Thursday, October 20, 2016



PERIMETRY PART 4

 

SCREENING PROGRAMS


SINGLE-LEVEL SUPRATHRESHOLD TEST:

A stimulus 2-6 dB higher than expected hill-of-vision [suprathreshold] is presented. The response is recorded as seen [normal] or not seen [defect]. This is also known as the “Threshold related strategy”.
TWO-LEVEL SUPRATHRESHOLD TEST:
These are also known as the “3-zone screening tests”. The VF is divided into 3 categories: Normal, relative defect and absolute defect.
ONE-LEVEL SUPRATHRESHOLD SCREEN:
After quantification of defects, this test is performed.
Thresholds are measured at the locations which are determined to be abnormal. A single-level threshold test (given above) assesses defects 6 dB higher than the hill of vision, therefore shallow defects and early progressive depression of the VF can be missed. The test can be very time consuming if many points in the field of vision are abnormal.


THRESHOLD PROGRAMS

These programs measure the DLS in the central 240 to 300.
These strategies include:

FULL THRESHOLD:
The sensitivity at a particular point in the visual field is measured in decibels (dB) by convention.
The DLS at every point in the VF is assessed using the 4-2 staircase or bracketing algorithm. The perimeter projects stimuli ranging from 0 dB (brightest) to 50 dB (dimmest) in a graded fashion to determine the threshold of a particular point and this technique is called the “staircase” or “bracketing strategy”. The stimulus strength is increased in 4 dB increments till it is perceived. Subsequently the strength is reduced by 2 dB till it is no longer perceived. The process involves a double-crossing of each point and tries to attain maximum accuracy in grading threshold of a particular point.

Testing starts with a suprathreshold [seen] or infrathreshold [not seen] stimulus. In a suprathreshold stimulus, the intensity of the stimulus is decreased in 4 dB steps until the stimulus is no longer seen (threshold crossed). The stimulus intensity is then increased in 2 dB steps until the threshold is crossed a second time and the stimulus is seen again. The Humphrey Perimeter uses the intensity of the last seen stimulus as threshold.
If the initial stimulus is close to the actual threshold the test can be done faster. Humphrey Perimeters use a “region growing technique” to determine the starting level for each point. After testing the fovea and blind spot location the threshold is measured at a single point in each quadrant. The threshold level measured at these 4 cardinal points is used to start testing the adjacent regions. If the thresholds are more than 5 dB from the expected values, the location is retested. The second result is printed below the first in parentheses.



FASTPAC:
The DLS is determined at every point in the VF. However, instead of the 4-2 bracketing strategy, the threshold is measured using 3 dB steps and the threshold is crossed only ONCE. 

SWEDISH INTERACTIVE THRESHOLD ALGORITHM (SITA):
By using SITA the test time is reduced by half. SITA-Standard and SITA-Fast capture twice the amount of data in the same time as a Humphrey Full Threshold and FastPac respectively.
Testing is started at the level which is near threshold. The time between presentations of stimuli is customized to the patient’s response time, making it more user-friendly. The strategy is able to calculate the false-positive and false-negative responses from the threshold measurements, eliminating the need for separate testing of the false-catch trials. Thus, the test time is reduced by 6% compared to a full threshold strategy.
SITA uses artificial intelligence and computer modeling, incorporating probability models of normal and glaucomatous VFs to provide more efficient testing of the VF. The testing is interactive. Thus, the response from the patient is used to predict the future responses. The data captured is incorporated in probability models of normal and glaucomatous VFs, normal age-corrected threshold values, patterns of glaucomatous damage and multiple frequency of seeing curves in normal and abnormal states. At the end of the test, the threshold is calculated based on all the data captured by the system. The program is also able to discard the likely false responses.
The program estimates the threshold expected at each point on a continuous basis and stops testing when the estimated error is less than a predetermined value.
SITA Standard takes longer to perform as the confidence limit is narrower for SITA standard than SITA Fast. Due to the lower error in SITA Standard, the resultant variability is low and it is more reliable than SITA Fast for future comparisons.
Although SITA-Standard is comparable in reliability to Full threshold strategies, these 2 strategies are not directly comparable. Defects appear shallower on SITA-Fast because of less patient fatigue.

