1. INTERPRETATION OF
AUTOMATED PERIMETRY

2. © Thomas R
Automated perimetry

3. © Thomas R
Automated perimetry
I. Perimetry logic
II. Identifying field defects
III. Criteria for glaucomatous defects
IV. Detecting glaucomatous progression
V. Advanced field defects

4. © Thomas R
Bracketing strategy
B
A

5. © Thomas R
Normal thresholds
• Mean threshold in disease-free fields
• In a given age group
• At a given location in the visual field
• Mean normal values are stored in the
automated perimeter and compared
against patient data

6. © Thomas R
Computers and ease of
interpretation
Sensitivity
+
Simple set of rules
Computer
Diagnosis

7. © Thomas R
Perimeter logic (1)
• Sensitivity determined at each location
• Normal range developed
• Normal range is arbitrary
– Includes the values of 95% of the
normal population

8. © Thomas R
Perimeter logic (2)
• ‘Abnormal’ values include the lowest
5% of those in normal individuals
• Therefore, 5% of normal individuals
will be labelled abnormal
‘Abnormal’ is not the same
as diseased

9. © Thomas R
Perimeter logic (3)
• General population – 100 tested
• 1% glaucoma; 99% normal
• Six will have abnormal tests:
• 1 glaucoma patient
• 5 normal individuals

10. © Thomas R
Perimeter logic (4)
• Clinic population – 100 tested
• 30% glaucoma; 70% normal
• 33 will have abnormal tests
• 30 glaucoma patients
• 3 normal individuals

11. © Thomas R
Interpretation is not child’s play
Automated perimeters still need interpretation

12. © Thomas R
Before interpretation …
… a few principles

13. © Thomas R
Rely on threshold tests
• First real evidence of glaucoma
• Detect scotoma
• Detect depression of the ‘hill’ of vision
• May predict visual loss

14. © Thomas R
Screening tests
• Screening
• Fishing
• Fatigue

15. © Thomas R
Interpreting decibel values is
just half the challenge …
• False positives
• False negatives
• Fixation
• Fluctuation
• Strategy
• Experience
• Technicians
• Artefacts

16. © Thomas R

17. © Thomas R
Optimising patient performance
• Choose the most appropriate investigation
– Test pattern and strategy
• Ensure the patient is comfortably positioned
– Support feet, back and arms
– Adjust chin rest
– Cover the other eye fully
• Provide careful instructions prior to the test
• Support the patient during the test
• Give feedback on test performance
SEAGIG. Asia Pacific Glaucoma Guidelines. 2003–2004.

18. © Thomas R
A word about the grey scale
• Never use the grey scale alone for
interpretation
• It is useful to educate the patient
and to identify false-positive
and false-negative errors

19. ‘White’ scotomas associated
with false positives
© Thomas R

20. © Thomas R

21. ‘Clover leaf’ pattern associated
with false negatives
© Thomas R

22. © Thomas R
Using the grey scale
• To educate the patient
• White scotomas with false positives
• Clover leaf pattern with false negatives
• Never interpret using the grey scale alone

23. © Thomas R
Questions
• Is there a field defect?
• Is it due to glaucoma?
• Is the defect progressing?

24. © Thomas R
Is the field abnormal?
• Without obvious defects, it is difficult
to make a decision based on the
first field
• Repeat examinations provide
definitive information
• Never make a diagnosis based on
the visual field alone

25. Interpret the field
systematically using
zones 1–8
© Thomas R

26. 2
© Thomas R
AGE 57 2
FIXATION LOSSES 0/24
FALSE POS ERRORS 0/14
FALSE NEG ERRORS 1/13
QUESTIONS ASKED 449
FOVEA: 33 DB
TEST TIME 13:59

27. • Just glance at the
grey scale and move
on to zones 4 & 5
• Never interpret using
the grey scale alone
3
© Thomas R

28. © Thomas R
• Point-by-point difference from the
expected value for age-related
normal individuals
• Reveals generalised depression
• Cannot confirm a scotoma
• Look at the number and pattern
of symbols
Zone 4: total deviation

29. © Thomas R
180° 0°
40 dB
0
30
20
10
90 60 30 0 30 60 90
Normal ‘hill’ of vision

30. © Thomas R
180° 0°
40 dB
0
30
20
10
90 60 30 0 30 60 90
Generalised depression

31. © Thomas R
180° 0°
40 dB
0
30
20
10
90 60 30 0 30 60 90
Generalised depression with
‘hidden’ localised scotoma

32. © Thomas R
180° 0°
40 dB
0
30
20
10
90 60 30 0 30 60 90
Pattern deviation plot: scotoma revealed
after adjusting for generalised depression

