Using InterWave Aberrometry to Measureand Improve the Quality of Vision inLASIK Surgery

Keith P. Thompson, MD,1 P. Randall Staver, MS,1 Jose R. Garcia,1 Stephen A. Burns, PhD,2

Robert H. Webb, PhD,2 R. Doyle Stulting, MD, PhD1

Objective: To compare visual outcomes in eyes undergoing aberrometry-guided (InterWave) LASIK withthose in eyes undergoing standard LASIK treatment based upon refractive measures.

Design: Single-center, comparative, interventional, consecutive case series.Participants: Four hundred two consecutive eyes undergoing LASIK were analyzed retrospectively. One

group, 106 eyes undergoing primary LASIK and 224 eyes undergoing LASIK enhancement, was treated withstandard LASIK treatment using a 5.5-mm optical zone, 1.5-mm transition zone laser with the settings deter-mined by manifest refraction. The second group, 44 untreated (primary) eyes and 28 previously treated(enhancement) eyes, received a multipass, multistage treatment in which the laser settings for each stage weredetermined by aberrometry measurements. Eyes with desired monovision (undercorrected) outcome and pre-operative hyperopia were excluded from the study.

Intervention: An aberrometry-guided laser treatment (InterWave LASIK) was compared with the standardLASIK treatment based upon the manifest refraction.

Main Outcome Measures: Uncorrected visual acuity (VA), manifest refraction, best spectacle-corrected VA(BSCVA), severity of halos, and root mean square (RMS) retinal blur area measured at 3 months postoperatively.

Results: Three months postoperatively there was no difference in uncorrected VA, BSCVA, refraction, orRMS retinal blur areas for pupil sizes of 3.5 mm between eyes treated by InterWave and those treated bystandard LASIK. However, InterWave LASIK reduced the retinal blur area by 48% (P�0.0103) and 58%(P�0.0004) in primary cases and 43% (P�0.0430) and 74% (P�0.0271) in enhancement cases, respectively, forpupil sizes of 4.5 and 6.5 mm relative to standard LASIK treatments. Patients undergoing InterWave-guidedtreatment reported less severity of halo (0.37 vs. 0.98 [P�0.016] for primary cases and 0.35 vs. 0.73 [P�0.04] forenhancement cases).

Conclusion: InterWave LASIK achieved acuity and refractive results equivalent to those of standard LASIKtreatment based upon refraction, but resulted in superior quality mesopic vision. Ophthalmology 2004;111:1368–1379 © 2004 by the American Academy of Ophthalmology.

Over 1 million LASIK procedures were performed in theUnited States in 2001. It has been projected that 3 millionLASIK procedures will be performed in 2006, surpassingthe number of cataract procedures performed annually.1

Although most patients are satisfied with their vision afterLASIK, some patients experience glare, halos, and ghostimages, particularly at night or in dim conditions. Visualaberrations after LASIK surgery have received increasedattention in the news media2 and scientific literature3,4 and

Originally received: May 23, 2002.Accepted: June 24, 2003. Manuscript no. 220346.1 Emory Vision, Atlanta, Georgia.2 Schepens Eye Research Institute, Boston, Massachusetts.

Supported in part by the National Eye Institute, Bethesda, Maryland (grantno.: EYO4395) (SAB).

The authors retain a financial interest in the intellectual property that isdescribed in the article.

Reprint requests to Jose R. Garcia, InView, 4170 Ashford DunwoodyRoad, Suite 300, Atlanta, GA 30319.

1368 © 2004 by the American Academy of OphthalmologyPublished by Elsevier Inc.

on internet sites dedicated to patients who have experiencedpoor visual outcomes after surgery.5

Many patients who complain of visual aberrations afterrefractive surgery have visual acuities of 20/20 and refrac-tions near plano, making it impossible to evaluate theirsymptoms with conventional measures of acuity or refrac-tion. Corneal topographers measure the optical power of thecorneal surface and provide a qualitative indication of thepresence of irregular astigmatism. Quantitative indices,such as the Topographic Irregularity Index developed byMaloney, have been reported to predict visual perfor-mance.6–8 However, corneal topographic measures alonecannot describe aberrations, because information about therefractive power of the posterior corneal surface, the crys-talline lens, and the focal length of the eye is lacking.Description of aberrations requires measuring the effect ofall of the eye’s optical components and their interrelation-ships on the quality of the retinal image.

Aberrometers, or wavefront sensors, are diagnostic in-

ISSN 0161-6420/04/$–see front matterdoi:10.1016/j.ophtha.2003.06.031

Thompson et al � InterWave Aberrometry and the Quality of Vision in LASIK Surgery

struments that measure the wavefront error of the eye.9 Thewavefront error is a mathematically derived description ofthe 3-dimensional shape of a wave of light after it is alteredby the eye’s optical system. The wavefront error is refer-enced in the x–y dimension to the pupillary plane and in thez dimension to anterior or posterior deviations from a the-oretical wavefront that produces a perfect, diffraction-lim-ited image on the retina. Unlike corneal topography, whichmeasures only the surface power of the cornea, the wave-front error describes how an image on the retina is alteredby all of the eye’s optical elements. Unlike spherocylindri-cal refractive measurements, the wavefront error is notrestricted to axisymmetric shapes; it can describe both re-fractive error (defocus) and higher order aberrations such ascoma and spherical aberration, which are common afterrefractive surgery.10 Aberrometry may provide more de-tailed information about visual performance than acuity andmore comprehensive information about the refractive powerof the eye than refraction. Aberrometry may be useful inrefractive surgery in measuring the quality of vision and increating laser treatments that minimize postoperative aber-rations.11,12

