SPIE Membership 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 Photonics West 2019 | Call for Papers

SPIE Defense + Commercial Sensing 2019 | Call for Papers

2018 SPIE Optics + Photonics | Register Today



Print PageEmail PageView PDF

Biomedical Optics & Medical Imaging

High resolution retinal images for better diagnosis

Long wavelength light permits in vivo imaging of the human eye at a cellular scale.

18 May 2009, SPIE Newsroom. DOI: 10.1117/2.1200905.1592

Eye diseases can often be detected at an early stage if cellular-scale images of the retina are available. However, it is not easy to obtain high-resolution images of photoreceptors—light-detecting cells—because of inherent flaws in the eye's own optics, so-called aberrations. The size of a photoreceptor is around a few microns, and consequently, the required resolution is less than 5μm. Present standard retinal imaging systems cannot provide such a high resolution.

‘Adaptive optics’ (AO) is based on a feedback loop comprising aberration measurement and dynamic wavefront correction to cancel the aberration of the eye. It has recently been introduced for a variety of high-resolution retinal imaging applications1–3 that have revealed microstructures of the retina. AO instruments have also shown a probe-wavelength-dependent contrast in the retinal microstructure.4 Until now, the retinal microstructure has been investigated with visible and near infrared light, up to 840nm, which is still sensible by the human eye despite not being in the visible (380–750nm) spectrum. We are working on high-resolution long-wavelength imaging to reveal more properties of the retina.

The 1μm wavelength we are using has several advantages. Not only are the eye's photoreceptors unaffected by this wavelength of light, but water, the principal component of the fluid in the eye, has a local minimum in its absorption of light at 1.05μm.5 In addition, longer wavelengths are less scattered by the retina, and they enable deeper penetration into the eye.6 Further, water has a relatively constant refraction index around 1μm, which minimizes wavelength-dependent blurring of images.

Figure 1. Schematic of adaptive-optics (AO) scanning laser ophthalmoscope with a probe wavelength of 1μm. The AO setup is made up of a deformable mirror (DM) and a Hartmann-Shack wavefront sensor (HSWS). (Inset) A vertically off-axis pair of spherical mirrors.

We have developed an AO scanning laser ophthalmoscope (AO-SLO) with a 1.04μm probe beam.7 The AO subsystem is a confocal microscope used to correct the retinal images obtained by the SLO subsystem for the eye aberration. An amplified spontaneous emission light source with a center wavelength of 1.04μm was used as a probe beam and introduced into an SLO based on reflection optics, as shown in Figure 1. The reflection setup reduces obstructive surface reflection from objects and wavelength-dependent aberrations. We use off-axis spherical mirrors to help reduce the size of the SLO so that the entire setup fits onto a 70cm×90cm compact optical bench. Although the tilted use of these mirrors generates a strong astigmatism (an asymmetric aberration), this is canceled by using a second pair which is tilted orthogonally to the others.

The AO subsystem uses a Hartmann-Shack wavefront sensor (HSWS) to measure the aberration and a deformable mirror (DM) to correct it. The HSWS consists of a standard silicon-based CCD camera and a 32×32 lenslet array. At 1.04μm, the sensitivity of this CCD is low, so we used an 840nm superluminescent diode to measure the aberration. The AO software determines and applies the optimal shape of the DM to compensate. It iterates the correction process with an 11Hz loop frequency, which results in dynamic correction of the eye aberration.

Figure 2. 300μm ×300μm patch of in vivohuman retina at 3 degree eccentricity measured without AO (a) and with AO (b). Retinal patches at 5 degree eccentricity with focus at the photoreceptor layer (c) and at superficial vessels (d).

Figure 3. A mosaic of retinal patches. The photoreceptors are clearly visualized up to an eccentricity of 10 degrees.

Using AO dramatically improves the resolution of the SLO, as shown in Figure 2(a) and (b). Individual photoreceptors are clearly visible as well as the improvement in resolution due to the AO correction. In addition, it is possible to focus on different layers of the eye's tissues by applying an additional defocus to the DM. This permits the investigation of not only the retina but also the choroid, a deeper part of the eye that cannot be viewed at shorter wavelengths. The focus is on the photoreceptor in Figure 2(c) and on superficial vessels in Figure 2(d). Fine, small vessels and individual blood cells in large vessels can easily be seen. Figure 3 shows a wide field image obtained as a mosaic of several retinal images and clearly shows the presence of photoreceptors at an eccentricity of 5–10 degrees.

To summarize, 1μm is a new and attractive band for retinal investigations. Another technique, named optical coherence tomography (OCT), has already begun to employ 1μm light in clinical studies to reveal eye diseases deep at the rear of the eye.8 We hope that the combination of 1μm AO and OCT may provide high-resolution images of the retina and the layer behind it, the choroid. Another attractive feature of using 1μm light is its invisibility, and we hope to introduce a 1μm aberrometer9 to replace the current 840nm one, which will enable high-resolution retinal imaging without any optical stimulation to the subject. This is expected to reveal more detailed physiology of the eye.

We would like to acknowledge technical support from Shuichi Makita and Daiki Tamada. This research is supported in part by Japan Society for Promotion of Science (contract KAKENHI 18360029).

Yoshiaki Yasuno, Kazuhiro Kurokawa
Computational Optics Group
University of Tsukuba
Tsukuba, Japan

Yoshiaki Yasuno received his PhD in 2001 for work on spatio-temporal optical computing from the University of Tsukuba, where he now leads the computational optics group. His main research interests are optical coherence tomography and ophthalmic imaging.

Kazuhiro Kurokawa is a graduate student at the University of Tsukuba. He has been working on AO-SLO in the computational optics group since 2007. In 2009, he received the Pascal Rol Award at SPIE Photonics West for the best paper in ophthalmic technologies.