- Email: [email protected]
Biometry of cataractous eyes using partial coherence interferometry
articles
Biometry of cataractous eyes using partial coherence interferometry Clinical feasibility study of a commercial prototype I Barbara Kiss, MD, Oliver Findl, MD, Rupert Menapace, MD, Matthias Wirtitsch, MD, Wolfgang Drexler, MD, Christoph K. Hitzenberger, MD, Adolf F. Fercher, MD ABSTRACT Purpose: To evaluate the clinical feasibility of the prototype version of a commercial partial coherence interferometry instrument (axial length measurement, ALM, Carl Zeiss Jena) for noninvasive, high-precision biometry in cataractous eyes. Setting: Department of Ophthalmology, Vienna General Hospital, and Institute of Medical Physics, University of Vienna, Austria. Methods: The preoperative axial length in 49 eyes of 37 cataract patients was measured with the commercial (ALM) and laboratory (PCI) prototypes of the partial coherence interferometry instrument, as well as with immersion ultrasound (IUS). Results: Axial length measurements with the ALM and PCI did not differ significantly (P ⫽ .23). Both prototypes assessed longer axial lengths than the IUS technique (P ⬍ .0001; median 203 m; range ⫺476 to ⫹635 m). The precision of the axial length measurement was 18 m, 28 m, and 54 m with the PCI, ALM, and IUS, respectively. Conclusions: Partial coherence tomography is a high-precision, high-resolution, noncontact biometric technique.The commercial PCI prototype is practical in clinical use, with improved comfort for patients, no need for anesthesia, and a reduced risk of infection. However, the difference between the PCI and IUS in axial length measurement must be considered when using the constants supplied by intraocular lens (IOL) manufacturers for IOL power calculations. J Cataract Refract Surg 2002; 28:224 –229 © 2002 ASCRS and ESCRS
T
he precise measurement of intraocular distances is an important task in modern ophthalmology. Recent studies demonstrate that obtaining accurate preopAccepted for September 19, 2001. Reprint requests to Wolfgang Drexler, MD, Institut fu¨r Medizinische Physik, Universita¨t Wien, Wa¨hringer Strasse 13, A-1090 Vienna, Austria. E-mail: [email protected]. © 2002 ASCRS and ESCRS Published by Elsevier Science Inc.
erative biometric data is the most critical step in accurate calculation of the intraocular lens (IOL) refractive power.1,2 Currently, biometric measurements are performed with the ultrasound A-scan echo-impulse technique (US technique), enabling the measurement of axial length with a longitudinal resolution of typically 150 to 200 m and an accuracy of approximately 100 to 150 m.3⫺6 Studies based on US biometry demonstrate 0886-3350/02/$–see front matter PII S0886-3350(01)01272-X
BIOMETRY OF CATARACTOUS EYES USING PC1
that 54% of the errors in the predicted refraction after IOL implantation are attributed to axial length measurement errors.1 An axial length measurement error of 100 m would result in a corresponding postoperative refractive error of 0.28 diopter.1,7 Applanation US is the most commonly used technique for ocular biometry.8 Since this technique requires direct contact between the transducer and the eye, the cornea is indented and the anterior chamber depth as well as the axial length are shorter than with the more accurate but also more uncomfortable water-immersion ultrasound (IUS) technique. With the latter, the transducer has no direct contact with the cornea. Significant differences of 0.14 to 0.36 mm have been reported between applanation US and IUS axial length measurements.9,10 In both US techniques, a possible mismatch between the measurement axis and the visual axis of the eye may influence the results. A potential alternative to conventional US is an optical biometry technique that was developed in the past decade.11⫺13 It is a special version of a noninvasive optical biomedical imaging technology called optical coherence tomography (OCT), which is based on an optical measurement technique known as low or partial coherence interferometry. This technique is analogous to conventional US pulse–echo imaging (US A- and Bmode) except that OCT uses infrared laser light instead of sound.14⫺17 A special dual-beam version of this interferometric technique—partial coherence interferometry—that eliminates the influence of longitudinal eye motion during measurement by using the cornea as a reference surface was used to perform axial length measurements in normal18 and cataractous eyes.19 Depending on the measured intraocular distance, precision values from 0.3 to 10 m have been reported.20 The applicability of partial coherence interferometry for measuring the axial length in cataractous eyes was shown in an extensive study of 196 eyes of 100 patients.19 Recently, this technique has been reported to perform accurate biometry in cataractous eyes, with the potential to improve the refractive outcome of cataract surgery by about 30% using the SRK II power formula.21 This optical biometry technique has also been used in a clinical setting to investigate its clinical feasibility for IOL measurements in pseudophakic eyes.22⫺24 All these studies were performed with a laboratory prototype of partial coherence interferometry. Recently, Zeiss
developed a commercial prototype using the optical biometry technique (axial length measurement, ALM). In addition to the high precision of axial length measurements with this modality, the unit is reported to be easy to use clinically. In this study, the commercial prototype (ALM) for measuring axial length in cataract patients was investigated by comparing the findings with those obtained with the laboratory prototype (PCI) as well as those obtained with IUS biometry. The handling of this first clinical prototype and its performance in a clinical routine was also assessed.
