|
 
 |
| COMMISSIONED ARTICLE |
|
| Year : 2014 | Volume
: 2
| Issue : 1 | Page : 31-59 |
|
Interpretation of autoperimetry
Barun K Nayak1, Sachin Dharwadkar2
1 P.D Hinduja Hospital, Mumbai
2 Samartha Clinic, Mumbai
| Date of Submission | 17-Oct-2013 |
| Date of Acceptance | 01-Nov-2013 |
| Date of Web Publication | 3-Dec-2013 |
Correspondence Address:
Barun K Nayak
P.D. Hinduja National Hospital & MRC, Veer Savarkar Marg, Mumbai- 400 016
 Source of Support: None, Conflict of Interest: None  | Check |
DOI: 10.4103/2320-3897.122659
Autoperimetry is an essential investigation for glaucoma management, which helps in the initial diagnosis as well as the follow up of glaucoma patients. The interpretation of autoperimetry is tricky and crucial. This article deals with the basics of autoperimtery explaining the various terminologies which are frequently used. This is followed by guidelines and algorithms for interpreting single field analysis. It also deals with the follow up strategies used in autoperimetry with emphasis on understanding the interpretation of "glaucoma progression analysis" (GPA) on Humphrey. This article will be of great help to comprehensive ophthalmologists as well as the post graduate student of ophthalmology, in understanding the intricacies of autoperimetry analysis which will be of great help in the management of glaucoma. Keywords: Autoperimetry, threshold, glaucoma hemifield test, visual field index, Blue on yellow perimetry
How to cite this article:
Nayak BK, Dharwadkar S. Interpretation of autoperimetry. J Clin Ophthalmol Res 2014;2:31-59 |
Automated perimetry (standard white on white) is, as on today, the gold standard for detecting the functional deficit and evaluation of progression of glaucomatous damage. The article aims at simplifying this difficult and painstaking art of conducting perimetry and interpreting the visual field printouts. The main focus is on the most commonly used Humphrey perimeter, however, a mention is also made of the other devices and their salient features. To begin with it would be helpful to get acquainted with the standard terminologies and definitions in relation to the subject.
What is visual field/perimetry/Autoperimetry?
Visual field: It is that portion of the external environment of the observer wherein the steadily fixating eye(s) can detect visual stimuli. It can be unilateral or bilateral.
Perimetry is defined as the measurement of the visual function (of differential light sensitivity) of the eye at topographically defined locations in the visual field. In short it is a visual field test.
When the test is automated using a computer it is called as autoperimetry.
Importance of perimetry
Glaucoma: Diagnosis and progression: Perimetry is fundamental test and the current gold standard in glaucoma diagnosis and management. Reproducibly demonstrating visual field loss on perimetry is the strongest contributor to glaucoma diagnosis. Even the most precise method for following up the disease and labeling progression is a visual field test. The ever increasing imaging modalities are at best complimentary to it as on today.
Neurology: Though replaced by neuroimaging in a big way for the part of diagnosis, perimetry remains the only functional test to assess effect of neurological conditions related to the optic pathway. It forms an inexpensive and noninvasive alternative for neuroimaging to document changes in visual function.
Retinal disease: It is at best an ancillary test as the stereo fundus examination holds the pride of place in retinal diagnosis. However, the peripheral field testing is usually more important in retinal disease as compared with glaucoma and neurology, for example, retinoschisis and retinitis pigmentosa.
Visual disability assessment: Field testing forms an important part for the certification for drivers and navigators. Special programs like the Estermann binocular test, which is available with the Humphrey machine, help in charting of these fields. Field of vision is one of the criterion in defining blindness.
Important terminologies
Threshold: For perimetric definition; it is that intensity of the stimulus, which has probability of detection on 50% of the occasions when presented to a particular point [Figure 1]. In other words if the threshold stimulus is presented 100 times at a location, the person will identify it as present 50 times and miss it 50 times. | Figure 1: Shows a plot of stimulus intensity against the percentage of "point seen". Threshold is the intensity with probability of 50% detection
Click here to view |
The fovea has the highest threshold (can detect the least intense stimulus) and the function progressively decreases towards the periphery where it is the least. The threshold is measured by the staircasing method of 4-2-1 decibel (dB) in Humphrey in the normal strategy [Figure 2]. Any intensity of stimulus higher than the threshold is called as suprathreshold stimulus. | Figure 2: 4-2-1 strategy for thresholding: There can be two possibilities at any given point when the stimulus is presented to it. The left side of the Figure shows, that if the initial stimulus is not seen then the stimulus intensity is increased in steps of four decibels till such a time it is detected. Subsequently, the intensity is reduced by interval of two decibel steps till such a time it is not seen again, from this point the intensity is again increased by 1 decibel till the point is seen again and this is recorded as the threshold for that point. In the right half of the Figure if the stimulus is initially seen, then the intensity is reduced by four decibels and the same sequence followed to detect the threshold
Click here to view |
Retinal sensitivity: It can be used interchangeably with threshold and has the same meaning.
Kinetic perimetry: Typical example is the Goldmannn perimetry and the test is performed actively by an examiner physically moving the target on the arc and bowl perimeter with the patient reciprocating on visualization of a test object.
Most importantly the intensity of the test stimulus cannot be varied but the color and the size can be modified. The contours thus derived for a particular stimulus specification are called as isoptres and are indicative of the patients' field of vision. However, as the stimulus intensity cannot be varied, it is a form of a suprathreshold test. In other words it can test the two dimensional extent of the field (X and Y axis) but cannot gauge the depth of the defect (Z axis).
Static perimetry: The test is performed by a machine in an automated fashion and the examiners role is limited to the passive observation of the patient's performance. It quantifies the patients' peripheral vision using efficient and standardized testing algorithms.
The most important difference from the kinetic test is the variation of the intensity at individual tested points predetermined in an area, which in turn, is a surrogate to the severity of ongoing visual loss at each individual point in the field.
Apostilb and dB: Apostilb is the unit of light intensity, whereas the dB is the unit of retinal sensitivity. Apostilb and the dB are inversely proportional to each other. Higher the apostilb value, lower will be the dB value. The apostilb value is an absolute measure but the dB value varies with the machine used (hence a relative value), that is, the dB value of the Humphrey does not match with the dB value of the Octopus. In Humphrey, the value of 0 dB is equivalent to 10000 apostilbs, whereas the 0 dB on the Octopus is equivalent to 1000 apostilbs. Hence the field charted on the two machines cannot be compared accurately and the data obtained is not interchangeable.