SHORT WAVELENGTH AUTOMATED PERIMETRY (SWAP):
In this test a blue stimulus (440 nm) is presented on a background of yellow illumination.
The yellow background desensitizes the green and red cones, while the blue stimulus activates the blue cones. Thus, the blue cones and their ganglion cell (K or kainocellular cells) connections are tested.
SWAP detects VF defects earlier compared to white-on-white perimetry. However, variability in the inter-test sessions is greater with SWAP. It is also affected by lens-induced changes in the elderly. This test also takes 15% longer time than full threshold testing. Interpretation of the VF results requires the SWAPac on the Humphrey Perimeter. The grey scale does not useful in interpretation of the SWAP results.
It is not known if SWAP detects changes early due to relatively earlier damage to the blue/yellow system or testing of only a subset of the visual system enables earlier detection of the VF defects, even though the damage is not selective. 

  
FREQUENCY DOUBLING TECHNOLOGY

This tests the larger diameter M (magnocellular pathway). These cells occupy 3% to 10% of the ganglion cell population. M-cells are particularly susceptible to glaucomatous damage and are damaged early in glaucoma. Full threshold FDT tests can be completed in less than 6 minutes per eye, while suprathreshold screening takes a minute.
In FDT, 17 regions in the central 200 or 19 regions of the 300 of the VF are tested. Each stimulus is a series of black and white bands flickering at a frequency of 25 Hz. A normal eye perceives the illusion of twice the number of bands more closely spaced. Glaucomatous eyes have preferential damage to the M-cells and they require higher contrast to detect the frequency doubling illusion.
FDT has a 97% sensitivity and specificity in detecting glaucomatous field defects.




Sunday, October 16, 2016

PERIMETRY PART 3


Early VF changes are typically of 2 types: 

(1) Retinal nerve fiber bundle (RNFB) defects
(2) Generalized depression of VF.

(1)The nature of RNFB defects corresponds to the topography of these fibers. Common findings include: (a) Nasal step: A nasal step (of Roenne) represents a difference in the sensitivity above and below the midline in the nasal field. These are usually located in between 200 to 300 of fixation and can further extend into the periphery. In 5% patients the nasal step is present outside the 300 and may be missed on conventional perimetry. A temporal wedge is less common with similar implications. (b) Deformation of the Blind Spot (BS): A homogenous enlargement may occur in conditions such as myopia or aged patients and is not a sign of glaucoma. A vertical deformation of the BS associated with an early arcuate scotoma is seen in glaucoma and known as “Seidel scotoma”. (c) Paracentral scotoma: Focal paracentral scotomas occur in central 200, frequently superonasally. Usually a paracentral scotoma becomes denser and larger with time. A scotoma smaller than 60 of the visual angle can be missed; this is due to the spots being displayed every 60 in the central area with the 30-2 program. In case of a paracentral scotoma close to fixation, it is worthwhile to measure the foveal threshold to evaluate a possible defect. (d) Arcuate defect: An arcuate defect or Bjerrum’s scotoma is a reliable early form of glaucomatous field loss. It draws a scotoma which surrounds the fixation point. Its nasal extreme can come within 100 of fixation or extend to the periphery. Most of the arcuate defects are continuous with the BS but some are disconnected from it. Most early scotomas (nasal step, paracentral scotoma) later enlarge and form an arcuate scotoma, extending from the BS to the median raphe. Arcuate scotomas from the superior and inferior fields may join to form a “Ring scotoma”. (e) Peripheral defects: Rarely, early defects may appear outside the 300. In such cases the scotoma may appear as a nasal step, vertical step or temporal sector defect. In 3-10% of the patients an abnormality of the peripheral field may be present.
Advanced glaucoma is characterized by a small island of central vision and an accompanying temporal island. The latter is usually extinguished prior to the central island.
(2)Diffuse depression of the VF: on automated perimetry a diffuse defect is noted when the MD (Mean Defect or Mean Deviation) is abnormal and PSD (Pattern Standard Deviation) remains within normal limits. [In Goldmann kinetic perimetry, diffuse loss corresponds to a generalized concentric isopter constriction] Often this diffuse loss may disappear on reduction of IOP. Although diffuse depression is considered an early sign of glaucoma, it is non-specific and may occur with media opacities.