Patients and Methods Preoperative Biometry The principle of the dual-beam version of partial coherence interferometry has been described.11⫺13,16,18⫺21 The differences between the ALM and the PCI prototypes are as follows: (1) The axial resolution of the ALM is approximately 100 m and of the PCI, 10 m. (2) The ALM emits laser light centered at 780 nm; with the PCI, it is centered at 855 nm. A single A-scan to measure axial length takes about 0.5 seconds with both prototypes. The differences in resolution and wavelength are due to the different light sources used in the 2 instruments. Both instruments use special optics (diffractive optics element) to enhance sensitivity, especially important for measuring cataractous eyes.25 Since the ALM and PCI yield optical distances, these have to be divided by the group refractive indices of the respective ocular media for the center wavelength to obtain geometric distances.21,26 For measurements performed with the IUS technique (Ocuscan, Alcon), sound velocities of 1532 m/s were taken for the aqueous and vitreous humor and 1641 m/s for the crystalline lens. Patients and Study Design The study protocol was approved by the Ethics Committee of the Vienna University School of Medicine. The nature of the study was explained, and all patients gave written consent to participate. Preoperatively, 2 examiners assessed 10 measurements of axial length with each technique in 49 eyes of 37 cataract patients. The mean age of the patients was 75 years (range 36 to 91 years). Axial length was assessed with the
J CATARACT REFRACT SURG—VOL 28, FEBRUARY 2002
225
BIOMETRY OF CATARACTOUS EYES USING PC1
ALM, PCI, and IUS on the same day. During in vivo measurements of the human eye, laser safety regulations were considered. For the ALM, the light source used in this study had a center wavelength of 1 ⬇ 780 nm with power of about 350 W at the cornea. For the PCI, the light source had a center wavelength of 1 ⬇ 855 nm with power of about 220 W at the cornea or an intensity of approximately 572 W/cm2 (averaged over a 7 mm aperture). This can be applied for about 28 minutes.27 The time for a single measurement of axial length is 0.5 seconds. By performing 8 to 10 longitudinal scans for statistical purposes, the maximum time of continuous illumination is about 4 seconds, well below the safety limits. Data Analysis Data are presented as the median and range, indicating minimum and maximum. The precision of the techniques is defined as the standard deviation of multiple consecutive measurements of axial length under investigation. For the comparison of means, paired t tests were applied to data that could be described by the normal distribution. Correlations were investigated using linear regression; P ⬍ .05 was considered the level of significance.
Figure 1. (Kiss) Correlation of axial length measurements between the PCI and ALM (N ⫽ 49).
Results Axial length measurements obtained with all 3 biometry techniques are shown in Table 1. Axial length measured with the ALM and PCI did not differ significantly (P ⫽ .23; median difference ⫺3 m; range ⫺150 to ⫹328 m; mean difference ⫺9 m) (Figure 1), with 90% of measurements differing less than 100 m. There was a significant difference in axial length between the PCI and IUS (P ⬍ .001) and the ALM and IUS Table 1. Axial length assessed with the ALM, PCI, and IUS. Data are presented as median and range (N ⫽ 49). Measurement
ALM
PCI
IUS
Median (mm)
23.500
23.509
23.195
Minimum (mm)
20.253
20.190
19.882
Maximum (mm)
26.450
26.445
27.371
1.433
1.451
1.554
Standard deviation (mm)
ALM ⫽ commercial prototype of partial coherence interferometry; PCI ⫽ laboratory prototype of partial coherence interferometry; IUS ⫽ immersion ultrasound
226
Figure 2. (Kiss) Correlation of axial length measurements between IUS and the ALM (N ⫽ 49).
(P ⬍ .001). Axial lengths assessed with ALM correlated well with those assessed with IUS (r ⫽ 0.991, P ⬍ .001) (Figure 2). With the PCI, the axial length measured longer by ⫹186 m (mean ⫹171 m; range ⫺926 to ⫹581 m) than with IUS; with ALM, it measured longer by ⫹202 m (mean ⫹180 m; range ⫺921 to
J CATARACT REFRACT SURG—VOL 28, FEBRUARY 2002
BIOMETRY OF CATARACTOUS EYES USING PC1
ure 6). The difference between PCI and ALM in axial length measurements did not correlate with axial length. No correlation between precision and axial length was observed among the 3 techniques, indicating that short or long eyes can be measured with similar precision. Six of 55 eyes (11%) could not be measured with ALM. Five patients had fixation problems in 1 eye because of a mature cataract or macular degeneration. One eye was extremely hyperopic; thus, the short axial length could not be detected with the ALM prototype.
Discussion
Figure 3. (Kiss) Differences between axial lengths measured with the ALM, PCI, and IUS presented as box plots. (Box plots are characterized by the minimum and the maximum; the box depicts the 25th and 75th percentiles, the small square represents the median, the star represents the mean, and circles are the outliers [N ⫽ 49]).
