Académique Documents
Professionnel Documents
Culture Documents
Ophthalmology Contemporary
Learning Objectives: After reading this lesson, the participant should be able to:
1. Describe the differences between the Swedish interactive threshold algorithm (SITA) and full-threshold strategies.
2. List the criteria used with glaucoma progression analysis to determine progressive visual field loss in patients with
glaucoma.
3. Describe the underlying principles and clinical applications of short wavelength automated perimetry and frequency-
doubling technology perimetry.
Visual function assessment is an integral component of LD 400 (Paradigm Medical Industries), Henson Pro (Tinsley
glaucoma evaluation and management. Advances in com- Medical Instruments), Oculus Centrefield (Oculus), Med-
puter technology have allowed more sensitive and repro- mont, and Ophthimus (Hi-Tech Vision). This article concen-
ducible visual field loss detection than was possible with trates on the HFA, one of the most commonly used perime-
manual perimetry.1,2 In recent years, strategies have been ters in clinical practices.
introduced to simplify the procedure for patients and to pro-
vide results in a more timely manner. In addition, emphasis Efficient Threshold Strategies
has been placed on the early detection of glaucomatous dam- Currently, standard automated perimetry (SAP), or
age and the ability to monitor progression in order to initiate white-on-white perimetry, is the most common form of
or modify therapy, as deemed necessary. This article discuss- visual field testing performed in glaucoma practices.
es new efficient threshold strategies such as the Swedish With this procedure, achromatic stimuli of various inten-
interactive threshold algorithm (SITA) and new approaches sities are projected on a white background to determine
to identify progressive visual field loss in patients with estab- threshold sensitivity values.
lished glaucoma, including the glaucoma progression analy- Traditionally, a staircase or bracketing technique has been
sis (GPA) software for the Humphrey Field Analyzer (HFA; employed to determine the retinal threshold measurement at
Carl Zeiss Meditec, Inc.). New perimetric tests also are dis- each test point on both Octopus perimeter and HFA. With
cussed, including short-wavelength automated perimetry the full-threshold (FT) algorithm, a presumed suprathresh-
(SWAP) and frequency-doubling technology (FDT) perime- old stimulus is presented, and its value then is increased or
try, which are designed to detect early glaucomatous damage decreased by increments of 4 dB until the patient’s threshold
before it manifests on conventional perimetry. is crossed.3,4 The instrument then recrosses the threshold
Many types of perimeters are used for visual field test- with 2-dB steps to further refine the accuracy of the thresh-
ing, including HFA, Octopus (Haag-Streit USA), Dicon old values. Although these double-reversal threshold meas-
urements are more accurate, they add significantly to the
Dr. Salim is Assistant Professor of Ophthalmology, and Dr. Netland is Siegal total testing time, taking at least 15 minutes per eye to com-
Professor of Ophthalmology, Hamilton Eye Institute, University of Tennessee plete the standard 30-2 program. In patients with glaucoma
with preexisting visual field loss, the test may be prolonged,
Health Science Center, Department of Ophthalmology, 930 Madison Avenue,
Suite 470, Memphis, TN 38163; E-mail: ssalim@utmem.edu.
The authors have disclosed that they have no significant relationships with resulting in considerable fatigue and reduced reliability. In
or financial interests in any commercial organizations pertaining to this educa-
addition, FT demonstrates significant variability with
increasing deficit depth, limiting the ability to determine
tional activity.
All faculty and staff in a position to control the content of this CME activi-
ty have disclosed that they have no financial relationships with, or financial progression in these damaged areas.5
An alternative strategy, the FASTPAC threshold program
interests in, any commercial companies pertaining to this educational activity.
Lippincott CME Institute, Inc., has identified and resolved all faculty and
staff conflicts of interest regarding this educational activity. (Humphrey), was introduced. This program determines the
The continuing education activity in Contemporary Ophthalmology is intended for ophthalmologists and other physicians
with an interest in the prevention, diagnosis, and treatment of ocular disorders.
1
COP16.qxd:Layout 1 7/9/08 9:59 AM Page 2
threshold with a single reversal using an itive responses are measured using “listen
increment of 3 dB. Although the FASTPAC time data,” in contrast to the use of tradi-
strategy cuts the total time in half compared tional catch trials. With this technique,
EDITOR
with FT, it has been reported to underesti- periods during the test when no positive
C. Stephen Foster, M.D.,
mate the severity of visual field defects.6 answers are expected are used to establish
FACS, FACR*
FT and FASTPAC, SITA, for use on the responses. One such period is the time
Ophthalmology
total testing time significantly compared The use of data from many of these peri-
Cambridge, MA
with traditional strategies, without com- ods compared with a limited number of
Founder and CEO
promising the accuracy and reliability of catch trials improves overall accuracy with
Ocular Immunology and
the data.7,8 SITA has demonstrated similar reduction in total testing time.
