SPIE Digital Library Get updates from SPIE Newsroom
  • Newsroom Home
  • Astronomy
  • Biomedical Optics & Medical Imaging
  • Defense & Security
  • Electronic Imaging & Signal Processing
  • Illumination & Displays
  • Lasers & Sources
  • Micro/Nano Lithography
  • Nanotechnology
  • Optical Design & Engineering
  • Optoelectronics & Communications
  • Remote Sensing
  • Sensing & Measurement
  • Solar & Alternative Energy
  • Sign up for Newsroom E-Alerts
  • Information for:
SPIE Defense + Commercial Sensing 2017 | Register Today

OPIE 2017

OPIC 2017




Print PageEmail PageView PDF

Biomedical Optics & Medical Imaging

Novel hyperspectral imager aids surgeons

Illuminating bodily tissues with a digital micromirror device enables non-invasive characterization of living tissue based on chemistry and morphology.
31 December 2008, SPIE Newsroom. DOI: 10.1117/2.1200812.1394

When performing open or endoscopic surgery, it is often difficult to differentiate between neighboring tissues. For example, when removing the gallbladder, it is important not to damage the common bile duct. If we could non-invasively distinguish the bile duct from surrounding arteries, the surgeon would know better where to cut.

Spectroscopy has been used for decades to characterize chemical and biological molecules based on their spectral signatures, that is, the way they reflect or absorb different wavelengths of light. It is well-documented that oxygenated tissue reflects different wavelengths of light at different intensities than deoxygenated tissue.1 In the same way, gallbladder and bile duct tissue has a different spectral signature than surrounding anatomical structures such as the liver and blood vessels.2 In hyperspectral imaging, we capture a series of images while scanning through wavelengths of light. Each processed image pixel corresponds to the spectrum for that point on the image. We then compare each spectrum to known spectral signatures to determine which tissue they match, or their level of oxygenation. Liquid crystal tunable filters (LCTF) have been used previously to scan through the wavelengths of light, generating only one processed image every 30s. We have developed a new hyperspectral imager which uses a programmable digital micromirror device (DMD) from Texas Instruments DLP® Products group to generate over three processed images per second.

Figure 1. DMD-based hyperspectral imager for medical imaging.

The primary limitation of the previous LCTF system is that only single bandpasses of narrow bandwidth light can be filtered at a time.3 The new system has a DLP DMD-based spectral illumination light source (OL 490, Optronic Labs).4 In the DMD source, broadband light is diffracted through a slit, reflected from a grid of digitally-controlled programmable micromirrors, and optically directed to a focal plane array (FPA, or digital camera) through a liquid light guide (LLG), similar to an optical fiber that has an increased capacity for transmitting light. Figure 1 shows the experimental setup. By controlling the state of each mirror individually, the source illuminates tissue with a programmed selection of wavelengths of light and the camera measures the reflection of that spectrum. The programmed spectrum may simply be a single bandpass—similar to the LCTF output—or a more complex combination of wavelengths that cannot be duplicated by the LCTF. Since the DMD hyperspectral imager uses mirrors instead of liquid crystals, the tuning time is faster and the wavelength range is broader. Using a micromirror array and DLP technology lets the user program the source illumination using predetermined complex spectra or color mixtures of light.

Figure 2 shows how the percentage of oxyhemoglobin in the blood can be detected. The pixel color illustrates the percentage of oxyhemoglobin. Red relates to high levels of oxyhemoglobin while yellow, green, and blue represent decreasing levels of oxyhemoglobin, respectively. The DMD hyperspectral imager is currently operated in two illumination modes to assess tissue oxygenation: ‘Sweep’ and ‘3shot’. Examples of both types of image are shown in Figure 3. In ‘Sweep’ mode, we scan 126 discrete wavelengths with 10nm bandwidth, which is similar to filtering broadband white light with our LCTF system. After acquiring the hyperspectral data (126 images), we separate out known spectral signatures of oxyhemoglobin (HbO2) and deoxyhemoglobin (deoxy-Hb) to color-code each pixel. One color-coded image is generated every 20s. Our ‘3shot’ method uses the DLP source to illuminate the tissue with color mixtures of light chosen to reveal differences between the known spectral signatures of HbO2 and deoxy-Hb.

Figure 2. An initial proof of principle of the hyperspectral imager's ability to show oxygenation. (left) A subject's index finger was temporarily bound with a rubber band to restrict blood flow. (right) When the rubber band is removed, a surge of blood returns to the finger causing hyperfusion.

The DMD hyperspectral imager can be used for real-time imaging in a wide variety of surgical and clinical applications. Because the light source is programmable and has a spectral range of 380–1600nm, any chemical or biological components with differing spectral signatures can be contrasted with this system. Currently, the system is used in animal studies during partial kidney removals to monitor the restricted blood supply in the kidney after arterial clamping (see Figure 3 and online video5). We expect to fine tune the system for use with standard medical equipment, such as endoscopes, so that it can be used in human surgery.

Figure 3. Raw and processed images from animal study of kidney clamping. ‘Sweep’ and ‘3shot’ images are color-coded according to the color bar shown in Figure 2, where red indicates high oxygenation, green indicates medium oxygenation, and blue indicates low oxygenation.

We would like to thank the UT Southwestern Medical Center Animal Research Labs for allowing us to image their surgeries, AAVA Technology for helping with software integration, Optronic Laboratories for their superb technical support, and Texas Instruments for funding the project.

Karel Zuzak, Robert Francis
Laboratory of Biomedical Imaging, Department of Bioengineering
The University of Texas at Arlington
Arlington, TX
Jack Smith
DLP Products Group
Texas Instruments
Plano, TX
Chad Tracy, Jeffrey Cadeddu
Department of Urology
University of Texas Southwestern Medical Center
Dallas, TX
Edward Livingston
GI/Endocrine Surgery
University of Texas Southwestern Medical School
Dallas, TX