ACTA OPHTHALMOLOGICA SCANDINAVICA 2006
Sensitivity and specificity of frequency-doubling technology, tendency-oriented perimetry, SITA Standard and SITA Fast perimetry in perimetrically inexperienced individuals Paulo de Tarso Ponte Pierre-Filho,1 Rui Barroso Schimiti,1 Jose´ Paulo Cabral de Vasconcellos1 and Vital Paulino Costa1,2 1
Glaucoma Service, Department of Ophthalmology, University of Campinas, Sa˜o Paulo, Brazil 2 Glaucoma Service, Department of Ophthalmology, University of Sa˜o Paulo, Sa˜o Paulo, Brazil
ABSTRACT. Purpose: To evaluate the sensitivity and specificity of the screening modes of frequency-doubling technology (FDT), tendency-oriented perimetry (TOP), SITA Standard (SS) and SITA Fast (SF) in perimetrically inexperienced individuals. Methods: One eye of 64 glaucoma patients and 53 normal subjects who had never undergone automated perimetry were tested with programs C-20-5 (FDT), G1 (TOP) and 24-2 (SS and SF). The gold standard for glaucoma was the presence of a typical glaucomatous optic disc appearance on stereoscopic examination (judged by a glaucoma expert), and intraocular pressure (IOP) > 21 mmHg. The test order among strategies was randomized for each subject. To define an abnormal visual field, we applied three criteria for SS and SF and two criteria for TOP and FDT, all of which have been previously described in the literature. Sensitivities and specificities among the different criteria were compared using the Cochran test. Results: Frequency-doubling technology showed the shortest mean test duration, followed by TOP, SF and SS (p < 0.05). Sensitivity ranges were 87.5–89.1% for SS, 92.2–93.8% for SF, 87.5–89.1% for TOP, and 82.8–85.9% for FDT (p = 0.34). Specificity ranges were 73.6–83% for FDT, 56.6–62.3% for TOP, 60.4–69.8% for SF and 66.0–71.7% for SS. The specificity obtained with criterion 2 for FDT (based on the presence of two or more abnormal locations regardless of the severity of abnormal points) was higher than those measured with the other strategies (p < 0.01). Conclusion: When testing individuals with no perimetric experience, moderate sensitivities and specificities should be expected, regardless of the strategy chosen. Key words: visual field – perimetry – sensitivity – specificity – glaucoma – screening
Acta Ophthalmol. Scand. 2006: 84: 345–350 Copyright # Acta Ophthalmol Scand 2006.
doi: 10.1111/j.1600-0420.2005.00639.x
Introduction Glaucoma is a significant cause of blindness worldwide. The estimated number of people suffering from glaucoma in the year 2000 was approximately 67 million, 10% of whom were bilaterally blind (Quigley 1996). The fact that glaucoma blindness is irreversible, but often develops after a slowly progressive asymptomatic stage, calls for the development of improved methods of early detection (Quigley 1996). Although early detection of glaucoma is fundamental to reducing progression to blindness, an optimal screening method has remained elusive. Screening methods need to be accurate, rapid, inexpensive, easily administered, reliable, and highly sensitive and specific (Katz et al. 1993; Quigley 1996). Although direct optic disc observation by glaucoma specialists is considered an ideal method, applying this method to large mass screening is impossible. As an alternative, screening methods using simple and rapid visual field (VF) tests have been suggested (Katz et al. 1993; Sponsel et al. 1995). Automated perimetry has been the gold standard for the detection and
345