Staging and phasing: Staging [Figure 3] is a hallmark of the Octopus perimeter and is not available on the Humphrey perimeter. | Figure 3: The complete test may be divided in various stages. After completion of each stage if the test cannot be continued further still inference can be drawn. Phasing is a combination of different strategies to get more information
Click here to view |
Normally in a single pass perimetric test program the result of the examination is available only at the end of the test, when it has run its full course. In the Octopus, the test runs in stages of modular steps, mutually independent. Each stage is tested without compromising the reliability and the accuracy of the other, the priority areas for a particular program being tested first. Hence even if only the first two stages are over, the report can be assessed; which is not the case with a single pass examination. In a single pass exam, a shorter duration program has to be adopted to give information in the same time frame.
Test phases can later be incorporated to complete the examination if required to provide more information. This is available on both the Humphrey as well as the Octopus.
Strategy: It is the testing algorithm chosen for a particular test depending on the defect expected, the patient profile and the physical ability, for example, full threshold, Swedish interactive thresholding algorithm (SITA) standard, SITA FAST, etc., on Humphrey.
Suprathreshold/screening tests: Both mean the same thing and they are qualitative tests. They are quick and just tell the presence or the absence of the defect. The suprathreshold intensity stimulus is used and the points are recorded as seen and not seen. It can be divided into:
Two zone: Differentiates the normal and the abnormal points.
Three zone: Abnormal points are further tested to see for a relative or an absolute defect by retesting of missed spots with the brightest stimulus available in the machine.
Quantify defect: Abnormal points are thresholded to see the extent of the defect.
Threshold testing: This is a quantitative test, wherein the preselected points, as per the program, are tested for the threshold values. It gives the numerical value in dB at each point tested and therefore the change over time can be accurately judged. It is, however, a time-consuming test.
SITA for the Humphrey and the TOP (tendency-oriented perimetry) for the Octopus are patented testing algorithms that drastically reduce the testing time of the threshold tests with a little compromise of information.
SWAP (short wave automated perimetry): Also known as the blue-on-yellow perimetry, it takes advantage of the redundancy of the visual system to isolate a subset of ganglion cells susceptible to early glaucomatous damage. It has a size 5 blue stimulus on a yellow background, the performance of the test being affected by the lens density. Though touted to pick up field defects earlier than the white on white perimetry, it has not become popular as it takes much longer than the white on white test.
Strategies that take a longer time, give more information, however, this advantage can be offset by the onset of fatigue that can affect the test-retest variability or the fluctuation; thus affecting the final result. Hence the appropriate strategy should be chosen depending on the patients' general ability, psychological makeup, and attention span to give the most representative results.
Program: The programs denote the location and the pattern of arrangement of test points in a particular area, for example, central 30-2, central 24-2, 10-2 for Humphrey and the program 32, G1x, M2x, STx, etc., for the Octopus perimeter.
Factors affecting threshold
- Stimulus size and intensity: The Humphrey machine can produce stimuli that can vary in brightness in the range over 5.1 log units between 0.08 and 10000 apostilbs. With the standard size 3 stimulus, the dimmest stimulus that can be perceived by a well trained observer is a little less than 40 dB. Hence the dimmest 10 dB range is beyond the perception of the human eye. The standard automated perimetry almost exclusively uses size 3 and in advanced field loss size 5. The other available sizes 1, 3, 4 are rarely used.
- Background illumination: It is an internationally accepted standard to choose 31.5 apostilb as the uniform background illumination for the bowl, the one that was started with then Goldmann machine. The rationale is that it is the minimum intensity for photopic (daylight) cone-related vision. The advantage of testing the photopic system is that it is more contrast than brightness oriented. Similarly under photopic conditions, the effect of lens color, pupil size, and lens transparency have minimal effect on the results. In scotopic conditions, absolute brightness becomes the dominant factor.
- Stimulus duration: Standard stimulus duration is 200 ms, long enough for visibility to be affected by slight changes in duration and adequately short for latency of eye movement.
- Stimulus location and disease: The more peripheral is the location of the stimulus, the more is the intensity required to perceive it, in normal individuals. In glaucoma, however, depending on the location of the defect the retinal sensitivity decreases and the abnormal points require more than the normal intensity for them to "be seen". This results in decrease in the dB value of that point as compared with normals.
Reliability parameters
There are a set of reliability parameters that are printed at the left upper corner of the printout, which indicate the test performance reliability of the patient. They are fixation losses, false positives, and the false negatives.
Fixation losses: They are calculated using the Heijl - Krakau method. After the localization of the blind spot in the initial part of the test; the machine on multiple occasions, lands a stimulus (5% of total stimuli), on the location of the blind spot. If the patient still responds to such stimulation it is considered as a fixation loss. Fixation losses of greater than 20% are indicative of unreliable field tests.
The other two methods are the gaze tracker (not in all models); the tracing of which is printed at bottom of the printout and the video eye monitor. An upward deflection on the graph indicates a saccade, whereas a downward deflection indicates a blink. "Blinks are the most significant part of the tracing to decide the quality of filled report." Caution should, however, be exercised as it may not always correlate especially if the patient moves even slightly after the gaze initialization stage of the test. The video eye monitor with an attentive perimetrist is a very robust indicator of the performance. Besides, the gaze track may not work well in patients with ptosis, miosis, and high refractive errors.
In case the fixation losses show an XX and you are sure that the patient was fixating well, the problem may well have been improper blind spot localization than the actual fixation errors. It can be verified by looking at the location of the blind spot on the printout obtained.
False positive error: This is a positive response by the patient even in absence of stimulus or to an audible "click" by the machine in the full threshold tests. This is referred to as the positive catch trial. In the shorter programs such as the SITA there are no catch trials but the anticipatory responses faster than the expected reaction time to the stimulus, are labeled as false positive. Up to 15% false positives are acceptable and defects can get masked in cases of high false positivity. If the false positives exceed this limit, an XX symbol will be seen with the message "Excessive high false positives".
False negative error: Some of the previously thresholded "seen" points are again presented with brighter stimuli and absence of response is considered as a false negative. Conventionally acceptable limit was 20% but in the SITA it is no more "flagged", as the false negatives can be indicator of fatigue as well as disease (especially at the edges of the scotomas where the points can exhibit variable thresholds). If the patient is inattentive, the defects can appear exaggerated.
It is important to note that the SITA strategies calculate the data at the completion of the whole test hence the values of the indices on the final printout should only be used. These values may differ from the values that were seen on the screen when the test was going on.
An XX mark is placed beside the reliability parameters that are outside the acceptable limits. If the short-term fluctuation (SF) is in the abnormal range a "P" value is supplemented on its side in the full threshold tests.
The Octopus also displays the reliability factor (RF) based on the false positive and the false negative error. RF up to 20 is acceptable.
Global indices
There are certain indices that are calculated after the completion of the threshold testing.
Mean sensitivity (MS): The average sensitivity of all the thresholded points. This index is seen in the Octopus but not in the Humphrey printout.
Mean deviation (MD): This is the average deviation from the normative data at all the tested points. It is called as the mean defect in the Octopus machine. It has a negative (-) sign. A small localized defect will show a small MD, whereas a generalized or an advanced defect will show a high MD. The value does not differentiate a generalized and a localized field loss. It also does not give the location of the defect.
