Verkicharla, Pavan K.*; Mallen, Edward A.H.†; Atchison, David A.‡
Myopia (short sightedness) is the major cause of poor vision in children and young adults. It has become one of the main areas of research in vision science, with the substantial increase in its prevalence worldwide for the past few decades.1 Recent research indicates that the retinal shape may be an important consideration in myopia progression.2,3 Magnetic resonance imaging can be used to investigate retinal shape, but this is expensive.4 Retinal shape can be estimated by measuring central and peripheral eye lengths, followed by some manipulation,4 but there is no device specifically designed for peripheral measurements. A simple device that is feasible, accurate, noncontact, and inexpensive will be of considerable benefit in myopia research.
Two recent commercial instruments, the IOLMaster (Carl-Zeiss Meditec AG, Jena, Germany) and the Lenstar (Haag Streit, Bern, Switzerland) contain a Michelson interferometer to create partial coherence and to compare the optical path lengths of two beams, one of which is reflected from the cornea and the other which travels into the eye and is reflected from one or more surfaces. The IOLMaster contains a diode laser–producing infrared radiation centered at approximately 780 μm, and the Lenstar contains a super luminescent diode–producing infrared radiation centered at approximately 820 μm. The diode laser or super luminescent diode sources have wide bandwidths with corresponding short coherence length so that strong interference signals occur only when the optical path lengths are similar rather than differing by multiples of wavelengths. The IOLMaster uses the partial coherence interferometry principle only for axial length measurements, whereas the Lenstar provides corneal thickness, anterior chamber depth, lens thickness, retinal thickness, and vitreous depth. The IOLMaster assumes a single refractive index within the eye (group refractive index, 1.3549), but the manufacturer of the Lenstar does not indicate what refractive index or indices are used. These instruments provide better resolution (0.01 to 0.02 mm) than ultrasound (0.10 mm) and magnetic resonance imaging (0.15 mm).5,6 Both instruments also obtain corneal curvature information, and the IOLMaster measures anterior chamber depth by a split beam technique.
There is consensus that these instruments have high repeatability and good agreement for on-axis axial length measurements,5–15 but these comparisons have not been made in the few studies using them for measuring peripheral eye lengths,16–19 and we will address this.
A total of seven healthy adults, consisting of two emmetropes (<±0.75 D) and five myopes (−0.75 to −6.25 D), with best corrected visual acuity of 6/5 or better were recruited for the measurement of central and peripheral eye lengths. The study was approved by the Human Research Ethics Committee of the Queensland University of Technology, and informed consent was obtained from all participants before taking measurements.
A post hoc power analysis showed that the seven participants provided a power of 0.81 against type 2 errors. This calculation was based on the assumption that a 0.09-mm difference in length between the IOLMaster and Lenstar instruments would be important. The value of 0.09 mm was chosen because it is approximately equivalent to a 0.25-D difference in refractive error, as determined from the Bennett-Rabbetts schematic eye.
After dilating the pupil with one drop each of 1.0% tropicamide and 2.5% phenylephrine, central and peripheral eye lengths were recorded using two instruments (IOLMaster V5 and Lenstar LS 900). The eye lengths were determined in 5° steps out to 30° along the nasal visual field, out to 35° along the temporal visual field, and out to ±30° along the vertical visual field. Measurements were not possible any further because the edge of the pupil obstructed the passage of the beam. All the measurements were performed by the same investigator and collected from right eyes except for one participant for whom the left eye was used.
Peripheral eye length measurements were obtained by using an external attachment, similar to that of Mallen and Kashyap,19 containing a goniometer, a 50/50 beam splitter, a Maltese cross fixation target at optical infinity, and a light emitting diode source (Fig. 1). The Maltese cross was aligned with the instrument fixation axis for on-axis measurements. The goniometer was moved over the base rail (movement along x, y, and z axes), until the Maltese cross target could be seen at all positions of goniometer rotation, thus ensuring that the effective position about which the target rotated corresponded with the center-of-rotation of the eye. For the horizontal field, the attachment was fixed to the top of the chin-rest frame of the IOLMaster or Lenstar instrument using a pair of right-angle retort clamps, whereas for the vertical field, the attachment was fixed to the side frame of the chin rest. For right eyes, rotation to the right side corresponded to the nasal visual field (temporal retina), and rotation to the left side corresponded to the temporal visual field (nasal retina). Similarly, upward rotation corresponded to the inferior visual field (superior retina), and downward rotation corresponded to the superior visual field (inferior retina).