The InterWave Scanner is a visual aberrometer devel-oped by researchers at Emory Vision, General ElectricCompany Research Center, Harvard, and the MassachusettsInstitute of Technology during the past decade.13,14 Anupdated, solid-state version of the InterWave Scanner suit-able for use in our refractive surgery referral practice atEmory Vision was built and registered with the Food andDrug Administration in April 2001 as a 510(k) exemptdevice.

In this study, we used the aberrometer’s wavefront errormeasurements to derive parameters of visual performanceand optical power specific to 3 pupillary zones (central,0–3.5 mm, and 3.5–4.5-mm and 4.5–6.5-mm annularzones), and we correlated them to acuity, refraction, andseverity of halos reported by patients after LASIK surgery.A method for customizing laser treatment using the zonaloptical power measurements determined by InterWave ab-errometry is described, and the visual results using InterWave-guided treatment are compared with those of conventionalLASIK treatment based upon refraction.

Materials and Methods

InterWave Scanner

The optical design of the InterWave Scanner has been describedpreviously.13 Briefly, the scanner employs the Scheiner principle(Fig 1) to measure the aberrations of the patient’s visual system.The instrument is equipped with a Badal optometer that correctsthe spherical equivalent (SE) refractive error of the eye from �15to �7 diopters (D). The images of the test spot and the alignmentreticule are delivered through 2 independent optical channels thatpass through 0.5-mm-diameter regions of the cornea. A thirdchannel (housing a charged coupled device camera) acts as apassive pupil-monitoring system that records the patient’s pupilimage, measures its diameter, and tracks its location during theentire measurement session.

Clinical Measurements

The scanner is approximately the size of a common autorefractorand is operated by a technician (Fig 2). InterWave aberrometry isperformed 20 minutes after the instillation of 1 drop of 1% cyclo-pentolate (Cyclogyl, Alcon Laboratories Ltd., Fort Worth, TX) toobtain measurements through a widely dilated pupil and to negatethe effects of accommodation. The patient’s task is to align the testspot within the center of the reticule using a joystick, and then toclick a button when he or she deems the alignment satisfactory.The device then moves the test spot to another location on thecornea, and the test sequence is repeated. At each spot, the scannerrecords the angles required to bring the test spot into alignment.This measurement is made for both the x and the y directions (anangle pair). The testing takes approximately 3 minutes per eye inmost subjects to measure 50 to 75 testing points.

Aberrometry Analysis

The angle pair measured at each testing point represents the localderivatives of the wavefront error of the visual system at eachpupillary locus. A least-squares procedure, written in MATLAB,15

was used to fit the slope measurements to the derivatives of theZernike polynomials using the method described by Webb.16 Weused the derived coefficients to provide estimates of the weight ofindividual aberrations and to reconstruct the overall wave aberra-tion of the eye. The measurement technique and analysis routineswere validated as described in He et al.17 From the wavefrontaberration measurements for the entire pupil, we derived 3 param-eters for the 3 pupillary zones described in the introduction (Fig2A): the Pupillary Zone Blur Anatomy, the zonal root mean square(RMS) retinal blur area, and the zonal optical power, as describedbelow.

InterWave Aberrometry ParametersPupillary Zone Blur Anatomy. The Pupillary Zone Blur Anat-

omy (Fig 2B, middle) is an alternative to the Point Spread Function(Fig 2B, left). It is an optical ray-trace of a grid of 100�100 rays,from the object (a point source located at infinity), through theentrance pupil of the eye, and to the focal plane. This depictionprovides a simultaneous estimate of vision quality color-coded foreach of 3 different pupil diameters.

We established a calibration circle around the center of the mapat 8.7 milliradians (Fig 2B), which is the angle subtended by thefull moon when �45° above the horizon, to add a familiar per-spective of blur size for clinicians and patients in describing theaberrations.

This figure is identical to an optical spot diagram, except thatlight passing through the central part and the paracentral andperipheral pupil is color-coded blue, green, and red, respectively.The Blur Anatomy cannot show the small-scale ripple associatedwith the diffraction and interference effects; however, it can cor-rectly reveal the large-scale intensity distribution through the raydensity in ray-hits per square milliradians at the focal plane. Thisis principally true when the wavefront error is large relative to thewavefront of light, which is true in most clinical situations. Inthese circumstances, blur diameter predictions calculated usingdiffraction-based methods or ray-based methods yield similarresults.

Although the Zone Blur Anatomy is not an exact representationof the intensity distribution, it has the following distinct advan-tages over the Point Spread Function: if an irregular defect in thewavefront causes a strongly distorted spot in the image quality, theBlur Anatomy can immediately reveal what portion of the pupil isresponsible for the defect. It also reveals the pupil diameter at

1369

Ophthalmology Volume 111, Number 7, July 2004

1

2

hery;

Thompson et al � InterWave Aberrometry and the Quality of Vision in LASIK Surgery

which the defect will affect vision. In the Point Spread Function,this pupillary dependence is lost. The Blur Anatomy shows theimage quality for several different pupil diameters simultaneouslyin a single graphical figure, and we have found it to be very usefulin the clinical environment for visualizing the blur pattern pro-duced by the eye’s defocus and higher order aberrations and forcorrelating to the patient’s reports of glare, halos, or ghost images.

