A Clinical Study of the Human Lens with a Dynamic Light Scattering Device

Exp. Eye Res. (2002) 74, 93±102 doi:10.1006/exer.2001.1106, available online at http://www.idealibrary.com on

A Clinical Study of the Human Lens with a Dynamic Light Scattering Device M A N U EL B . D AT ILE S I II a, R A FAT R . A N S A R I b

AND

G E O R G E F. R E E D c

a

Cataract Section, Ophthalmic Genetics and Visual Function Branch, National Eye Institute, National Institutes of Health, Bldg. 10 Rm. 10N226, MSC 1860, Bethesda, MD 20892-1860, U.S.A., bNASA Glenn Research Center, 21000 Brookpark Road, MS 333-1, Cleveland, OH 44135, U.S.A. and cBiometry Branch, Division of Epidemiology and Clinical Research, National Eye Institute, National Institutes of Health, Bldg. 31, Rm. 6A52, MSC 2510, Bethesda, MD 20892-2510, U.S.A. (Received Milwaukee 18 April 2001 and accepted in revised form 10 September 2001) A study was conducted to determine the potential usefulness and repeatability of a new dynamic light scattering (DLS) device for clinical studies of the human lens and early cataract. Studies using the cold cataract model showed this new device to be more sensitive than the Scheimp¯ug cataract imaging system in detecting the earliest cataractous changes. A miniaturized clinical DLS device developed by NASA using ®ber optic probes was mounted on a Keratoscope (Optikon 2000), which has a 3-dimensional aiming system for accurate repeated sampling of the same area of the lens. A test/retest study was then conducted on the nuclear region of the lenses of 12 normal eyes. After a full, dilated eye examination, DLS data were obtained using the new device on the same eyes twice, 30±60 min apart. Particle size distributions and mean log particle size data were obtained. The mean percent differences between the larger and smaller of the test±retest pairs was 6.4 % (range 0.05±10.8 %); the between-test S.D. was 0.116. Actual numerical margin of error was +0.023. In addition, the mean coef®cient of variation was 4.2 % (range 0.3±7.3 %). A useful clinical end point obtained from data produced by the device was the mean log particle size. These results suggest that the DLS will be useful in the detection and study of the beginning and earliest stages of cataract formation in humans. Key words: cataract; lens; dynamic light scattering; corneal topography; clinical study.

1. Introduction Although 8±10 million cataract operations are performed each year worldwide, there remain 100 million eyes legally blind from cataracts today (Foster, 1999). It is estimated that a delay in cataract formation of about 10 years will reduce the prevalence of disabling cataracts by 45 % (Kupfer, 1984). In animal studies (Hu, Datiles and Kinoshita, 1983; Hu et al., 1984) it has been shown that in the early stages of cataract formation, there is a point beyond which the cataract becomes irreversible. Hence, for successful reversal treatment of cataracts by future anti cataract drugs, the earlier a cataract is detected, the better the chances of a drug being successful in stopping or reversing cataract formation. However, the methods in clinical use for quantifying cataract remain quite primitive and elusive at the level of early detection (Datiles, 1992). Various methods employed rely mainly on optical methods such as slit lamp clinical grading, grading of cataract photographs and Scheimp¯ug and Retroillumination photography, online digitization, and image analysis. Visual function tests have also been used, such as the Snellen acuity test, contrast sensitivity and glare testing. Unfortunately, by the time lens opacities are grossly optically visible and visual function becomes abnormal, the cataract may already be quite advanced and 0014-4835/02/010093‡10 $35.00/0

irreversible. In fact, a recent clinical trial of an anti cataract drug (Zigler and Datiles, 2000) cited the preexistence of signi®cant cataract in the lenses of the sample population as one of the reasons the interim trial results were equivocal, before the trial was ®nally discontinued. Hence, it is important that means be found that could detect the cataract early enough before the molecular changes become irreversible. A new device, using the technique of quasielastic or dynamic light scattering (QELS or DLS), is employed to detect the earliest molecular changes in the lens much earlier than the optical methods and visual function methods available today. Benedek and co-workers at MIT (Benedek, 1971; Tanaka and Benedek, 1976; Benedek et al., 1987) were the ®rst to use the DLS to study cataractogenesis. The DLS monitors the Brownian movement of the crystallins and shows the distribution of diffusion coef®cients (Benedek et al., 1987; Datiles, Podgor and Edwards, 1988; Bursell et al., 1989; Laethem et al., 1991; Thurston et al., 1997). However, because only a small portion of the lens is sampled during each measurement in order to maintain coherence conditions at the detector and hence good signal to noise, it was dif®cult to return to the same spot repeatedly (Datiles et al., 1988). This would be critical when testing a drug and following lens changes over time.

