TENDENCY ORIENTED PERIMETRY VERSUS FASTPAC IN PATIENTS WITH NEURO-OPHTHALMOLOGICAL DEFECTS

JOSE MORALES, KEN FREEDMAN and CALEB SAWYER

Department of Ophthalmology and Visual Sciences, Texas Tech University Health Sciences Center, Lubbock, TX, USA

Abstract

Twenty-five patients consulting the neuro-ophthalmology service underwent automated static perimetry in both eyes with the Fastpac 32 (Humphrey perimeter) and tendency oriented perimetry (TOP) 32 program (Octopus 1-2-3). Both tests were given the same day. Excellent clinical agreement was found between both programs in 17 patients, moderate in six, and poor in two. Mean testing time per eye was TOP: 2:29 (SD ± 0.16) minutes; Fastpac: 9:54 (SD ± 1.10) minutes. Correlation coefficient was 0.92 (STEYX 2.52) for mean deviation-mean defect (MD). Topographical correlation was good particularly in cases with a well-defined defect. TOP ‘smoothed’ the edges of sharp scotomas and produced less profound scotomas than Fastpac, as described previously. TOP is capable of detecting abnormalities and accurately mapping well-defined neuro-ophthalmological field defects in less than a third of the time used by the Fastpac strategy. The smoother edges and shallower scotomas observed with TOP do not seem to impair the ability to make a topographical diagnosis of well-defined lesions. Neurological patients might benefit from a much shorter testing time when examined with automated perimetry.

Introduction

Manual (Goldmann) kinetic perimetry in the assessment of neuro-ophthalmological patients used to offer the flexibility of targeting the examination to certain areas of the field and shortening the examination time. Automated static perimetry (ASP) is replacing Goldmann perimetry more and more, because it offers the advantage of providing threshold quantification, it is fairly reproducible and reduces examiner variability. However, one of the main disadvantages of ASP is its lengthy examination times, which contribute to fatigue and unreliable results. This may represent more of a problem in patients with neuro-ophthalmological disease. Shorter techniques have been attempted in the past, such as suprathreshold testing1 or modified standard full threshold testing (Fastpac, 24-2 grid). Recently developed tendency oriented perimetry (TOP) can decrease testing time by as much as 80% compared to traditional full-threshold bracketing strategies.2-4 This is primarily due to the fact that significantly fewer stimuli

Address for correspondence: Jose Morales, MD, 3601 Fourth Street STOP 7217, Lubbock, TX 794307217, USA. Email: jose.morales@ttuhsc.edu

Perimetry Update 2002/2003, pp. 121–128

Proceedings of the XVth International Perimetric Society Meeting, Stratford-upon-Avon, England, June 26–29, 2002

edited by David B. Henson and Michael Wall

© 2004 Kugler Publications, The Hague, The Netherlands

are presented than with full-bracketing strategies. The purpose of the present study was to compare the utility of TOP, in patients with neuro-ophthalmological disease, to the Fastpac strategy.

Methods

A series of 25 patients were recruited from the neuro-ophthalmology service of our department. These patients had a wide variety of conditions and disease locations (optic nerve head, pre-chiasmal, chiasmal and post-chiasmal). In addition to a complete neuro-ophthalmological evaluation, each patient underwent ASP testing on the HFA-II Analyzer with the Fastpac 30-2 program and on the Octopus 1-2-3 with the TOP-32 program. Both eyes were tested in all patients, except one who happened to be monocular. The testing sequence consisted of HFA-II/Fastpac 30-2 testing followed by Octopus 1-2-3/TOP 32 testing after a rest period of at least ten minutes. No effort was made to select cooperative patients, presence of well-delineated lesions, or even a precise final diagnosis.

Analysis included topographical agreement, comparison of mean deviation-mean defect, time taken by the test, and correlation with clinical diagnosis.

Factors considered when defining agreement as excellent, moderate or poor were as follows:

•agreement on whether an abnormality existed;

•if asymmetric abnormalities existed between both eyes, whether agreement existed on which eye was more affected; and

•perfect agreement on type of topographic defect suggested.

The excellent agreement qualifier was given if the topographical representation was practically identical; moderate agreement when the same type of defect was suggested by both strategies but the extent of involvement was different, and poor agreement when different types of defects were suggested or when a diagnosis based on the topographic information was not very evident.

Results

Forty-nine pairs of fields from 25 patients were available for examination. A variety of disorders were represented (traumatic, inflammatory, degenerative, vascular) as well as different location of the abnormality (pre-, postor chiasmatic level). Defects were classified as hemianopic in nature in 13 patients, non-specific in five, nerve fiber bundle type in four, horizontal hemifield losses in two, and ceco-central in one.