Uveitis Foundation
test–retest variability compared with FT.5 The SITA is available as SITA Standard
Cambridge, MA
It has been suggested that SITA may be (SS) and SITA Fast (SF), with mean test
ASSOCIATE EDITOR
seen by age-matched normal individuals. normal. SITA can be performed for the cen-
Research/Education Coordinator
Subsequent presentations of the target are tral 10-2, 24-2, and 30-2 programs (Figure
Massachusetts Eye Research
difference used with FT. In addition, SITA fluctuation and corrected pattern standard
Richard L. Anderson, M.D.
does not involve double threshold deter- deviation values are not recorded. The
Salt Lake City, UT
determine short-term fluctuation. All data tional strategies, and this has been attrib-
Thessaloniki, Greece
analysis is performed with StatPac statis- uted to improved sensitivity detection from
William Ayliffe, M.D.
tical software (StatPac, Inc.). reduced fatigue due to shorter test time.5,9
London, England
SITA employs a new method to estimate SF, although requiring shorter test time,
Dimitri T. Azar, M.D.
Chicago, IL
2
COP16.qxd:Layout 1 7/9/08 9:59 AM Page 3
is used, rather than SF, for long-term monitoring of patients lished with an average of two tests. Up to 14 follow-up
with glaucoma. tests may be compared with the baseline. If the examiner
does not choose baseline fields, the GPA automatically
Glaucoma Progression Analysis selects baseline tests, with preference for SITA tests. If
Detection of progressive visual field loss in patients FT tests are used as baseline, follow-up tests must be
with glaucoma continues to be a challenge for clinicians. either SS or SF. Once a certain number of SITA tests are
Several methods have been used to judge progression, available, a new baseline should be established with one
including point-wise comparison, glaucoma change prob- of the SITA versions and the previous baseline with FT
ability analysis (GCPA) for FT tests, various algorithms should be deleted. GPA verifies examination reliability
used in multicenter clinical trials, and analysis of global and excludes unreliable tests from analysis. GPA also
indices, such as mean deviation. 10 provides mean deviation graph and linear regression
Nouri-Mahdavi et al. compared point-wise linear regres-
11 analysis to help verify the similarity between two base-
sion, GCPA, and the Advanced Glaucoma Intervention line visual fields.
Study (AGIS) method to predict visual field progression Three types of printouts are available with GPA, includ-
using the longitudinal visual field data from AGIS. For both ing the single field analysis with GPA summary (Figure 2),
point-wise linear regression and GCPA analyses, the criteri- GPA baseline printout, and GPA follow-up printout. The
on for progression was defined as the presence of three or follow-up printout contains four columns of data for each
more test locations that demonstrated worsening. These test, including the graytone, pattern deviation, deviation
points were not necessarily contiguous, and the changes were from baseline, and progression analyses (Figure 3). The
confirmed on three consecutive visual fields. For the AGIS deviation from the baseline plot compares the pattern
criteria, progression was defined as an increment of four or deviation of a test point to the average of the pattern of
more points greater than the baseline value, with confirma- deviation values from the baseline tests. The change at
tion requiring three subsequent field tests. The authors con- each test point is recorded in dB notation. A zero value
cluded that all algorithms had low false-prediction rates and means no change from the baseline, whereas a –4 value
that different methods might be more appropriate at different indicates that the current tested point is 4 dB lower than
stages of follow-up. For example, point-wise linear regres- the pattern deviation for the same point in the baseline.
sion is sensitive to the number of points available for analy- The progression analysis plot demonstrates statistical sig-
sis; therefore, it performs well once a number of visual fields nificance of the decibel changes and uses a set of symbols to
are available for interpretation. GCPA, on the other hand, identify a point that exceeds the expected variability. A single
may predict outcomes better during early follow-up. white dot represents no change. An empty triangle indicates a
A new change analysis program, GPA, has been intro- point that has changed compared against the previous base-
duced on the HFA 700 series.12 This software helps to line test. A half-filled triangle indicates that change at that test
3
COP16.qxd:Layout 1 7/9/08 9:59 AM Page 4
point has been documented on two consecutive tests. A filled functional loss may improve.17–19 These theories have led to
triangle indicates presence of change at that test point on three the development of new perimetric tests, such as SWAP and
consecutive tests. All triangles identify points that have FDT perimetry.