94

Recently, one of the authors (Ansari et al., 1996) developed a miniaturized version of the DLS using ®ber optic probes developed for space experiments. This device is employed in this study. To determine ®rst whether the DLS device was able to detect the earliest changes in cataract formation, a study was performed comparing the DLS device to the Scheimp¯ug Imaging (SI) System (Ansari et al., 1998; Ansari and Datiles, 1999, Ansari, Datiles and King, 2000). The cold cataract model was used so that the calf eyes being studied were ®xed in position in front of the SI Camera as well as the DLS device, which was mounted on the SI using a swinging arm. This allowed the authors to return repeatedly to the same area of the lens in vitro, as the temperature of the calf eyes was lowered and a cold cataract appeared (Ansari et al., 1998, 2000). A brief summary is as follows: Fig. 1 shows the experimental set up. It consists of a Zeiss Scheimp¯ug Video Camera System (attached on line to the NEI image analysis system (Datiles, 1992) (SI), a compact DLS mounted on a custom designed swing arm holder, a custom built sample assembly for inducing cold cataract, temperature monitoring sensors, water circulation unit, and a laptop computer containing a

M . B . D AT I L E S I I I E T A L .

BI-9000 digital correlator card. Six fresh calf eyes (ages 14±16 weeks) were used. Each calf eye was placed inside the custom built eye holder made of plexiglass with a cooling jacket mounted on an aluminum manifold. A top cover is machined to an optically smooth ®nish, also surrounded by a cooling jacket, exposing 15.2 mm of the front of the calf eye. The cover also keeps the cornea moist. The cooling jackets allow the eye to achieve equilibrium quickly. Using a refrigerated water circulation unit, cooling water is circulated through both the upper and lower cooling jackets, controlling the eye temperature. Through access holes, the lens temperature is continuously monitored at the front and back of the eye using thermocouples. Both temperatures are allowed to equilibrate with each other for 10 min before a data point is determined and data from both SI and DLS obtained. Both SI and DLS data were obtained from the same point in the lens by mounting the DLS probe on a special swing arm attached to the SI unit (Fig. 1). Fig. 2 shows a plot of the scattered light intensity as a function of temperature from the cold cataract study. For the DLS, the plot represents the total scattered light intensity whereas for the SI System the plot represents

F IG . 1. Instrument set up for the cold cataract study. On the right is the Zeiss Scheimp¯ug Video Camera Imaging System (attached on line to the NEI image analysis system, not shown) (`SI'). The Dynamic Light Scattering probe (`DLS') is mounted on a customized swing arm, which is mounted on a special compartment in front of the SI initially built in to hold the Ulbricht Sphere, a device used to calibrate the SI image as needed. The DLS device could then be swung into place or away as needed to obtain DLS or SI images of the same area of the lens nucleus of the calf eye. The enlarged area on the left shows the customized eye holder for the calf eye described in detail in the text. The water cooled eye holder has a transparent front to allow imaging of the cornea and lens. The temperature in the lens is monitored by the stainless steel sheath, closed end grounded, `K' thermocouples entering through access holes in the front and back. Data were obtained at each data point after the temperature equilibrated between front and back for 10 min. The temperature was lowered until a visible cold cataract appeared.