Pattern standard deviation (PSD): This index gives an idea about the resemblance of the patients' field to the shape of hill of vision. It has a positive sign. Low PSD indicates a normal shape of the hill, whereas a high value indicates a disturbed shape of the hill. Localized defect will give a high PSD, whereas a generalized defect will give a low PSD. Hence paradoxically the PSD will improve with the generalization of the defect in advanced field loss. This indicator consequently loses its teeth in advanced disease, in which case the patient can be followed up with the MD.
It is called as loss variance (LV) in the Octopus.
SF: Also referred to as the intra-test variability. It is available only with the full threshold printouts. Ten preselected points are thresholded twice and the variation in the thresholds is represented as a number. SF > 3 is considered as an indicator of unreliable result, but is also seen in advanced disease.
Long-term fluctuation: Also known as inter-test variability, should be kept in mind while interpreting the multiple tests over time, however, no machine provides any measure for long-term fluctuation.
Corrected pattern standard deviation (CPSD) or corrected loss variance (CLV): It is the PSD or LV corrected for the SF in the Humphrey and the Octopus, respectively.
P-value (probability value): All the global indices are supplemented with a probability (P < x%) on the side, indicating that less than x% of the normal population has figure like this or in other words there is an x% chance that the index would be seen in normal. Lower the P value beside the global index the higher chance of it being abnormal. Therefore if no P value is given beside a global index, it can be considered normal.
Interpretation of the global indices : All the indices are considered together for interpretation and are inter-related. MD is used for determination of the stage of glaucoma damage. PSD and the CPSD are important for the diagnosis of early glaucoma.
Cataract alone will show a high MD but a normal PSD and CPSD.
Early glaucoma without cataract will show close to zero MD but high PSD and CPSD with a low P value.
In combination of both cataract and glaucoma or in advanced glaucoma, MD, PSD, CPSD all will be abnormal with a low P value.
Glaucoma hemifield test
This is useful in the diagnosis of early glaucoma and is available only on the Humphrey. It basically relies on the fact that the glaucomatous defect occurs on either side of the horizontal midline never crossing it and is unlikely to be symmetrical across the horizontal meridian. Thresholds derived at the five sets of points, which are mirror image along the horizontal meridian as shown [Figure 4] are compared and the results are displayed as follows:
- Glaucoma hemifield test (GHT) within normal limits
- GHT borderline
- GHT outside normal limits
 | Figure 4: This Figure shows the sets of points compared in the glaucoma hemifi eld test. The sets labeled A, B, C, D, E are compared with their mirror images A1, B1, C1, D1, E1 across the horizontal midline and a message based on the analysis is printed next to the GHT result on the perimetry printout
Click here to view |
Two additional messages are also flashed if that is the case:
- Generalized decrease in sensitivity or abnormally high sensitivity
- Low patient reliability
The positive GHT does not mean always glaucoma hence, a clinical correlation is required. However, in case of very early glaucoma, a negative GHT can definitely rule out any glaucomatous field defect.
Visual field index: Visual field index (VFI) is a single number that summarizes each patient's visual field status as a percentage of the normal age-corrected sensitivity. It was originally designed to approximately reflect the rate of ganglion cell loss. It is derived from the pattern deviation plot and is centre weighted, considering the high density of the retinal ganglion cells in the central retina. This implies that while calculating, the central points are given more weightage than the peripheral. This index is less affected than the MD by factors that cause a general reduction in sensitivity like cataract, miosis, and refractive error.
It is given as a percentage and the minimum value is 0 for a blind field and 100% for a normal individual. It is also plotted against time to obtain a trend analysis and slope on similar lines as the MD value, after five reliable fields covering at least 2 years. Glaucoma progression analysis (GPA) provides a projection of the linear regression line into the future, if the width of the calculated 95% confidence interval for VFI slope is found to be acceptably small - no larger than VFI value of +/- 2.5%. Otherwise it gives a message that the confidence interval is too large for the calculation and does not calculate the slope.
The projected goal predicts the future trend provided the existing rates continue and there is no alteration of therapy. This indeed translates into what "COULD" happen if the existing trend were to continue. Projections never exceeds 5 years and are never longer than the measured follow up period and are demarcated by two vertical lines each at the start and the end of this projection. The vertical bar on the right of this graph indicates the change in percentage VFI over time considering the same slope.
At the bottom of the VFI plot the slope with the confidence limits and the rate of progression with its statistical significance is given. This is the most useful piece of information in the trend analysis that can translate into alteration of therapy, when used in conjunction with the life expectancy and physical state of the patient.
In the USA a reasonable goal of therapy is to maintain at least 50% VFI in the better seeing eye. An MD of -22 dB corresponds to approximately a VFI of 30%.
Parts of a printout: The single field analysis of Humphrey printout is shown in [Figure 5]. | Figure 5: Single field analysis printout of Humphrey General information: (part A) shows the general information regarding the patient's detail and the test detail.
Reliability parameters: (part B) gives all the reliability parameters — Fixation loss, false positive, and false negative results. The unacceptable errors are also marked with two crosses (xx) in front of them. Just below the reliability parameters, total test duration is printed.
Threshold numerical values™: (part C) shows the raw data of the patient. This is the actual values at each tested points as per the patient's response.
All other parts of the printout are calculated and generated by the computer from these data. If any point is thresholded twice, then those points are represented by two Figures.
Grayscale: (part D) The numerical value is represented here in gray scale. This is useful mainly to explain to the patients. We should not interpret on grayscale alone.
Total deviation plot: (part E) The numerical value of threshold is compared with the age matched normative data and the difference in value at each point is printed in numbers. Lower than normal value is printed with — sign and points with higher than normal value is printed without any sign.
Probability plot of total deviation plot: (part F) gives the probability of each deviation being normal or abnormal. All dot signs are considered as normal, whereas all other symbols denotes the different P-value (e.g., <5% <2% <1% <0.5%). The detail of each symbol is given below the probability plot. In general, darker the symbol, more chances of it being abnormal.
Pattern deviation plot: (part G) gives the total deviation plot after correcting it for the generalized fi eld defect and is called as pattern deviation plot. The localized defect will be more prominent in this plot. Hence, the study of this plot is very important for early diagnosis of glaucoma.
Probability plot of pattern deviation plot: (part H) depicts the probability of pattern deviation plot being abnormal. This plot is very important for the detection of early glaucomatous fi eld defect.
Global Indices: (part I) gives the value of various global indices for the patient. If the value is given without any P-value, then it should be considered as normal. However, if the value is given with a P-value (e.g., <10% <5% <2% <1% <0.5%) then it may be abnormal. In general lower the P-value, more chances of it being abnormal.
GHT: (PART J) gives the outcome of GHT. The message of low reliability as well as of generalized depression or elevation is fl ashed here.