Zonal Root Mean Square Retinal Blur Area. The zonal RMSretinal blur area is a calculation of the area of the blur pattern onthe retina, in square microns, produced by light passing througheach of the pupillary zones. The RMS is a standard statisticalmethod used to compute a mean where the variable in question hasboth positive and negative values; it is the square root of the sumof squares divided by the number of observations. The RMS retinalblur area is a parameter derived from InterWave measures thatdepicts the size of the retinal blur caused by a point source of lightat infinity. The smaller the RMS retinal blur, the sharper andcrisper the retinal image. The larger the RMS retinal blur, the morediffuse the retinal image and the more likely that the patient willreport a halo or glare.

Zonal Optical Power. We determined the sphere, cylinder,and axis that minimized the RMS retinal blur area for each zone.Zonal-optimized spherocylindrical corrections could be easily cor-related to the patient’s refraction, and the zonal optical powerswere used to program the laser settings in a multipass, multistageInterWave LASIK treatment.

Each zone was independently optimized using a conventionaloptimization algorithm that input the sphere, cylinder, and axis ofa hypothetically applied correction. The algorithm determined theresultant RMS blur area as an output figure of merit. The resultingzonal optical power predicts the best spherocylindrical correctionfor each zone, independent of the other zones.

For example, the zonal optical power profile of an eye with arefraction of �1.50 �0.50 �180 is shown in Figure 2D. Thisprofile shows that the optical power changes from �1.45 �0.66�178 in zone 1 (3.5 mm diameter) to �2.66 �0.86 �3.3 in theannual zone 3 (4.5–6.5 mm diameter). This reflects increasingmyopia in the periphery of the eye’s optical system consistent withthe presence of spherical aberration.

Correlation of Pupillary Zonal Root Mean SquareRetinal Blur Areas to Acuity and to Severity ofHalos

InterWave aberrometry was performed on 1424 eyes being eval-uated for LASIK. The refraction of the eyes varied from an SE of�14.25 D to �6.63 D, with astigmatism up to �5.75 D. The zonal

Figure 1. Top, The Scheiner principle used in InterWave aberrometrypinholes. Each hole coincides with a different location in the corneal plathe light source viewed through the pinholes would form 2 images on thethe angle of the beam to be changed while maintaining the same positionrepresents the slope of the wavefront at that location. Position 1 � perip

Figure 2. Pupillary zones and zonal parameters used in InterWave aberromstabilizers. An eye tracker monitors patient fixation in the x, y, and z axeto 6.5-mm annular zones. B, The Point Spread Function shows the imageof the image perceived created by a point source of light at infinity, color coat an 8.7-milliradian (mR) diameter, the same diameter of the full moon 4square (RMS) retinal blur area. This is a measure of optical performance asby a bundle of light rays passing through each of the 3 pupillary apeSpherocylindrical optical power by pupillary zone. These measurementssmallest RMS blur area for each zone. WeSC � wavefront error without4™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™

RMS retinal blur area for the 3 pupillary zones was calculated andcompared with uncorrected visual acuity (VA) measured by theEarly Treatment Diabetic Retinopathy Study chart and with the SEof the manifest refraction using an analysis of variance regressionmodel (PROC REG, SAS 8.018).

For 177 eyes that were examined with the InterWave Scannerat 3 months after LASIK, we asked the patient to rank the severityof their halos at night (0 � no halo, 4 � severe halo). The zonalRMS retinal blur areas for the 3 pupillary zones were derived fromaberrometry and compared with the patient’s score of halo sever-ity. A statistical regression model was used to determine therelationship between the severity of halos and the zonal RMSretinal blur area for each pupillary zone.

Standard LASIK TechniqueConventional LASIK treatment is based upon a single spherocy-lindrical measurement determined by the manifest refraction, thecycloplegic refraction, or a consensus between the two. The re-fractive setting is entered into the laser’s computer that controls thediameter of the optical and transition zones and the spherocylin-drical treatments that are applied to those zones. The surgeon mayuse a nomogram or a look-up table to modify the manufacturer’slaser settings based upon the type of surgery to be performed andprior results.

In 1997, we began LASIK treatments with an EC-5000 excimerlaser (Nidek Co., Fremont, CA) using a first-generation nomogramdeveloped by Salah and Waring that used a 5.5-mm optical zoneand a 7.0-mm transition zone.19 Huang and Stulting determinedthe coupling effects between the spherical and cylindrical compo-nents by regression analysis20 and created a second-generationnomogram in 1998. Patients treated with the second-generationnomogram obtained more accurate postoperative refractions, fewerovercorrections, and improved unaided Snellen acuities relative topatients treated with the first-generation nomogram (Invest Oph-thalmol Vis Sci 40[suppl]:S588, 1999). Since that time, the sec-ond-generation nomogram has been used as a standard in ourLASIK proceedures.