Test duration

Fastpac mean duration for the test was 9:54 (S.D. ± 0.05) minutes, while TOP mean duration was only 2:29 (± 0.01) minutes. Range was 8:04-14:04 minutes for Fastpac and 2:09-3:16 minutes for TOP.

Global indices comparison

Correlation coefficient for global MD values between both programs was 0.92 (e.e. 2.52 dB; Fig. 1). Correlation coefficient for square root of loss of variance (sLV) and pattern standard deviation (PSD) was 0.85 (e.e. 1.51 dB). Mean global deviationdefect was estimated at 11.1 by Fastpac and 9.8 by TOP.

Fig. 1. Scattergram to show the mean defect-mean deviation correlation between Fastpac-32 and TOP-32.

Topography comparison

Clinical topographical correlation between both strategies, utilizing the criteria outlined above, demonstrated an excellent agreement in 17 cases, moderate agreement in six cases, and poor agreement in two cases.

The phenomenon of smoothing of the edges of sharp scotomas with TOP, already mentioned in previous studies,3,4 was confirmed during this study. This difference did not impair the capacity to make a diagnosis with TOP. We also noted that, with TOP, there was easier identification of vertical or horizontal midline respecting defects when considering the grayscale of comparisons and the comparisons value graph than when considering the grayscale of values alone (Fig. 2).

Best topographical agreement was noted between Fastpac and TOP in cases with characteristic neurological pattern scotomas such as congruent hemianopias, quadrantanopias and altitudinal defects (Fig. 3; example I). Moderate agreement was noted in some incomplete hemianopic defects or diffuse abnormalities where topographic diagnosis was still similar, but the extent of field involvement varied (Fig. 4; example II). Agreement was less consistent in cases with less characteristic changes (Fig. 5; example III).

Discussion

Automated static perimetry has several advantages over manual Goldmann kinetic perimetry: quantitative analysis, reduction of operator variability, and reproducibility

Fig. 2. Display of grayscale, comparison of values and value numeric scale demonstrating that, with TOP paying attention to these other displays, it makes it easier to identify the neuro pattern than with grayscale alone.

of technique. Until recently, an important disadvantage was its lengthy duration, which can be more of a problem with patients with neuro-ophthalmological deficits. Techniques with shorter test times than the standard full threshold technique have been utilized in the past, including suprathreshold testing, decreased set of points (24-2 test), testing only both sides of the vertical midline, or a modified threshold technique, such as Fastpac. The problem with suprathreshold testing is that it does not save that much time, and the advantage of quantitative information is lost. Testing both sides of the vertical midline gives very limited information, and has never enjoyed widespread use. We were interested in comparing the results of a more traditional thresholding technique such as Fastpac, which has frequently been combined with a reduced set of points (24-2) in neuro-ophthalmology patients. We wanted to assess the recently developed TOP strategy in this kind of population, and to determine whether

→

Fig. 4. Example II: patient with incomplete bitemporal defect. Moderate correlation with both strategies agreeing on the existence of an abnormality and its bilaterality, but not on the extent of defect, which is smaller with TOP than with Fastpac. Topographical diagnosis of bitemporal defect is still feasible.

Fig. 3. Example I: excellent correlation with both strategies agreeing on the degree of abnormality and its topographical representation.

Fig. 5. Example III: poor correlation with Fastpac suggesting an inferior nasal defect OD and TOP plotting a non-specific inferior defect of different configuration. This patient’s diagnosis was optical atrophy from optic neuritis.

the saving in time with TOP would compromise the capability of rendering an accurate topographic diagnosis. We used the 30-2 grid of Fastpac instead of the 24-2 grid, which is frequently used, because it includes exactly the same points as the TOP-32 grid, thus allowing point-by-point comparison.

The correlation coefficient for mean defect-deviation was 0.92 with an XY standard error of 2.52. The correlation coefficient from previous publications, including comparison of a variety of patients and normals, are: 0.97 for 32 grid comparison of TOP and full threshold,3 or 0.95 for G1X/G1X-TOP grid comparison.2 Although the correlation coefficient found in this sample was lower than those reported previously, we still believe this is a very acceptable correlation coefficient level. We need to take into consideration that, in this study, we were comparing results not only from two different perimetric strategies, but also from two different perimeters. Moreover, a neuroophthalmological group of patients, including hemianopic defects where TOP will smooth the edges of the scotomas, will affect the correlation coefficient.

Mean deviation-defect value was 1.2 dB lower with TOP than with Fastpac. This smaller value in global indices has been consistently and repeatedly found in different studies comparing shorter strategies with traditional bracketing strategies. It is not clear whether this is the result of an underestimation on the part of the shorter tests or whether a shorter test results in better values because of decreased fatigue.