For measurements along the peripheral visual field, participants rotated their eye to fixate the target without head movement, thus requiring realignment of the instrument. Participants were asked to blink completely before each measurement. The instrument’s alignment mire on cornea was maintained in clear focus, and it moved toward the pupil margin as the field angle increased. A minimum of four consecutive measurements were recorded at each position, and means were calculated.
For intersessional repeatability determination, measurements were obtained at two different sessions. Measurements were made at the same time but on different days for four participants. For the other three participants, the measurements were made in 1 day with a gap of a few hours between measurements. The order of instruments in a session was assigned randomly. Measurements were recorded along the horizontal visual field (temporal to nasal), followed by the measurements along the vertical visual field (superior to inferior).
For determining intrasessional repeatability for each instrument, each participant/session/visual field position was represented by the SD of the first four measurements. The intrasessional repeatability was given by the mean of these SDs across seven participants, two sessions, and visual field positions (14 for the horizontal visual field and 13 for the vertical visual field).
For determining intersessional repeatability for each instrument, each participant/visual field position was represented by the difference between the mean values of the two sessions. The intersessional repeatability was given by the SD of these differences across participants and visual field positions.
For determining the agreement between the two instruments, each participant/visual field position was represented by the mean difference between instruments across the two sessions. The agreement was given by the mean and the SD of the differences across participants and visual field positions.
One weakness about the aforementioned approach is that multiple positions from each participant are treated as independent observations. Bland and Altman20 have a method for investigating agreement between methods with multiple observations for individual participants. This can be applied here, treating different peripheral positions as if they are different observations for which the underlying quantity is varying. As compared with considering each position for a participant as independent, the SDs (or 95% prediction limits) increase by 2% or less for intrasessional, intersessional, and interinstrument analyses, which is small and can be ignored.
In addition to the above analyses, repeated-measures analyses of variance (ANOVA) were conducted on eye lengths, with participants as the repeated measures. A first ANOVA was conducted for intrasessional SDs, with session (sessions 1 and 2), instrument (IOLMaster and Lenstar), and visual field position as within-participant factors. A second ANOVA was conducted for absolute intersessional differences, with instrument and visual field position as within-participant factors. A third ANOVA was conducted for differences between instruments, with session and visual field position as within-participant factors. These three ANOVAs were conducted for the horizontal and vertical visual fields separately and for combined data; because results were similar for the three approaches, only results for the combined data are mentioned.
Fig. 2 shows eye length measurements as a function of visual field position for both the IOLMaster and Lenstar for a myopic participant (−0.75-D correction). This shows intrasessional variability (error bars represent SDs) and intersessional variability (difference between the two plots on each part of the figure).
For the IOLMaster, the repeatabilities were 0.04 ± 0.04 mm along the horizontal and vertical visual fields. Corresponding results for the Lenstar were 0.02 ± 0.02 mm along both the horizontal and vertical visual fields. The difference between the two instruments was significant in the corresponding ANOVA (F1,6 = 19.1, P = 0.005). The IOLMaster and the Lenstar had intrasessional SDs of 0.02 and 0.01 mm, respectively, at the center of the visual field. The SDs were greater away from the center, with maximum values for the IOLMaster of 0.07 mm (at 20°, 25° temporal and 10°, 30° superior field positions) and for the Lenstar of 0.06 mm (at 15° temporal field position corresponding to the optic disc). The increased intrasessional variation away from the center was supported by the significant effect of visual field position (F26,156 = 4.2, P < 0.001).
Fig. 3 shows Bland-Altman plots of intersessional repeatability. Different symbols are given for different participants. The intersessional repeatabilities for the IOLMaster for the horizontal and vertical visual fields were ±0.11 and ±0.08 mm, respectively; corresponding repeatabilities for the Lenstar were ±0.05 and ±0.04 mm. The difference between the two instruments was marginally significant in the corresponding ANOVA (F1,6 = 5.8, P = 0.05).
The intersessional repeatability increased from the center toward the peripheral visual field for both the IOLMaster and the Lenstar. Both instruments had repeatabilities of 0.03 mm at the center of the field, increasing for the IOLMaster to approximately 0.20 mm (15° temporal, 30° temporal, and 30° superior field positions) and increasing for the Lenstar to approximately 0.08 mm (15° temporal and 20°–30° nasal field positions). The increased intersessional variation away from the center was supported by the significant effect of visual field position (F26,156 = 2.4, P < 0.001).