changed by an amount that is significant at the p < 0.05 level. Retinal ganglion cells have been classified into several
X indicates a point that has data out of range for analysis. types, including parvocellular, magnocellular, and konio-
When applicable, the GPA alert appears. When significant cellular.20–22 The most abundant of these are the parvocel-
change is seen in the same three or more points on two con- lular cells, constituting about 80% of total cell population.
secutive tests, the GPA interprets the pattern as “possible pro- The parvocellular pathway transmits information mostly
gression.” When significant degradation is seen in three or about color and form. The magnocellular and koniocellular
more points on three consecutive tests, the GPA interprets the cells represent the remaining 20%. The magnocellular path-
pattern as “likely progression.” These three points do not have way is responsible for transmitting information regarding
to be clustered to document progression. flicker or motion, whereas the koniocellular pathway is
Although GPA simplifies and standardizes analysis for involved with transmission of short or blue wavelength.
change in glaucomatous visual fields, its significance SWAP selectively measures the blue wavelength func-
should be interpreted in the full context of a patient’s clini- tion by using a large size (V) blue stimulus (440 nm) on a
cal situation when determining the effectiveness of an yellow background (100 cd/m2), taking advantage of the
existing therapy or a need for more aggressive intervention. known glaucoma-induced color vision deficit.23 The yel-
Bengtsson et al. introduced a new perimetric index, the
13
low background desensitizes the green and red cones,
glaucoma progression index (GPI), to calculate the rate of sparing blue cones. The transmission of this short wave-
glaucomatous progression. GPI is calculated based on sig- length is thought to be mediated by the koniocellular path-
nificantly depressed points in the pattern deviation probabil- ways. SWAP is available on HFA II series and Octopus 1-
ity maps and is recorded in percentage after these values are 2-3 perimeters. With the exceptions of a different color
4
COP16.qxd:Layout 1 7/9/08 9:59 AM Page 5
a single or serial field overview can be printed with the 16. Glovinsky Y, Quigley HA, Dunkelberger GR. Retinal ganglion cell loss is
standard format of HFA printouts. size dependent in experimental glaucoma. Invest Ophthalmol Vis Sci
1991;32:484-491.
Conclusion
17. Johnson CA, Adams AJ, Casson EJ, Brandt JD. Progression of early glau-
comatous visual field loss for blue-on-yellow and standard white-on-
Automated perimetry has significantly contributed to
white automated perimetry. Arch Ophthalmol 1993;111:651-656.
18. Johnson CA, Adams AJ, Casson EJ, Brandt JD. Blue-on-yellow perimetry
our understanding of visual function deficits in glaucoma can predict the development of glaucomatous visual field loss. Arch
patients. Since its inception nearly three decades ago, Ophthalmol 1993;111:645-650.
advances in computer technology with more complex
19. Johnson CA, Samuels SJ. Screening for glaucomatous visual field loss
with frequency-doubling perimetry. Invest Ophthalmol Vis Sci 1997;
stimuli and efficient strategies have provided more accu- 38:413-425.
rate and reproducible results as well as the flexibility to 20. Kaplan E, Benardete E. The dynamics of primate retinal ganglion cells.
compare these results over time. However, there is still a Prog Brain Res 2001;134:17-34.
21. Curcio CA, Allen KA. Topography of ganglion cells in human retina. J
tremendous challenge to find an ideal perimetric test that Comp Neurol 1990;300:5-25.
is reliable, efficient, and truly reflects underlying neural 22. Johnson CA. Selective versus nonselective losses in glaucoma. J Glaucoma
damage and progression. Although new technologies 1994;3:532-544.
23. Sample PA. Short-wavelength automated perimetry: its role in the clinic
seem promising, further refinement in clinical perimetry and for understanding ganglion cell function. Prog Retin Eye Res
is an important task and will continue to be an active 2000;19:369-383.
focus of research for the ophthalmic community. 24. Johnson CA, Adams AJ, Casson EJ, Brandt JD. Progression of early glau-
comatous visual field loss for blue-on-yellow and standard white-on-
white automated perimetry. Arch Ophthalmol 1993;111:651-656.
REFERENCES 25. Demirel S, Johnson CA. Incidence and prevalence of short wavelength
1. Katz J, Tielsch JM, Quigley HA, et al. Automated perimetry detects visu- automated perimetry deficits in ocular hypertensive patients. Am J
al field loss before manual Goldmann perimetry. Ophthalmology Ophthalmol 2001;131:709-715.