33. © Thomas R
• Reveals focal defects
after adjusting for
overall depression
(or elevation) of the
hill of vision
• Confirms a scotoma
::
::
Zone 5: pattern deviation

34. Examples of total and pattern
deviation plots in different situations

35. © Thomas R
Normal ‘hill’ of vision

36. © Thomas R
‘Normal’ hill of vision with
localised scotoma
SEAGIG. Asia Pacific Glaucoma Guidelines. 2003–2004.
180° 0°
40 dB
0
30
20
10
90 60 30 0 30 60 90
‘Normal’ hill of vision with
localised scotoma

37. © Thomas R
Generalised depression with
‘hidden’ localised scotoma

38. © Thomas R
Generalised depression

39. © Thomas R

40. © Thomas R
MD –2.18 dB
PSD 4.63 dB; p < 1%
SF 1.24 dB
CPSD 4.44 dB; p < 0.5%
• All the information
from all the points
tested is reduced to
single numbers
Global indices
MD, mean deviation; PSD, pattern standard deviation; SF, short-term fluctuation;
CPSD, corrected PSD.

41. • Both MD and PSD
are derived from the
total deviation plot
• However, they
provide different
types of information
© Thomas R

42. © Thomas R
• Average of all the numbers
in the total deviation plot
• Indicates overall deviation
of the visual field from
normal
• Positive numbers indicate
an ‘elevated’ field
• Negative numbers indicate
a ‘depressed’ field
Global indices: mean deviation (1)
MD –2.18 dB
PSD 4.63 dB; p < 1%
SF 1.24 dB
CPSD 4.44 dB; p < 0.5%

43. © Thomas R
• Provides similar
information to total
deviation
• Cannot confirm the
presence of a scotoma
Global indices: mean deviation (2)
MD –2.18 dB
PSD 4.63 dB; p < 1%
SF 1.24 dB
CPSD 4.44 dB; p < 0.5%

44. © Thomas R
• Also derived from the
total deviation plot
• Indicates the degree
to which the numbers
differ from each other
• Highlights ‘roughness’
or ‘pot-holes’ in the hill
of vision
Global indices:
pattern standard deviation (1)
MD –2.18 dB
PSD 4.63 dB; p < 1%
SF 1.24 dB
CPSD 4.44 dB; p < 0.5%

45. © Thomas R
Global indices:
pattern standard deviation (2)
MD –2.18 dB
PSD 4.63 dB; p < 1%
SF 1.24 dB
CPSD 4.44 dB; p < 0.5%
• Provides similar
information to the
pattern deviation
• Calls attention to
scotomas

46. © Thomas R
28
28 29 33 32 32
32
30
30
33
32
29 31
28
30
29
29
29
21
26
2728293332
31
24
29
31
30
2928
26
29
29
27
26
26
25
28 29 32 32 32
32
29
30
32
31
29 31
25
28
29
25
20
27
26
272803434
32
29
32
33
30
3032
25
27
29
28
23
29
(31)
(32)
(32) (30)
(31)
(30)
(33)
(30) (31)
(33)
• Intra-test error in
threshold determination
• Standard deviation of
10 predetermined
points that are each
tested twice
Global indices:
short-term fluctuation

47. © Thomas R
Global indices: corrected
pattern standard deviation
• CPSD is PSD corrected for the SF
– If SF is due to unreliability,
then CPSD is better
– If SF is due to pathology,
then PSD is better

48. © Thomas R
MD
Total
deviation plot
PSD
Pattern
deviation plot
Generalised depression
Can suspect a scotoma
Review of key points
Local irregularity
Confirms scotoma

49. Glaucoma Hemifield Test
© Thomas R

50. © Thomas R
Zone 7: Glaucoma Hemifield Test
44 5
3
2
1

51. © Thomas R
GHT, Glaucoma Hemifield Test.

52. © Thomas R

53. 8
© Thomas R

54. • Never rely on the
grey scale alone to
make a diagnosis
• Never rely on the
visual field alone to
make a diagnosis
• Always correlate
with the clinical
findings
© Thomas R

55. © Thomas R
Questions
Is there a field defect?
• Is it due to glaucoma?
• Is the defect progressing?

56. © Thomas R
Glaucomatous defects
• Characteristics of glaucomatous defects:
– Asymmetrical across the horizontal midline*
– Located in the mid-periphery*
(5–25 degrees from fixation)
– Reproducible
– Not attributable to other pathology
– Localised
– Correlating with the appearance of the optic disc
and neighbouring areas
* Applicable to early/moderate cases.
SEAGIG. Asia Pacific Glaucoma Guidelines. 2003–2004.