Agreement Between IOLMaster and Lenstar
Fig. 4 shows Bland-Altman plots of agreement between the two instruments. As for Fig. 3, different symbols are given for different participants. The agreements between the instruments were 0.01 mm and 0.02 mm for the horizontal and vertical visual fields, respectively, with SDs of ±0.07 mm for both visual fields. These results indicate that the instruments are in good agreement. The differences between the two instruments ranged from 0.02 mm at the center of the visual field to 0.04 mm along the horizontal field (5°, 10°, 25° nasal and 30° temporal) and 0.06 mm along the vertical field (30° inferior), but there was no statistically significant difference between the instruments at any field position, and ANOVA did not show an effect of field position (F26,156 = 1.0, P = 0.47). For one participant (square boxes on the left side of plots), the IOLMaster had greater measures than the Lenstar for most positions, whereas for another participant (filled triangles on the right side of plots), the reverse was the case.
The differences between the instruments changed significantly with axial length, with the Lenstar giving larger measurements of axial length than the IOLMaster for longer eyes; the slopes in Fig. 4 are approximately −0.016 (P < 0.001).
Other Peripheral Measurements
As well as the eye length measurements, the internal eye distances were noted with the Lenstar. It allowed corneal and retinal thicknesses across the visual field but did not record anterior chamber depth and lens thickness measurements beyond approximately ±10° and ±5° along both the horizontal and visual fields, respectively, except for one participant for whom it gave measurements out to ±25° and ±10°. Fig. 5 shows corneal thicknesses across the vertical field for the two sessions. The expected increased corneal thickness toward the periphery occurred, with good repeatability along both the horizontal and vertical fields. Fig. 6 shows retinal thicknesses for the two sessions of the same participant as in Fig. 5; intrasessional repeatability appears to be poor for some visual field positions. The Lenstar software depends on the investigator to determine the retinal thickness manually (internal limiting membrane and retinal pigment epithelium) from reflectance signal plots.
For measuring peripheral eye lengths along the horizontal and vertical visual fields, we assessed the intrasessional and intersessional repeatability of IOLMaster and Lenstar partial coherence interferometry instruments and the agreement between the instruments. Intrasessional repeatability was 0.04 mm for the IOLMaster and 0.02 mm for the Lenstar. Intersessional repeatabilities were ±0.11 and ±0.08 mm for the IOLMaster for the horizontal and vertical visual fields, respectively; corresponding repeatabilities for the Lenstar were ±0.05 and ±0.04 mm. Repeatabilities worsened away from fixation. Agreements between the instruments were good at 0.01 ± 0.07 mm and 0.02 ± 0.07 mm for the horizontal and visual fields, respectively, with no significant influence of visual field position, but the lengths with the Lenstar became greater than those with the IOLMaster as axial length increased (rate of approximately 0.016 mm/mm).
The intrasessional and intersessional repeatabilities of both instruments were excellent. The latter is particularly of note because the external device had to be reattached before each session with each instrument, and the eye lengths across the visual field ranged between 0.3 and 2.1 mm for our participants. The smaller (better) intrasessional repeatability with Lenstar compared with IOLMaster may be partly caused by a difference in the recording method because each Lenstar measurement is the average of 16 scans. The Lenstar also had the better intersessional repeatability.
The on-axis intrasessional repeatability of 0.02 mm for the IOLMaster is better than that reported by Santodomingo-Rubido et al,8 whereas the 0.01 mm for the Lenstar is at the lower end of 0.01 to 0.04 mm repeatabilities in other studies.7,9–11,21,22 In the only previous investigation of off-axis repeatability with the Lenstar, for five positions along the horizontal field, Schulle and Berntsen21 reported repeatabilities of 0.03 to 0.05 mm, similar to those obtained here. The on-axis intersessional repeatabilities of 0.03 mm for both instruments are within the 0.02- to 0.04-mm range reported for the IOLMaster5,12,13 and poorer than the 0.01 mm reported for the Lenstar in two studies7,10 but similar to that reported by Schulle and Berntsen.21 The latter reported repeatabilities of 0.025 and 0.06 mm at two off-axis positions, similar to those obtained here.