⫹635 m). Eighty-six percent of the differences in axial length measurements were within 400 m (Figure 3). To confirm the good correlation between the ALM and PCI and the ALM and IUS, the measuring agreement was assessed by plotting the difference between the methods against their mean (Figure 4). The precision of axial length measurements with the PCI (median 18 m; range 5 to 116 m; mean 26 m) (Figure 5) was significantly better than with the ALM (median 26 m; range 10 to 156 m; mean 36 m) (P ⬍ .04). The precision of the ALM measurements was significantly better than with IUS (median 54 m; range 4 to 211 m; mean 64 m) (P ⬍ .001) (Figure 5). There was no correlation in axial lengths for each individual within the measurements for the different techniques (Fig-
Previous studies demonstrate the applicability of optical biometry based on partial coherence interferometry for measuring the axial length in cataractous eyes.19,21 Accurate biometry with this technique was recently demonstrated to improve the refractive outcome of cataract surgery by about 30% using the SRK II power formula.21 Further improvement can be obtained by using third-generation IOL power formulas.28 In this study, we compared the laboratory prototype of dual-beam partial coherence interferometry with the commercial prototype of this optical biometry technique. The precision of the PCI was 18 m, which is worse than we have demonstrated,21 probably because of the use of a specific optic to enhance sensitivity. Hence, different peaks originating from the retinal pigment epithelium/Bruch’s membrane/choriocapillaris interface cannot always be clearly separated. We showed that both optical devices measure similar axial lengths. The correlation between both techniques was excellent (r ⫽ 0.999). The precision was
Figure 4. (Kiss) Agreement of axial length measurements between the ALM and PCI (left, N ⫽ 49) and the ALM and IUS (right, N ⫽ 49). J CATARACT REFRACT SURG—VOL 28, FEBRUARY 2002
227
BIOMETRY OF CATARACTOUS EYES USING PC1
Figure 5. (Kiss) Precision of axial length measurements of the ALM, PCI, and IUS presented as box plots. The precision is defined as the mean of the standard deviations of 10 consecutive measurements. (Box plots are described in Figure 3.)
Figure 6. (Kiss) Comparison of differences in axial lengths between the PCI and ALM (solid triangle) and the PCI and IUS (open circle) in all eyes (N ⫽ 49).
significantly better with the laboratory prototype, perhaps because of the different bandwidths of the super luminescent diodes, resulting in resolutions of 12 m and 100 m for the laboratory and commercial prototype, respectively. Compared with IUS, the optical technique measures axial length longer by approximately 200 m. The median difference between PCI and IUS was 186 m and between ALM and IUS, 203 m. The systematic difference between optical and acoustic techniques may be due to the different reflection sites in the retina.19,21 Ultrasound measures up to the inner limiting membrane, whereas with PCI the maximum interference pat228
tern from the retina is detected at the interface with the retinal pigment epithelium. Furthermore, the mismatch of the beam axis and the visual axis during US measurements may cause a deviation in axial length measurements between biometry methods. This agrees with the results of our previous study comparing partial coherence interferometry with applanation US and IUS techniques.21 Nevertheless, we observed that the regression line of the correlation of IUS with the ALM did not have a gradient of 1 so the deviation is more obvious with shorter and longer axial eye lengths. This means that long eyes are measured to be shorter with the optical techniques than with the IUS technique. Immersion US uses a different velocity for the lens and the vitreous and aqueous humor. In this study, we used the mean group refractive index with the optical techniques.19 In long eyes, this assumption may cause the deviation. Correction factors should then be used.19 The precision of the PCI was also significantly better than that of the US technique. However, the PCI also caused some outliers. The reason could be fixation problems of the patients, so measurements were not repeated at the same point on the retina. With the ALM, we were not able to measure in a few eyes (6 of 55 [11%]). Five patients had fixation problems because of mature cataract or macular degeneration. In conclusion, the ALM and PCI measured similar axial lengths. In contrast to US, optical techniques have a higher measurement precision. Measurements can be carried out with more comfort for the patient and without corneal contact, which minimizes the risk of infection. The assessment of axial length with the ALM is efficient, easy to use, quick to learn, and adequate for the clinical routine.
References 1. Olsen T. Sources of error in intraocular lens power calculation. J Cataract Refract Surg 1992; 18:125–129 2. Boerrigter RMM, Thijssen JM, Verbeek AM. Intraocular lens power calculations: the optimal approach. Ophthalmologica 191; 1985:89 –94 3. Binkhorst RD. The accuracy of ultrasonic measurement of the axial length of the eye. Ophthalmic Surg 1981; 12:363–365 4. Schachar RA, Levy NS, Bonney RC. Accuracy of intraocular lens powers calculated from a-scan biometry with the Echo-Oculometer. Ophthalmic Surg 1980; 11:856 – 858 5. Olsen T. The accuracy of ultrasonic determination of
J CATARACT REFRACT SURG—VOL 28, FEBRUARY 2002
BIOMETRY OF CATARACTOUS EYES USING PC1
6.