57. © Thomas R
Criteria for glaucomatous
defects (1)
Pattern deviation plot
• ≥ 3 non-edge points
with p < 5%
• One point with p < 1%
• Cluster in arcuate area

58. © Thomas R
Criteria for glaucomatous
defects (2)
CPSD or PSD
depressed
with p < 5%

59. © Thomas R
Criteria for glaucomatous
defects (3)
Abnormal GHT

60. © Thomas R
Three criteria for glaucomatous
defects*
1. Pattern deviation plot
– ≥ 3 non-edge points
with p < 5%
– One point with p < 1%
– Cluster in arcuate area
2. CPSD or PSD
depressed with p < 5%
3. Abnormal GHT
*Anderson DR, Patella VM. Automated Static Perimetry. 2nd Edn. St Louis: Mosby, 1999.

61. • Try interpreting
this visual field,
going from
zones 1–8
© Thomas R

62. 2
2
Visual acuity should correlate
with the foveal threshold
© Thomas R

63. • Continue
interpreting
this visual field:
zones 3–8
• Remember:
no more than a
glance at the
grey scale
© Thomas R

64. © Thomas R
Revision: typical cataract

65. © Thomas R
Revision: typical glaucoma

66. © Thomas R
Revision: glaucoma and cataract

67. © Thomas R
Does this patient have
glaucoma? (1)
Only if the defects are repeatable and correlate with disc and clinical findings

68. © Thomas R
Does this patient have
glaucoma? (2)
Only if the defects are repeatable and correlate with disc and clinical findings

69. © Thomas R
Questions
Is there a field defect?
Is it due to glaucoma?
• Is the defect progressing?

70. © Thomas R
Principle
• Is there a field defect?
• Is it due to glaucoma?
• Is the defect progressing?
– Compare to selected baseline
– Discard learning fields from baseline
– Recognise ‘false’ progression

71. © Thomas R
False progression
• Learning curve
• Long-term fluctuation
• Artefacts
• Patient factors
• Pupil size

72. Pupil: 1 mm
© Thomas R

73. Pupil: 2.5 mm
© Thomas R

74. © Thomas R
Detecting change
• Change analysis – box plot
• Overview programme
• Glaucoma progression analysis™
(GPA™)
1. Select appropriate baseline
2. Discard learning fields from baseline

75. © Thomas R
Overview programme
• Sequential series of fields for the same
patient over a period of time
• Has all the single field information,
including total and pattern deviation plots
• Tells us at a glance what is happening
and allows us to deduce WHY it is
happening

76. Fluctuation over time
© Thomas R

77. Overview: the patient developed a cataract, which was
extracted. Note that the pattern deviation plot remains clear.
© Thomas R

78. Overview: glaucoma is progressing. Both the total and pattern
deviation plots show worsening.
© Thomas R

79. © Thomas R
Overview
programme shows
progression
Full threshold
SITA standard
SITA, Swedish Interactive
Threshold Algorithm.

80. © Thomas R
Overview
programme shows
progression
• SITA is different
from full threshold
• Can't compare
apples to oranges
• Fields may fluctuate

81. © Thomas R
Glaucoma Progression Analysis™*
• GPA™ is now in clinical use
• Change is based on the pattern deviation plot
• Compatible with both SITA and full threshold
(baseline only)
*Carl Zeiss Meditec.

82. © Thomas R

83. GPA™
Right eye:
baseline
© Thomas R
GPATM, Glaucoma Progression
AnalysisTM.

84. GPA™
Right eye:
follow-up
© Thomas R
GPATM, Glaucoma Progression
AnalysisTM.

85. © Thomas R
3 or more points deteriorate in at least 2 consecutive tests
© Thomas R

86. 3 or more points deteriorate in at least 3 consecutive tests
© Thomas R

87. GPA™
Left eye:
baseline
© Thomas R
GPATM, Glaucoma Progression
AnalysisTM.

88. GPA™
Left eye:
follow-up
© Thomas R
GPATM, Glaucoma
Progression
AnalysisTM.

89. © Thomas R

90. © Thomas R
Diagnosis of visual field
progression
• Different for research purposes
– Set criteria in isolation
• Clinical follow-up scenario
– Other criteria (IOP, disc changes) to consider
– A corresponding repeatable change is sufficient
– If in doubt, REPEAT
• Baseline fields are not constant
– Select accordingly