Gaze tracker tracing: (part K) This gives us the graphic record of patients fixation. The upward defl ections indicate saccades and the downward defl ections indicate blink. Frequent blink makes the field report unreliable
Click here to view |
The statistical package that is available with the Humphrey device is called as STATPAC. The analysis of the data acquired is presented in five formats
- Single field analysis
- Change analysis
- Overview printout
- Glaucoma change probability (GCP, with the full threshold tests)
- GPA (with the SITA tests)
How to read a printout: GRADES (mnemonic for the interpretation flow chart)
G = General information
R = reliability
A = abnormal or normal field
D = defects, after analysis of the field defect should be named/classified
E = evaluate. Once the defect has been identified, one should try and correlate clinically and evaluate about the patient's disease status.
S = subsequent evaluation. This is applicable in case repeat fields are done after some time to evaluate the progression (stable, deterioration, or improvement) of the field defect.
Single field analysis [Figure 6]: The following sequence should be followed always | Figure 6: Follow the sequence of "GRADES". Points to note are:
Reliable response, pattern deviation probability plot, and total deviation probability show significantly defective points in upper temporal field.
This is also confirmed on grayscale plot by dark area and low dB value on raw data plot. The pattern deviation probability plot is slightly better than of total deviation plot indicating slight amount of element of generalized depression. Global INDICES, (MD, PSD, and CPSD) in abnormal range with low P-value, GHT "outside normal limits", the Anderson's criteria is also met with. Suggestive of glaucomatous change, subject to clinical correlation.
However, since this Figure also shows some depressed points on the right side of the vertical midline giving a suspicion of heminopic defect.
To confirm this fact when RP Mills criteria are applied to it the top three points immediately adjacent to the vertical midline show a unidirectional shift in sensitivity (decreased toward the right side).
The corresponding points in the next pair of columns adjacent to them show the shift in same direction in the top two points but the third point shows a reverse change in sensitivity. Hence the RP Mills criteria are not met; still this fact needs reconfi rmation in repeat testing. If RP Mills criteria is met with in subsequent testing, a neurological investigation will be called for. For this case even the repeat test also did not meet the RP Mill's criteria
Click here to view |
General information Reliability indices → Grayscale→ total deviation → pattern deviation → global indices → hemifield test result → RAW DATA → VFI.
Andersen's criteria for glaucomatous field defect: These criteria (in relation to a Humphrey printout only) are helpful in the diagnosis of early glaucoma and are as follows:
- Abnormal GHT
- Three or more nonedge points of the 30-2 printout, contiguous and with a P < 5%, out of which at least 1 has a P < 1%
- CPSD should be abnormal and should have a P < 5%
Obviously these are made with the 30-2 full threshold printout of the Humphrey in mind that was the standard strategy in use earlier and less frequently today.
All these defects should be reproducible and should be demonstrated on two successive tests.
As an extrapolation of these criteria to apply to the SITA test, CPSD needs to be replaced with PSD, as the SITA does not calculate SF and hence the CPSD. These are also known as the modified Andersen's criteria.
Similarly in 24-2 programs, the criteria number 2 should be read as any three points rather than 3 nonedge points, as the edge points in the 30-2 are already discarded in 24-2 except the nasal step area.
R P Mill's Criteria: For subtle hemianopic defect (suggestive of neurological disorder) must be looked at the end of all field interpretation. Otherwise it may be misinterpreted as early glaucomatous defect. For this purpose, one should analyze the raw data and look for the following points:
- First compare the dB value of adjacent rows on the either side of the vertical meridian. At least three adjacent pairs should show unidirectional difference in sensitivity.
- The corresponding points pairs on the next column adjacent to the first column should also show difference in sensitivity in the same direction.
- At least a difference of 2dB is significant and is suggestive of early hemianopic defect.
General guidelines
a. For conducting a field test: Good counseling and table time spent with the patient, by the surgeon and the perimetrist is invaluable in obtaining a good quality response.
A dimly lit dedicated room without too much thoroughfare is ideal for field testing.
Perimetrist with a good understanding of the subject and willing to continuously monitor the patient through the entire test time, on the video monitor and externally, is mandatory.
b. For interpreting a field report: There is a learning curve in autoperimetry. Sometimes it may extend up to three tests or more.
It is difficult to get a first test reliable but some of the people do well to begin with, hence it may not necessarily be ignored.
A field defect should be reproducible, to be considered significant; especially if contemplating a change in therapy. Droopy eye lids can cause defect in the upper area.
Field defects are always to be correlated clinically with disc and retinal nerve fiber layer changes.
A defect that is bizarre and not clinically correlating can be discarded.
Fields not following the reliability criteria can be discarded if they correlate to the perimetrists opinion about the patients' performance.
False negatives are to be interpreted with caution as they can indicate inattentiveness as well as disease.
Abnormal patterns like clover leaf, white scotomas, concentrically contracted fields (as in miotic pupils) and rim artifacts should be kept in mind and such fields should be repeated with the necessary changes. Droopy eyelids can cause defect in the upper area.
Potential sources of error
Wrong patient, wrong eye, wrong entry of date of birth.
Inattention to the change in refraction of the patient over time.
Uncomfortable positioning and physical disability can affect performance and should be watched out for.
Performing the test in an over lit, noisy, and crowded room can affect the thresholds of seeing.
Pitfalls in autoperimetry
Learning curve should be kept in mind. The test being a psychophysical one, some patients can never manage it well at all.
Always compare the same program. However it is worthy to note that the SITA30-2 can be used as a baseline for the SITA 24-2 but no interchange of the strategy is allowed (i.e., Full threshold to SITA, etc.) for the follow up.
No follow up programs are available for the 10-2 and the macula, where we have to rely on the overview prints and the quadrant totals.
No eyeballing comparisons are allowed between the field reports and ideal comparisons are machine software derived; by retesting the patient on the same machine every time.
In standard white on white test with stimulus size 3 the test retest variability increases as the defect depth increases, hence the test falters somewhat when it is required to be the most accurate. This is due to the limitations of the property of target size and the currently available technology. This has to be kept in mind as the results and reproducibility in advanced cases can be quite a problem.
Follow up and its analysis: Why is it so important?
This is probably the most important and time tested use of the perimeter, that is, to accurately predict glaucoma progression. At present it is the only diagnostic modality used to guide and modify therapy, whereas imaging modalities are under improvement.
The early manifest glaucoma trial (EMGT) has shown beyond doubt that all the glaucoma patients eventually progress irrespective of their value of intraocular pressure (IOP) being "under control" on follow up visits. Thus progression is a rule in glaucoma, than an exception. The rates of progression vary widely even under careful and close follow up and risk factors alone cannot predict the rates accurately each time. [1],[2] While some of the patients will progress very slowly and need only minimal therapy, an important minority (approximately 1 in 6) depending on the type of practice will progress at rates that will quickly lead to visual disability if not checked. [3]
Perimetry is hence given due consideration for detecting existing field loss and rate of progression, which along with quality of life (QOL), and life expectancy is use in altering the therapeutic decisions.