ACTA OPHTHALMOLOGICA SCANDINAVICA 2006 monitoring of functional loss in glaucoma for the past three decades (Fankhauser et al. 1972; Bebie et al. 1976). Numerous attempts have been made to improve the sensitivity, specificity and speed of data acquisition, in order to help reduce the fatigue induced by the test (Bengtsson et al. 1997; Bengtsson & Heijl 1998a; Johnson 2002). The Humphrey Swedish Interactive Threshold Algorithm (SITA) has greatly reduced testing time without reducing data quality, compared with the full-threshold (FT) strategy (Bengtsson et al. 1997). The SITA includes two different programs: SITA Standard (SS) and SITA Fast (SF). The SS strategy test time is nearly 50% shorter than that of FT (Bengtsson et al. 1997), while SF is even more rapid than SS (Bengtsson & Heijl 1998a). Another fast strategy called tendency-oriented perimetry (TOP) was introduced in Octopus perimeters to obtain an approximation of the VF threshold in a much shorter period of time than traditional strategies. The TOP strategy is based on testing each position only once, but interpolating the information to the surrounding areas (Morales et al. 2000; Gonzales de la Rosa et al. 2003). Frequency-doubling technology (FDT) presents stimuli on a black-andwhite video monitor. The stimulus is a low spatial frequency sinusoidal grating that undergoes a 25-Hz counterphase flicker. Seventeen VF locations are evaluated in a short testing time (Kalaboukhova & Lindblom 2003). Furthermore, FDT may detect functional loss in glaucoma patients earlier than standard white-on-white perimetry (Quigley 1998; Medeiros et al. 2004). The purpose of our study was to examine the sensitivity and specificity of four rapid VF tests (FDT, TOP, SF and SS) in normal subjects and glaucoma patients who had never undergone automated perimetry.
Materials and Methods The study was approved by the Ethics Committee of the University of Campinas. Written informed consent was then obtained from consecutive subjects selected from February to
346
November 2004. After a complete ophthalmological examination, including visual acuity (VA) measurement, slit-lamp biomicroscopy, Goldmann applanation tonometry, gonioscopy and optic disc evaluation (with the pupils dilated and using a 78-dioptre lens), patients were included in the study if appropriate criteria were met. One eye per subject was randomly included in the analysis. All recruited subjects were 18 years or older, had not previously undergone automated perimetry, had corrected VA of 20/50 and a spherical equivalent of 5 D. Subjects were excluded if they presented with a history of systemic or ocular disease other than glaucoma that might interfere with visual field results. We also excluded pseudophakic eyes and those with significant cataract greater than moderate lens opacification, according to the Lens Opacity Classification System III (Chylack et al. 1993). The diagnosis of glaucoma was based on the presence of typical glaucomatous optic disc damage (i.e. cup : disc ratio 0.6, localized rim loss, optic disc haemorrhage, or cupto-disc asymmetry > 0.2), and intraocular pressure (IOP) > 21 mmHg. Normal subjects were recruited from volunteers among spouses, friends, relatives or accompanying persons of patients and had IOP < 21 mmHg, optic disc examination graded as normal and no family history of glaucoma. Visual field data were not used as inclusion criteria for either group, as this would have generated an important source of bias, artificially increasing the sensitivity and specificity of VF tests (Garway-Heath & Hitchings 1998). After appropriate instruction, all participants were tested with the FDT C-20-5 program (Zeiss Humphrey Systems, Dublin, California, USA), the Octopus 301 G1-TOP program (Interzeag, Schlieren, Switzerland), and the SS and SF 24-2 program using the Humphrey Field Analyser II (model 750, Version A-10; Zeiss Humphrey Systems, Dublin, California, USA) with appropriate refractive correction (except for in FDT), during a single visit on the same day. The order of tests was randomly chosen for each subject, with an interval of approximately 15–30 mins between each test. The technician