N E W D E V I C E TO S T U DY C ATA R AC TO G E N E S I S

optical density units (O.D.U.). As can be seen, the DLS device detected earliest changes with the cold cataract model at 178C, whereas the SI System detected changes only at the 98±108C range. Note that in the NEI-Zeiss Scheimp¯ug System, an ultra clean sample of water shows a density of 0.01 O.D.U. In addition, Fig. 3 showed that as the cold cataract appeared and became larger and denser, the crystallin size distribution detected by the DLS device also shifted from the lower molecular size particles to the higher molecular size particles. As previously pointed out, experimental evidence for this shift representing the formation of bigger protein aggregates or complexes as a cataract progresses, has been described by other authors such as Jedziniak et al. (1973), Spector, Li and Sigelman (1974) and Tanaka and Benedek (1976). These initial studies therefore were useful in demonstrating the sensitivity of the DLS device in detecting the earliest changes in cataract formation compared to other optical methods such as the SI System. The authors therefore proceeded to studies in human lenses on normal volunteers to determine the device's usefulness for in vivo studies and to obtain reproducibility data on the device. Herein a further modi®cation to the DLS apparatus is described: mainly the incorporation of the DLS device into a Keratoscope (Keratron by Optikon) to allow the use of a 3dimensional aiming system, to enhance the capability of returning repeatedly to the same sampling area in the lens for clinical use.

95

2. Materials and Methods Twelve eyes were studied from seven normal healthy volunteers. On two volunteers (AE and JK), only one eye each was used: one volunteer (AE) had a viral, upper respiratory tract infection and could only do the left eye and the other volunteer (JK) had increased tearing on the left eye and only the right eye was tested. Both eyes were used in ®ve of the volunteers. These volunteers signed informed consents and were all part of a NEI-IRB approved study protocol. All tenets of the Helsinki Declaration were followed. All eyes underwent complete eye examination. Pupils were dilated using 1 % Mydriacyl and 2.5 % Phenylephrine eye drops. All the eyes were determined to be normal by slit lamp clinical grading, with a nuclear opalescence of 0.5 or less using the Lens Opacities Classi®cation System II (LOCS II) (Chylack et al., 1989). Six eyes came from volunteers below 40 years of age and six eyes from volunteers over 40 years of age. Six volunteers were male, one female. The same observer (MBD) performed all DLS measurements in all volunteers. DLS Device The new DLS device consists of a specially developed DLS probe (Ansari et al., 1996, 2000) mounted on a corneal topography device (Keratron Corneal Analyzer by Optikon 2000 S.P.A., via del Casale de Settibagni 13-00138 Rome, Italy) (Fig. 4).

F IG . 2. Data obtained from the cold cataract study, comparing how early the DLS and SI detected cataractous changes as the temperature of the lens was lowered. The DLS data are shown as scattered light intensity in photons per sec. SI data consist of mean density of the nucleus of the lens in terms of O.D.U. (Transparency of an ultra clear sample of water in the Zeiss/NEI Scheimp¯ug System ˆ 0.01 O.D.U.) Note that the DLS detects the earliest changes as early as at 178C and more dramatically at 148C. On the other hand the SI detects the earliest changes much later at the 9±108C level showing the greater sensitivity of the DLS in detecting early cataractous lens changes than the SI. (With permission from Proc. SPIE).

96

M . B . D AT I L E S I I I E T A L .

F IG . 3. Another comparison between DLS and SI data obtained from the cold cataract study. DLS data shown as protein particle size distribution on the left column. SI data shown as Scheimp¯ug slit lamp images of the calf lenses taken with the NEI-Zeiss SI System, with the labels below the images showing the temperature at that data point, and the mean nuclear density in O.D.U. Note that as the temperature drops and the cold cataract appears, there is a shift from low molecular weight to high molecular weight proteins. (With permission from Proc. SPIE).

The DLS probe projects a low power (80 mW) laser beam (much lower power compared with the laser used to scan grocery items in the supermarket which uses up to 5 mW) into the lens of a patient and the scattered light is collected from a small volume of the lens by an avalanche photodiode detector (APD) connected with the probe (Fig. 5). The laser used is a solid-state laser at a wavelength of 670 nm. The beam is defocused at the retina so that the maximum intensity does not exceed American National Standards Institute (ANSI) standards. (Max. ANSI standard ˆ 47.464 J cm 2, DLS maximum power density at retina ˆ 0.881 J cm 2.) The scattered light was collected from a focused volume of 2  10 5 mm3 in the region of the lens and then analysed and stored. A correlation function reduces the scattered intensity data obtained which can then

be used to characterize the particle size in the lens (Benedek et al., 1987; Bursell et al., 1989; Van Laethem et al., 1991; Ansari et al., 1996; Thurston et al., 1997; Ansari et al., 1998). The volunteers were examined using the new instrument. A yellow ®xation target in the center of the cone-like end of the video Keratoscope (VK) was used to ®xate the patient. The VK is equipped with a CCD camera that allows the operator to align the pupil to the center of a built in crosshair on the image acquisition screen in the back of the device (Fig. 6) similar to a small TV screen (xy position). The entire camera is moved using a joystick. On the cone-like end of the VK are two ®ber optic position sensors [Figs (4) and (5)]. These send a signal (beeping sound) when the apex of the cornea is intercepted by the infrared beam passing the two sensors (z position). At