In the absence of effective changes in therapy the past rates of progression have found to be predictive of the future rates. [4]
Use of rates of progression in therapeutic decisions has been recommended in decision making in the European glaucoma society (EGS) guidelines, the goal of treatment being avoidance of visual disability. This underlines the need to understand progression and follow it immaculately.
Techniques which allow us to understand progression
Change analysis printout [Figure 7]: This report can include a maximum of 16 tests and is presented in the form of a box plot analysis of tests, a summary of the global indices and linear regression analysis of MD, all on one page. | Figure 7: Change analysis printout: This is the change analysis printout of the right eye, it has the following parts: (1) Eight fields each represented in the form of 'Box plot' with the date of test written next to it. (2) It has one plot each for SF, MD, PSD, and CPSD. The bottom of the chart shows the slope of the MD with its confi dence limits.
In the given box plot the 2nd testing shows improvement as compared to the first (probably the learning curve). The 3rd test shows marked downward shift of "box plot" with a normal shape indicating generalized depression of field (development of posterior subcapsular cataract was confirmed clinically). The 4th "box plot" shows shifting of box upwards with normal shape (Following cataract surgery). The 6th test shows marked worsening of field, which was not confirmed on 7th testing; 8th testing indicates worsening field. The next testing will confirm this deterioration.
Change analysis also gives separate graphs of SF, MD, PSD, and CPSD arranged in chronological order. The reliable response is denoted by "0" where as unreliable response is marked with "X" on these graphs. Study of these graphs gives fairly good idea about the progression of these indices in follow up examinations.
Lower end of the printout also gives the MD slope with 95% confidence
Click here to view |
Box plot: The box plot is a modified histogram that gives a summary of TOTAL DEVIATION test values for each test with reference to the age-related STATPAC database, but without reference to the location on the field. It is basically a distribution of all the point thresholds around their mean and how much they deviate from it. It is in the shape of a box located on a line the arms of the line and the length of the box varying according to the extent of distribution of the points around their mean values. The "box plot" is charted from the "total deviation plot". The central dark line in the box is the mean of all deviation and should represent the 50 percentile point. The upper end of box represents the 85 percentile point (i.e., only 15% points show more than this deviation on positive side) and the lower end of box represents the 15 percentile point (i.e., only 15% points show more than this deviation on negative side). The upper end of upper tail represents 100 percentile point of total deviation plot and the lower end of lower tail represents the 0 percentile point of total deviation plot.
In generalized depression the total box plot shifts down without any change in shape from the normal.
In small localize defect the position of the box plot remains the same as that of normal, the central box with the central dark line of the mean remains normal but the lower limit gets elongated below.
In large defects (approximately involving more than 15% of the points tested) the box and the dark line of the mean shifts down with the elongation of the box and elongation of its tail.
The change analysis gives a general idea about the change in the field in subsequent testing, however, it does not give an idea about progress at each location.
Each test date will be indicated on top of the box plot including the program used.
Summary of Global indices: The indices are summed up in the lower part of the page in the same chronological order as the box plot. They include the MD and PSD in the SITA and the SF and CPSD in the full threshold test printouts in addition. The graph consists of two dashed reference lines indicating P < 5% and P < 1% limits; for example, if the index value is lower than the 1% dotted line, it means that the index value is significant at the 1% level or in other words, less than 1% of the normal population has value as large or larger than that found in the test.
Linear regression of the MD: Using the values of the MD from all the test results the STATPAC performs a regression analysis and prints the result as "MD slope significant/not significant". It also mentions the value in dB per year and qualifies its slope with a "P" value; for example, if the result indicates "MD slope significant", the lower the P value the more likely it is that MD has changed in direction of the slope. Both the magnitude and the significance level of the slope are equally important. If the slope is not significant at largest P value that the STATPAC considers (5%), the machine declares "slope not significant".
The STATPAC linear regression analysis appears on the Change analysis as well as the GPA printout. The STATPAC automatically modifies the field selection to consider for the regression analysis when learning effects are present. In the GPA though it indicates the fields that need to be corrected it does not delete the fields on its own. Using identical tests for the Change analysis and the GPA will make sure that the slope is identical for both printouts, unless the former is affected by media opacities.
While interpreting the statistical values it is important to remember what they represent! The probability statements reflect the distribution of the particular change in general population.
The meaning of P < 0.5 % means only that the defect can be seen in 5% of normal population and nothing more.
It DOES NOT MEAN that there is only a 5% chance that the field result is normal.
The positive predictive value of the report depends on the prevalence of the defect in the population studied. The probability that a given result is abnormal depends on the relative prevalence in the population of the defects caused by disease versus the prevalence of the same field "defect" in normals.
This is probably the most significant point for citing the importance of the regional database and representativeness of the normative data.
There are situations when the STATPAC does not give correct information. It can be due to limitations of the software or the software being applied to wrong set of data.
It has to be borne in mind that it is the clinical correlation that is the most robust and the evaluation of the defect identified, is ultimately the clinicians call.
Overview printout: [Figure 8] A-C are the example of overview printout.  | Figure 8: a-c show the overview printouts of the left eye incorporating fields done from 2004 to 2012. The initial three fields are done with the 30-2 full threshold program and the remaining four are done on the 30-2 SITA standard. The PSD in the SITA standard have worsen from 10.3 to 12.27 and the VFI has decreased from 82% to 71%. This is also reflected in the increase in the number of solid black dots on the probability map. (However, the gray scale also shows a component of clover leaf pattern setting in hence field should be supervised by an expert before deciding about the treatment.)
Click here to view |
The two basic analytical techniques for progression are the: (a) Event analysis and (b) Trend analysis
Event-based analysis: It tells us whether any change has occurred and whether it is statistically significant. The change probability maps of the GPA event analysis indicate statistical progression on fields in Humphrey and is based on pattern deviation maps. It is applicable only to SITA standard and SITA fast group of tests. The critical difference between the two, however, is the inability of GPA to calculate progression from fields with MD value less than 20 dB due to limitations of the machine algorithms that calculate the PSD. This, however, does not affect the older GCP, which is dependent upon the total deviation values. It should be remembered, however, that the test-retest variability of the depressed points on tests based on stimulus size 3 increase drastically with increasing depth of the defect, so much so that we may not end up losing much information even we do not have analysis beyond the 20 dB range. The GPA alert is the plain language event analysis that applies the EMGT progression criteria to the GPA change probability maps. [5]
It will display a message of "possible progression" if two successive fields indicate worsening, in same 3 or more points as compared with the selected baseline (of two fields) and if the similar changes are seen in three successive fields, the message displayed is "likely progression". The corresponding symbols for the each points on the map are an empty, half filled, and a solid filled triangle for the points that have worsened on 1, 2, and 3 successive tests, respectively.