The distribution of the deviation level demonstrated that, with TOP, it was less frequent to observe deeper deviations, partly because of the ‘softening’ of the edges of the scotomas, but also we think this is related to a different dynamic range between algorithms, with Fastpac having a wider dynamic range. The same explanation is valid for the reduced number of ‘above normal’ deviations. The intermediate defects (above 7.5 dB) were more frequent with TOP, probably due to the ‘softening of the edges’ in heminanopsias and quadrantanopsias .

It is not surprising that the best agreement regarding topographic diagnosis between both strategies occurred in cases where the scotoma was well defined and respected the vertical or horizontal mid-line better. With either instrument or strategy, it is easier to make a topographical diagnosis in such instances. TOP tends to represent scotoma as shallower and with smoother edges. Therefore, a scotoma plotted by Fastpac will appear smaller and shallower with TOP. Since there were no instances of cases with a clear topographical diagnosis by standard strategy that were missed by TOP, we believe that TOP is adequate for use as an initial tool. If there is a suggested abnormality, TOP can be repeated, or a more traditional threshold technique can be utilized if better definition, depth, and extent of the defect are needed. In neuro-ophthalmo- logical patients, it is often only the topographical location of a lesion that is needed.

We did not have the opportunity to include cases of pseudotumor cerebri with relatively stable, well-defined, and reproducible enlargements of the blind spot in this series. The two cases of pseudotumor cerebri included had generalized depression of the visual field which showed moderate agreement between strategies. We were not able to compare blind spot size between the two strategies.

More recently, other shorter strategies have been developed by Humphrey, such as the SITA algorithms.5 Although it would be interesting to compare them with TOP, a disadvantage with SITA strategies is that test time increases when there are significant visual field abnormalities or when the patient is slower. A recent report on the use of SITA in neuro-ophthalmological patients confirmed that testing time is longer and time savings are less in neuro-ophthalmological patients compared to normals, and attributed this finding to less optimal starting values in damaged fields.6 One advantage we found with TOP is that it consistently took less than 3.0 minutes, regardless of the condition, while still giving a significant amount of information. Testing time with SITA Standard has been noted to be dependent on the degree of abnormality and the consistency of the patient, ranging from 4:75 to 9:27 minutes in neuro-oph- thalmological disorders.6 Testing time also varied significantly with Fastpac in the present study from a low of 8:04 to a high of 14:04 minutes, the higher end getting closer to the time needed for a full threshold strategy.

We believe that TOP is a good alternative for the diagnosis of neuro-ophthalmo- logical defects. These kinds of patients may benefit from the shorter testing time, although it is important to bear in mind that TOP smoothes the edges of sharp scotoma. The utility of reviewing the corrected probability representation and the comparison of values grayscale enhance TOP’s usefulness in neuro-ophthalmological patients.

Acknowledgment

The authors and their families have no commercial or proprietary interest in the equipment or software described, and have no financial interest in any of the products described. Dr Morales has received speaker honoraria from Interzeag.

References

1.Siatkowski RM, Byron LL, Anderson DR, Feuer JF, Halikman AB: Automated suprathreshold static perimetry screening for detecting neuro-ophthalmologic disease. Ophthalmology 103:907-917, 1996

2.González de la Rosa M, Martinez A, Sanchez M, Mesa C, Cordovés L, Losada MJ: Accuracy of the tendency oriented perimetry (TOP) in the Octopus 1-2-3 Perimeter. In: Wall M, Heijl A (eds) Perimetry Update 1996/1997, pp 119-123. Amsterdam: Kugler Publ, 1997

3.Morales J, Weitzman ML, Gonzalez de la Rosa M: Comparison between tendency oriented perimetry (TOP) and Octopus threshold perimetry. Ophthalmology 107:134-142, 2000

4.Gonzalez de la Rosa, Morales J, Dannheim F, Papst N, Seiler T, Matsumoto C, Lachkar Y, Mermoud A, Prunte C: G1-tendency oriented perimetry versus Standard G1 bracketing in glaucoma. Eur J Ophthalmol 2002 (accepted for publication)

5.Bengtsson B, Olsson J et al: A new generation of algorithms for computerized threshold perimetry, SITA. Acta Ophthalmol Scand 75(4):368-375, 1997

6.Wall M, Punke SG, Stickney TL, Brito CF, Withrow KR, Kardon R: SITA Standard in optic neuropathies and hemianopsias: a comparison with full threshold testing. Invest Ophthalmol Vis Sci (42)2:528537, 2002