N E W D E V I C E TO S T U DY C ATA R AC TO G E N E S I S

97

F IG . 4. Photographs showing how the DLS device was mounted in the Keratron Corneal Topography device. The ®gure on the left shows the location of the two ®ber optic DLS probes. `Light out' denotes the low power laser source, whereas `Light in' denotes the APD. The APD collects the light scattering in the 2  10 5 mm3 focal point of the laser and sends it to the DLS data acquisition and analysis system (Fig. 5). The ®gure on the right shows the video camera used to position the eye in the xy position. Shown as well are the locations of the two ®ber optic position sensors which detects the third or anteroposterior or `Z', dimension, the apex of cornea (see also IR light beam in Fig. 5). (With permissions from Proc. SPIE).

this preset position, the DLS probe's focal point is automatically focused 4 mm behind the cornea, into the lens nucleus. At this point (when the pupillary center is within the crosshair in the monitor screen and the cornea is signaled to be in the predetermined location), the operator then triggers the DLS device by simply pushing a button to obtain a measurement in the lens. The video image of the cornea and the pupil is recorded for archiving purposes in the hard drive of the computer every time a DLS measurement is obtained. This image can be correlated with DLS data that is stored in a separate database, and processed as described above. One unique feature of this device is that during the alignment procedure the laser beam is turned off and the alignment is completed within a few seconds using the joystick. Then the DLS device is turned on for 5 sec and the scattered light is collected. The laser automatically shuts off in 5 sec. Measurements are made in the nucleus of the lens. 3. Results Fig. 7(A) and (B) show the size distribution of protein crystallins using the technique of exponential sampling (Stock and Ray, 1985), obtained from each eye, with the ®rst test results superimposed on the second test results. The distributions are discrete,

taking values according to the observed particle sizes, which occur not in continuum but in discontinuous steps. For further analysis of the data, it is assumed that the distribution will give more weight (or frequency) to higher sized values as the genesis or development of lens opacities proceeds. Thus a reasonable approach for simplifying the data to a single value is to calculate the mean particle size; for, as a lens develops opacities or cataracts it is expected that the mean particle size will increase. However, the range of particle sizes in a given distribution may cover more than one order magnitude, and it is well recognized that the arithmetic mean in such cases is unduly dominated by the larger values. For example, in the series 10, 10, 10, 10, 10, 10, 10, 10, 10, 1000, the ratio between the smallest and largest is 100; and the mean is 109, which is 10 times greater than most numbers in the series. The single large value distorts the fact that most values are of a more modest magnitude than the calculated mean. A simple way to reduce the in¯uence of the larger values is to transform the data to a logarithmic scale. The log base 10 of the series 1, 1, 1, 1, 1, 1, 1, 1, 1, 3, with mean 1.2. This mean is only 20 % greater than most numbers in the series and therefore reduces the in¯uence of the largest number. For this reason the index chosen for summarizing the particle size

98

M . B . D AT I L E S I I I E T A L .

F IG . 5. A schematic representation of how the DLS device was mounted inside the Keratron and where the focal point of the laser (and the sampling) is in the lens. The `IR light beam' is an infrared beam coming from two ®ber optic position sensors. These send a signal to the operator when the apex of the cornea is intercepted by the IR light beam (z position). At this preset position, the DLS probes' focal point is automatically focused 4 mm behind the cornea, into the lens nucleus.

