor of 50, refining the technique to obtain "pretty pictures" of biological samples, and developing endoscopic delivery systems for the fiberoptic probe. The researchers have also demonstrated preliminary results based on optical as opposed to computer processing of the image data, but Zuluaga described the latter as a relatively long-term goal. "Application-wise, it's fine to have a computer," he said.

Hassaun Jones-Bey

REFERENCE

  1. 1. A. F. Zuluaga and R. Richards-Kortum, Opt. Lett. 24(8), 519 (1999).

More in Research