If the first 2 fields are consistent and do not show a substantial learning effect (<2 dB MD) they can be used for future comparison. Screening tests can only detect scotomas deeper than 6 decibels.

If there is significant difference between the first and subsequent VFs, the possibilities are:
1.       Learning effect: Some patient’s need some experience to reliably do the test. The VFs should be repeated until a consistent result is obtained. Then, the last 2-3 VFs can then be used as baseline.
2.       Marked true variability: If the patient seems reliable but the fields are truly variable, we can use the average of 3 fields as a baseline by creating a Master field.
3.       Poor reliability: If the VFs vary widely and apparently randomly, attempt should be made to re-educate the patient.

If the first VF is markedly depressed, the 30-2 test (which normally uses a size III target) can be done with a Size V target. Marked constriction (the points mainly falling in the central 300) indicates that a 10-2 VF should be done. This program tests only the central 100 using points separated by 20. The 10-2 field provides higher resolution data points, identifies scotomas closer to fixation and reduces patient anxiety by focusing on the seeing areas.
If the VF is both constricted AND depressed a 10-2 program using a Size V target can be done. Increasing the stimulus size leads to spatial summation and provides more data points to follow.

VISUAL FIELD INTERPRETATION:

The interpretation of a VF involves 3 steps=

1.       Recognition of artifact.

2.       Determination of reliability

3.       Assessment of damage


1.       Two common artifacts are caused by (a) Upperlid: producing a flat, superior scotoma (b) Lens rim defect: These appear if the corrective lens is too far from the eye or is not centered. The defects are usually “sharply demarcated absolute scotomas”. 

2.       Reliability estimates depend on 4 main signs: 

i.                     False positive rate 
ii.                   False negative rate
iii.                  Fixation loss
iv.                 Short-term fluctuation

i.                  False positive rate (FPR): If the patient presses the response button when the machine makes only a noise and no target is shown, it is recorded as ‘false-positive”. Usually, the perimeter will signal low reliability if there are 33% or more false-positives (More than 20% according to AAO). However, some authors regard even 1 (or more practically 2 or more) False-positives per VF as a sign of unreliability. The patient should be re-educated about the procedure, asked not to respond in haste and to expect stimuli which may not be visible (which could happen in normal individuals also). If the FPR is more than 20%, it may mask or minimize an actual scotoma and may result in the VF having impossibly high threshold values, producing what are called “white scotomas”.
ii.      False-negative rate (FNR): A “false-negative” response indicates the failure to respond to a stimulus which is 9 dB brighter than the one seen previously in the same location. False negatives might indicate areas of depressed sensitivity and a brighter stimulus may not be truly visible, indicating a reliable field. This could happen from the stimuli falling on the edges of deep scotomata, where short-term fluctuation is quite variable. However, a high FNR usually suggests the field is not as depressed as suggested by the test result. It could happen due to the patient’s tiredness or inattentiveness, indicating an unreliable field.  A FNR of more than 33% suggests test unreliability (AAO).
Patients with fatigability may have a high FNR alongwith a “clover-leaf” or constricted VF. The cloverleaf pattern happens due to good early responses but poor responses later during the test session. It could also suggest “malingering”.
iii.      Fixation losses: The patient is expected to focus at the fixation point throughout the procedure. The Humphrey perimeter monitors fixation by locating the BS and then presenting an occasional maximum stimulus onto it. If the patient responds to it, the machine records it as a fixation loss. A high number of fixation losses may indicate that the centre of the BS was mislocated. A high FPR will also give a high fixation loss rate as well. If the FPR, FNR and short-term fluctuation is low, we can ignore a high fixation loss rate. Similarly if the 2 baseline fields are very similar, we can discard the high fixation loss rate.
iv.            Short-term fluctuation (SF): It is not a reliability parameter but gives useful reliability information. A low (<2 dB) SF indicates good reproducibility on retests at 10 preset points and usually indicates a reliable field. A high (>3 dB) SF indicates a poorer reproducibility, indicating poor patient reliability or field damage with true variability. (Damaged fields are more variable than normal fields). The perimeter retests the 10 points mentioned above and inconsistent responses lead to a high SF value. However, this could occur due to the retest points falling on the edges of a scotoma. Therefore look at the numeric printouts of the double determinations and determine if the patient responded consistently in the areas of best vision.
GLAUCOMA HEMIFIELD TEST:
Glaucomatous damage on a single VF test can be assessed by the Glaucoma Hemifield Test (GHT). This compares 5 corresponding zones in the upper and lower hemifields.


