Download Nuclear magnetic resonance microscopic ocular imaging for

Investigative Ophthalmology & Visual Science, Vol. 30, No. 7, July 1989
Copyright © Association for Research in Vision and Ophthalmology
Nuclear Magnetic Resonance Microscopic Ocular
Imaging for the Detection of Early-Stage Cataract
Chang B. Ann,* Janet A. Anderson,^ Sung C. Juh,* Inja Kim,t William H. Garner, f and Zang-Hee Cho*§
A nuclear magnetic resonance (NMR) microscopic ocular imaging was performed at 7.0 Tesla to
investigate its usefulness in the detection of early-stage cataracts. For this study, galactose cataracts
were generated in experimental rabbits through diet (35% galactose), and enucleated eyes were imaged
at various times after initiation of the diet. In previous studies using a 0.6 Tesla conventional magnetic
resonance imager (MRI), the contrast between normal and cataractous tissues in the lens was not well
defined, mainly due to the partial volume effect coming from the limitation of resolution and signalto-noise ratio (SNR). With resolution of 60 X 60 X 80 nm, early localized precataractous tissue
changes were clearly observed after 5 days diet. Precataractous tissue changes were seen histologically
but no visible evidence of lens change was detected by the conventional slit lamp biomicroscope at this
time. Substantially elongated spin-spin relaxation times (T2) in localized cataractous tissues (72.4
± 8.8 msec) were consistently observed compared with those in normal lens region (16.1 ± 3.2 msec);
however, the changes of the spin-lattice relaxation time (Ti) were not significant: Some ocular NMR
microscopic images with corresponding histological photographs are demonstrated to show the potential of NMR microscopy. Invest Ophthalmol Vis Sci 30:1612-1617,1989
Magnetic resonance imaging (MRI) has begun to
find clinical relevance in ophthalmology.1"8 The
noninvasive characteristics plus inherent high resolution and various chemical contrasts made NMR
imaging a unique diagnostic modality, and now
NMR imaging begins to supercede existing diagnostic
technologies, such as x-ray and computerized tomography (CT). Since the early 1980s, clinical and physiological applications of NMR imaging have been
largely investigated.9"11 Although several reports have
already shown the potential of NMR imaging in the
study of eye,1"8 NMR imaging of the eye has been
limited, mainly due to the resolution.
Previously we reported a study of galactose cataract
using a conventional whole-body MRI operated at
0.6 Tesla.12 With this system, however, the detection
of early-stage localized cataractous change was limited by the large size of the picture element (pixel)
compared to the size of precataractous tissue. Since
the image intensity in each digitized pixel is the integrated spin signal within the unit volume, if the pixel
size or resolution is not sufficiently fine, a pixel containing both cataractous and normal tissues appears
ambiguously, thereby reducing contrast and detectability. The pixel size and the image contrast are also
associated with the signal-to-noise ratio (SNR) of the
employed system, which is mainly determined by the
strength of the main magnetic field. The limitation of
the conventional MRI in the reduction of echo time
(TE) (about 25 msec) was another factor degrading
the accuracy of the measured T2 values, especially in
the lens, where T2 is pretty short.
In the study reported here, most of these problems
were solved by the introduction of a new high-field
NMR imaging and microscopy.1314 Galactose cataract in the rabbit was chosen for the animal model.
Due to the increased fraction of free water to bound
water in the cataractous tissue, the spin-spin relaxation time (T2) increases exceedingly. Using this
NMR characteristic of cataract, several imaging pulse
schemes were investigated to achieve maximal contrast between normal and cataractous tissues. The
application of NMR imaging to the detection of
early-stage cataract can be useful for understanding
the etiology of cataract as well as for the design of
preventive measures.
From the Departments of *Radiological Sciences and fOphthalmology, University of California, Irvine, Irvine, California.
J Current address: Ophthalmology Research, Sharp Cabrillo
Hospital, San Diego, California.
§ Also at Department of Electrical Science, Korea Advanced Institute of Science, Seoul, Korea.
Presented in part at the 1988 meeting of the Association for
Research in Vision and Ophthalmology, Sarasota, Florida.
Supported in part by the National Eye Institute grant
EY-07941-01.
Submitted for publication: August 19, 1988; accepted January 9,
1989.