InterWave LASIK TechniquePop and Payette reported a multipass, multistage laser techniquefor photorefractive keratectomy that improved results over asingle-pass treatment.21 Pop and Payette divided the laser treat-ment into several stages and applied treatment in a sequentialmanner. We adopted Pop and Payette’s multipass, multistagetechnique that used distinct laser treatment settings for pupillaryzones of 3.5, 4.5, and 6.5 mm, as determined by the zonal optical

disk with 2 holes to view a distant point source of light through bothP). If there were any errors in refraction or the presence of aberrations,

a, rather than one. RP � retinal plane. Bottom, InterWave optics allowse cornea. The angle required to null the aberrations at each pupil positionposition 2 � center.

Adjustments are made in patient positioning using chin rest and foreheadThe 3 pupillary zones: 0 to 3.5 mm, central, and 3.5- to 4.5-mm and 4.5-sity in the retinal plane. The Pupillary Zone Blur Anatomy: a depictiony the pupillary zone that the light rays passed through. The circle is drawnmore above the horizon as viewed by an observer on earth. C, Root meantion of pupil size. It describes the RMS blur area in square microns created

s. The smaller the blur area, the sharper and brighter the image. D,be the best combination of sphere, cylinder, and axis that results in thent’s spherocylindrical correction.™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™

uses ane (Cretin

on th

etry.s. A,

intended b5° ora funcrture

descripatie™™™

1371

Ophthalmology Volume 111, Number 7, July 2004

powers from InterWave aberrometry. A transition zone with anouter diameter of 8.0 mm was used. No information from themanifest or cycloplegic refraction was used to program the laser inthe InterWave LASIK technique (Table 1, Fig 3).

Table 1. Comparison of Standard and InterWave LASIKTechniques

Standard LASIK InterWave LASIK

Laser settingsbased upon

Manifest andcycloplegic refraction

Zonal optical power derivedby aberrometry wavefrontmeasurements

No. of stages 1 3Zones 5.5-mm optical, 7.0-mm

transition3.5-, 4.5-, and 6.5-mm

optical zones; 8.0-mmtransition zone

Description Single treatmentapplied across thepupil with transitionzone

Three treatment zonestailored to correctvariation in opticalpower across 3 zones of3.5-, 4.5-, and 6.5-mmdiameter

Figure 3. A, The individual aberrometry measurements are recorded by therror is divided into 3 pupillary zones of 0- to 3.5-mm, 3.5- to 4.5-mm, ato determine the best spherocylindrical optical power. D, Each zone is thmanner described by Pop and Payette, and treatment is delivered to the evariations in the optical power across the pupil.

1372

InterWave LASIK versus Standard LASIK inPrimary and Enhancement Surgery

Forty-four untreated (primary) eyes and 28 previously treated(enhancement) eyes underwent InterWave LASIK treatment atEmory Vision. We compared these results at 3 months postop-eratively to those of a consecutive series of 106 patients under-going primary LASIK and 224 eyes undergoing LASIK en-hancement surgery with our current standard (single stage)LASIK technique during the same period (August–September2001) by the same surgeon (KPT). Patients undergoing en-hancement surgery were initially treated with a myopic LASIKprocedure using a 5.5 optical zone and a 7.0 transition zoneusing our second-generation nomogram with a Nidek EC-5000laser. The analysis for the enhancement cases was furthercontrolled for the original (untreated) manifest refraction of thepatient and the time since the initial treatment. The 4 cohortsexcluded eyes with a desired monovision outcome (undercor-rected) and those with preoperative hyperopia. Informed con-sent, approved by the Emory University institutional reviewboard, was obtained, and institutional review board approvalwas obtained for this research. On their 3-month visit aftertreatment, acuity, refraction, and InterWave aberrometry were

erWave Scanner, and the wavefront error is calculated. B, The wavefront- to 6.5-mm diameters. C, Each zone undergoes an optimization routinedependently programmed into the Nidek EC-5000 excimer laser in thea multipass, multistage treatment that tailors the laser correction to the

e Intnd 4.5en inye in

Thompson et al � InterWave Aberrometry and the Quality of Vision in LASIK Surgery

recorded. At that visit, patients were asked to quantify theseverity of halos at night using a scale of 0 (none) to 4 (intense).A regression model adjusting for preoperative refraction andsurgically induced refractive changes in outcomes between thecohorts was used to analyze the results.

Results

Correlation of Root Mean Square Retinal BlurArea to Acuity and to Severity of Halos

Figure 4 shows the correlation for each of the zonal RMS retinalblur areas (for pupillary apertures of 0–3.5 mm, 3.5–4.5 mm, and4.5–6.5 mm) to the uncorrected VA. The horizontal axis representsthe denominator (20/x) of the Snellen acuity in a logarithmic scale,and the vertical axis represents the RMS retinal blur area for eachof 3 zones. The dashed line represents the mean of the actual

Figure 4. The zonal root mean square (RMS) retinal blur area for 3 pupRMS blur area is predictive of acuity; the outer zones are not.

