This article is best viewed in a PDF Format.
|
Some very basic effects of polarised light can be detected with the use of a simple polarisation filter. Where light is reflected from surfaces the percentage of light reflected will vary for light radiating parallel and at right angles to the plane of incidence. At the so-called Brewster angle, the component of light parallel to the plane of incidence is at a minimum. This effect can be observed when reflections from a window are observed through a rotating polarisation film. The intensity of reflections will vary with angular position of the polarising firm.
About polarised light
Specific optical materials can demonstrate different properties, with light polarised in different planes. In particular, the property of bifringence occurs where components of light of each polarisation travel at different velocities. The respective rays tend to be described as the ordinary ray, which vibrates in a plane defined by the incident ray and normal to the surface, and the extraordinary ray, which is at right angles to this. This phenomenon is clearly demonstrated by a calcite crystal. However, within the eye the specific ordering of molecular structures can introduce some degree of demonstrable bifringence. This indicates that polarisation effects may have the potential to provide additional clinical information within the framework of existing diagnostic screening/imaging procedures. From the definitive review of polarisation effects in the eye, Bour2 indicated the major contribution to bifringence reside in the cornea/lens structures, although this is not significantly used for diagnostic analysis of corneal structures.
Polarisation effects and imaging technologies in the eye
An increasing number of ocular imaging technologies use light that may have highly specific polarisation parameters to optimise the sensitivity of the specific technology. Such imaging techniques include those referenced in Table 1.
However, investigation by Prieto et al3 of basic parameters such as the eye’s wave aberrations indicated that no measurable change in aberrations was noted for the two directions of light polarisation. The authors do note there were basic issues with measurement accuracy of analysis systems used.
There is a growing interest in instrumentation to detect the polarisation properties of the retinal nerve fibre layer (RNFL) as a method to give rapid diagnostic indication of potential deterioration associated with glaucoma. The instrumentation used for this technique is described as a scanning laser polarimeter. Such systems can provide rapid measurement of the thickness of this layer by derivation of a corresponding bifringence map. In this context, bifringence is associated with the alignment of nerve fibre bundles in the RNFL.
Knighton et al4 provides a useful description of a model of scanning laser polarimeter system from the context of optical interfaces involved (Figure 1). The area centred on the fundus is represented as having a radial bifringent structure. The light reflected from the retina is assumed to preserve polarisation attributes. Presumably as part of the scanning mode of the system the path difference of light at each pixel is effectively determined for light with polarisation in the radial direction and at right angles to this direction.
There is also potentially the effect of variation of bifringence as a function of wavelength. Current scanning laser polarimeter systems tend to use a fixed measurement wavelength of 780nm. Little work has been done, however, in determining how polarisation effects vary as a function of wavelength. Of particular relevance would be the conversion factor in RNFL thickness of retardation to thickness. Also, some aspects of bifringence have been associated with structures under stress.
Where different polarisations travel at different velocities, one measure of bifringence is the equivalent path difference between the two beams after having travelled a specific distance. This has been expressed by some investigators as nm per micron where the value in nm is the air path difference required for the ‘slow’ beam to catch up with the ‘fast’ beam after travelling through 1 micron of tissue. (Note 1 nm = 10-9 m.)
In the diagram in Figure 2, where d1 is the distance travelled by the fast beam and d2 is the equivalent distance travelled by the slow beam, then the percentage difference in light velocities, Vdiff%, is given by:
Vdiff% = 100 x (v1-v2)/v1 (1)
= 100 x (1-v2/v1) (2)
If path lengths d1 and d2 are travelled in same time and v2 is the velocity of light of ‘slow’ ray and v1 is the velocity of light of ‘fast’ ray:
d2/v2 = d1/v1 (3)
Vdiff% = 100 x (1-d2/d1) (4)
= 100 x (1 - (d1- ?)/d1) (5)
= 100x ?/d1 (6)
= 0.067
where ? = 0.67 nm and d1 = 1 micron (1,000 nm).
Thus the level of bifringence is relatively small. In making measurements using such scanning laser polarimeter systems, however, allowance is required to take account of bifringence originating in the cornea. Devices termed corneal compensators have been included in scanning laser polarimeter systems used to date. The use of systems to provide variable degrees of compensation is indicated as preferable to provide levels of correction that may more closely match that of patients being scanned.
The application of scanning laser polarimeters has progressed through careful modes of exploration where variations of measurement performance within the normal population and in specific disease conditions have been studied. One study by Katsanos et al5 has studied the effect of subfoveal choroidal neovascularisation (CNV) on images of the macula undertaken using scanning laser polarimetry. One finding was that the magnitude of the ‘macular ratio’ was significantly higher in the CNV group compared to the normal group. This parameter measures the ratio of maximum to minimum retardation. It was also observed that while the classic ‘bow tie’ polarisation pattern was observed in all 23 normal subjects, this was observed only in seven of the CNV group.