F IG . 6. Photograph of the combined DLS and Keratron instruments. The on line computer performs the analysis of the data acquired by the DLS probes which have been mounted inside the Keratron. Note the computer monitor and the TV display on the back of the Keratron showing the eye/pupil image of a volunteer.

distribution is the mean log10-transformed particle size. For each eye, the means of the log10 particle sizes at the ®rst and second measurement were calculated and displayed in Table I. Table I summarizes the average particle size measurements from each eye for the ®rst and second

(repeat) tests. Displayed also are the ratios between the larger and the smaller measurements. From the ratios of larger to smaller test measurements we see that test±retest % differences ranged from 0.5 to 10.8 % with a mean of 6.4 %. From the application of a variance components analysis (Snedecor and Cochran, 1980), the between-test S.D. was estimated

N E W D E V I C E TO S T U DY C ATA R AC TO G E N E S I S

99

F IG . 7. Caption over the page.

TABLE I Summary of DLS data obtained from 12 normal eyes. The `®rst' and `second' columns represent results of ®rst and second tests in terms of the mean log particle size for each particular eye (see text) OBS

Eye

Volunteer

Age

1 2 3 4 5 6 7 8 9 10 11 12

OD OS OD OS OD OS OD OS OD OS OD OS

J.K. J.K G.D G.D. K.J. A.E. D.J. D.J. S.D. S.D. M.T. M.T.

20 20 26 26 35 20 47 47 56 56 70 70

First* 2.5004 2.0335 2.8329 2.5245 2.600 2.5070 2.4182 2.6744 2.5030 2.5054 3.7681 2.6137

Second** 2.6985 2.2538 2.6171 2.3854 2.5366 2.7470 2.5538 2.5208 2.6778 2.6135 3.7505 2.7581

Ratio** 1.07923 1.10834 1.08246 1.05831 1.02519 1.09573 1.05607 1.06093 1.06984 1.08246 1.00469 1.05525

CV ( %)*** 5.4 7.3 5.6 4.0 1.8 6.5 3.9 4.2 4.8 3.0 0.3 3.8

*Indices are means of log10 particle size at ®rst and second measurement. **`Ratio' represents ratios between the larger and the smaller measurement for each eye. Mean % test±retest differences was 6.4 % (range 0.5±10.8 %). By variance components analysis (Snedecor and Cochran, 1980) the between-test±retest S.D. was 0.116. ***CV % (coef®cient of variation in %) is an additional measure of repeatability and refers to variation in the measurements from test to retest.

100

M . B . D AT I L E S I I I E T A L .

F IG . 7. (a) and (b) Distribution of protein particle sizes, obtained using the DLS device, on the 12 normal eyes. First and second sets of data for each eye are superimposed on each other. Patient initials and ages are indicated. (a) Data from younger eyes (18±40 years). (b) Data from older eyes (41±75 years).

to be 0.116. The coef®cient of variation (variation of measurements from test to test) ranged from 0.3 to 7.3 % with a mean of 4.2 %.

4. Discussion The between-test S.D. is a measure of variation in measurements when the underlying mean log particle size is unchanged (i.e. it is a measure of repeatability). Thus, assuming a normal distribution, 95 % of the time two repeated measurements with no underlying change will be within 1.96 S.D., or 0.23 U apart. So 0.23 may be considered an estimate of the lower threshold of change. For future studies to correlate mean log particle size with lens opacities, a change in the mean log particle size cannot be considered to have occurred unless it is at least as large as 0.23. This ®gure needs to be con®rmed by larger studies.

An interesting case is patient M.T. (70 years old). The right eye of this patient had very early nuclear changes, compatible with a LOCSII clinical grading of 0.5, compared with the left eye of 0.0. Fig. 7 shows that the size distribution of his lens proteins in the right eye is shifted to the right, whereas the left clearer lens has the size distribution similar to the other, normal lenses. Since the `normal' range is stipulated to be between 0.0 and 0.5 in nuclear opalescence using LOCSII, he was still considered normal, but clearly there is something changing in the molecular level as shown by the DLS measurement. The authors plan to conduct a follow up of this patient. Early studies on the DLS device were mainly cross sectional studies to demonstrate the changes in the lens with aging (Thurston et al., 1997) in various grades or severities of cataracts (Tanaka and Benedek, 1976; Benedek et al., 1987; Datiles et al., 1988;

