Some results of spectrometric investigation of the eye

D. Schweitzer

University of Jena, Department of Ophthalmology/Experimental Ophthalmology
Bachstr.18, 07740 Jena, Germany

Introduction

The goal of spectrometric investigations is the measurement of parameters, characterizing the metabolism. Furthermore it can be distinguished between sizes which change in a short time like blood flow, blood velocity or oxygen saturation and parameters which are the result of the long time effect of alterations in the metabolism like the concentration of pigments or exudates.

Measuring the reflectance or fluorescence spectra at selected sites one gets spectra over a distinguished spectral range. From such spectra the kind of the substances and its optical density can be determined. Examples are the measurement of the fundus pigmentation, the optical density of xanthophyll or the oxygen saturation. Under certain circumstances the optical density of pigments in different fundus layers can be determined simultaneously.

If the fundus is illuminated by a specific monochromatic light, pathologic alterations appear in an optimal contrast in the 2-dimensional image. In such monochromatic images the area of e.g. hard exudates can be measured before and after laser coagulation as an objective control of the effect of the therapy.

Methods and results

At the Department of Ophthalmology/Experimental Ophthalmology of the University of Jena, two different experimental arrangements were developed for spectrometric investigations at the eye.

Fig. 1: Scheme of the confocal Jena Imaging spectrometer.

The Imaging Spectrometry is based on the principle of a fundus camera. A slit like field is illuminated at the fundus by a 10ms xenon flash. This illuminated field is confocally imaged onto the entrance slit of polychromator. In the exit plane of the polychomator a cooled multi channel plate (MCP) connected with a charge coupled detector array (CCD) registers the spectral dispersed light of each illuminated fundus site. Figure 1 shows the scheme of the Jena Imaging Spectrometer.

As the spectra of different locations are measured at the same time, local differences in the distribution of pigments can be measured very exactly. The result of a spectral measurement is an imaging spectrum. One coordinate corresponds to location of the measurement, the second coordinate is the wavelength, and the third coordinate in the measured intensity. Figure 2 is an imaging spectrum of a cross section through a retinal vessel.

Figure 2: Imaging Spectrum of a cut through a vessel.

Figure 3: Calculation of the oxygen saturation by the approximation of a measured extinction spectrum by a model function

From such a measurement the oxygen saturation is calculated. As the reflectance is measured at a large number of wavelengths, the oxygen saturation can be calculated also in case of noisy signals. The approximation of an extinction spectrum of a vessel by a model function (fig.3) corresponds to an averaging over the wavelength.

Simultaneously with the oxygen saturation the diameter of vessels can be measured using the algorithm, developed by Dr. Vilser at the University of Jena . That means, the only required parameter for the determination of supply and of consumption of oxygen is the velocity of blood. Measurements of the oxygen saturation in vessels of different diameters result in no correlation between the oxygen saturation and the vessel diameter. That means the used model function can be applied in the diameter range between 50mm and 250mm. The mean arterio-venous difference in the oxygen saturation of neighboring vessels was determined as 34.5+- 11. This mean value is in good correspondence with the value of the brain (30%).

Figure 4: Simultaneous measurement of oxygen saturation in an artery and vein together with the measurement of the vessel diameter

 

A further promising application of imaging spectrometry is the determination of the optical density of pigments in single fundus layers. For this calculation a model function of the radiant transport is used, taking the anatomical structure of the fundus into account. Independent variables in such an approximation are the absorption spectra and scattering spectra of the used pigments or the tissues. During the approximation process, a measured reflectance spectrum is adapted by the model function by optimization of the parameters corresponding to the optical density of single layers. The deconvolution of the reflectance spectra of different fundus sites results in a kind of „functional tomography" determining the local distribution in the optical density of xanthophyll, melanin, or choroidal blood.

