1. Head-to-Head Comparative Study of Two Optical
Biometric Devices in Modern Cataract Surgery
8 Shareef Mahdavi • SM2 Strategic • Pleasanton, CA 7
Today's cataract surgeon has adopted non-contact optical
biometry as the standard of care in performing IOL calculations. Figure 1: IOLMaster vs. Lenstar
While modern formulae incorporate multiple variables as part Inability to Capture
of their calculations, Axial Length and Keratometry readings
McNemar
continue to be the inputs that are most influential in determin- n=124 IOLMaster Lenstar Test P-value*
ing IOL power. Some of the newer generation formulas such as
Holladay 2 and Olsen now incorporate more elements to help AL 8 6.5% 6 4.8% 0.625
increase accuracy. Ks 10 8.1% 14 11.3% 0.481
Given the competitive nature of device manufacturers to ACD 11 8.9% 8 6.5% 0.581
demonstrate technical superiority of their instruments, surgeons LT – – 19 15.3% –
are often left confused in understanding which specific device *
There was no significant statistical difference between the two devices
will best meet their needs in cataract surgery. One surgeon,
Amin Ashrafzadeh, MD of Modesto, California (“Dr. Ash”) sure that the measurements I’m getting are serving to opti-
wanted to answer this question for himself and undertook the mize my surgical outcomes.” Both devices measure axial
task of performing a prospective comparison of two non-contact length, surface curvature via keratometry, white-to-white,
biometers, the IOLMaster 500 (Carl Zeiss Meditec) and the and anterior chamber depth. In addition, the Lenstar mea-
Lenstar LS 900 (Haag- sures lens thickness.
Streit) across a series Beyond the series
of 124 consecutive eyes "Our concern as surgeons is in taking axial length of cataract patients,
scheduled for cataract measurements that are too short. Short AL readings translate both devices were
surgery. His goal was to to hyperopic surprise for both the surgeon and patient.” also used to measure
understand the impact 20 eyes of 10 young,
of biometry on predicted agile patients to deter-
cataract surgery outcomes by comparing the accuracy and mine the impact on workflow by timing how long it takes
repeatability of measurements between the devices. to capture measurements on each device. A statistician was
SM2 Strategic was asked by Carl Zeiss Meditec to summa- employed to conduct tests of significance on the results.
rize the analysis and findings of Dr. Ash and his team. Neither device was able to capture every measure on every
What follows is an eye. Figure 1 shows
overview of the study Figure 2: Difference in Axial Length Measurements that each device was
and its implications for IOLMaster vs. Lenstar (n = 115) unable to obtain at
surgeons trying to deter- 70
least one component
mine which device they 63 of the available cal-
should use with their 60
culations on between
cataract patients. 5% and 11% of eyes;
Lenstar Lenstar
there was no statisti-
Frequency
measures measures
LONGER SHORTER
The Study cal difference in the
20
Dr. Ash utilized 13
capture rate between
both devices equally to the devices. It is also
10
perform pre-operative 3 4 5
3
important to note
1 2 1 2 1 1 1 1 0 2 1 0 1 0 0 0 0 0 0 0 0 1 1 0 0 1 1 0 1 0 2
calculations on a series 1 1 0 0 1
that these data are
0
of 124 eyes prior to Axial Length
not cumulative (e.g.,
More
1.4
0.1
0.2
0.3
0.5
0.6
0.8
0.9
1.1
1.2
1.3
1.5
1.6
1.8
1.9
2.1
2.2
2.3
2.5
2.6
2.8
2.9
3.1
3.2
3.3
3.5
1.0
2.0
3.0
-0.5
-0.4
-0.3
-0.2
-0.1
0.4
0.7
1.7
2.4
2.7
3.4
0
Difference in mm
cataract surgery. “Each Change in Predicted some eyes could get
Refractive Outcome -1 D <-1 D <1 D +1 D +2 D +3 D +4 D +5 D +6 D +9 D +10 D +11 D +12 D
of these devices offers (Dioptors) K reading but not
highly sophisticated ... ... .. ... axial length).
technology,” noted Dr.
Incidence .. . ... ... .. .. .. .. .
... . ..
...
... .. . .. .. . . .
(out of 115 eyes)
.