Reprint requests: C. B. Ahn, PhD, Department of Radiological
Sciences, University of California, Irvine, Irvine, CA 92717.
1612
Downloaded From: http://iovs.arvojournals.org/pdfaccess.ashx?url=/data/journals/iovs/933378/ on 05/08/2017
NMR MICROSCOPIC IMAGING FOR THE DETECTION OF CATARACT / Ahn er ol
No. 7
1613
Materials and Methods
Experimental System
A general concept of NMR microscopy and a detailed description of the employed 7.0 Tesla NMR
microscopy system can be found elsewhere.14 In this
section, only a brief description specifically related to
ocular imaging is given. A specially designed radiofrequency (r.f.) coil and gradient coil set were developed for eye imaging. The employed r.f. coil shown in
Figure 1A is a single-turn solenoidal coil made of
copper sheet (diameter = 1.5 cm, width = 2 cm),
which provides maximal r.f. homogeneity as well as
high sensitivity. A modified Golay gradient coil is
employed, which can generate gradientfieldsof up to
40 gauss/cm when derived by the gradient amplifier
made for the NMR microscopy system. The integrated imaging probe (right part) is shown in Figure
1B and C; it will be placed at the center of the vertical
magnet with field strength of 7.0 Tesla (proton resonance frequency = 300 MHz).
r.f. Coil for NMR
A Ocular Imaging
7.O Tfcftln NMR
Microscopic Iroaelng Probe
Imaging Methods
A conventional spin echo sequence was employed
with repetition times (TR) of 1 sec, 2 sec, and 4 sec.
The used echo times (TE) were 12 msec, 40 msec, and
80 msec. By considering physical eye size (~ 1.5 cm),
image matrix (256 X 256), and measurement time
(10-30 min), the pixel size was chosen as 60 X 60
X 80 urn. Among experiments with various TRs and
TEs, three imaging methods appeared to be most
promising, namely: (1) short-TE and long-TR sequence resulting in image intensity close to proton
density; (2) long-TE and long-TR sequence generating T2-weighted image, and (3) calculated T2 map
from the above two sequences with same TRs but
with different TEs. These imaging methods will be
used mainly in further discussion of the experimental
results.
Animal Models
Galactose cataract in the rabbit was chosen for the
animal model. Galactose cataracts are formed by osmotic changes in the lens fibers.15 Galactose is converted to dulbicol in the lens fibers. The presence of
alcohol in the lens fibers creates hypertonicity that is
corrected by an influx of water. The water influx is
followed by electrolyte changes, membrane permeability breakdown and eventually a large influx of
sodium because of the increased lens hydration. The
increased relaxation time in the cataractous tissue results from the increased fraction of free water to
bound water due to the loss of water binding sites on
the aggregated proteins.8
Fig. 1. NMR microscopy system for ocular imaging. (A) Singleturn solenoidal r.f. coil (diameter = 1.5 cm, width = 2 cm). (B)
Integrated imaging probe (right part). (O Top view of the imaging
probe. The central object is the r.f. coil, which is surrounded by the
gradient coil.
In this high resolution NMR imaging and microscopy study of cataract, young (4-week) albino New
Zealand rabbits were fed on a diet consisting of 35%
galactose to induce cataractogenesis. Control rabbits
of the same age were also maintained in the same
facility and were fed a normal diet. NMR microscopic imagings of both control and cataractous animals were performed from 1 day to 4 weeks following
Downloaded From: http://iovs.arvojournals.org/pdfaccess.ashx?url=/data/journals/iovs/933378/ on 05/08/2017
1614
INVESTIGATIVE OPHTHALMOLOGY b VISUAL SCIENCE / July 1989
Vol. 30
Control / Rabbit Eye
Control / Rabbit Eye
Proton Density Image ( 2000/12 msec ) T 2 Weighted Image ( 2000/40 msec
T2 Map of Rabbit Eye / Control
D
Fig. 2. NMR microscopic images of the control rabbit eye. (A) TR/TE = 2000/12 msec. (B) TR/TE = 2000/40 msec. (C) T2 map. (D)
Histological image of the equatorial section of the lens.
the initiation of the diet. The development of lens
opacities was also examined by a slit-lamp biomicroscope before the NMR scanning. Each rabbit was sacrificed by intravenous injection of a T-61 euthanasia
solution (Hoechst-Roussel, Somerville, NJ) and eyes
were immediately enucleated. One fresh enucleated
eye was placed in a specially designed eye holder to fit
in the r.f. coil of the microscopy system and scanned
within one-half hour of the enucleation. The other
eye was embedded in paraffin and stained with hemotoxilin/eosin, and then sent to a histology laboratory to examine microscopic structural changes. All
investigations described in this paper were carried out
in accordance with the ARVO Resolution on the Use
of Animals in Research.