Table 2. Correlation of Zonal Root Mean

Zone Diameters

All Cases

Coefficient P

1 0–3.5 mm 0.34 0.082 3.5–4.5 mm 0.20 0.193 4.5–6.5 mm 0.22 0.03

population, and the solid line represents the predictive relationshipfor each zone. The results show that the RMS retinal blur area forthe 0- to 3.5-mm and 3.5- to 4.5-mm zones are good predictors ofthe uncorrected VA (P�0.0001 and R2 � 0.35, and P�0.0001 andR2 � 0.36, respectively). To a lesser degree, the RMS retinal blurarea of the 4.5- to 6.5-mm zone is predictive of VA (P�0.0001 andR2 � 0.21).

Table 2 shows the regression coefficients, P values, and R2 forthe RMS retinal blur area for each of the pupillary zones as apredictor of the severity of halos reported by the patient. The RMSblur area of the 4.5- to 6.5-mm zone was predictive of the haloseverity (P�0.05). The blur area of the inner 2 zones was notpredictive (P�0.05). Neither the unaided Early Treatment DiabeticRetinopathy Study acuity nor the manifest refraction was predictive ofhalo severity (P�0.05). These results were analyzed further for thesubset of patients with unaided 20/20 acuity postoperatively. For thisgroup, we also found the RMS blur area in the outer zone to be highlypredictive of the report of halos, as shown in Table 2.

zones versus Snellen acuity. N � 794 eyes. As expected, the inner zone

re Retinal Blur Area to Severity of Halos

Patients with 20/20 Acuity

R2 Coefficient P R2

0.14 0.23 0.23 0.060.05 0.33 0.07 0.220.24 0.20 0.03 0.37

illary

Squa

1373

Ophthalmology Volume 111, Number 7, July 2004

Correlation of Zonal Optical Power to ManifestRefraction

Figure 5 shows the relationship between the SE of the opticalpower of zone 1 (0–3.5-mm diameter) determined by InterWaveaberrometry and the SE of the patient’s manifest refraction. Thecorrelation coefficient was 0.965.

InterWave LASIK 3-Month Results

The general demographics and preoperative measurements of pa-tients undergoing primary and enhancement treatment with thestandard and InterWave LASIK techniques are shown in Table 3.Table 4 lists the preoperative refractive error and the time betweenprimary and enhancement surgery.

Figure 6 shows the mean postoperative acuity obtained forpatients undergoing primary and enhancement surgery with Inter-Wave and standard LASIK techniques. Both treatment modalitiesproduced similar results in average unaided VA and in averagebest spectacle-corrected VA (BSCVA) (P�0.05). No eyes lost �2lines of BSCVA. After enhancement surgery, 81% of patientsobtained 20/20 acuity or better.

InterWave aberrometry was used to compute zonal opticalpower (SE) as a function of pupillary diameter. These values areshown in Figure 7 for the 2 treatment methods for both primaryand enhancement surgery. We found that there was no differencein the residual refractive error in the first 3 mm of pupil diameter;

Figure 5. The spherical equivalent of the zone 1 optical power versus th

1374

however, with pupil sizes larger than 3 mm, InterWave LASIKprovided a more uniform correction of refractive error across theentire pupil, and it corrected more of the refractive error in theperiphery of the pupil. At a pupillary diameter of 7 mm, InterWaveLASIK eyes showed an average of 1.4 D less residual myopia thaneyes undergoing standard LASIK treatment.

We compared the reported severity of halos after surgery (0 �none, 4 � severe) between the 2 treatment modalities. Patientswho underwent standard treatment reported an average halo se-verity score of 0.98, whereas the patients who underwent Inter-Wave LASIK had an average halo severity score of 0.37 (P�0.02),as shown in Figure 8. Fifty-three percent of the patients whoreceived standard treatment reported no halos, versus 68% for thepatients who underwent InterWave LASIK. We also found that thepercentage of patients who reported halo severity of 3 or 4 wassignificantly lower for the patients who underwent InterWaveLASIK (1.3%), versus the patients receiving standard LASIK(12.9%, P�0.05).

The RMS retinal blur area for each pupillary zone is shown inFigure 9 for patients undergoing primary and enhancement surgerywith both types of treatment. In primary surgery, there was nodifference in retinal blur areas produced by the 2 treatments forpupil sizes of up to 3.5 mm. However, InterWave LASIK treat-ments reduced retinal blur area by 48% (P�0.01) and 58%(P�0.0004), respectively, for pupil sizes of 4.5 and 6.5 mmrelative to standard LASIK treatments. We found similar results inpatients undergoing enhancement surgery.

ient’s manifest refraction (MR).

e pat

Thompson et al � InterWave Aberrometry and the Quality of Vision in LASIK Surgery

Discussion

InterWave Aberrometry

Several aberrometers have recently become commerciallyavailable. Most are based upon the Shack–Hartmannmethod, in which a beam of light is reflected from theposterior eye wall and captured by an array of small lensesthat focus the reflected light onto a charged coupled devicechip for analysis.22 During the past decade, we designed aninstrument that measures the aberrations of the entire visualsystem, from the cornea to the visual cortex. Unlike aber-rometers that reflect light from the posterior eye wall, theInterWave Scanner incorporates a patient feedback task tolocate the retinal plane and to account for any neural pro-cessing of aberrations by the intraretinal layers and thehigher visual pathways. In this sense, the process performedby the InterWave scanner is similar to a manifest refraction,whereas the process performed by Shack–Hartmann aber-rometers is similar to that performed by an autorefractom-eter. The InterWave Scanner’s Badal slide neutralizes theeye’s spherical error, the dominant component of the wave-