Ash. “I want to make

2. Finding #1: Axial Length Measurements “Our concern as surgeons is in taking axial length mea-
Differ Between the Devices surements that are too short,” commented Dr. Ash. “Short
The study first analyzed the difference in axial length AL readings translate to hyperopic surprise for both the sur-
measurements on the same eye for a series of 115 eyes that geon and patient.” The impact of the difference in readings is
had axial length measurements from each device. Using the shown in the bottom half of Figure 2. A demarcation line is
formula, shown at each 0.3 mm of Axial Length Difference along the
AL measurement (difference) = IOLMaster – Lenstar, X-axis. Each 0.3 mm step represents approximately 1 diopter
two-thirds of the eyes (67%) fell within +/- 0.1 mm of one of refractive change as described earlier. In this series of eyes,
another. According to if the surgeon chose to
Hossein Zahed, a MS Figure 3A: Right and Left Eye Symmetry in Each Instrument use the Lenstar reading,
in Statistics, the use 29% of eyes (33 of 115)
of a range around the 35% would be at risk for a
desired mean differ- AL Difference = Right Eye minus Left Eye hyperopic surprise rang-
ence of zero (which 30% ing from 1 to 12 diop-
would indicate exact IOLMaster: 89% within 0.3mm ters. Conversely, using
25%
agreement between the the IOLMaster reading
Lenstar: 56% within 0.3mm
devices’ measurements) 20%
would yield only 2 of
is a more appropriate 115 eyes at risk for myo-
real world application 15% pic surprise of 1 diopter.
of statistical analysis.
In clinical practice,
10% Finding #2:
the impact of improper No impact of learning
5%
axial length measure- curve on axial length
ment on predicted 0% measurement
-1.7
-1.4
-1.1
-0.8
-0.5
-0.2
-0.1
-0.4
-0.7
-1.0
-1.3
-1.6
-1.9
-2.2
-2.5
-2.8
-3.1
-3.4
-3.7
-4.0
refractive outcome is The study analysis
that, “most clinicians Axial Length Difference in mm investigated the impact
adopt the rule of thumb of technician learning
that one millimeter curve and whether or
of axial length equals "In this series, if the surgeon chose to use the Lenstar not “operator error”
three diopters of refrac- could be a factor in the
reading, 33 of 115 eyes would be at risk for hyperopic
tive power,” said Dr. resulting axial length
Ash. “While the actual surprise ranging from 1 to 12 diopters." measurement on either
ratio changes for upper device. One would
and lower range IOLs, expect that if a learn-
this rough guideline Figure 3B: Axial Length Difference Through Time ing curve was present,
provides an easy assess- then readings would be
ment tool.” Left Eye Right Eye more variable early in
4
“Because of my long- the study and less vari-
term use of IOLMaster 3.5 able later in the study.
and its wide acceptance, 3 This was not the case.
it serves as the bench- 2.5
Differences in readings
mark for comparison between the devices
Difference in mm
2
in this study.” Using +/- were present through-
0.1 mm as statistically 1.5 out the course of the
the same measurement, 1 study and peaked at
Lenstar’s Axial Length various points without
.5
measurements are any identifiable pattern.
0
shorter than IOLMaster The sole technician that
in 31% of eyes (36 of -.5 took the readings had
115) and longer than -1 ten years experience
0 10 20 30 40 50 60
IOLMaster in 2% of Observation Number (62 patients)
working with ophthal-
eyes (2 of 115). mic diagnostic devices.

3. Finding #3: Intra-patient axial left eyes and right eyes is shown
length measurement differences Figure 4: Mean Keratometric Difference
in Figure 3B.
exist (eye symmetry) = Mean IOLM K - Mean LS K
The next question in the 60 Finding #4: Keratometric
study pertained to right/left eye 50
50
measurements are in general
symmetry of axial length mea- 40
32
agreement between the devices
surement in the same patient. 30
When considering all 103 eyes
20
Clinical experience in cataract that had keratometric measure-
10 7
surgery indicates that most pairs 2 2
4
1 1 1 1 0 1 1 ment on both instruments, 82
0
of eyes, while not identical, (79.6%) had mean K’s within
0
5
0
5
0
5
0
5
0
5
0
5
6
0.5
0.2
-
0.2
0.5
0.7
1.0
1.2
1.5
1.7
2.0
2.2
2.2
to
<-
-
o+
o+
o+
o+
o+
o+
o+
o+
o+
>+
should be relatively close to one 0.25 D; 91 (88.3%) within 0.5 D
24
to
1t
6t
1t
6t
1t
6t
1t
6t
1t
-0.