Results
Ten rabbits were placed on the galactose diet and
examined 1, 2, 3, 4, 5, 7, 10, 14, 21 and 28 days from
the initiation of the diet. Two control rabbits were
also examined. Figure 2 shows the NMR microscopic
images of the control rabbit eye with (A) TR/TE
= 2000/12 msec and (B) TR/TE = 2000/40 msec.
The calculated T2 map from Figures 2A and B is
shown in C, and the corresponding histological photograph is shown in D. In the T2 map, the displayed
range of T2 is from 0 (black) to 128 msec (white). For
example, the T2 values in the vitreous or anterior
chamber are greater than 128 msec (appearing white),
while the average T2 value in cornea is 19.6 msec
(appearing dark gray). A better differentiation between lens cortex and nucleus is made in the T2 map
(Fig. 2C). The T2 values in lens cortex were reported
to be longer than T2 in the lens nucleus,516 which is
also consistent with our measurement in the T2 map.
The experimentally obtained images after 5 days'
diet are shown in Figure 3, (A) TR/TE = 2000/12
msec, (B) TR/TE = 2000/40 msec, (C) T2 map, and
(D) histological image. Note the appearance of cataract at the equatorial region of the lens in NMR microscopic image (marked with a black arrow in Fig.
Downloaded From: http://iovs.arvojournals.org/pdfaccess.ashx?url=/data/journals/iovs/933378/ on 05/08/2017
No. 7
NMR MICROSCOPIC IMAGING FOR THE DETECTION OF CATARACT / Ahn er ol
1615
Early Detection
of Cataract
5 Days Diet / Rabbit Eye
5 Days Diet / Rabbit Eye
Proton Density Image ( 2000/12 msec ) T2 Weighted Image ( 2000/40 msec )
T2 Map of Rabbit Eye / 5 Days Diet
D
Fig. 3. NMR microscopic image of the rabbit eye after 5 days1 diet. (A) TR/TE = 2000/12 msec. <B) TR/TE = 2000/40 msec. Early-stage
cataract can be observed at the equatorial region. (C) Tj map. (D) Histological photograph of the lens at the equatorial section.
3B), which had not yet been detected by conventional
slit-lamp biomicroscope. Due to the increase of T 2 ,
the cataractous region appears bright in the T2weighted image, while the normal lens region appears
dark due to the fast decay of the signal with short T2
values. The increased T2 values are also demonstrated in the T2 map in Figure 3C. Other experimentally obtained eye images are shown in Figure 4 (after
2 weeks' diet). At this stage, the lens opacity could be
seen by both conventional slit-lamp biomicroscope as
well as by NMR microscope. With extremely long
echo time (80 msec), the signal from the cataractous
tissues only remained as shown in Figure 4C. Also,
fully formed vacuoles are seen in the histological
photograph of Figure 4E.
The average T2 value of the cataractous tissues in
the lens was 72.4 msec with a standard deviation of
8.8 msec, while the average T2 in normal lens (over
both nucleus and cortex regions) was 16.1 msec with
a standard deviation of 3.2 msec. A summary of examinations of the 12 rabbits by the NMR microscope, conventional slit-lamp biomicroscope and histology is given in Table 1.
Discussion
With the introduction of high-field NMR microscopy, it was possible to detect early-stage cataract (5
days after diet) before any other evidence could be
observed by slit-lamp biomicroscopy. It has been
noted that lens clarity as examined by slit-lamp biomicroscope is not a reliable index in detecting earlystage cataract. The importance of detecting earlystage cataract lies in the fact that the progression of
galactose cataract can be reversed by early treatments.17
The role of high-resolution NMR imaging in the
detection of cataract is evident. For example, in conventional MRI, imaging resolution is about 1 X 1 X 4
mm with a magnetic field of about 0.5-2.0 Tesla,
which is much larger than the size of cataractous tissue (see NMR images of cataract lens in Figs. 3B and
4C). In our experiment, however, it was possible to
get a microscopic resolution (60 X 60 X 80 fim) due
to the SNR improvements by employing a high magnetic field (7.0 Tesla) and a small r.f. coil (diameter
= 1.5 cm).