conducting VF examinations was masked to the underlying diagnosis. In the FDT C-20-5 program, the test stimuli are presented at contrast levels corresponding to the normal 5% of probability level. If the stimulus is detected, no further testing is performed in that location. If the stimulus is missed, a repeat presentation at the 5% probability level is made at the same location. If it is missed again, a stimulus at the 2% probability level is presented, and if this is missed, a 1% probability level stimulus is presented. The result at each location is labelled to one of the following values: p 5% (normal), p < 5% (mild defect), p < 2% (moderate defect), p < 1% (severe defect) (Johnson & Samuels 1997; Joson et al. 2002; Anderson & Johnson 2003). Two different criteria (previously described in the literature) were used to define an abnormal FDT test: (1) the presence of at least one abnormal location (Burnstein et al. 2000), and (2) the presence of two or more abnormal locations regardless of the severity of abnormal points (Quigley 1998). The TOP strategy tests each point only once, but each point is affected by the responses of the surrounding points to reach the final threshold approximation. The TOP examinations were performed with an Octopus 301 perimeter connected to and controlled by an external personal computer (PC). The TOP test was considered abnormal if: (1) the mean defect (MDe) > 2 dB and/or the loss variance (LV) > 6 dB, and (2) there were at least seven points (three of them contiguous) with a reduction in sensitivity 5 dB in the corrected comparisons graphic (Sponsel et al. 1995; Morales et al. 2000; Wadood et al. 2002). The new SITA strategies use models of normal and typically altered VFs. The threshold values and measurement errors are constantly estimated during the test, which is interrupted when measurement errors have been reduced to a predetermined level (Bayesian posterior probability function) (Bengtsson et al. 1997). Sita Standard and Sita Fast were performed using a size III white stimulus on a white background (31.5 apostilbs or 10 cd/m2). Results for SS and SF were considered abnormal if:
ACTA OPHTHALMOLOGICA SCANDINAVICA 2006 (1) the glaucoma hemifield test (GHT) was borderline or outside normal limits; (2) the pattern deviation probability map showed a cluster of three or more non-edge points deviating at p < 5%, with one of them deviating at p < 1%, or (3) the pattern standard deviation (PSD) was increased to values deviating at p < 5% (Wadood et al. 2002). Sita Standard and Sita Fast tests were considered reliable if fixation losses were < 20%, and false-positive and false-negative answers were < 33%. Tendency-oriented perimetry VFs with > 33% false-positive and/or false-negative responses or a reliability factor > 15% were considered unreliable. Frequency-doubling technology was considered unreliable when falsepositive responses and/or fixation errors were > 33.3%. After all examinations had been performed, glaucoma subjects were further classified according to the mean deviation (MD) obtained with SS as having early (MD > 6 dB), moderate ( 6 dB MD > 12 dB), or severe (MD 12 dB) glaucoma (Iwasaki et al. 2002). Initially, a global analysis was performed, comparing the results of all tests, regardless of the order in which they were applied. A second analysis, which included the results of only the first examination of each patient, was performed in order to exclude the possible learning effect observed in subsequent examinations. The sensitivity and specificity of the above two criteria for FDT, two criteria for TOP, and three criteria for SS and SF were compared. Statistical analyses were performed with SAS Version 8.20 (SAS Institute Inc., Cary, North Carolina, USA). Demographic and clinical characteristics between the two groups were compared using chi-squared and twotailed Student’s t-tests for categorical and continuous variables, respectively. Differences among proportions were determined using the chi-squared test, Fisher’s exact test or the Cochran test as appropriate. Times for test completion (which did not include the time spent entering data, instructing the patient and printing out the result) were compared using the Friedman test or Student’s t-test. A p-value < 0.05 was considered statistically significant. Data are reported as mean SD where applicable.