49
.0
.2
.5
.7
.0
.2
.5
.7
.0
another in axial length as well (Figure 4). When considering
-0.
+0
+0
+0
+0
+1
+1
+1
+1
+2
as refractive error. The threshold the value of the astigmatism in
adopted by most surgeons is a difference that is within 0.3 patients with ≥ 0.75 D of astigmatism, 65% were within
mm or one diopter. Using this threshold, the IOLMaster 0.25 D and 84% within 0.5 D of one another. In addition,
found 89% of eyes measured within 0.3 mm of one 70% were within 10 degrees of axis of one another.
another. The Lenstar found
56% of eyes measured Figure 5: Impact of ACD on Predicted Finding #5: Anterior
within the same parameter. Spherical Equivalent Chamber Depth has
The distribution of the modest impact on
Pt.# 1 2 3 4 5 6 predicted refractive
right/left eye differences for
AL 23.87 23.91 23.98 24.01 22.36 21.58 outcome
each device are shown in
Figure 3A. K1 44.35 43.27 43.10 42.56 42.72 40.71 It is well known that
INPUTS
Next, differences were K2 51.53 47.67 43.95 43.16 44.64 46.46 the greatest impact to IOL
examined between right ACD 3.25 3.42 2.80 2.87 2.66 2.53 calculations is related to
eye measurements on both WTW 11.9 11.6 11.0 11.1 11.6 11.7 AL’s then K’s followed by
devices and separately the the ACD. Six sample eyes
ACD CALC
ADJUSTED IOL
Holladay 2 14.0 D 18.0 D 20.0 D 21.0 D 26.0 D 29.5 D
left eye differences. Using the 3.00 D -0.05 D -0.13 D +0.20 D +0.64 D +0.31 D +0.26 D with true measurements
same range (+/-0.1 mm) as were chosen to represent
4.00 D +0.17 D +0.15 D +0.50 D +0.94 D +0.71 D +0.70 D
in Finding #1 above, the for- the range of IOL powers
Difference 0.22 D 0.28 D 0.30 D 0.30 D 0.40 D 0.44 D
mulae are as follows: based on the IOLMaster
- OD Axial Length Difference = IOLMaster AL OD - Lenstar AL OD data. Different elements were examined to see how much
- OS Axial Length Difference = IOLMaster AL OS - Lenstar AL OS impact they had on the predicted refractive outcome. The
IOL power was calculated based upon the IOLMaster
measurements and analysis was based on Holladay 2 IOL
"The IOLMaster found 89% of eyes measured Consultant software (Figure 5).
within 0.3 mm of one another. The Lenstar found The two instruments had good correlation in ACD
results with 81.3% of the ACD’s within 0.3 mm. The
56% of eyes measured within the same parameter."
IOLMaster was slightly shorter vs. Lenstar, (3.06 vs.
3.26 mm), but had a tighter Standard Deviation (0.527
In right eyes, the magnitude of the difference was always vs. 0.686). When looking at the sample eyes, even when
within 2 mm (range: -0.5 to a 1.0 mm difference in the
+1.5), with 42% of eyes being ACD resulted in the maxi-
Figure 6: Impact of Lens Thickness on
+/- 0.1 mm. Much greater mal impact with the 29.5 D
Predicted Spherical Equivalent
variability was seen in the implant, it resulted in about
left eyes, both in the overall Pt.# 1 2 3 4 5 6 a 0.44 D predicted refrac-
range (0 to 4 mm) and with Holladay 2 14.0 D 18.0 D 20.0 D 21.0 D 26.0 D 29.5 D tive error, and 0.22 D in the
only 26% of left eyes having 2.50mm -0.01 D -0.02 D +0.09 D -0.09 D +0.08 D -0.08 D 14.0 D implant. While still
a difference in axial length an important element, it has
“No Info” +0.01 D +0.00 D +0.13 D -0.06 D +0.15 D 0.00 D
of +/- 0.1 mm. A scatter plot modest impact on the final
5.50mm +0.05 D +0.05 D +0.16 D -0.01 D +0.19 D +0.04 D
diagram showing the differ- IOL power calculation.