Downloaded From: http://iovs.arvojournals.org/pdfaccess.ashx?url=/data/journals/iovs/933378/ on 05/08/2017
1616
INVESTIGATIVE OPHTHALMOLOGY 6 VISUAL SCIENCE / July 1989
2 Weeks Diet / Rabbit Eye
Proton Density Image ( 2000/12 msec
Vol. 30
2 Weeks Diet / Rabbit Eye
T 2 Weighted Image ( 2000/40 msec )
2 Weeks Diet / Rabbit Eye
Heavily T 2 Weighted Image ( 4000/80 msec ) T2 Map of Rabbit Eye / 2 Weeks Diet
Fig. 4. NMR microscopic images of the rabbit eye after 2
weeks' diet. (A) TR/TE = 2000/12 msec. (B) TR/TE
= 2000/40 msec. (C) TR/TE = 4000/80 msec. Only cataractous regions are seen in the heavily T;-weighted image.
(D) T2 map. (E) Htstological photograph of the equatorial
section of the lens.
From a series of proton density images (Figs. 2A,
3A and 4A), some changes of hydration densities between images are observable; however, the difference
between normal and cataractous tissues in each
image is not clear. In contrast to the proton density
images, the T2-weighted images (Figs. 2B, 3B, 4B and
C) show high contrast between normal and cataractous tissues. Thus, the T2-weighted imaging seems the
most promising single-acquisition technique in the
detection of early-stage cataract. Some contrast enhancement and, more importantly, a quantitative
tissue characterization can be achieved if one calculates the T2 map. However, this technique requires at
least two acquisition steps with different echo times.
Throughout these experiments, no distinct contrast
was observed by T[ relaxation, which might be re-
Downloaded From: http://iovs.arvojournals.org/pdfaccess.ashx?url=/data/journals/iovs/933378/ on 05/08/2017
No. 7
NMR MICROSCOPIC IMAGING FOR THE DETECTION OF CATARACT / Ahn er ol
Table 1. A summary of the detectability of rabbit
cataracts by NMR microscopy, slit-lamp
biomicroscopy and histology
Cataract detectability
NMR
Rabbit status
Control (2)
1 days' diet
2 days' diet
3 days' diet
4 days' diet
5 days' diet
7 days' diet
10 days'diet
14 days' diet
21 days' diet
28 days' diet
microscopy
Slit lamp
Histology
N
N
Q
Q
N
N
N
N
N
Q
D
D
D
D
N
Q (small vesicles)
Q (small vesicles)
D (vacuole)
D (vacuole)
D (vacuole)
D (large vacuole)
D (large vacuole)
D (large vacuole)
D (large vacuole)
D
D
D
D
D
D
D—detectable, Q—questionable, and N—nondetectable.
lated to the minor changes of hydration density in
galactose cataract, as noted in the above proton density images. Similar observations were reported recently.8
Histological examination of the lenses revealed the
presence of small vesicles in the equatorial region as
early as 2 days after initiation of the galactose diet.
These vesicles appeared to coalesce into larger vacuoles, still predominantly in the equatorial region by
day 10. By this time (day 10), equatorial vacuoles
were seen by slit-lamp biomicroscope. These histological observations agreed well with the images obtained by the NMR microscope.
The current work can be applicable in both theoretical and experimental studies of cataractogenesis,
cataract pharmacology and cataract animal model. In
addition to the high SNR and resolution provided by
the use of a high magnetic field, NMR microscopy
has an inherent high spectral resolution that may be
used for further studies in chemical spectroscopic
imaging and localized spectroscopy for the early detection of metabolic changes caused by cataract.
Key words: NMR microscopy, NMR imaging, galactose
cataract, resolution
Acknowledgments
The authors thank Karen Mundweiler, Allergan Inc., for
the histological preparations, and Arlene Gwon, MD, Allergan Inc., for assistance in the interpretation of the histological preparations.
