Article Text


Colour imaging using a scanning laser ophthalmoscope
  1. Institute of Ophthalmology, University College London

    Statistics from

    New imaging techniques have opened up new possibilities for visualising the living human eye and can provide quantitative measurements of fundus features. One of the most important developments has been based on the scanning laser ophthalmoscope (SLO) which has the property of providing optical sections through the use of confocal optics.1 2 This can reveal three dimensional aspects of structures which would otherwise be invisible owing to the contrast degrading effects of overlying elements. The confocal SLO has had a major impact on research in ophthalmology and continues to break new ground in advancing the limits of seeing and measuring properties of the eye. The first widely used SLO was the Rodenstock device which provided facilities for confocal imaging, fluorescein and indocyanine green angiography, and psychophysical measurements such as microperimetry. These used a variety of lasers of different wavelengths to give real time video images. In comparison with photographic techniques, the spatial resolution of video images is inherently lower. This is because the horizontal scan lines which make up the image are limited to approximately 600 compared with the equivalent of 4000 or more for photographs. The mechanical limitations of the spinning polygon mirror mean that this limit will not be easily improved. Despite this limit the confocal optics of the SLO can result in improved resolution compared with photographs because it allows optical sectioning, and so the contrast degrading effects of structures above or below the depth of interest can be reduced. The depth resolution can be of the order of several hundred micrometres. In this way small vessels or other structural detail obscured by the scattering of neural tissue in the optic nerve can be seen and some of the contrast degrading effects of ocular media opacities can be bypassed. New opportunities continue to appear so that we now are able to see and measure the autofluorescence associated with lipofuscin3and quantify the structure of the lamina cribrosa4 using a Zeiss prototype laser scanning ophthalmoscope.

    Perhaps the most successful and widely used implementation of the SLO is the Heidelberg retina tomograph (HRT) which is aimed at quantifying the three dimensional shape of the optic nerve head using of a series of optical sections combined with the appropriate computer analysis.5 It is evident that this measurement ability is not possible using photographic methods. Although the presentation of the images makes very effective use of colour, this is not intended to be realistic and the use of a single laser makes true colour rendition impossible. The images used are of even lower lateral resolution than conventional video images, having only 256 × 256 picture elements (or pixels). This illustrates that, despite the lack of lateral resolution, the depth resolution along with the computer analysis allow the SLO to provide information about fundus structures which would otherwise not be available.

    Despite these recent advances, one of the limitations of SLO images has been the lack of realistic colour in the images. Important information such as pallor of the disc or subtle aspects of the colour appearance of drusen, diabetic retinopathy and other fundus features are lost with SLO imaging. Although the depth resolution and the resistance to contrast degrading effects provides an advantage over conventional imaging, the lack of colour has been a severe handicap. This has now been overcome, as shown in the paper by Manivannan and colleagues in this issue (p 342) which presents the first realistic colour images using the SLO.

    In principle, the method is straightforward. The combination of three lasers at appropriate wavelengths to form a colour image is analogous to the formation of colour television images using red, green, and blue phosphors. The use of multiple lasers in an SLO to provide colour information has been recognised before.6 7 However, in practice, achieving realistic colour rendition with an SLO requires careful selection of the appropriate wavelengths for the lasers and overcoming difficult problems in image registration which could distort the image.

    This is the first time that such realistic three colour images have been produced and the results are impressive. In comparing the colour SLO images with the photographic images the clarity and colour appearance is striking, particularly considering the difference in lateral resolution. In some cases additional detail may be visible in the SLO images and further experience using depth resolved colour imaging will elucidate the role of this new technique.

    Clearly there are some limitations and differences in comparison with colour photography, including the inherent difference in lateral resolution common to all SLOs. In addition, the wavelengths of the lasers used would be expected to have an effect. Each laser emits at a single wavelength while the photographic method in combination with the white light flash gives a broad spectral response. As a consequence some differences could arise in comparing subtle colour rendition. In particular, with yellowing of the lens, the short wavelength blue laser could contribute a reduced blue component in comparison with the photographic method. Whether these are significant effects remains to be seen. The advantages of confocal optics of the SLO may outweigh any colour differences.

    The use of realistic colour information certainly improves our appreciation of the SLO images. Whether the advantages outweigh the extra cost of an SLO compared with other systems is not known. It may be that the role of the SLO is primarily in quantification of images and in this the use of three simultaneous wavelengths could provide an advantage. For example, the qualitative description of “disc pallor” is amenable to exact measurement with the use of multiple wavelengths and other approaches may prove fruitful. The future role of a full colour SLO remains to be seen but the results presented here are a remarkable visual demonstration of the continuing expansion of new possibilities in imaging.


    View Abstract

    Request permissions

    If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

    Linked Articles