7.
8.
9.
10.
11. 12.
13.
14. 15.
16. 17.
18.
19.
20.
axial length in pseudophakic eyes. Acta Ophthalmol 1989; 67:141–144 Bamber JC, Trstam M. Diagnostic ultrasound. In: Webb E, ed, The Physics of Medical Imaging. Bristol, Philadelphia, PA, Adam Hilger, 1988; 319 –388 Olsen T. Theoretical approach to intraocular lens calculation using Gaussian optics. J Cataract Refract Surg 1987; 13:141–145 Leaming DV. Practice styles and preferences of ASCRS members—1998 survey. J Cataract Refract Surg 1999; 25:851– 859 Olsen T, Nielsen PJ. Immersion versus contact technique in the measurement of axial length by ultrasound. Acta Ophthalmol 1989; 67:101–102 Artaria LG. Messung der Bulbusla¨nge mit verschiedenen Ultraschall-Gera¨ten. Klin Monatsbl Augenheilkd 1986; 188:492– 494 Fercher AF, Roth E. Ophthalmic laser interferometry. Proc SPIE 658, 1986; 48 –51 Fercher AF, Mengedoht K, Werner W. Eye length measurement by interferometer with partially coherent light. Opt Lett 1988; 13:186 –188 Fercher AF, Hitzenberger CK, Juchem M. Measurement of intraocular optical distances using partially coherent laser light. J Mod Opt 1991; 38:1327–1333 Huang D, Swanson EA, Lin CP, et al. Optical coherence tomography. Science 1991; 254:1178 –1181 Huang D, Wang J, Lin CP, et al. Micron-resolution ranging of cornea anterior chamber by optical reflectometry. Lasers Surg Med 1991; 11:419 – 425 Fercher AF. Optical coherence tomography. J Biomed Opt 1996; 1:157–173 Fercher AF, Hitzenberger CK, Drexler W, et al. In vivo optical coherence tomography (letter). Am J Ophthalmol 1993; 116:113–114 Hitzenberger CK. Optical measurement of the axial eye length by laser Doppler interferometry. Invest Ophthalmol Vis Sci 1991; 32:616 – 624 Hitzenberger CK, Drexler W, Dolezal C, et al. Measurement of the axial length of cataract eyes by laser Doppler interferometry. Invest Ophthalmol Vis Sci 1993; 34:1886 –1893 Drexler W, Baumgartner A, Findl O, et al. Submicrometer precision biometry of the anterior segment of the hu-
21.
22.
23.
24.
25.
26.
27.
28.
man eye. Invest Ophthalmol Vis Sci 1997; 38:1304 – 1313 Drexler W, Findl O, Menapace R, et al. Partial coherence interferometry: a novel approach to biometry in cataract surgery. Am J Ophthalmol 1998; 126:524 –534 Findl O, Drexler W, Menapace R, et al. High precision biometry of pseudophakic eyes using partial coherence interferometry. J Cataract Refract Surg 1998; 24:1087– 1093 Findl O, Drexler W, Menapace R, et al. Accurate determination of effective lens position and lens⫺capsule distance with 4 intraocular lenses. J Cataract Refract Surg 1998; 24:1094 –1098 Findl O, Drexler W, Menapace R, et al. Changes in intraocular lens position after neodymium:YAG capsulotomy. J Cataract Refract Surg 1999; 25:659 – 662 Mo¨ ller B, Rudolph G, Klopffleisch A, et al. Application of diffractive optics for axial eye length measurements using partial coherence interferometry. In: Birngruber R, Fercher AF, Sourdille P, eds, Lasers in Ophthalmology IV. Proc SPIE 2930, 1996; 175–182 Drexler W, Hitzenberger CK, Baumgartner A, et al. Investigation of dispersion effects in ocular media by multiple wavelength partial coherence interferometry. Exp Eye Res 1998; 66:25–33 American National Standards Institute. Safe Use of Lasers, ANSI Z 136.1. New York, NY, American National Standards Institute, 1986 Findl O, Drexler W, Menapace R, et al. Improved prediction of intraocular lens power using partial coherence interferometry. J Cataract Refract Surg 2001; 27:861⫺867
From Universita¨ tsklinik fu¨ r Augenheilkunde, Allgemeines Krankenhaus Wien (Kiss, Findl, Menapace, Wirtitsch), and Institut fu¨ r Medizinische Physik, Universita¨ t Wien (Drexler, Hitzenberger, Fercher), Vienna, Austria. Presented in part at the XVIIth Congress of the European Society of Cataract & Refractive Surgeons, Vienna, Austria, September 1999. None of the authors has a financial interest in any product mentioned. Dr. Donnerhacke and colleagues, Zeiss Jena, provided the ALM prototype; H. Sattmann, Ing, and L. Schachinger, Institute of Medical Physics, University of Vienna, constructed the electronics and software of the instrument and provided technical support, respectively.