Results A total of 117 eyes, 53 normal and 64 glaucomatous, were enrolled in this study. Among the 64 glaucomatous eyes, 21 (32.8%), 12 (18.8%) and 31 (48.4%) were classified as having early, moderate and severe glaucoma, respectively. Demographic data for the study subjects are listed in Table 1. The mean age in the glaucoma group (57.6 13.1 years, range 33–81 years) was significantly higher than in the normal group (45.9 13.8 years, range 21–75 years) (p < 0.0001). Black people accounted for 24.8% of the individuals and were more frequent in the glaucoma group (p < 0.05). The cup : disc ratio was significantly higher among glaucoma patients than in the normal group (p < 0.0001). The mean test duration was 6.58 1.30 mins for SS, 4.18 1.06 mins for SF, 3.00 0.94 mins for TOP, and 1.29 0.57 mins for FDT. The mean test time was significantly lower for FDT (p < 0.0001). Glaucomatous patients had longer test times and more unreliable visual fields than normal subjects for all tests, even when we controlled for differences in age (Tables 2 and 3). The Cochran test showed significant differences in reliability indices among the different tests, with TOP presenting more unreliable results in both normal subjects (p ¼ 0.019) and glaucoma patients (p < 0.0001) (Table 3). The sensitivity and the specificity of each criterion for each test are shown in Table 4. When all tests were evaluated using the previously described criteria, regardless of the order in which they were performed, sensitivities ranged as follows: 82.885.9% for FDT; 87.589.1% for TOP; 92.293.8% for SF, and 87.589.1% for SS. There was no statistically significant difference between the sensitivities obtained with each perimetric test (p ¼ 0.34). Specificities varied as follows: 73.6 83% for FDT; 56.662.3% for
TOP; 60.469.8% for SF, and 66.0 71.7% for SS. The specificity obtained with criterion 2 for FDT (based on the presence of two or more abnormal locations regardless of the severity of abnormal points) was higher than those measured with the other strategies (p < 0.01). There was no significant difference regarding the distribution of early, moderate and severe glaucoma among patients undergoing the different examinations as a first test (p ¼ 0.223) (Table 5). Table 6 shows the sensitivity and specificity of each test considering only the first examinations of each patient. When only the first tests were evaluated using the criteria described previously, sensitivities were 78.6% for FDT, 94.1% for TOP, 89.5% for SS, and in the range of 92.7100% for SF. There was no statistically significant difference between the sensitivities obtained with each perimetric test (p ¼ 0.63). Specificities were 76.9% for FDT and in the range of 53.9 61.5% for TOP, 57.171.4% for SF and 76.984.6% for SS. There was no statistically significant difference between the specificities obtained with each perimetric test (p ¼ 0.65).
Discussion Automated perimetry has become an essential component in the successful diagnosis and follow-up of glaucoma. Some studies have been designed to evaluate the sensitivity and specificity of different visual field tests in perimetrically experienced patients (Burnstein et al. 2000; Wadood et al. 2002; Kalaboukhova & Lindblom 2003). It has been shown that individuals with no perimetric experience may exhibit a learning effect, thus reducing test accuracy and reliability (Joson et al. 2002; Budenz et al. 2002; Schimiti et al. 2002). When performing large-scale, mass screening for glaucoma, researchers often face a number of individuals
Table 1. Demographic data.
Number (n) Sex (M/F) Age (years SD) Race (Black/White/Asians) Cup : disc ratio
Normal subjects
Glaucoma patients
p-value
53 25/28 45.9 13.8 8/44/1 0.25 0.07
64 27/3 57.6 13.1 21/43/0 0.81 0.14
0.59 < 0.0001 0.03 < 0.0001
347
ACTA OPHTHALMOLOGICA SCANDINAVICA 2006
Table 2. Average time (minutes) required for each test (ANCOVA corrected for age). Total (n ¼ 117) FDT TOP SF SS
1.29 3.00 4.18 6.58
0.57 0.94 1.06 1.30
Normal subjects (n ¼ 53) 0.92 2.64 3.57 5.86
0.33 0.57 0.79 0.95
Glaucoma patients (n ¼ 64)
p-value
1.61 0.55 3.3 1.08 4.73 0.96 7.18 1.28
< 0.0001 0.0007 < 0.0001 < 0.0001
FDT ¼ frequency-doubling technology; TOP ¼ tendency-oriented perimetry; SF ¼ Humphrey SITA Fast; SS ¼ Humphrey SITA Standard. Table 3. Percentage of reliable visual field tests in normal subjects and glaucoma patients. (corrected for age by logistic regression).