1617
References
1. Atlas SW, Bilaniuk LT, Zimmerman RA, Hackney DB, Goldberg HI, and Grossman RI: Orbit: Initial experience with surface coil spin-echo MR imaging at 1.5 T. Radiology 164:501,
1987.
2. Mafee MF, Putterman A, Valvassori GE, Capek V, and
Campos M: Orbital space-occupying lesions: Role of magnetic
resonance imaging and computerized tomography: A review of
145 cases. In The Radiologic Clinics of North America: Imaging in Ophthalmology, Part I, Mafee MF, guest editor. Philadelphia, W. B. Saunders Company, 1987, pp. 529-560.
3. De Keizer RJW, Vielvoye GJ, and De Wolff-Rouendaal D:
Nuclear magnetic resonance imaging of intraocular tumors.
Am J Ophthalmol 102:438, 1986.
4. Aguayo J, Glaser B, Mildvan A, Cheng H-M, Gonzalez RG,
and Brady T: Study of vitreous liquifaction by NMR spectroscopy and imaging. Invest Ophthalmol Vis Sci 26:692, 1985.
5. Neville MC, Paterson CA, Rae JL, and Woessner DE: Nuclear
magnetic resonance studies and water "ordering" in the crystalline lens. Science 184:1072, 1974.
6. Racz P, Tompa K, and Pocsik I: The state of water in normal
and senile cataractous lenses studied by nuclear magnetic resonance. Exp Eye Res 28:129, 1979.
7. Greiner JV, Kopp SJ, Sanders DR, and Glonek T: Organophosphates of the crystalline lens: A nuclear magnetic resonance spectroscopic study. Invest Ophthalmol Vis Sci 21:700,
1981.
8. Cheng H-M, Yeh LI, Barnett P, Miglior S, Eagon JC, Gonzalez
G, and Brady TJ: Proton magnetic resonance imaging of the
ocular lens. Exp Eye Res 45:875, 1987.
9. Cho ZH, Oh CH, Kim YS, Mun CW, Nalcioglu O, Lee SJ, and
Chung MK: A new nuclear magnetic resonance imaging technique for unambiguous unidirectional measurement of flow
velocity. Appl Phys 60:1256, 1986.
10. Ahn CB, Lee SY, Nalcioglu O, and Cho ZH: An improved
nuclear magnetic resonance diffusion coefficient imaging
method using an optimized pulse sequence. Med Phys 13:789,
1986.
11. Ahn CB, Lee SY, Nalcioglu O, and Cho ZH: The effects of
random directional distributed flow in nuclear magnetic resonance imaging. Med Phys 14:43, 1987.
12. Wong EK, Cho ZH, Gardner BP, Ahn CB, Kim I, Jo JM, Juh
SC, and Anderson JA: In vivo magnetic resonance imaging
studies of galactose cataracts. ARVO Abstracts. Invest Ophthalmol Vis Sci 28(Suppl):80, 1987.
13. Cho ZH, Ahn CB, Juh SC, Anderson JA, Kim I, Wong EK,
and Garner WH: High field NMR microscopic imaging of the
early stages of galactose cataract formation. ARVO Abstracts.
Invest Ophthalmol Vis Sci 29(Suppl):3l8, 1988.
14. Cho ZH, Ahn CB, Juh SC, Lee HK, Jacobs RE, Lee S, Yi JH,
and Jo JM: NMR microscopy with 4 ^m resolution: Theoretical study and experimental results. Med Phys 15:815, 1988.
15. Kinoshita JH: Mechanisms initiating cataract formation.
Proctor Lecture. Invest Ophthalmol 13:713, 1974.
16. Gomori JM, Grossman RI, Shields JA, Augsburger JJ, Joseph
PM, and DeSimeone D: Ocular MR imaging and spectroscopy: An ex vivo study. Radiology 160:201, 1986.
17. Unakar NJ, Genyea C, Reddan JR, and Reddy VN: Ultrastructural changes during the development and reversal of galactose cataracts. Exp Eye Res 26:123, 1978.
Downloaded From: http://iovs.arvojournals.org/pdfaccess.ashx?url=/data/journals/iovs/933378/ on 05/08/2017