Table 3. General Demographics of Patients UndergoingPrimary and Enhancement Surgery, Standard versus InterWave

LASIK

Primary Surgery Enhancement Surgery

StandardLASIK

InterWaveLASIK

StandardLASIK

InterWaveLASIK

N 106 44 224 28Uncorrected Visual Acuity (20/—)

Mean 314 291 50.43 52.71SE 4 14 2.64 4.88Minimum 12 12 16 16Maximum 800 400 400 100

Sphere (D)Mean �5.47 �4.90 �1.22 �1.20SE 0.08 0.28 0.04 0.13Minimum �16 �10.5 �4.00 �2.50Maximum �1 �1 0.25 0.50

Cylinder (D)Mean 0.75 0.48 0.35 0.25SE 0.02 0.05 0.02 0.06Minimum 0 0 0.00 0.00Maximum 5.75 2 1.75 1.25

Best-corrected visual acuity (20/—)Mean 19.99 18.65 19.67 19.36SE 0.34 0.43 0.24 1.13Minimum 10 12 10 12Maximum 200 40 40 32

MRSE (D)Mean �5.10 �4.66 �1.04 �1.07SE 0.08 0.28 0.04 0.11Minimum �14.75 �10.25 �3.13 �2.13Maximum �0.53 �0.63 0.88 0.50

AgeMean 40 38 41.74 37.86SE 0.27 0.95 0.66 1.91Minimum 20 20 20 20Maximum 65 58 67 60

D � diopters; MRSE � manifest refraction spherical equivalent; SE �standard error.

front error in most eyes. This allows the dynamic measure-ment range of the instrument to be used to measure higherorder aberrations. The sequential nature of the test reducesthe effect of transient dry spots and debris in the tear filmthat can cause errors with instantaneous reflection-basedmethods. Finally, we designed the instrument with a mea-surement range of 20 D, which permits measurement ofhighly aberrated eyes without errors induced by imagecrossover, which have been reported by investigators usingShack–Hartmann aberrometry.23 These design advantagescame at the expense of a testing time that requires approx-imately 3 minutes per eye, versus only a few seconds for aShack–Hartmann aberrometer. In addition, an InterWavescan requires an alert, attentive, and cooperative patient toperform the testing task.

We then used the wavefront error measurements fromthe InterWave scanner to graphically depict visual aberra-tions, to predict visual performance, and to determine theoptical (spherocylindrical) power within 3 pupillary zonesin patients undergoing LASIK surgery.

InterWave Parameters

After LASIK surgery, many patients report that their visionis satisfactory in daytime conditions, but that they experi-ence halos, glare, ghost images, and other aberrations atnight or in dim conditions. The Blur Anatomy map depictsthe aberrations experienced by the patient, relates them topupil size, and compares their size and configuration to afamiliar astronomical object. Figure 10 shows visual aber-rations in the right eye of one of the authors (KPT) beforeand after InterWave LASIK.

We validated that the zonal RMS retinal blur area (anInterWave-derived metric) was predictive of visual perfor-mance for each pupillary aperture by correlating the RMSretinal blur area to acuity, refraction, and the severity ofhalos. As expected, we found that the RMS retinal blur areafor the 2 inner zones was most predictive of unaided Snellenacuity and the SE of the manifest refraction. The RMS blurarea of the outer (4.5–6.5 mm) zone was predictive of theseverity of halos reported by patients who had undergoneLASIK surgery. Even in a subgroup of postoperative pa-

Table 4. General Demographics of Patients UndergoingEnhancement Surgery, Standard versus InterWave LASIK

StandardLASIK

InterWaveLASIK

Original preoperative manifest refraction(diopters)

Mean �5.57 �5.85SE 0.16 0.49Minimum �13.25 �11.50Maximum �3.38 �2.25

Time to enhancement (days)Mean 222.55 205.29SE 9.25 20.29Minimum 85.00 75.00Maximum 834.00 460.00

SE � standard error.

1375

Va) a

Ophthalmology Volume 111, Number 7, July 2004

tients with 20/20 unaided acuity, the RMS retinal blur areaof zone 3 was predictive of halo severity. Therefore, thezone 3 RMS retinal blur area can be used as a uniquebenchmark to refine the quality of mesopic vision because itwas predictive of the patient’s severity of halos, whereasacuity and refraction were not. Even though there was acorrelation between the patient’s halo severity and the pu-pillary zonal blur area, a cause and effect relationship wasnot established in our study. Factors other than those westudied (age, preoperative manifest refraction, and mesopicpupil size) could have caused a correlation between halo

Figure 6. Mean postoperative acuity in primary LASIK and enhancementmodalities produced similar results in average unaided visual acuity (UC

Figure 7. Postoperative zonal optical power (spherical equivalent [SE] vsthe 2 groups in the first 3 mm of pupil diameter; however, with pupil sizerefractive error across the entire pupil. InterWave LASIK treatments coInterWave LASIK eyes showed an average of 2.4 diopters less residual m

1376

severity and the RMS retinal blur area. However, determina-tion of a cause–effect relationship will require further study.