FDT TOP SF SS
Normal subjects (n ¼ 53)
Glaucoma patients (n ¼ 64)
p-value
96.2% 83.0% 94.2% 98.1%
92.2% 56.2% 93.7% 93.7%
0.79 0.06 0.44 0.29
FDT ¼ frequency-doubling technology. TOP ¼ tendency-oriented perimetry. SF ¼ Humphrey SITA Fast. SS ¼ Humphrey SITA Standard.
with no previous perimetric experience. In this study, we compared the diagnostic performance of the FDT screening mode C-20-5, the Octopus G1-TOP program, and the SS and SF program 24–2 for the Humphrey Field Analyser II in subjects undergoing automated perimetry for the first time. The new perimetric tests FDT, TOP, SF and SS may offer important advantages over standard automated perimetry because of the reduced time required to administer the test, especially for elderly people (Bengtsson & Heijl 1998b; Budenz et al. 2002; Delgado et al. 2002; Iwasaki et al. 2002; Schimiti et al. 2002; Wadood et al. 2002; Kalaboukhova &
Lindblom 2003). In this study, SS had the longest test duration and FDT was the fastest (p < 0.0001). The shorter time required for FDT may be explained by the 17 test locations for FDT as compared with the 54 points evaluated with TOP, SF and SS. Moreover, FDT employs a suprathreshold screening strategy, whereas the other strategies aim to determine the threshold of each tested location. Surprisingly, the TOP strategy showed the lowest number of reliable tests in both normal subjects and glaucoma patients. Different methodologies used to derive reliability indices may explain discrepancies between reliability estimates.
Table 4. Sensitivities and specificities with their respective confidence intervals for each criterion of each test. Test*
FDT (1) FDT (2) TOP (1) TOP (2) SF (1) SF (2) SF (3) SS (1) SS (2) SS (3) p-value
Sensitivity
Specificity
%
(95% CI)
%
(95% CI)
85.9 82.8 87.5 89.1 93.8 92.2 92.2 89.1 89.1 87.5 0.34
(74.5–93.0) (70.9–90.7) (76.3–94.1) (78.2–95.1) (84.0–98.0) (82.0–97.1) (82.0–97.1) (78.2–95.1) (78.2–95.1) (76.3–94.1)
73.6 83.0 56.6 62.3 60.4 67.8 69.8 66.0 66.0 71.7 < 0.01
(59.4–84.3) (69.7–91.5) (42.4–69.9) (47.9–74.9) (63.0–82.9) (53.6–79.7) (55.5–81.3) (51.6–78.1) (51.6–78.1) (57.4–82.8)
* Refer to text for details of criteria for each test. CI ¼ confidence interval.
348
In addition to being rapid, a screening test needs to be associated with high sensitivity and specificity. Table 7 lists the sensitivities and specificities obtained with all tested instruments in previous studies with perimetrically experienced individuals. Although comparisons with these studies are limited by different sample characteristics and different inclusion criteria, it is evident that their sensitivities and specificities are generally higher than those reported herein (Tables 4 and 6), which suggests that perimetrically inexperienced individuals do not perform as well as experienced patients. Despite the widespread adoption of the SITAs for detection and follow-up of glaucomatous visual field defects, relatively little has been published on the sensitivity of these two algorithms when the gold standard for glaucoma is based on optic disc examination. Although SS was not designed specifically for screening purposes, its short test duration prompted us to include it in this comparison. In one of the few studies that compared FT and SS in 80 normal individuals with no perimetric experience, using criteria described by Anderson & Patella (1992) to define abnormality, the specificity of SS varied from 60% to 77.5% when each single criterion was analysed, and was 50% when the fulfilment of any criteria was enough to define abnormality (Schimiti et al. 2002). These findings were confirmed in our study, where specificities obtained with SS ranged from 66% to 71.7%. It was interesting to observe that, although SS provides more information than suprathreshold strategies such as the FDT screening program, its sensitivity and specificity were not significantly different from the other tests. With any diagnostic test, evaluation of its sensitivity and specificity provides important information about its usefulness, especially with regard to screening. In this study, we found moderate sensitivities and specificities for all strategies. We also observed that criterion 2 of FDT (based on the presence of two or more abnormal locations regardless of the severity of abnormal points) was associated with better specificity than other previously described criteria for other instruments. Similarly, the sensitivities and specificities obtained with each instrument when only first examinations
ACTA OPHTHALMOLOGICA SCANDINAVICA 2006
Table 5. Distribution of early, moderate and severe glaucoma among patients according to the first examination. Severity of glaucoma
FDT (n ¼ 14)
TOP (n ¼ 17)
SF (n ¼ 14)
SS (n ¼ 19)
Early Moderate Severe
2 5 7
6 1 10
4 4 6
9 2 8
p ¼ 0.223.