GPA is based on data obtained from the pattern deviation plots and hence less affected by media problems like cataract. This approach has been proven effective than expert analysis. The sensitivity and the specificity of this approach was 96% and more than 89%, respectively. However, when applied to the 24-2 test patterns the median time to detect progression marginally increased from 33 to 37 months. [4]
- GPA recognizes any change from baseline examinations (average of first two reliable tests) in the subsequent visual field examinations, if it is more than the test-retest variability, which also varies as per the stage of glaucoma.
- Statistically while evaluating the change probability maps one should expect that 5% of the points will be falsely flagged on the basis of chance alone, due to random test variability. Hence it is important to remember that reproducible change is a must to document true worsening.
- Credible change must be seen at multiple test locations.
The main drawback, however, is that the change probability maps cannot be applied to fields with a MD value less than 20 dB as the mathematical model for calculating the pattern deviation does not perform reliably below this level of damage.
Trend analysis: It gives us the rate of change and its statistical significance or in simple words how quickly the patient status is changing.
It is complimentary to the event analysis, serves different purpose, and gives us the complete picture to aid in therapeutic decision making. Nowadays the preferred approach is using the linear regression analysis of the VFI over time. It is displayed automatically on the summary or the full GPA printout when sufficient number of reliable tests are available. It has limited the role of the regression analysis of the MD in early and moderate disease, as the MD is affected by the media clarity among other variables.
Establishing the baseline
It is of fundamental importance in the therapy. The baseline can be a selection of two field tests that are reliable and representative and by and large similar besides being obtained within a short span of time, say a couple of months. The selection can be done by the examiner, in lieu of which the perimeter by default chooses the first two "reliability indices based" reliable tests as the baseline. It is an important part of event-based analysis. However, it is recommended that the examiner at least examines the selected baselines using them for analysis.
Please note that the baseline requires to be changed at the time a therapeutic intervention (i.e., antiglaucoma surgery or addition of medication) is carried out. Event-based analysis is a relative concept comparing reports at two points in time and hence using the original baseline even after intervention will always show worsening, unless a new baseline is created with the fields done when the intervention was initiated. This concept needs to be clearly kept in mind.
The printout of the GPA is available in four formats ranging from the multipage review of the patients entire field history to the abbreviated summary that appears as a small part of HFAs single field analysis report. The GPA summary report among them is the recommended printout for Glaucoma management. The full GPA detailed report is used usually only if a major therapeutic intervention is contemplated.
The trend analysis, however, can include as many as 15 tests on a single printout and can be used for the calculation of the rates of progression. It can be seen on two places in the printouts; the change analysis (regression analysis of the MD) page and the VFI trend. The two, however, have different significance, but in common, both can be plotted only after five valid tests.
The MD slope is calculated from the MD plot and is not centre weighted, hence affected by pupil size, media opacities like cataract, etc., whereas the VFI is calculated using the pattern deviation, which is centre weighted and less affected by the above. It is noteworthy, however, that if all the variables affecting the MD are removed the MD plot will closely follow the VFI plot.
Another fundamental difference to be understood is that, due to the different nature of the calculation, the value of MD in a blind field will depend on the strategy used, whereas the VFI of the blind field will always be 0. These plots give us the slope and their statistical significance, in other words the rate of progression and have to be put in perspective to correlate with the life expectancy in taking therapeutic decisions.
[Figure 9] A-C are the print out of glaucoma change probability and [Figure 10] A-C, [Figure 11] are the printout of GPA. [Figure 12] is the GPA summary printout, whereas [Figure 13] is single field analysis GPA printout. [Figure 14] is a printout of SITA SWAP test.  | Figure 9: a-c, GCP Glaucoma change probability. This is another method of analysis for follow up fi eld tests. The fi rst two (or the 2nd and 3rd test, in case of 1st field being unreliable) tests are averaged out for baseline field. Figure 9A represents the printout of "glaucoma change probability – baseline", which gives "grayscale" and total deviation probability plot" of these two tests with MD graph and MD slope analysis findings of all the tests done so far.
All subsequent tests are compared with the averaged out baseline two tests and analyzed. The printout is given as "glaucoma change probability – Follow up [Figure 9b and c]. It is arranged in chronological order from top to bottom. Each row represents one test. The first column is "grayscale" of that test, 2nd column is "total deviation probability plot" (compared with normative data), 3rd column is "Deviation from baseline" (average of the initial 2 tests) and 4th column gives the "change probability plot". Each point is depicted with a dot sign (no change significantly), dark triangle (significant deterioration from baseline, P value <5%), or empty triangle (significant improvement from baseline, P value <5%). Hence in this method point by point comparison from the baseline is analyzed. GHT, MD, and MD change is also written for each test.
The first test on Figure 9B (which is the 3rd test of patient) shows signifi cant deterioration of all point (all dark triangle) due to development of cataract, which become normal after the cataract surgery as shown in the 2nd test onwards in Figure (i.e., 4th test onwards of this patient)
This printout facility may not be available after the installation of the GPA software or in the new machines as they are based on pattern deviation maps
Click here to view |
 | Figure 10 a-c: GPA Baseline (Figure 10a): This printout shows the GPA baseline printout for the central 30-2 threshold test of the left eye. The first part of the printout shows the two selected baseline fields (which were full threshold in this case) printed like an "overview" with the testing dates. The second part shows the plots of the VFI with its slope and confidence limits.
GPA follow up: (Figure 10b and c) This is the GPA follow-up printout for the left eye. 1) It consists of five follow-up tests in this case, each showing the gray scale, pattern deviation, deviation from baseline and progression analysis. In addition at the bottom of each test the global indices, GHT and the GPA alert are shown. The first test is marked as low test reliability due to the excessive fixation losses. The second follow up onwards, the testing strategy has been changed to SITA standard. The second tests show a mixture of symbols on the progression analysis plot the key to which is given at the bottom of the page. From the fourth test done on July 25, 2012 onwards, since few points have deteriorated more than three occasions GPA alert of 'likely progression' is shown this has been verified on subsequent testing (November 17, 2012). GPA – last three follow up.
Click here to view |
 | Figure 11: This printout shows the results of the last three 30-2 SITA standard tests conducted from November 5 to November 17, 2012 for the left eye and comparison to the baseline. The first field shows a "GPA alert" indicating "possible progression" followed by the second indicating the "likely progression", which is reconfirmed by the third/subsequent test. At this time the full follow up printout should be examined in view of considering change of therapy.