In 2004 Carl Zeiss Meditec acquired Laser Diagnostic Technologies – the company that had undertaken the ground-breaking development of scanning laser polarimetry technology. During the evolution of scanning laser polarimeter technology, a range of studies have assessed errors introduced as a result of fixed corrections of bifringence arising from the cornea and the lens. An excellent resume of scanning laser polarimetery technology can be found at the Carl Zeiss Meditec website.
Features of GDx VCC (variable corneal compensation) system
The GDx VCC scanning laser polarimeter represents the current state-of-the-art technology. Table 2 summarises some of the key characteristics of the system.
Earlier models of the GDx system, such as the NFA and Access utilised fixed values of bifringence compensation for the cornea/lens contribution – typically of 60nm. Specific studies such as those of Knighton6 have indicated a broad variation in both direction and degree of corneal bifringence values where within a group of 73 normal subjects, 80 per cent of retardance values were distributed between 40nm to 140nm.
Currently two techniques are provided within the VCC system for variable corneal compensation. Initially the eye is imaged without compensation so that the retardation map contains elements of both the cornea/lens and that the RNFL. Assuming that the bifringence of the macular region is uniform due to the physical ordering of the cellular layers, the contribution of bifringence of the cornea/lens can be inferred from identification of the ‘bow tie’ axis of bifringence observed in the uncompensated scan. Where this feature is missing on account of macular pathology, then an averaging algorithm within the macular region can construct an equivalent value. The macular area used in cases of ‘macular pathology’ relies on accurate patient fixation.
Clinical applications
One of the key uses of the scanning laser polarimeter is for the diagnosis of early onset of glaucoma, where thinning of the nerve fibre layer can be detected before detectable changes in vision. One feature included in the VCC model is a ‘normative database’ of both normals and glaucoma patients as established within six centres within the US. The increased use of the VCC technology is developing a more mature set of clinical references.7,8,9
It has been established that thinning of the RNFL is a feature of ageing – with loss of around 5,000 ganglion cells per year as part of normal ageing. The onset of glaucoma, however, is associated with a more rapid loss of thickness of the RNFL. Various diagnostic modes are available in the GDx VCC system to classify the potential onset of disease. These include a deviation map to characterise changes in thickness between examinations (TNSIT graph) and a nerve fibre indicator (NFI) which is a predictive parameter based on a neural network descriptor.
Summary
The specific property of bifringence within the eye (especially that of the RNFL) has enabled valuable measurements of its thickness to be determined. This has provided a valuable clinical tool to monitor disease progression in glaucoma and in particular in identification of early onset of the disease.
A look at the clinical use of the GDx will appear in Optician in the next few weeks.
References
-
Gabor Horvath, Dezso Varju, Dezsv Varjz. Polarized Light in Animal Vision, ISBN 3540404570, Springer, Berlin, 2004.
-
Bour IJ. Polarised light and the eye, In Charman WN, ed Visual Optics and Instrumentation, Boca Raton, Fl:CRC Press, 1991, 310-325.
-
Prieto PM, Vargas-MartinF, McClellan JS, Burns SA, Effect of the polarization on ocular wave aberration measurements. J Opt Soc Am A Opt Image Sci Vis, 2002, 19, 809-14.
-
Knighton RW, Huang X, Greenfield DS, Analytical model of scanning laser polarimetry for retinal nerve fiber layer assessment, Invest Ophthalmology & Visual Science, 2002, 43, 383-392.
-
Katsanos A, Kothy P, Papp A, Hollo G. Influence of subfoveal choroidal neovascularisation on macular imaging with scanning laser polarimetry of the retinal nerve fibre layer, Eye, 2005, 19, 117–122.
-
Knighton RW, Huang XR. Linear Birefringence of the Central Human Cornea, Investigative Ophthalmology & Visual Science, 2002, 43, 82-86.
-
Reus NJ, Lemij HG. Diagnostic accuracy of the GDx VCC for glaucoma. Ophthalmology, 2004, 111,1860-5.
-
Mohammadi K, Bowd C, Weinreb RN, Medeiros FA, Sample PA, Zangwill LM. Retinal nerve fiber layer thickness measurements with scanning laser polarimetry predict glaucomatous visual field loss. Am J Ophthalmol, 2004, 138, 592-601.
-
Medeiros FA, Zangwill LM, Bowd C, Sample PA, Weinreb RN. Use of progressive glaucomatous optic disk change as the reference standard for evaluation of diagnostic tests in glaucoma. Am J Ophthalmol, 2005, 139,1010-8.
? Douglas Clarkson is development and quality manager, Walsgrave Hospital, Coventry