Several studies have already reported the excellent agreement between the instruments for on-axis length measurements. Mean differences were reported as 0.00 to 0.04 mm, with some studies, but not others, finding significant differences.10,11,14,15,23–26
We found that the measurements became greater for the Lenstar than for the IOLMaster with an increase in eye length (0.016 mm/mm). We analyzed the results of three studies of on-axis length10,11,14,25 (personal communications) and confirmed this trend only for the study of Buckhurst et al10 for which the slope was 0.010 mm/mm.
The good agreement between IOLMaster and Lenstar for central and peripheral eye length measurements along both the horizontal and vertical visual fields indicates that the instruments should give similar results, but the Lenstar is expected to give slightly greater ranges of lengths across the visual field, for example, an approximately 0.034-mm increase in the participant with the largest range of 2.1 mm.
With recent myopia research indicating the important role of peripheral retina in the development and progression of myopia (see review by Verkicharla et al4), eye length measurements obtained with partial coherence interferometry have been used for the determination of retinal shape by directing obliquely beams into the eye at oblique orientations.16–19,27 Schmid16,17,27 used a customized partial coherence interferometry instrument with children and estimated retinal steepness by subtracting on-axis measurements from peripheral measurements. Mallen and Kashyap19 used a modified IOLMaster with adult participants. Using simple equations involving corneal shapes as conicoids and assuming undeviated ray paths within the eye, they represented retinal shapes as Cartesian coordinates. Atchison and Charman28 performed a theoretical investigation of the partial coherence interferometry technique and indicated that it can give reasonably accurate results for retinal shape out to ±30° visual field if the incident beam is directed toward the center of curvature of the anterior cornea (corneal-direction method), the relevant method with the IOLMaster and Lenstar instruments. They concluded that estimates that are uncorrected for optical distortion (bending of light within the eye and lack of knowledge about how instruments convert optical path lengths to real lengths) must be interpreted with caution and that estimates can be improved with more sophisticated ray tracing because more information about the eye’s biometry is available.
The ease of peripheral measurements was similar for the two instruments. The average time to obtain a measurement set, including the adjustments of the external attachment, was 40 min for the IOLMaster and 50 min for the Lenstar. This difference is partly because of the different technology used by the instruments. The Lenstar uses a proprietary “intelligent detection system” that enables the instrument to take measurements only when the eye is stable—if the patient blinks or loses fixation, the instrument waits until the patient’s fixation returns. The IOLMaster does not consider eye movement and displays the reading immediately, with the investigator accepting or rejecting readings based on the signal-to-noise ratio.
Although this study has not established the validity of using peripheral eye length measurements for determining retinal shape, it does show that such measurements with two commercial partial coherence interferometers are similar and repeatable.
Pavan K. Verkicharla
School of Optometry and Institute of Health
and Biomedical Innovation
Queensland University of Technology
60 Musk Ave
Kelvin Grove Queensland 4059
This work was supported by Australian Research Council Discovery grant DP110102018, by Australian Research Council Linkage grant LP100100575, and by a Queensland University of Technology–Institute of Health and Biomedical Innovation Visiting Research Fellowship to Edward Mallen. We thank JJ Nel and Yossef Strasberg from Device Technologies for the loan of a Lenstar instrument. We also thank James Wolffsohn, Hosein Nowroozzadeh, and David Goldblum for analyzing or providing data for the analysis of instrument comparisons.
Received May 6, 2012; accepted November 28, 2012.
1. Pan CW, Ramamurthy D, Saw SM. Worldwide prevalence and risk factors for myopia. Ophthalmic Physiol Opt 2012; 32: 3–16.
2. Charman WN, Radhakrishnan H. Peripheral refraction and the development of refractive error: a review. Ophthalmic Physiol Opt 2010; 30: 321–38.
3. Smith EL 3rd. Prentice Award Lecture 2010: a case for peripheral optical treatment strategies for myopia. Optom Vis Sci 2011; 88: 1029–44.
4. Verkicharla PK, Mathur A, Mallen EA, Pope JM, Atchison DA. Eye shape and retinal shape, and their relation to peripheral refraction. Ophthalmic Physiol Opt 2012; 32: 184–99.
5. Kimura S, Hasebe S, Miyata M, Hamasaki I, Ohtsuki H. Axial length measurement using partial coherence interferometry in myopic children: repeatability of the measurement and comparison with refractive components. Jpn J Ophthalmol 2007; 51: 105–10.
6. Lam AK, Chan R, Pang PC. The repeatability and accuracy of axial length and anterior chamber depth measurements from the IOLMaster. Ophthalmic Physiol Opt 2001; 21: 477–83.