GHT is specific for glaucoma; therefore an abnormal GHT on a minimum of 2 fields suggests the presence of glaucoma.
MINIMUM CRITERIA FOR DIAGNOSING GLAUCOMATOUS DAMAGE 

(1)    Abnormal GHT on atleast 2 fields.
(2)     A cluster of 3 or more non-edge points in a location typical for glaucoma; all of which are depressed on the pattern deviation plot at a p <5 % level and one of which is depressed below p <1 % level.
(3)    Corrected pattern standard deviation which occurs in less than 5% of normal fields on two consecutive fields.


CLASSIFICATION OF DEFECTS

1)      EARLY DEFECT: Damage is neither extensive nor near fixation. The following 3 conditions should be met:
A.      MD is less than -6 dB
B.      On the PSD less than 25% of the points are depressed below the 5 % level and less than 10 points are depressed below 1% level.
C.      No point in the central 50 has a sensitivity of less than 15dB

2)      MODERATE DEFECT: Damage can be significant but there should be no profound damage to the central field. The following conditions should be met:
A.      MD less than -12 dB
B.      On PSD, less than 50% points are depressed below the 5% level and less than 20 points are depressed below the 1% level.
C.      No point in the central 50 has a sensitivity of 0 dB.
D.      Only 1 hemifield may have a point with sensitivity of <15 dB within 50 of fixation.

3)      SEVERE DEFECT: It is characterized by anyone of the following:
A.      MD greater than -12 dB.
B.      On PSD, more than 50% points are depressed below the 5% level or more than 20 points are depressed below the 1% level.
C.      Any point in the central 50 has a sensitivity of 0 dB.
D.      There are points within the central 50 with sensitivity <15 dB in both hemifields.

VISUAL FIELDS MAY DETERIORATE IN THE FOLLOWING WAYS:

1.                                                                   1). A new defect may develop in the previously normal area (normal refers to points on the Master Field within 4 dB of expected on the defect depth plot. A cluster of 3 or more non-edge points, each of which declines ≥5 dB compared to baseline on 2 consecutive fields or a single non-edge point which declines ≥10 dB compared to baseline on 2 consecutive fields.
2.                                                                   2). A preexisting defect may deepen. A cluster of 3 non-edge points, each of which declines ≥10 dB compared to baseline on 2 consecutive fields. The confirming points may differ if they are part of a contiguous cluster.
3.                                                                   3). A preexisting defect may expand. Atleast 2 previously normal points within the central 150, each of which declines ≥10 dB on 2 consecutive fields.
4.                                                                   4). The entire field may develop decreased sensitivity. Generalized depression can be done by inspecting all the fields available, but a recent field can also be compared with baseline and show a decline of all points by 3 dB on 2 occasions.

A SERIES OF FIELDS MAY SHOW THE FOLLOWING CHANGES:

a)      Short-term fluctuation remains stable (indicating reliable fields).

b)      MD gradually and steadily declines (around 1% level and is not explained by media opacities or pupillary size).

c)       PSD and corrected PSD (CPSD) increase (indicating increase in the irregularity of the VF, with the decline not being generalized but only in some areas of the field). The decline is usually observed over 5 VFs.