J CATARACT REFRACT SURG—VOL 28, FEBRUARY 2002
229
Biometry of cataractous eyes using partial coherence interferometry Clinical feasibility study of a commercial prototype I Barbara Kiss, MD, Oliver Findl, MD, Rupert Menapace, MD, Matthias Wirtitsch, MD, Wolfgang Drexler, MD, Christoph K. Hitzenberger, MD, Adolf F. Fercher, MD ABSTRACT Purpose: To evaluate the clinical feasibility of the prototype version of a commercial partial coherence interferometry instrument (axial length measurement, ALM, Carl Zeiss Jena) for noninvasive, high-precision biometry in cataractous eyes. Setting: Department of Ophthalmology, Vienna General Hospital, and Institute of Medical Physics, University of Vienna, Austria. Methods: The preoperative axial length in 49 eyes of 37 cataract patients was measured with the commercial (ALM) and laboratory (PCI) prototypes of the partial coherence interferometry instrument, as well as with immersion ultrasound (IUS). Results: Axial length measurements with the ALM and PCI did not differ significantly (P ⫽ .23). Both prototypes assessed longer axial lengths than the IUS technique (P ⬍ .0001; median 203 m; range ⫺476 to ⫹635 m). The precision of the axial length measurement was 18 m, 28 m, and 54 m with the PCI, ALM, and IUS, respectively. Conclusions: Partial coherence tomography is a high-precision, high-resolution, noncontact biometric technique.The commercial PCI prototype is practical in clinical use, with improved comfort for patients, no need for anesthesia, and a reduced risk of infection. However, the difference between the PCI and IUS in axial length measurement must be considered when using the constants supplied by intraocular lens (IOL) manufacturers for IOL power calculations. J Cataract Refract Surg 2002; 28:224 –229 © 2002 ASCRS and ESCRS
T
he precise measurement of intraocular distances is an important task in modern ophthalmology. Recent studies demonstrate that obtaining accurate preopAccepted for September 19, 2001. Reprint requests to Wolfgang Drexler, MD, Institut fu¨r Medizinische Physik, Universita¨t Wien, Wa¨hringer Strasse 13, A-1090 Vienna, Austria. E-mail: [email protected]. © 2002 ASCRS and ESCRS Published by Elsevier Science Inc.
erative biometric data is the most critical step in accurate calculation of the intraocular lens (IOL) refractive power.1,2 Currently, biometric measurements are performed with the ultrasound A-scan echo-impulse technique (US technique), enabling the measurement of axial length with a longitudinal resolution of typically 150 to 200 m and an accuracy of approximately 100 to 150 m.3⫺6 Studies based on US biometry demonstrate 0886-3350/02/$–see front matter PII S0886-3350(01)01272-X
BIOMETRY OF CATARACTOUS EYES USING PC1
that 54% of the errors in the predicted refraction after IOL implantation are attributed to axial length measurement errors.1 An axial length measurement error of 100 m would result in a corresponding postoperative refractive error of 0.28 diopter.1,7 Applanation US is the most commonly used technique for ocular biometry.8 Since this technique requires direct contact between the transducer and the eye, the cornea is indented and the anterior chamber depth as well as the axial length are shorter than with the more accurate but also more uncomfortable water-immersion ultrasound (IUS) technique. With the latter, the transducer has no direct contact with the cornea. Significant differences of 0.14 to 0.36 mm have been reported between applanation US and IUS axial length measurements.9,10 In both US techniques, a possible mismatch between the measurement axis and the visual axis of the eye may influence the results. A potential alternative to conventional US is an optical biometry technique that was developed in the past decade.11⫺13 It is a special version of a noninvasive optical biomedical imaging technology called optical coherence tomography (OCT), which is based on an optical measurement technique known as low or partial coherence interferometry. This technique is analogous to conventional US pulse–echo imaging (US A- and Bmode) except that OCT uses infrared laser light instead of sound.14⫺17 A special dual-beam version of this interferometric technique—partial coherence interferometry—that eliminates the influence of longitudinal eye motion during measurement by using the cornea as a reference surface was used to perform axial length measurements in normal18 and cataractous eyes.19 Depending on the measured intraocular distance, precision values from 0.3 to 10 m have been reported.20 The applicability of partial coherence interferometry for measuring the axial length in cataractous eyes was shown in an extensive study of 196 eyes of 100 patients.19 Recently, this technique has been reported to perform accurate biometry in cataractous eyes, with the potential to improve the refractive outcome of cataract surgery by about 30% using the SRK II power formula.21 This optical biometry technique has also been used in a clinical setting to investigate its clinical feasibility for IOL measurements in pseudophakic eyes.22⫺24 All these studies were performed with a laboratory prototype of partial coherence interferometry. Recently, Zeiss
developed a commercial prototype using the optical biometry technique (axial length measurement, ALM). In addition to the high precision of axial length measurements with this modality, the unit is reported to be easy to use clinically. In this study, the commercial prototype (ALM) for measuring axial length in cataract patients was investigated by comparing the findings with those obtained with the laboratory prototype (PCI) as well as those obtained with IUS biometry. The handling of this first clinical prototype and its performance in a clinical routine was also assessed.