Table 6. Sensitivities and specificities with their respective confidence intervals for only the first examinations of each patient. Test*
FDT (1) FDT (2) TOP (1) TOP (2) SF (1) SF (2) SF (3) SS (1) SS (2) SS (3) p-value
n
27 27 30 30 28 28 28 32 32 32
Sensitivity
Specificity
%
(95% CI)
%
(95% CI)
78.6 78.6 94.1 94.1 100.0 92.9 92.9 89.5 89.5 89.5 0.63
(48.9–94.3) (48.9–94.3) (46.0–93.8) (69.2–99.7) (73.2–99.3) (64.2–99.6) (64.2–99.6) (65.5–98.2) (65.5–98.2) (65.5–98.2)
76.9 76.9 53.9 61.5 57.1 64.3 71.4 84.6 76.9 84.6 0.65
(46.0–93.8) (46.0–93.8) (26.1–79.6) (32.3-84-9) (29.7–81.2) (35.6–86.0) (42.0–90.4) (53.7–97.3) (46.0–93.8) (53.7–97.3)
* Refer to text for details of criteria for each test. CI ¼ confidence interval.
were analysed did not appear to be significantly different (p > 0.05). Our study has important limitations. The ideal strategy to compare sensitivities and specificities of different instruments includes the construction of ROC curves, or the comparison of sensitivities at a fixed specificity. In this study, we aimed to compare the sensitivities and specificities of different instruments using criteria that had been previously described in the literature. This does not allow for the construction of ROC curves. Some have employed ROC curves derived from one data point (Wadood et al. 2002), but this represents an approximation that is not appropriate. Similarly, setting the specificity at a fixed value and comparing sensitivities is very difficult when categorical data are used as criteria. Because it is a clinic-based study including a proportion of glaucomatous cases that is significantly higher than that expected in the general population, our findings may not be applicable to the general population. Furthermore, our exclusion criteria selected a population free of other
Table 7. Sensitivities and specificities of FDT, TOP, SS, and SF in previously published studies. Author
Test
Burnstein et al. (2000) Quigley (1998) Kalaboukhova & Lindblom (2003) Patel et al. (2000) Trible et al. (2000)
FDT FDT FDT FDT FDT
Wadood et al. (2002)
TOP-G1 TOP-G1 SF 24-2 SF 24-2 SF 24-2 FDT C-20-1 FDT C-20-1 TOP-G1 TOP-G1 SS 30-2 SF 30-2 SS 24-2 SF 24-2 SF 24-2 SF 24-2 TOP-G1 TOP-G1
Morales et al. (2000) Shekar et al. (2000) Sharma et al. (2000) King et al. (2002)
C-20-1 C-20-1 C-20-5 C-20-1 C-20-1
Criteria*
Disease definition
Sensitivity
Specificity
FDT (1) FDT (2) FDT score† > 1 Own algorithm FDT (1)
Abnormal HFA fields Abnormal HFA fields Abnormal optic nerves Abnormal HFA fields Clinical diagnosis and abnormal HFA fields Abnormal optic nerves
85.7% 91% 91.7% 80% 39%, 86% and 100%‡
83.3% 94.0% 87.8% 93.0% 95.0%
97% 94% 98.5% 95.7% 94.2% 91.4% 80% 94% 89% 95.1% 92.7% 93.2% 88.3% 86.4% 89.2% 85.2% 84.7%
68.0% 82.0% 67.8% 75.0% 82.1% 96.4% 100% 87.0% 90.0% NA NA 79.0% 93.8% 82.4% 80.0% 86.7% 76.5%
TOP (1) TOP (2) SF (1) SF (2) SF (3) FDT (1) FDT (2) TOP (1) TOP (2) SS (1) SS (1) SS (1) SF (1) SF (2) SF (3) TOP (1) TOP (2)
Abnormal Octopus 1-2-3 Abnormal HFA fields Abnormal HFA fields Clinical diagnosis
* Refer to text for details of criteria for each test. † The FDT score was calculated as follows: normal areas (p 5%) were assigned a value of 0; areas with p < 5% were graded as 1; those with p < 2% as 2, and those with p < 1% as 3 on the basis of FDT grey-scale printout. ‡ For early, moderate, and severe glaucoma. FDT ¼ frequency-doubling technology; TOP ¼ tendency-oriented perimetry; SF ¼ Humphrey SITA Fast; SS ¼ Humphrey SITA Standard; HFA ¼ Humphrey field analyser; NA ¼ not available.