1995;102:21. 26. Remky A, Lichtenberg K, Elsner AE, Arend O. Short wavelength automated
2. Agarwal HC, Gulati V, Sihota R. Visual field assessment in glaucoma: perimetry in age related maculopathy. Br J Ophthalmol 2001;85:1432-1436.
comparative evaluation of manual kinetic Goldmann perimetry and auto- 27. Keltner JL, Johnson CA. Short-wavelength automated perimetry in neuro-
mated static perimetry. Indian J Ophthalmol 2000;48:301. ophthalmologic disorders. Arch Ophthalmol 1995;113:475-481.
3. Anderson DR, Patella VN. Automated Static Perimetry, 2nd ed. St. Louis, 28. Remky A, Arend O, Hendricks S. Short-wavelength automated perimetry
Mosby, 1999. and capillary density in early diabetic maculopathy. Invest Ophthalmol Vis
4. Allingham RR, ed. Shields Textbook of Glaucoma, 5th ed. Philadelphia: Sci 2000;41:274-281.
Lippincott Williams & Wilkins, 2005. 29. Horn FK, Brenning A, Jünemann AG, Lausen B. Glaucoma detection with
5. Wild JM, Pacey IE, Hancock SA, Cunliffe IA. Between-algorithm, between- frequency doubling perimetry and short-wavelength perimetry. J
individual, differences in normal perimetric sensitivity: full threshold, FAST- Glaucoma 2007;16:363-371.
PAC, and SITA. Invest Ophthalmol Vis Sci 1999;40:1152-1161. 30. Bengtsson B. A new rapid algorithm for short-wavelength automated
6. Spry PGD, Johnson CA. Advances in Automated Perimetry. Vol 20, Module 10, perimetry. Invest Ophthalmol Vis Sci 2003;44:1388-1394.
in: Focal Points 2002: Clinical Modules for Ophthalmologists. San Francisco: 31. Bengtsson B, Heijl A. Normal intersubject threshold variability and nor-
American Academy of Ophthalmology, 2002. mal limits of the SITA SWAP and full threshold SWAP perimetric pro-
7. Bengtsson B, Heijl A. Evaluation of a new perimetric strategy, SITA, in grams. Invest Ophthalmol Vis Sci 2003;44:5029-5034.
patients with manifest and suspect glaucoma. Acta Ophthalmol Scand 32. Anderson AJ, Johnson CA. Frequency-doubling technology perimetry.
1998;76:368-375. Ophthalmol Clin N Am 2003;16:213-225.
8. Olsson J, Bengtsson B, Heijl A, Rootzén H. An improved method to esti- 33. Maddess T, Henry GH. Performance of nonlinear visual units in ocular
mate frequency of false positive answers in computerized perimetry. Acta hypertension and glaucoma. Clin Vis Sci 1992;7371-7383.
Ophthalmol Scand 1997;75:181-183. 34. White AJ, Sun H, Swanson WH, Lee BB. An examination of physiologi-
9. Artes PH, Iwase A, Ohno Y, Kitazawa Y, Chauhan BC. Properties of peri- cal mechanisms underlying the frequency-doubling illusion. Invest
metric threshold estimates from full threshold, SITA standard, and SITA Ophthalmol Vis Sci 2002;43:3590-3599.
fast strategies. Invest Ophthalmol Vis Sci 2002;43:2654-2659. 35. Ferreras A, Polo V, Larrosa JM. Can frequency-doubling technology and
10. Vesti E, Johnson CA, Chauhan BC. Comparison of different methods for short wavelength automated perimetries detect visual field defects before
detecting glaucomatous visual field progression. Invest Ophthalmol Vis standard automated perimetry in patients with preperimetric glaucoma? J
Sci 2003;44:3873-3879. Glaucoma 2007;16:372-383.
11. Nouri-Mahdavi K, Hoffman D, Ralli M, Caprioli J. Comparison of meth- 36. Medeiros FA, Sample PA, Weinreb RN. Frequency doubling technology
ods to predict visual field progression in glaucoma. Arch Ophthalmol perimetry abnormalities as predictors of glaucomatous visual field loss.
2007;125:1176-1181. Am J Ophthalmol 2004;137:863-871.