From the InterWave wavefront measurements, we de-rived the zonal optical power for pupillary apertures of 0 to3.5 mm central, 3.5 to 4.5 annular, and 4.5 to 6.5 mmannular zones. Whereas the RMS retinal blur area is pre-dictive of visual performance specific to the 3 pupillaryapertures, the zonal optical power is predictive of refractivepower specific to the 3 pupillary apertures. As expected, thezonal optical power for the inner zone (0–3.5 mm) washighly predictive of the SE of the manifest refraction. Vari-

ry. Standard LASIK treatment versus InterWave LASIK. Both treatmentnd in average best spectacle-corrected visual acuity (BC Va).

l diameter). There was no difference in the residual refractive error fromer than 3 mm, InterWave LASIK provided a more uniform correction ofd more of the refractive error in the periphery of the pupil. At 7 mm,.

surge

. pupis largrrecteyopia

Thompson et al � InterWave Aberrometry and the Quality of Vision in LASIK Surgery

ations in the optical power across the pupil indicate thepresence of higher order aberrations. For example, myopiathat increases in the outer pupillary zone is an indication ofthe presence of spherical aberration, common in most my-opic eyes. The zonal optical power measurements were usedto create laser settings for the Nidek EC-5000 laser in amultipass, multistage treatment method that we termedInterWave LASIK.

InterWave LASIK

In conventional LASIK surgery, the manifest (cycloplegic)refraction is used to program the laser, which delivers thedesired treatment to a single optical and transition zone.However, refraction provides only an average measure ofthe spherocylindrical error across the pupil; it does not

Figure 8. Mean severity of halos, 3 months after surgery. Patients reportedInterWave LASIK reported less severity of halos than patients underenhancement (reduction of 52%) surgery.

Figure 9. From the root mean square blur angle, we compute the averagea distant point source produced by the eye’s wavefront error for each of 3areas produced by the 2 treatments for pupil sizes of 3.5 mm (as expected)58% (P�0.0004), respectively, for pupil sizes of 4.5 and 6.5 mm comparedretinal blur area by 43% (P�0.0430) and 74% (P�0.0271), respectively,enhancement surgery. Sample size: standard LASIK, 106 eyes; InterWave

measure the variation in optical power across the pupil,which is the cause of visual aberrations. Because the Inter-Wave LASIK technique tailors the laser dosing to the vari-ations in optical power across 3 discrete pupillary zones asmeasured by aberrometry, we expected to decrease thepatient’s postoperative visual aberrations as compared withpatients undergoing the standard LASIK treatment method.

Three months after surgery, InterWave LASIK patientsundergoing primary and enhancement surgery achievedacuity and refractive results identical to those of patientsundergoing standard LASIK treatment. After enhancementsurgery, 81% of patients achieved 20/20 acuity. This indi-cates that the laser settings derived from InterWave aber-rometry were at least as accurate as the manifest and cyclo-plegic refractions that we use in our standard LASIKtreatment.

everity of halos after surgery (0 � none, 4 � severe). Patients undergoingstandard treatment in both primary surgery (reduction of 62%) and

(bars) at the retinal plane of the blur (relative size, circles) produced bysizes. For primary and enhancement cases, there was no difference in blurever, InterWave LASIK reduced retinal blur area by 48% (P � 0.01) andtandard LASIK treatments in primary surgery. InterWave LASIK reducedpil sizes of 4.5 and 6.5 mm compared with standard LASIK treatment inIK, 44 eyes.

the sgoing

areapupil. Howwith sfor puLAS

1377

perio

Ophthalmology Volume 111, Number 7, July 2004

However, patients undergoing InterWave LASIKachieved a higher quality of vision than did patients whounderwent standard LASIK treatments in both primary andenhancement surgery, as measured by a reduction in themean severity of halos and a reduction in the zone 3 RMSretinal blur area. The RMS blur area predicts that a pointsource of light would be, on average, 4 times as bright inmesopic conditions in patients undergoing InterWave en-hancement surgery as that in patients who underwent stan-dard enhancement surgery.

Were the results of InterWave LASIK attributed solely tothe increase in optical or transition zone size? We separatelyanalyzed 23 patients who underwent the standard LASIK

Figure 10. The preoperative and postoperative blur anatomy for the aPreoperative blur anatomy. Note the oval-shaped blur pattern characteristby the manifest refraction. Note that there is little improvement in image qThe image quality for all 3 zones has improved over preoperative valuescorrection; DX � x direction (right is nasal); DY � y direction (up is su

1378

technique using a 6.5-mm optical zone and an 8.0-mmtransition zone (N3 nomogram), the same diameters that areused in InterWave LASIK. Patients undergoing InterWaveLASIK had significantly less severity of halos and a smallerRMS retinal blur area in zone 3 than patients undergoingtreatment with standard LASIK (P�0.04 and P�0.03, re-spectively). This suggests that the improvements in visualperformance were a result of the aberrometry-guided lasertreatment, which corrected for variations in optical poweracross the pupil, and not simply due to the increased diam-eter of the optical and transition zones.