Patients and Methods Preoperative Biometry The principle of the dual-beam version of partial coherence interferometry has been described.11⫺13,16,18⫺21 The differences between the ALM and the PCI prototypes are as follows: (1) The axial resolution of the ALM is approximately 100 m and of the PCI, 10 m. (2) The ALM emits laser light centered at 780 nm; with the PCI, it is centered at 855 nm. A single A-scan to measure axial length takes about 0.5 seconds with both prototypes. The differences in resolution and wavelength are due to the different light sources used in the 2 instruments. Both instruments use special optics (diffractive optics element) to enhance sensitivity, especially important for measuring cataractous eyes.25 Since the ALM and PCI yield optical distances, these have to be divided by the group refractive indices of the respective ocular media for the center wavelength to obtain geometric distances.21,26 For measurements performed with the IUS technique (Ocuscan, Alcon), sound velocities of 1532 m/s were taken for the aqueous and vitreous humor and 1641 m/s for the crystalline lens. Patients and Study Design The study protocol was approved by the Ethics Committee of the Vienna University School of Medicine. The nature of the study was explained, and all patients gave written consent to participate. Preoperatively, 2 examiners assessed 10 measurements of axial length with each technique in 49 eyes of 37 cataract patients. The mean age of the patients was 75 years (range 36 to 91 years). Axial length was assessed with the
J CATARACT REFRACT SURG—VOL 28, FEBRUARY 2002
225
BIOMETRY OF CATARACTOUS EYES USING PC1
ALM, PCI, and IUS on the same day. During in vivo measurements of the human eye, laser safety regulations were considered. For the ALM, the light source used in this study had a center wavelength of 1 ⬇ 780 nm with power of about 350 W at the cornea. For the PCI, the light source had a center wavelength of 1 ⬇ 855 nm with power of about 220 W at the cornea or an intensity of approximately 572 W/cm2 (averaged over a 7 mm aperture). This can be applied for about 28 minutes.27 The time for a single measurement of axial length is 0.5 seconds. By performing 8 to 10 longitudinal scans for statistical purposes, the maximum time of continuous illumination is about 4 seconds, well below the safety limits. Data Analysis Data are presented as the median and range, indicating minimum and maximum. The precision of the techniques is defined as the standard deviation of multiple consecutive measurements of axial length under investigation. For the comparison of means, paired t tests were applied to data that could be described by the normal distribution. Correlations were investigated using linear regression; P ⬍ .05 was considered the level of significance.
Figure 1. (Kiss) Correlation of axial length measurements between the PCI and ALM (N ⫽ 49).
Results Axial length measurements obtained with all 3 biometry techniques are shown in Table 1. Axial length measured with the ALM and PCI did not differ significantly (P ⫽ .23; median difference ⫺3 m; range ⫺150 to ⫹328 m; mean difference ⫺9 m) (Figure 1), with 90% of measurements differing less than 100 m. There was a significant difference in axial length between the PCI and IUS (P ⬍ .001) and the ALM and IUS Table 1. Axial length assessed with the ALM, PCI, and IUS. Data are presented as median and range (N ⫽ 49). Measurement
ALM
PCI
IUS
Median (mm)
23.500
23.509
23.195
Minimum (mm)
20.253
20.190
19.882
Maximum (mm)
26.450
26.445
27.371
1.433
1.451
1.554
Standard deviation (mm)
ALM ⫽ commercial prototype of partial coherence interferometry; PCI ⫽ laboratory prototype of partial coherence interferometry; IUS ⫽ immersion ultrasound
226
Figure 2. (Kiss) Correlation of axial length measurements between IUS and the ALM (N ⫽ 49).
(P ⬍ .001). Axial lengths assessed with ALM correlated well with those assessed with IUS (r ⫽ 0.991, P ⬍ .001) (Figure 2). With the PCI, the axial length measured longer by ⫹186 m (mean ⫹171 m; range ⫺926 to ⫹581 m) than with IUS; with ALM, it measured longer by ⫹202 m (mean ⫹180 m; range ⫺921 to
J CATARACT REFRACT SURG—VOL 28, FEBRUARY 2002
BIOMETRY OF CATARACTOUS EYES USING PC1
ure 6). The difference between PCI and ALM in axial length measurements did not correlate with axial length. No correlation between precision and axial length was observed among the 3 techniques, indicating that short or long eyes can be measured with similar precision. Six of 55 eyes (11%) could not be measured with ALM. Five patients had fixation problems in 1 eye because of a mature cataract or macular degeneration. One eye was extremely hyperopic; thus, the short axial length could not be detected with the ALM prototype.
Discussion
Figure 3. (Kiss) Differences between axial lengths measured with the ALM, PCI, and IUS presented as box plots. (Box plots are characterized by the minimum and the maximum; the box depicts the 25th and 75th percentiles, the small square represents the median, the star represents the mean, and circles are the outliers [N ⫽ 49]).