349
ACTA OPHTHALMOLOGICA SCANDINAVICA 2006 ocular diseases (such as cataract), which is not the case in populationbased screening. Finally, our sample size was relatively small, resulting in wide confidence intervals and possibly reducing the chances of finding differences between strategies. Nevertheless, we were able to determine that fast strategies are associated with reduced performance in individuals undergoing VF examinations for the first time. When testing individuals with no perimetric experience, a situation typically observed during population screening, moderate sensitivities and specificities are expected, regardless of the strategy chosen. These findings suggest that relying on a single rapid VF test in a screening setting is not appropriate. Although these tests certainly represent an advance in the functional evaluation of glaucomatous patients, the confirmation of initial results and/or the employment of additional tests (including the analysis of structural damage) are required for accurate diagnosis.
References Anderson AJ & Johnson CA (2003): Frequency-doubling technology perimetry. Ophthalmol Clin North Am 16: 213–225. Anderson DR & Patella VM (1992): Automated Static Perimetry. St. Louis, Missouri: Mosby 10–15. Bebie H, Frankhauser F & Spahr J (1976): Static perimetry. Acta Ophthalmol (Copenh) 54: 325–338. Bengtsson B & Heijl A (1998a): SITA Fast, a new rapid perimetric threshold test: description of methods and evaluation in patients with manifest and suspect glaucoma. Acta Ophthalmol Scand 76: 431–437. Bengtsson B & Heijl A (1998b): Evaluation of a new perimetric threshold strategy, SITA, in patients with manifest and suspect glaucoma. Acta Ophthalmol Scand 76: 268–272. Bengtsson B, Olsson J, Heijl A & Rootze´n H (1997): A new generation of algorithms for computerized threshold perimetry: SITA. Acta Ophthalmol Scand 75: 368–375. Budenz DL, Rhee P, Feuer WJ, McSoley J, Johnson CA & Anderson DR (2002): Sensitivity and specificity of the Swedish Interactive Threshold Algorithm for glaucomatous visual field defects. Ophthalmology 109: 1052–1058. Burnstein Y, Ellish NJ, Magbalon M & Higginbotham EJ (2000): Comparison of frequency-doubling perimetry with Humphrey visual field analysis in a
350
glaucoma practice. Am J Ophthalmol 129: 328–333. Chylack LT, Wolfe JK & Singer DM et al. (1993): The Lens Opacities Classification System III. Arch Ophthalmol 111: 831–836. Delgado MF, Nguyen NT & Cox TA et al. & the American Academy of Ophthalmology Ophthalmic Technology Assessment Committee 2001–02 Glaucoma Panel (2002): Automated Perimetry. A report by the American Academy of Ophthalmology. Ophthalmology 109: 2362–2374. Fankhauser F, Koch P & Roulier A (1972): On automation of perimetry. Graefes Arch Clin Exp Ophthalmol 184: 126–150. Garway-Heath DF & Hitchings RA (1998): Sources of bias in studies of optic disc and retinal nerve fibre layer morphology. Br J Ophthalmol 82: 986. Gonzales de la Rosa M, Morales J & Dannheim F et al. (2003): Multicentre evaluation of tendency-oriented perimetry (TOP) using the G1 grid. Eur J Ophthalmol 13: 32–41. Iwasaki A, Sugita M & Glaucoma Screening Project Study Group (2002): Performance of glaucoma mass screening with only a visual testing using frequency-doubling technology perimetry. Am J