12. Carl Zeiss Meditec Inc. Humphrey Field Analyzer II—I Series User’s 37. Spry PGD, Johnson CA. Variability components of standard automated
Guide. Dublin, CA: Carl Zeiss Meditec, 2003. perimetry and frequency doubling technology perimetry. Invest Ophthalmol
13. Bengtsson B, Heijl A. A visual field index for calculation of glaucoma rate Vis Sci 2001;42:1404-1410.
of progression. Am J Ophthalmol 2008;145:343-353. 38. Cioffi GA, Mansberger S, Spry P, Johnson C, Van Buskirk EM. Frequency
14. Kerrigan-Baumrind LA, Quigley HA, Pease ME, Kerrigan DF, Mitchell RS. doubling perimetry and the detection of eye disease in the community.
Number of ganglion cells in glaucomatous eyes compared with threshold visu- Trans Am Ophthalmol Soc 2000;98:195-202.
al field tests in the same persons. Invest Ophthalmol Vis Sci 2000;41:741-748. 39. Wall M, Neahring RK, Woodward KR. Sensitivity and specificity of fre-
15. Quigley HA, Dunkelberger GR, Green WR. Chronic human glaucoma quency of doubling perimetry in neuro-ophthalmic disorders: a compari-
causing selectively greater loss of large optic nerve fibers. Ophthalmology son with conventional automated perimetry. Invest Ophthalmol Vis Sci
1988;95:357-363. 2002;43:1277-1283.
7
COP16.qxd:Layout 1 7/9/08 9:59 AM Page 8
1. Which one of the following statements regarding threshold 6. Which one of the following statements regarding the inter-
strategies used in automated perimetry is true? pretation of GPA is false?
A. The full-threshold (FT) algorithm determines the thresh- A. A small open triangle identifies a change that was not seen
old by double reversal using 4-dB and 2-dB increments. on the previous follow-up test.
B. FASTPAC strategy determines threshold by single B. A filled triangle identifies change at p < 0.05 that is repeat-
reversal in 3-dB increments. ed in three consecutive tests
C. The Swedish interactive threshold algorithm (SITA) uses C. GPA alert appears when significant degradation is seen
the technique of visual field modeling to estimate thresh- in three or more points on two or three consecutive tests
old values by extrapolating information from adjacent test D. Three points must be clustered to satisfy the criterion
points. for progression.
D. All of the above
7. Which of the following statements regarding retinal ganglion
2. Which one of the following statements regarding SITA is false? cells is false?
A. SITA compromises the accuracy and reliability of the visu- A. There is considerable redundancy in the visual system.
al field defect by employing a “best guess” approach to B. Magnocellular cells are involved in the doubling per-
threshold estimation. cept of frequency doubling.
B. SITA strategies generally have higher values for differen- C. Parvocellular cells are sparsely distributed.
tial light sensitivity compared with traditional algorithms. D. The koniocellular pathway transmits short wavelength.
C. SITA Standard takes only half as much time as FT
8. Which of the following statements regarding SWAP is/are
perimetry.
true?
D. SITA Fast takes only half as much time as FASTPAC
A. SWAP may detect visual field deficits up to 5 years earlier
perimetry.
than SAP by selectively targeting a subset of ganglion cells.
3. Which one of the following statements regarding SITA strat- B. SWAP uses a yellow background to allow adaptation of
egy is false? long and medium wavelength mechanisms.
A. SITA can be used for short-wavelength automated perime- C. SWAP is more affected by ocular media opacities than SAP.
try (SWAP) testing. D. All of the above
B. SITA is available only for central 24-2 and 30-2 programs.
9. Which one of the following statements regarding frequency-
C. SITA does not calculate short-term fluctuation.
doubling technology (FDT) is false?
D. SITA is available only on HFA II.
A. FDT is more affected by inherent and acquired color vision
4. Which one of the following statements regarding glaucoma deficits than SWAP.
progression analysis (GPA) is false? B. FDT demonstrates less test–retest variability compared
A. GPA simplifies analysis to document change in glauco- with standard automated perimetry (SAP).
matous visual fields. C. FDT is a faster procedure than SWAP.
B. GPA uses data of expected range of variability from D. FDT poorly defines hemianopic defects compared with
established patients with glaucoma. SAP.
C. GPA provides simple plain-language alert.
10. Which one of the following statements regarding FDT is true?
D. GPA does not correct for media opacities.
A. Only screening tests can be performed.
5. Which one of the following statements regarding GPA is true? B. Reliability indices are omitted given its short test duration.
A. Only FTs can be used for baseline tests. C. FDT is more resilient to refractive error compared with SAP.
B. Only SITA versions can be used for follow-up tests. D. FDT utilizes the parvocellular pathway sensitive to color
C. GPA does not verify unreliable tests. and form.
D. GPA can be used with the SITA-SWAP test.