⫹635 m). Eighty-six percent of the differences in axial length measurements were within 400 m (Figure 3). To confirm the good correlation between the ALM and PCI and the ALM and IUS, the measuring agreement was assessed by plotting the difference between the methods against their mean (Figure 4). The precision of axial length measurements with the PCI (median 18 m; range 5 to 116 m; mean 26 m) (Figure 5) was significantly better than with the ALM (median 26 m; range 10 to 156 m; mean 36 m) (P ⬍ .04). The precision of the ALM measurements was significantly better than with IUS (median 54 m; range 4 to 211 m; mean 64 m) (P ⬍ .001) (Figure 5). There was no correlation in axial lengths for each individual within the measurements for the different techniques (Fig-
Previous studies demonstrate the applicability of optical biometry based on partial coherence interferometry for measuring the axial length in cataractous eyes.19,21 Accurate biometry with this technique was recently demonstrated to improve the refractive outcome of cataract surgery by about 30% using the SRK II power formula.21 Further improvement can be obtained by using third-generation IOL power formulas.28 In this study, we compared the laboratory prototype of dual-beam partial coherence interferometry with the commercial prototype of this optical biometry technique. The precision of the PCI was 18 m, which is worse than we have demonstrated,21 probably because of the use of a specific optic to enhance sensitivity. Hence, different peaks originating from the retinal pigment epithelium/Bruch’s membrane/choriocapillaris interface cannot always be clearly separated. We showed that both optical devices measure similar axial lengths. The correlation between both techniques was excellent (r ⫽ 0.999). The precision was
Figure 4. (Kiss) Agreement of axial length measurements between the ALM and PCI (left, N ⫽ 49) and the ALM and IUS (right, N ⫽ 49). J CATARACT REFRACT SURG—VOL 28, FEBRUARY 2002
227
BIOMETRY OF CATARACTOUS EYES USING PC1
Figure 5. (Kiss) Precision of axial length measurements of the ALM, PCI, and IUS presented as box plots. The precision is defined as the mean of the standard deviations of 10 consecutive measurements. (Box plots are described in Figure 3.)
Figure 6. (Kiss) Comparison of differences in axial lengths between the PCI and ALM (solid triangle) and the PCI and IUS (open circle) in all eyes (N ⫽ 49).
significantly better with the laboratory prototype, perhaps because of the different bandwidths of the super luminescent diodes, resulting in resolutions of 12 m and 100 m for the laboratory and commercial prototype, respectively. Compared with IUS, the optical technique measures axial length longer by approximately 200 m. The median difference between PCI and IUS was 186 m and between ALM and IUS, 203 m. The systematic difference between optical and acoustic techniques may be due to the different reflection sites in the retina.19,21 Ultrasound measures up to the inner limiting membrane, whereas with PCI the maximum interference pat228
tern from the retina is detected at the interface with the retinal pigment epithelium. Furthermore, the mismatch of the beam axis and the visual axis during US measurements may cause a deviation in axial length measurements between biometry methods. This agrees with the results of our previous study comparing partial coherence interferometry with applanation US and IUS techniques.21 Nevertheless, we observed that the regression line of the correlation of IUS with the ALM did not have a gradient of 1 so the deviation is more obvious with shorter and longer axial eye lengths. This means that long eyes are measured to be shorter with the optical techniques than with the IUS technique. Immersion US uses a different velocity for the lens and the vitreous and aqueous humor. In this study, we used the mean group refractive index with the optical techniques.19 In long eyes, this assumption may cause the deviation. Correction factors should then be used.19 The precision of the PCI was also significantly better than that of the US technique. However, the PCI also caused some outliers. The reason could be fixation problems of the patients, so measurements were not repeated at the same point on the retina. With the ALM, we were not able to measure in a few eyes (6 of 55 [11%]). Five patients had fixation problems because of mature cataract or macular degeneration. In conclusion, the ALM and PCI measured similar axial lengths. In contrast to US, optical techniques have a higher measurement precision. Measurements can be carried out with more comfort for the patient and without corneal contact, which minimizes the risk of infection. The assessment of axial length with the ALM is efficient, easy to use, quick to learn, and adequate for the clinical routine.
References 1. Olsen T. Sources of error in intraocular lens power calculation. J Cataract Refract Surg 1992; 18:125–129 2. Boerrigter RMM, Thijssen JM, Verbeek AM. Intraocular lens power calculations: the optimal approach. Ophthalmologica 191; 1985:89 –94 3. Binkhorst RD. The accuracy of ultrasonic measurement of the axial length of the eye. Ophthalmic Surg 1981; 12:363–365 4. Schachar RA, Levy NS, Bonney RC. Accuracy of intraocular lens powers calculated from a-scan biometry with the Echo-Oculometer. Ophthalmic Surg 1980; 11:856 – 858 5. Olsen T. The accuracy of ultrasonic determination of
J CATARACT REFRACT SURG—VOL 28, FEBRUARY 2002
BIOMETRY OF CATARACTOUS EYES USING PC1
6.