N E W D E V I C E TO S T U DY C ATA R AC TO G E N E S I S

Bursell et al., 1989; Thurston et al., 1997). Follow up studies to track or monitor the changes that occur over time were dif®cult to conduct because there were no adequate clinical aiming systems that would allow the investigator to study and precisely return to the same area of the lens over time. In addition, there were no studies on the sensitivity of the device in testing the earliest changes in comparison with the available sensitive methods of cataract detection and imaging. Studies conducted at NEI in collaboration with NASA, (Ansari et al., 1998, 2000; Ansari and Datiles, 1999; mentioned earlier) using this device demonstrated its sensitivity in picking up the earliest changes in cataract formation. The present study, on the other hand, further explores the usefulness of this DLS device in a clinical setting. After trial and error studies mounting the DLS probe with various clinical instruments to study the lens (such as the Slit Lamp Microscope, the SI System and the Retroillumination System), the probe was mounted onto the VK and it was found that its 3dimensional aiming system greatly enhanced the ability to return to the same area of the lens repeatedly. As shown in this study, the reproducibility of the device is good. This will now allow use of the device for longitudinal follow up of each cataractous change in the human lens. A ®xed antero posterior distance was used between the DLS probe and the lens nucleus for this study for uniformity. However, using A-scan ultrasonography, the distance between the cornea to the anterior lens capsule and the posterior lens capsule can be obtained for each patient and the precise and repeatable sampling of various parts of the lens is now possible. The authors are further modifying this device to allow the examiner to change the focus of the DLS probe to allow sampling of various areas in the clear as well as cataractous lens, and other parts of the eye. Previous studies also described the time autocorrelation function pro®les as well as protein diffusivities. However, much discussion centered on the fast and slow components of the autocorrelation function, but what these components represent clinically is still a subject of further study. A recent paper by Dhadwal and Wittpen (2000) describes a different way of using the DLS data obtained from clear and cataractous human lenses. The authors proposed the use of a derived global `Cataract Index' (C.I.), based on a single sampling of the nuclear region of the lens (®ve for clear to 10 000 for opaque lenses) to represent the status of the entire lens. This remains to be clinically tested. Regarding the DLS device used, the NASA-developed device used in this study uses lower laser power and shorter measurement time. It was found that high quality data are obtained when the patient is relaxed. Lower light intensity (lower laser power) resulted in less glare; and shorter data acquisition time (shorter laser exposure time) resulted in shorter periods for the

101

patient to keep the eye and the rest of the body still while keeping the eyes wide open and staring at the ®xation target. These helped the patient to be more relaxed. In this study, the log10 transformation was applied to the particle sizes in order to reduce the in¯uence of larger particle sizes on the resultant mean size. The authors therefore propose the mean of the log particle sizes as an end point for clinical studies. Other data obtained from this device, such as distribution of particle size, may become cumbersome to use for statistical analysis for large longitudinal (follow up) clinical studies. In conclusion, studies performed so far have demonstrated that the DLS device is sensitive in detecting the earliest changes in cataract formation by 2±3 orders of magnitude over presently used clinical systems and that mounting the DLS on a VK with 3-dimensional aiming system gives clinically repeatable data. In addition, the mean particle size derived from the protein distribution for each area sampled within each eye can be used as a simple main end point for clinical studies using the device. # 2002 US Government Acknowledgements This study was supported by a NEI-NASA Interagency Agreement: National Eye Institute, National Institutes of Health, US Public Health Service, Bethesda, MD 20892 and National Aeronautic & Space Administration, John Glenn Research Center, at Lewis Field, Cleveland, OH, 44135. The authors would also like to thank Ms. Shirley J. Quander for typing the manuscript and Mrs. Doretha Leftwood (NEI) and Mr. James King (NASA) for technical assistance.

References Ansari, R. R., Datiles, M. and King, J. (2000). A new clinical instrument for the early detection of cataract using dynamic light scattering and corneal topography. SPIE 3908, 38±49. Ansari, R. R. and Datiles, M. B. (1999). Use of dynamic light scattering and Scheimp¯ug imaging for the early detection of cataracts. Diabetes Technol. Ther. 1;2, 159±68. Ansari, R. R., Datiles, M. B., King, J. F. and Leftwood, D (1998). Measuring lens opacity: combining quasielastic light scattering with Scheimp¯ug imaging system. SPIE 3246, 35±42. Ansari, R. R., Suh, K. I., Arabshai, A., Wilson, W., Bray, T. L. and Delucas, L (1996). A ®ber optic probe for monitoring protein aggregation, nucleation, and crystallization. J. Crys. Growth 168, 216±26. Benedek, G. B. (1971). Theory of Transparency of the Eye. Applied Optics 10;3, 459±73. Benedek, G. B., Chylack, L. T., Libondi, T., Magnante, P. and Pennett, M. (1987). Quantitative detection of the molecular changes associated with early cataractogenesis in living human lenses using quasielastic light scattering. Curr. Eye Res. 6;12, 1421±32. Bursell, S. E., Baker, R. S., Weiss, J. N., Haughton, J. F. and Rand, L. I. (1989). Clinical photon correlation spectroscopy evaluation of human lenses. Invest. Ophthalmol. Vis. Sci. 49, 241±58.