Ophthalmol 134: 529–537. Johnson CA (2002): Recent developments in automated perimetry in glaucoma diagnosis and management. Curr Opin Ophthalmol 13: 77–84. Johnson CA & Samuels SJ (1997): Screening for glaucomatous visual field loss with frequency-doubling perimetry. Invest Ophthalmol Vis Sci 38: 413–425. Joson PJ, Kamantigue ME & Chen PP (2002): Learning effects among perimetric novices in frequency-doubling technology perimetry. Ophthalmology 109: 757–760. Kalaboukhova L & Lindblom B (2003): Frequency-doubling technology and highpass resolution perimetry in glaucoma and ocular hypertension. Acta Ophthalmol Scand 81: 247–252. Katz J, Tielsch JM, Quigley HA, Javitt J, Will K & Sommer A (1993): Automated suprathreshold screening for glaucoma: the Baltimore Eye Survey. Invest Ophthalmol Vis Sci 34: 3271–3277. King AJ, Taguri A, Wadood AC & AzuaraBlanco A (2002): Comparison of two fast strategies, SITA Fast and TOP, for the assessment of visual fields in glaucoma patients. Graefes Arch Clin Exp Ophthalmol 240: 481–487. Medeiros FA, Sample PA & Weinreb RN (2004): Frequency-doubling technology perimetry as predictors of glaucomatous visual field loss. Am J Ophthalmol 137: 863–871. Morales J, Weitzman ML & Gonzales de la Rosa M (2000): Comparison between
tendency-oriented perimetry (TOP) and Octopus threshold perimetry. Ophthalmology 107: 134–142. Patel SC, Friedman DS, Varadkar P & Robin AL (2000): Algorithm for interpreting the results of frequency-doubling perimetry. Am J Ophthalmol 129: 323–327. Quigley HA (1996): Number of people with glaucoma worldwide. Br J Ophthalmol 80: 389–393. Quigley HA (1998): Identification of glaucoma-related visual field abnormality with the screening protocol of frequencydoubling technology. Am J Ophthalmol 125: 819–829. Schimiti RB, Avelino RR, Kara-Jose´ N & Costa VP (2002): Full-threshold versus Swedish Interactive Threshold Algorithm (SITA) in normal individuals undergoing automated perimetry for the first time. Ophthalmology 109: 2084–2092. Sharma AK, Goldberg I, Graham SL & Mohsin M (2000): Comparison of the Humphrey Swedish Interactive Threshold Algorithm (SITA) and full-threshold strategies. J Glaucoma 9: 20–27. Shekar GC, Naduvilath TJ, Lakkai M, Jayakumar AJ, Pandi GT, Mandal AK & Honovar SG (2000): Sensitivity of Swedish Interactive Threshold Algorithm in Humphrey visual field testing. Ophthalmology 107: 1303–1308. Sponsel WE, Ritch R, Stamper R, Higginbotham EJ, Anderson DR, Wilson MR & Zimmerman TJ (1995): Prevent Blindness America visual field screening study. The Prevent Blindness America Glaucoma Advisory Committee. Am J Ophthalmol 120: 699–708. Trible JR, Schultz RO, Robinson JC & Roth TL (2000): Accuracy of glaucoma detection with frequency-doubling perimetry. Am J Ophthalmol 129: 740–745. Wadood AC, Azuara-Blanco A, Aspinall LL, Taguri A & King AJ (2002): Sensitivity and specificity of frequency-doubling technology, tendency-oriented perimetry, and Humphrey Swedish Interactive Threshold Algorithm Fast perimetry in a glaucoma practice. Am J Ophthalmol 133: 327–332.
Received on February 24th, 2005. Accepted on November 14th, 2005. Correspondence: Paulo de Tarso P. Pierre-Filho MD Rua Visconde de Maua´ 470/500 Fortaleza-CE CEP 60130-160 Brazil Tel: þ 55 88 3613 2119 Email:
[email protected].