7. Shammas HJ, Hoffer KJ. Repeatability and reproducibility of biometry and keratometry measurements using a noncontact optical low-coherence reflectometer and keratometer. Am J Ophthalmol 2012; 153: 55–61.
8. Santodomingo-Rubido J, Mallen EA, Gilmartin B, Wolffsohn JS. A new non-contact optical device for ocular biometry. Br J Ophthalmol 2002; 86: 458–62.
9. Cruysberg LP, Doors M, Verbakel F, Berendschot TT, De Brabander J, Nuijts RM. Evaluation of the Lenstar LS 900 non-contact biometer. Br J Ophthalmol 2010; 94: 106–10.
10. Buckhurst PJ, Wolffsohn JS, Shah S, Naroo SA, Davies LN, Berrow EJ. A new optical low coherence reflectometry device for ocular biometry in cataract patients. Br J Ophthalmol 2009; 93: 949–53.
11. Rohrer K, Frueh BE, Walti R, Clemetson IA, Tappeiner C, Goldblum D. Comparison and evaluation of ocular biometry using a new noncontact optical low-coherence reflectometer. Ophthalmology 2009; 116: 2087–92.
12. Sheng H, Bottjer CA, Bullimore MA. Ocular component measurement using the Zeiss IOLMaster. Optom Vis Sci 2004; 81: 27–34.
13. McDaniel JG, Mutti DO. Repeatability and comparability of IOL master and ultrasound measures of ocular axial length in young adults. Invest Ophthalmol Vis Sci 2002; 43: E-Abstract 356.
14. Salouti R, Nowroozzadeh MH, Zamani M, Ghoreyshi M. Comparison of the ultrasonographic method with 2 partial coherence interferometry methods for intraocular lens power calculation. Optometry 2011; 82: 140–7.
15. Zhao J, Chen Z, Zhou Z, Ding L, Zhou X. Evaluation of the repeatability of the Lenstar and comparison with two other non-contact biometric devices in myopes. Clin Exp Optom 2012; 96: 92–9.
16. Schmid GF. Axial and peripheral eye length measured with optical low coherence reflectometry. J Biomed Opt 2003; 8: 655–62.
17. Schmid GF. Variability of retinal steepness at the posterior pole in children 7–15 years of age. Curr Eye Res 2003; 27: 61–8.
18. Gray LS, Strang NC, Seidel D, Brawley F, Howard K. Corneal asphericity, ocular aberrations and retinal shape in hyperopia, myopia and emmetropia. Invest Ophthalmol Vis Sci 2005; 46: E-Abstract 5600.
19. Mallen EA, Kashyap P. Technical note: measurement of retinal contour and supine axial length using the Zeiss IOLMaster. Ophthalmic Physiol Opt 2007; 27: 404–11.
20. Bland JM, Altman DG. Agreement between methods of measurement with multiple observations per individual. J Biopharm Stat 2007; 17: 571–82.
21. Schulle KL, Berntsen DA. Repeatability of on- and off-axis eye length measurements using the Lenstar. Optom Vis Sci 2013; 90: 16–22.
22. Bjelos Roncevic M, Busic M, Cima I, Kuzmanovic Elabjer B, Bosnar D, Miletic D. Intraobserver and interobserver repeatability of ocular components measurement in cataract eyes using a new optical low coherence reflectometer. Graefes Arch Clin Exp Ophthalmol 2011; 249: 83–7.
23. Rabsilber TM, Jepsen C, Auffarth GU, Holzer MP. Intraocular lens power calculation: clinical comparison of 2 optical biometry devices. J Cataract Refract Surg 2010; 36: 230–4.
24. Hoffer KJ, Shammas HJ, Savini G. Comparison of 2 laser instruments for measuring axial length. J Cataract Refract Surg 2010; 36: 644–8.
25. Jasvinder S, Khang TF, Sarinder KK, Loo VP, Subrayan V. Agreement analysis of LENSTAR with other techniques of biometry. Eye (Lond) 2011; 25: 717–24.
26. Holzer MP, Mamusa M, Auffarth GU. Accuracy of a new partial coherence interferometry analyser for biometric measurements. Br J Ophthalmol 2009; 93: 807–10.
27. Schmid GF. Association between retinal steepness and central myopic shift in children. Optom Vis Sci 2011; 88: 684–90.
28. Atchison DA, Charman WN. Can partial coherence interferometry be used to determine retinal shape? Optom Vis Sci 2011; 88: 601–7.