Click here to view |
 | Figure 12: GPA summary: This printout shows the baseline tests (top row), VFI plot (middle row) and (bottom row) the current test all in one page. The current follow up printout shows us the deviation from the baseline the progression analysis plot and the "GPA alert" and also includes the summary of the global indices
The VFI plot shows the VFI values plotted against time, the various testing strategy are marked as different symbols as shown, the empty box indicating the full threshold, the solid boxes indicating the SITA test after completing 2 years follow up. The data is extrapolated and projected further for the next 5 years to predict the change in VFI function. The rate of progression is given as a percentage at 95% confidence limits with its statistical significance below this plot
Click here to view |
The basic knowledge provided in this article can help understand any other perimeter printout easily. Printout of some other perimeters, which are in use, is given for the better understanding; [Figure 15] is single field analysis of Appa perimeter (Field explorer), [Figure 16] is single field printout from Medmont perimeter and [Figure 17] is its progression analysis. [Figure 18] and [Figure 19] are the single field analysis from Octopus perimeter. | Figure 13: Single field analysis GPA: In addition to the single field test report this printout has an additional box (A) that includes the data from the GPA
It contains (I) The date of the baseline and previous follow up exams for reference as shown and the strategy used therein. (II) Shows the event analysis comparing the current fi eld with the baseline documented in I and gives a "GPA alert" (III) which is "likely progression" as shown in this Figure. The part IV in the box shows the key for the symbols used in part II
Click here to view |
 | Figure 14: This is a printout of 24-2 threshold test with SITA SWAP. Gray scale is deceptive. Total deviation probability plot shows some defective points, however, the pattern deviation probability plot is normal. Global indices are given in a box (reminder for SWAP test). From glaucoma point of view the field should be considered as normal.
Click here to view |
{Figure 14} | Figure 15: This is the single field analysis of Appa Perimeter (Field explorer) (Printing format of 7 in 1). This should be read in the same manner as Humphrey. Threshold values are given as "Raw data" "compnorm" is total deviation of Humphrey, pattern filter (GH) is pattern deviation of Humphrey. Global indices are calculated overall as well as each quadrant wise. All are self explanatory. Follow the principle of "GRADES" while reading the printout.
Click here to view |
 | Figure 16: Single field printout right eye from medmont perimeter. General information about the test is given on top. The reliability parameters and global indices are given at the bottom. "Level" is the actual threshold values in numbers ("O" indicates that patient has seen the OdB stimulus, whereas "No" indicates, patient cannot see OdB stimulus at all) and grayscale, "Age normal deviation" is the "total deviation" of Humphrey and "Patient HOV" is the "Pattern deviation of Humphrey". They are also given with the probability symbols. Indices are given at the bottom. Mean defect and PSD of Humphrey are overall defect and pattern defect of Medmont, respectively. The values on Humphrey and Medmont are different due to difference in calculation strategy. Cluster analysis of Medmont is same as glaucoma hemifield test of Humphrey
Click here to view |
 | Figure 17: This printout is of progression analysis in addition to the single field analysis. A baseline is the average of up to three results (by default 2). In the Difference from Baseline, the present exam and the two immediately previous exams are compared point wise to the baseline exam. If there is no significant deterioration in the threshold value of a given point between the present exam and the baseline exam, it is displayed as a dot on the plot. If the selected point has deteriorated significantly in the present exam only, it is displayed as an open square. If the selected point has deteriorated progressively in the current exam and the immediately previous exam, it is displayed as a half filled in square. If the selected point has deteriorated progressively in the current exam and the two immediately previous exams, it is displayed as a filled in square
Click here to view |
 | Figure 18: This is the printout of single field analysis of "Octopus". Follow the principles of "GRADES", "values" are the actual threshold dB of the patients locations tested, which is also depicted as gray scale. "Comparison" and "corrected comparison" are total deviation and pattern deviation of Humphrey, respectively, with probability plots. Test parameters, reliability indices, and global indices are given at the bottom of the printout.
Defect curve (Bebic curve) is drawn from total deviation
Click here to view |
 | Figure 19: The single field analysis on "Octopus" can also be printed in the form of Humphrey printout. This is the printout of the same test as given in Figure 17. Please note that this does not give the outcome of "glaucoma hemifield test" as this test is the copyright of Humphrey
Click here to view |
Follow up examination frequency
The progression can be charted earliest from the third field onwards, by applying the baseline of the first two satisfactory fields obtained. The time frames in which these three reports need to be acquired depend on the stage of the disease, age of the patient, and anticipated rate of progression, in short, are individualized.
Here it has to be remembered that no change in the strategy is acceptable and all tests need to be done on the same perimeter, same place.
The frequency of testing hence forth can follow the rule of Bayesian mathematics and a concept called "Adaptive testing" fields with suspected progression only, need validation or falsification. It implies that patient can be monitored at low frequency as far as baseline is made over a short period of time and frequency of test increased when progression is suspected. The next scheduled test is performed sooner if progression is suspected based on event analysis. Three tests in the first year in early glaucoma is a reasonable start.
Convincing the patient for such repeated tests can be quite a daunting task requiring table time and patients understanding of the disease to ensure a good compliance. Comparing field test in glaucoma to blood sugar test in diabetes and blood pressure (BP) monitoring in a hypertensive, can be good corollaries to explain to the patient about the status field testing enjoys in glaucoma therapy today and the need for a regular follow up test! Disconnecting the chances of going blind, to what is seen in the field report, can also help allay the psychological fears associated with this test and improve performance.
The progression algorithms are the GCP for the full threshold programs and the Guided progression analysis (GPA) for the SITA in the Humphrey device and the Peri trend in the Octopus machine are explained with examples as shown:
This forms the core of the Glaucoma therapy.
When would you like to shift the strategy/program/target size
Full threshold 30-2 tests are ideal and give maximum information. They are, however, time consuming and patients can develop fatigue and can compromise performance. The most commonly used 24-2 SITA standard test reflects a good compromise between time and information obtained.
A patient who cannot sit for long for physical reasons can be relegated to the 24-2 SITA fast test where the progression algorithms are still available, but should not be a strategy of first choice. This is the last strategy (i.e., requiring the least time) that has a progression software back up.
(Herein lies the advantage of the staging and phasing methods of the Octopus machine, wherein useful information can also be obtained from an interrupted test.)
The pattern deviation plots cannot be calculated for the patients with severely depressed fields. A similar comment will appear on the field printout in advanced disease. In such cases a larger (in terms of area tested) field test can be shifted to a smaller one with more points tested in the same area to derive more information; that is, a 24-2 (that has points 6° apart in a 24° field) can be shifted to a 10-2 test (where two points are 2° apart). In the bargain, however, we lose the use of the GPA (it cannot be used for tests less than the 24-2 and an MD worse than 20 dB) and have to follow the patients on the absolute MD value and the overview prints thereafter.
If the patient worsens even further than the 10-2; the macula program can be used, where the quadrant totals are the surrogate of the residual field and visual function.
In patients with vision so poor as not to see the fixation spot, the fixation target can be changed to the small or the large diamond. (Especially in those with central scotoma/foveal pathology.)