Difference 0.06 D 0.07 D 0.07 D 0.08 D 0.11 D 0.12 D
ence between the devices for

4. Finding #6: Lens Thickness Has Minimal Impact They undergo extensive testing on other instruments prior to
on Predicting Refractive Outcome the optical biometry reading. Even with highly cooperative
Of significant interest is the role of lens thickness in patients as used in this study, the added time of the Lenstar
determining the final refractive error following cataract sur- leads to exhaustion.”
gery. The study used the same recommended IOL powers
in the previous analysis from Figure 5 and applied a range Discussion
of lens thickness measurements that cover the full range as Dr. Ashrafzadeh and his team undertook an ambitious
measured by the Lenstar device (average: 4.53 mm; low 2.74 project, motivated by a single question on the mind of virtu-
mm; high 5.41 mm). For analysis purposes, this range was ally all cataract surgeons: “Which device will give me better
adjusted to 2.50 mm and 5.50 mm as shown in Figure 6. results in order to obtain the best visual outcome possible
The Holladay 2 Formula was then for my patients? What I found
applied to show the impact on was that in 2 of every 3 eyes the
"Regardless of the thickness of the lens,
the predicted spherical equivalent instruments perform identically”
for different lens thicknesses (left the resulting difference (1/8th of a diopter he commented. But his concern
column) across the sample eyes or less) is below the limit of subjective was in the volatility of the data;
IOL Powers (top row). Because discrimination of most patients in visual “should I trust the right eye or left
IOLMaster does not calculate LT, eye, IOLMaster or Lenstar?”
the Holladay Formula makes its acuity tests and, more importantly, in In the less forgiving world of
own assessment (shown in the everyday visual tasks." refractive cataract surgery, Dr.
row labeled “No Info”); for each Ash stated emphatically, “I cannot
patient example shown, the “No afford to make a guess.” The data
Info” calculation from Holladay Figure 7: IOLMaster vs. Lenstar Efficiency from this study serve to refute the
2 falls within range calculated by claims that Keratometry is more
the low and high lens thicknesses Average Time to accurate on the Lenstar or that
of 2.50 mm and 5.50 mm, respec- Measure Two Eyes DIFFERENCE differences in Anterior Chamber
tively. An example from Figure 6 N =10 Subjects Depth and Lens Thickness mea-
shows the impact of various lens IOLMaster 97.3 SECONDS* surements have a meaningful
thicknesses on a 14 Diopter lens (1:37 FASTER) outcome on the final refractive
TIME: 1:41 | RANGE (85-127 SECS.)
RANGE (74-113)
implant. At 2.5 mm, the predicted outcome. In this series of 124
spherical equivalent is 0.01 D Lenstar eyes, none of those hypotheses
of myopia while at 5.5 mm it is TIME: 3:18 | RANGE (176-220 SECS.) held true. The different measure-
0.05 D of hyperopia. The resulting ments of Axial Length are a more
difference between these is 0.06 D 0 50 100 150 200
significant and troubling find-
of predicted spherical equivalent. *P-Value = <0.00001 SECONDS* ing. In the narrower context of
Even at the highest power implant performing cataract surgery on
of 29.5 diopters, the difference is 0.12 D. Regardless of the previous LASIK patients, those with prior hyperopic treat-
Lens Thickness of the patient, the resulting difference of ments are excluded from multifocal IOL’s due to visual qual-
1/8th diopter (or less) is below the limit of subjective dis- ity issues. “Risk of significant hyperopic surprise will need
crimination of most patients in visual acuity tests and, more to be addressed with Excimer laser leading to the same exact
importantly, in everyday visual tasks. profile that would have excluded the patient in the first place
and the visual quality issues I do not wish to contend with.”
Finding #7: Significant difference in impact In a broader sense, having 29% of eyes at risk for hyper-
on workflow between the devices opic surprise is bad for cataract surgeons and their patients.
A key consideration for any device is the impact on clinic Of course, most surgeons would re-check and re-measure
workflow. Measurements were taken on 20 eyes of young, Axial Length findings prior to inducing multiple diopters of
agile patients. The results are shown in Figure 7, with the hyperopia. However, that need to re-do measurements goes
average time to measure both eyes of each subject taking against the very reason surgeons invest in these devices in the
3:18 on the Lenstar and 1:41 on the IOLMaster. The differ- first place, which is to save time and reduce errors.
ence of 97 seconds was highly significant both statistically In the end, Dr. Ash concluded that “both devices have
and practically in its impact on clinic workflow. “The lon- their unique appeal, but only one device can be counted on
ger time required for the Lenstar test needs to be put into to avoid surprise and to maintain workflow, and that is the
context, especially with premium IOL patients in my clinic. IOLMaster.”
© Copyright 2012, SM2 Strategic. All Rights Reserved.