102

Chylack, L., Leske, M. C., McCarthy, D., Khu, P. M., Kashiwagi, T. and Sperduto, R. (1989). Lens Opacities Classi®cation System II. Arch. Ophthalmol. 107, 991±7. Datiles, M. B. (1992). Clinical Evaluation of Cataracts. In Duane's Clinical Ophthalmology. (Tasman, W. and Jaeger, E., Eds.), Vol. 1. Ch. 73B, 1±5. Philadelphia, PA, U.S.A.: Lippincott Co., Inc. Datiles, M. B., Podgor, M. and Edwards, P.. Reproducibility of early cataract detector (Kowa ECD 2000). Ophthalmic Surgery (1988). 19;9, 664±6. Dhadwal, H. and Wittpen, J. (2000). In vivo dynamic light scattering characterization of the human lens: cataract index. Curr. Eye Res. 20;6, 502±10. Foster, A. (1999). CataractÐA Global Perspective: output, outcome and outlay. Eye 3, 449±53. Hu, T.-S., Datiles, M. B. and Kinoshita, J. H.(1983). . Reversal of galactose cataract with sorbinil in rats. Invest. Ophthalmol. Vis. Sci. 24, 690±4. Hu, T.-S., Merola, L., Kuwabara, T. and Kinoshita, J. H. (1984). Prevention and reversal of galactose cataracts in rats with topical sorbinil. Invest. Ophthalmol. Vis. Sci. 25, 603±5. Jedziniak, J., Kinoshita, J. H., Yates, G. M., Hocker, L. O. and Benedek, G. B. (1973). On the presence and mechanism of formation of heavy molecular weight aggregate in human normal and cataractous lenses. Exp. Eye Res. 15, 185±9. Kupfer, C. and the Bowman lecture. (1984). The conquest of cataract: a global challenge. Trans. Ophthalmol. Soc. U.K. 104, 1.

M . B . D AT I L E S I I I E T A L .

Snedecor, G. W. and Cochran, W. G. (1980). Statistical Methods, 13. Second edn, 235±53. Ames, IA, U.S.A.: Iowa State University Press. Spector, A., Li, S. and Sigelman, J. (1974). Age-dependent changes in the molecular size of human lens proteins and their relationship to light scattering. Invest. Ophthalmol. 13, 795±8. Stock, R. H. and Ray, W. H. (1985). Interpretations of photon correlation data: A comparison of analysis and methods. J. Polym. Sci., Polym. Phys. Ed. 23, 1393±447. Tanaka, T. and Benedek, G. B. (1976). Observation of protein diffusivity in intact human and bovine lenses with application to cataract. Invest. Ophthalmol. Vis. Sci. 14;6, 449±56. Thurston, G. M., Hayden, D. L., Burrows, P., Clarke, J. I., Taret, V. G., Kandel, J., Courogen, M., Peetermans, J. A., Bowen, M., Millar, D., Sullivan, K. M., Storb, R., Stern, G. and Benedek, G. B. (1997). Quasielastic light scattering study of the living human lens as a function of age. Curr. Eye Res. 16, 197±207. Van Laethem, M., Babusiaux, B., Neetens, A. and Clauwaerts, J. (1991). Photon correlation spectroscopy of light scattered by eye lenses in vivo conditions. Biophys. J. 59, 433±44. Zigler, J. S. and Datiles, M. B. (2000). Pathogenesis of Cataracts. In Duane's Clinical Ophthalmology. (Tasman, W. and Jaeger, E., Eds.), 72B, 13, Philadelphia, PA, U.S.A.: Lippincott Co., Inc.