In compatible machines (only the higher versions) an option of using stimulus size 5 is available to carry on testing even further if required. This can be done alone or in addition to the change in fixation target. These tests do not, however, have any cushion of the statistical packages.
Pre test counseling in perimetry
Can take time initially but it is time well spent in the end as it can decrease testing times and improve reliability and reproducibility, shortening the learning curve.
The following are the facts told before the first and subtly reminded before every test.
"This test gives us an idea of the "ability to see on the sides and nothing more, (this does not tell us whether or not you will lose vision)".
This will need to be repeated from time to time just like measuring BP in hypertensive patients and checking sugar levels in diabetics to see if things are under control. We will inform you from time to time the schedule you will need to follow.
The basic idea of a good test is that you should be as comfortably seated as possible with one eye tested at a time but with both eyes open, but one under cover.
Please blink normally and try and time it at the moment you see a stimulus (spot of light). Blinking does not mean missing a point. Please time the blink with pressing the button so that you do not miss a point. Pause the test if you need to speak or feel tired.
You are not supposed to see the stimulus at all the time (as per the machine program) or at fixed intervals so press the buzzer only if you see one.
Some stimuli can be bright or some can be dim, press the button irrespective of the change in intensity as far as you see it. No one can see all the stimuli in the entire test, hence not to get worried if anytime you feel that you do not see one."
How to choose a perimeter for your practice?
While choosing to buy a perimeter one has to answer to himself/herself a few fundamental questions:
- Aim of buying a perimeter - dedicated glaucoma practice or for screening for field defects; glaucomatous or otherwise.
- Validated database - whether the instrument selected has a validated and regional database available/inbuilt for comparison. This is fundamental, as the field thresholds are compared with age matched normals.
- Validated progression analysis - if the perimeter is being bought for dedicated glaucoma work, does it have the mettle to statistically compare successive fields and bring out a result. After all the aim of glaucoma therapy is to identify and prevent progression. Eyeballing in fields can be dangerous and nonscientific method of treatment, leave alone the difficult nature of the task.
- Price and the service backup in the region of practice: It is a more practical issue, but worthy of consideration if all the above are decided upon and satisfied.
It is noteworthy, however, that if an ophthalmologist needs a perimeter for dedicated glaucoma practice, one has to sadly put the budget considerations on the backburner as only a very few satisfy the criteria as on today.
The cost cutting can be done only on the features beyond the licensed progression software! (Such as the head and vertex monitor, pupillometry, SWAP and the variable stimulus size option that can easily double up the cost.)
For those who want it for screening purposes, a plethora of options are available to suit every pocket size.
The other perimeters that have undergone a clinical trial vis-a-vis the Humphrey is the medmont perimeter and the frequency doubling perimetry (FDP) matrix, for which comparative data is available and can help in the decision making process while buying for oneself as a screening tool today. [6],[7],[8]
Out of the two, the Matrix FDP has more studies to its credit with the (receiver operating characteristic (ROCs) curve) for the PSD equivalent or better than the SAP in some studies.
| References | |  |
| 1. | Anderson DR, Drance SM. Natural history of normal tension Glaucoma. Ophthalmology 2001;108:247-53. 
[PUBMED] |
| 2. | Heijl A, Bengtsson B. Natural history of open angle Glaucoma. Ophthalmology 2009;110:2271-6.
|
| 3. | Ahrlich KG, De Moraes CG, Teng CC, Prata TS, Tello C, Ritch R, et al. Visual field progression differences between normal tension and exfoliative high tension Glaucomas. Invest Ophthalmol Vis Sci 2010;51:1458-63.
[PUBMED] |
| 4. | Heijl A, Bengtsson B, Chauhan BC, Lieberman MF, Cunliffe I, Hyman L, et al. A comparison of visual field progression criteria of 3 major Glaucoma trialsin early manifest Glaucoma trial patients. Ophthalmology 2008;115:1557-65.
|
| 5. | Leske MC, Heijl A, Hyman L, Bengtsson B. Early manifest Glaucoma trial, design and data. Ophthalmomogy 1999;106:2144-53.
|
| 6. | Liu S, Lam S, Weinreb RN, Ye C, Cheung CY, Lai G, et al. Comparison of standard automated perimetry, frequency-doubling technology perimetry, and short-wavelength automated perimetry for detection of glaucoma. Invest Ophthalmol Vis Sci 2011;52:7325-31.
[PUBMED] |
| 7. | Landers J, Sharma A, Goldberg I, Graham SL. Comparison of visual field sensitivities between the Medmont automated perimeter and the Humphrey field analyser. Clin Experiment Ophthalmol 2010;38:273-6.
[PUBMED] |
| 8. | Landers J, Sharma A, Goldberg I, Graham S. A comparison of global indices between the Medmont Automated Perimeter and the Humphrey Field Analyzer. Br J Ophthalmol 2007;91:1285-7.
[PUBMED] |
[Figure 1], [Figure 2], [Figure 3], [Figure 4], [Figure 5], [Figure 6], [Figure 7], [Figure 8], [Figure 9], [Figure 10], [Figure 11], [Figure 12], [Figure 13], [Figure 14], [Figure 15], [Figure 16], [Figure 17], [Figure 18], [Figure 19]
| This article has been cited by | | 1 |
Agreement between the newly developed OCT glaucoma staging system and the standardized visual field glaucoma staging system 2 |
|
| Elshimaa A Mateen Mossa, Heba Khallaf, Khulood Muhammad Sayed | | European Journal of Ophthalmology. 2021; : 1120672121 | | [Pubmed] | [DOI] | | | 2 |
Progression to blindness in 20 years among patients with glaucomatous visual field loss in a tertiary hospital in the Philippines |
|
| Cathleen Camille Cabrera,Joseph Anthony Tumbocon,John Mark de Leon | | BMJ Open Ophthalmology. 2020; 5(1): e000424 | | [Pubmed] | [DOI] | | | 3 |
A Smart Context-Aware Hazard Attention System to Help People with Peripheral Vision Loss |
|
| Ola Younis,Waleed Al-Nuaimy,Fiona Rowe,Mohammad Alomari | | Sensors. 2019; 19(7): 1630 | | [Pubmed] | [DOI] | | | 4 |
Development of an LCD-Based Visual Field System |
|
| Jin-Ho Joo,Jihyoung Lee,Heecheon You,Jaheon Kang | | Journal of Korean Medical Science. 2018; 33(3) | | [Pubmed] | [DOI] | | | 5 |
INTERPRETATION OF AUTOPERIMETRY |
|
| Mishra P,Sridevi V,Parth R,Ramya M,Abbin G M,Nagalakshmi V S | | Journal of Evolution of Medical and Dental Sciences. 2014; 3(38): 9847 | | [Pubmed] | [DOI] | |
|
 |
|
|
|
Search Pubmed for