7.
8.
9.
10.
11. 12.
13.
14. 15.
16. 17.
18.
19.
20.
axial length in pseudophakic eyes. Acta Ophthalmol 1989; 67:141–144 Bamber JC, Trstam M. Diagnostic ultrasound. In: Webb E, ed, The Physics of Medical Imaging. Bristol, Philadelphia, PA, Adam Hilger, 1988; 319 –388 Olsen T. Theoretical approach to intraocular lens calculation using Gaussian optics. J Cataract Refract Surg 1987; 13:141–145 Leaming DV. Practice styles and preferences of ASCRS members—1998 survey. J Cataract Refract Surg 1999; 25:851– 859 Olsen T, Nielsen PJ. Immersion versus contact technique in the measurement of axial length by ultrasound. Acta Ophthalmol 1989; 67:101–102 Artaria LG. Messung der Bulbusla¨nge mit verschiedenen Ultraschall-Gera¨ten. Klin Monatsbl Augenheilkd 1986; 188:492– 494 Fercher AF, Roth E. Ophthalmic laser interferometry. Proc SPIE 658, 1986; 48 –51 Fercher AF, Mengedoht K, Werner W. Eye length measurement by interferometer with partially coherent light. Opt Lett 1988; 13:186 –188 Fercher AF, Hitzenberger CK, Juchem M. Measurement of intraocular optical distances using partially coherent laser light. J Mod Opt 1991; 38:1327–1333 Huang D, Swanson EA, Lin CP, et al. Optical coherence tomography. Science 1991; 254:1178 –1181 Huang D, Wang J, Lin CP, et al. Micron-resolution ranging of cornea anterior chamber by optical reflectometry. Lasers Surg Med 1991; 11:419 – 425 Fercher AF. Optical coherence tomography. J Biomed Opt 1996; 1:157–173 Fercher AF, Hitzenberger CK, Drexler W, et al. In vivo optical coherence tomography (letter). Am J Ophthalmol 1993; 116:113–114 Hitzenberger CK. Optical measurement of the axial eye length by laser Doppler interferometry. Invest Ophthalmol Vis Sci 1991; 32:616 – 624 Hitzenberger CK, Drexler W, Dolezal C, et al. Measurement of the axial length of cataract eyes by laser Doppler interferometry. Invest Ophthalmol Vis Sci 1993; 34:1886 –1893 Drexler W, Baumgartner A, Findl O, et al. Submicrometer precision biometry of the anterior segment of the hu-
21.
22.
23.
24.
25.
26.
27.
28.
man eye. Invest Ophthalmol Vis Sci 1997; 38:1304 – 1313 Drexler W, Findl O, Menapace R, et al. Partial coherence interferometry: a novel approach to biometry in cataract surgery. Am J Ophthalmol 1998; 126:524 –534 Findl O, Drexler W, Menapace R, et al. High precision biometry of pseudophakic eyes using partial coherence interferometry. J Cataract Refract Surg 1998; 24:1087– 1093 Findl O, Drexler W, Menapace R, et al. Accurate determination of effective lens position and lens⫺capsule distance with 4 intraocular lenses. J Cataract Refract Surg 1998; 24:1094 –1098 Findl O, Drexler W, Menapace R, et al. Changes in intraocular lens position after neodymium:YAG capsulotomy. J Cataract Refract Surg 1999; 25:659 – 662 Mo¨ ller B, Rudolph G, Klopffleisch A, et al. Application of diffractive optics for axial eye length measurements using partial coherence interferometry. In: Birngruber R, Fercher AF, Sourdille P, eds, Lasers in Ophthalmology IV. Proc SPIE 2930, 1996; 175–182 Drexler W, Hitzenberger CK, Baumgartner A, et al. Investigation of dispersion effects in ocular media by multiple wavelength partial coherence interferometry. Exp Eye Res 1998; 66:25–33 American National Standards Institute. Safe Use of Lasers, ANSI Z 136.1. New York, NY, American National Standards Institute, 1986 Findl O, Drexler W, Menapace R, et al. Improved prediction of intraocular lens power using partial coherence interferometry. J Cataract Refract Surg 2001; 27:861⫺867
From Universita¨ tsklinik fu¨ r Augenheilkunde, Allgemeines Krankenhaus Wien (Kiss, Findl, Menapace, Wirtitsch), and Institut fu¨ r Medizinische Physik, Universita¨ t Wien (Drexler, Hitzenberger, Fercher), Vienna, Austria. Presented in part at the XVIIth Congress of the European Society of Cataract & Refractive Surgeons, Vienna, Austria, September 1999. None of the authors has a financial interest in any product mentioned. Dr. Donnerhacke and colleagues, Zeiss Jena, provided the ALM prototype; H. Sattmann, Ing, and L. Schachinger, Institute of Medical Physics, University of Vienna, constructed the electronics and software of the instrument and provided technical support, respectively.
J CATARACT REFRACT SURG—VOL 28, FEBRUARY 2002
229