Figure 5: Local distribution of melanin, choroidal blood, and of xanthophyll, determined by connected deconvolution of an imaging spectrum

 

The application of the imaging spectrometer for locally resolved fluorescence measurements requires only the insertion of an excitation filter and of a cut off filter. In comparison with the reflected light, the fluorescence light of the fundus is very weak. Fig.6 demonstrates the local distribution of natural fluorophores of the fundus along a line of 1.5mm.

Figure 6. Local resolved fluorescence spectrum of a human fundus

For the spectral measurement of the anterior part of the eye an additional lens, corresponding to the human lens, is mounted in front of the ophthalmoscope. In this way the measuring slit is imaged at the lens. Now fluorescence spectra of the lens, of the iris, or of the sclera are measurable. Figure 7 is an imaging spectrum of an intraocular artificial lens, measured in situ. The bright fluorescence at 590nm seems to be caused by a cellular growing at the lens.

Figure 7: Locally resolved fluorescence of an artificial lens in situ

The second method for spectrometric investigation of the fundus is the mapping spectrometry. The wavelengths under whose illumination alterations appear with an optimal contrast were determined by imaging spectrometry. For detection of the 2-dimensional distribution of fundus substances like hard exudates or xanthophyll, a laser scanner ophthalmoscope with free selectable wavelength was developed. In this arrangement an Ar laser pumps a system of 3 dye lasers. Both 7 wavelengths of the pump laser and the a continuously tunable wavelength range between 540nm and 680nm are available. The intensity at all wavelengths is controlled by means of an acousto-optical modulator. Figure 8 shows the scheme of the laser scanner ophthalmoscope with free selectable wavelength.

Figure 8: Scheme of the laser scanner ophthalmoscope with free selectable wavelength.

The available wavelength for the monochromatic fundus illumination are given in figure 9.

Figure 9: Available dye laser wavelength range

The reduction in the area of hard exudates was measured as effect of laser coagulation in diabetic retinopathy using this arrangement. As an automatic tie point detection and image registration was also developed, an application was possible in the clinical routine. The example in figure 10 shows a reduction in the area of hard exudates 2 weeks and further 3 month after coagulation in rings circling the fovea. In the study for each case the relative area of exudates and its alteration were determined.

before treatment
2 weeks after treatment
3 months after treatment

Figure 10: Example for the reduction of hard exudates after coagulation in diabetic retinopathy

 

The detection of the optical density and of the local distribution of xanthophyll is an interesting application of mapping spectrometry. As a new information changes of both parameters were found as the result of laser coagulation in diabetic retinopathy.

As a protecting effect of xanthophyll is assumed, the detection of a reduced optical density of xanthophyll might be an early sign of a risk factor in getting ARMD.

Conclusions

Spectrometric methods allow a noninvasive and touchless investigation of parameters both in differential diagnosis and in the control of the effect of a patient-specific therapy. Especially fluorescent methods might be a key for a better understanding of pathomechanisms.

References:

Schweitzer D, Hammer M, Scibor M: Imaging Spectrometry in Ophthalmology- Principle and Application in Microcirculation and in Investigation of Pigments. Ophthalmic Research 1996, 28(suppl 2):37-44

Schweitzer D, Leistritz L, Hammer M, Scibor M, Bartsch U, Strobel J: Calibration - free measurement of the oxygen saturation in retinal vessels of men. in Parel JM, Ren Q, Joos KM(eds) Ophthalmic technologies. V. Progress in Biomedical Optics. SPIE 1995; Vol. 2393: 210-218

Schweitzer D, Kalve B, Leistritz L, Wagner M, Hammer M: Experimental results with an wavelength-tunable laser scanner ophthalmoscope. 6. International Meeting on Laser Scanner Ophthalmoscopy, Tomography and Microscopy, Paris, France Feb. 13-14, 1997

Münch C, Vilser W, Senff I: Adaptiver Algorithmus zur automatischen Messung von Gefäßweiten. Biomed. Technik, 1995;40: 322-325