91. Don’t forget to discard
‘learning’ fields from
baseline
© Thomas R

92. © Thomas R
Follow-up of advanced
field defects

93. Advanced field defect
Why is the pattern
deviation plot not
showing a defect?
© Thomas R

94. Not enough points with
sensitivity to produce the
pattern deviation plot
© Thomas R

95. Follow-up with a 10–2 programme –
now there are enough sensitive points
to produce a pattern deviation plot
© Thomas R

96. Advanced defect
and/or low sensitivities –
follow-up with a size V
target
Disadvantage: we lose
statistical help for
interpreting the total and
pattern deviation plots
© Thomas R

97. © Thomas R
More advanced defects: follow
with macular programme

98. Macular programme in
advanced glaucoma
© Thomas R

99. Size V target: macular split
Macular split (0 dB) next to the fovea
with a size V target may predict ‘wipe out’
© Thomas R

100. © Thomas R
Recent developments: SITA
• Asks smart questions
• Gold standard
• More abnormal points on pattern
deviation
• Shallower defects
• Significant because of less variability

101. SITA is interpreted in
the same 8 zones as
previously described
© Thomas R
SITA, Swedish Interactive
Threshold Algorithm.

102. SITA uses the same
criteria to identify a
glaucomatous field
defect
© Thomas R
SITA, Swedish Interactive
Threshold Algorithm.

103. Applying the skills
Does this field fulfil
the criteria for a
glaucomatous defect?
Does this patient
have glaucoma?
© Thomas R

104. Not unless the field
defect correlates with
clinical findings
Never diagnose
based on the visual
field ALONE
© Thomas R

105. © Thomas R
Automated perimetry: warning
Sophisticated techniques and elaborate
data printouts should not seduce us into
a false sense of security or a misplaced
belief in the validity or reliability of
automated perimetry*
*Zalta AH. Ophthalmology 1989; 96: 1302–11.

106. INTERPRETATION OF
OCTOPUS FIELDS

107. © Thomas R
Test parameters – Octopus vs.
HFA
4–2 dB bracketing
strategy
SITA standard
SITA fast
4–2–1 dB bracketing
strategy
Dynamic
Tendency oriented
perimetry (TOP)
Test strategies
0–40 dB0–40 dBMeasuring range
Goldmann I–V
200 ms
10,000 asb
Goldmann III and V
100 ms
4800 asb
Stimulus size
Stimulus duration
Luminance for 0 dB
10 cd/m2 (31.5 asb)10 cd/m2 (31.4 asb)Background luminance
Aspherical bowlDirect projectionBowl type
HFA 700 seriesOctopus 300Parameter
Fankhauser F et al. Automated Perimetry: Visual Field Digest. 5th
Edn. Köniz: Haag-Streit AG, 2004.

108. [[Credit line to be added]]
Probability
plots
Comparison
tables
Grey scale
Patient data
and refraction
Strategy and
test parameters
Actual values
Bebie (defect)
curve
Deviation
Global indices
RP: permission
requested

109. © Thomas R
Octopus global indices
• MS Mean sensitivity
– Average of all measured values
• MD Mean defect
– Average of all values corrected for age
• LV Loss variance
– Equivalent to PSD
• SF Short-term fluctuation
• CLV ‘Corrected’ loss variance
– Equivalent to corrected PSD
• RF Reliability factor

110. © Thomas R
Is the visual field abnormal?
• Octopus criteria for a visual field defect1
– MD greater than 2 dB
– LV greater than 6 dB
– At least 7 points with sensitivity decreased
by ≥ 5 dB, three of them being contiguous
• How do these compare to HFA criteria?
1. Morales J et al. Ophthalmology 2000; 107: 134–42.

111. © Thomas R
HFA criteria for glaucomatous
defects*
1. Pattern deviation plot
– ≥ 3 non-edge points
with p < 5%
– One point with p < 1%
– Cluster in arcuate area
2. CPSD or PSD
depressed with p < 5%
3. Abnormal GHT
*Anderson DR, Patella VM. Automated Static Perimetry. 2nd Edn. St Louis: Mosby, 1999.

112. Comparison of Octopus and
HFA fields from a single patient
© Sihota R

113. © Thomas R
Patient data, strategy and test
parameters
© Sihota R

114. © Sihota R
Grey scale

115. © Thomas R© Sihota R
Octopus: comparison tables
Phase I Phase 2 Mean
# 59 59 59
MS 21.8 18.6 20.2
MD 6.8 10.1 8.5
LV 46.6 73.2 51.0
CLV 42.2
SF 4.9
RF 3.1

116. © Thomas R© Sihota R
GHT Outside normal limits
MD –7.58 dB; p < 0.5%
PSD 6.30 dB; p < 2%
SF 2.27 dB; p < 10%
CPSD 5.75 dB; p < 1%
HFA: total and pattern deviation