The InterWave LASIK technique required an average of40 seconds more treatment time and an approximately 20%

(KPT) who underwent InterWave LASIK as part of this study. Left,residual myopic astigmatism. Top right, Predicted blur anatomy providedy. Bottom, Postoperative blur anatomy after InterWave LASIK treatment.the image quality for a simple spherocylindrical treatment. CC � with

r); SC � uncorrected.

uthoric ofualitand

Thompson et al � InterWave Aberrometry and the Quality of Vision in LASIK Surgery

greater depth of tissue per diopter than the single-zonetreatment nomogram. Yet, for most patients with myopia upto 8 D with a preoperative corneal thickness of more than500 �m, the additional ablation did not exceed the 250-�mlimit to the posterior stroma, and adequate tissue was avail-able for enhancement treatment. Although the benefit ofsignificant improvement in the quality of vision may beworth the extra tissue required, surgeons must balance theimprovement in quality of vision gained against the de-creased tissue available for enhancement and the possibleincreased risk for ectasia inherent in deeper ablations. Ourresults were obtained without the benefit of refining thelaser nomogram. It is likely that visual outcomes could befurther improved by modifying the laser settings by apply-ing Huang’s method20 of regression analysis to each treat-ment zone.

This study was performed before Food and Drug Admin-istration approval of a laser for custom ablation in theUnited States, and the Nidek EC-5000 laser that was usedis limited to creating spherocylindrical ablations. Non-axisymmetric aberrations such as coma cannot be correctedwith this laser. In the future, flying spot laser systems willpermit treatment of higher order, non-axisymmetric aberra-tions. With these lasers, we can use the wavefront errorfrom the aberrometer directly to determine the laser ablationpattern. Thus, we view InterWave LASIK with the EC-5000as a simple method for customizing treatment with aber-rometry that will be supplanted as more sophisticated flyingspot laser technology becomes approved and available.24

Nevertheless, the techniques described in this study wereused effectively to improve the quality of vision for ourpatients.

References

1. Scerra C. Refractive market keeps eyes focused. OphthalmolTimes 2001;26:1, 64–5.

2. Braham L. Laser eye surgery: take a second look. BusinessWeek May 20, 2002:140.

3. Hersh PS, Steinert RF, Brint SF, Summit PRK-LASIK StudyGroup. Photorefractive keratectomy versus laser in situ kera-tomileusis. Comparison of optical side effects. Ophthalmol-ogy 2000;107:925–33.

4. Oshika T, Klyce SD, Applegate RA, et al. Comparison ofcorneal wavefront aberrations after photorefractive keratec-tomy and laser in situ keratomileusis. Am J Ophthalmol 1999;127:1–7.

5. Paulson T. Foundation helps people deal with laser eye sur-gery mishaps. Seattle Post-Intelligencer July 9, 2001. Avail-able at: http://seattlepi.nwsource.com/local/30560_laser09.shtml. Accessed August 1, 2001.

6. Maloney RK, Bogan SJ, Waring GO III. Determination of

corneal image-forming properties from corneal topography.Am J Ophthalmol 1993;115:31–41.

7. Dingeldein SA, Klyce SD, Wilson SE. Quantitative descrip-tors of corneal shape derived from computer-assisted analysisof photokeratographs. Refract Corneal Surg 1989;5:372–8.

8. Wilson SE, Klyce SD. Quantitative descriptors of cornealtopography. A clinical study. Arch Ophthalmol 1991;109:349–53.

9. Howland HC. Ophthalmic wavefront sensing: history andmethods. In: MacRae SM, Krueger RR, Applegate RA, eds.Customized Corneal Ablation: The Quest for Super Vision.Thorofare, NJ: Slack Inc.; 2001.

10. Marcos S. Aberrations and visual performance following stan-dard laser vision correction. J Refract Surg 2001;17:S596–601.

11. Applegate RA. Limits to vision: can we do better than nature?J Refract Surg 2000;16:S547–51.

12. Williams DR, Yoon GY, Guirao A, et al. How far can weextend the limits of human vision? In: MacRae SM, KruegerRR, Applegate RA, eds. Customized Corneal Ablation: TheQuest for Super Vision. Thorofare, NJ: Slack Inc.; 2001:11–32.

13. Webb RH, Penney CM, Thompson KP. Measurement of oc-ular wavefront distortion with a spatially resolved refractom-eter. Appl Opt 1992;31:3678–86.

14. Burns SA, Marcos S. Measurement of image quality with thespatially resolved refractometer. In: MacRae SM, KruegerRR, Applegate RA, eds. Customized Corneal Ablations: TheQuest for Super Vision. Thorofare, NJ: Slack Inc.; 2001:201–9.

15. MATLAB [computer program]. Version 5.2. Natick, MA: TheMathworks, Inc.; 1998.

16. Webb RH. Zernike polynomial description of ophthalmic sur-faces. Ophthalmic Vis Opt 1992;3:38–41.

17. He JC, Marcos S, Webb RH, Burns SA. Measurement of thewave-front aberration of the eye by a fast psychophysicalprocedure. J Opt Soc Am A Opt Image Sci Vis 1998;15:2449–56.

18. SAS [computer program]. Version 8.0. Cary, NC: SAS Insti-tute; 2001.

19. Salah T, Waring GO III, el-Maghraby A, et al. Excimer laserin-situ keratomileusis (LASIK) under a corneal flap for myo-pia of 2 to 20 diopters. Am J Ophthalmol 1996;121:143–55.

20. Huang D, Stulting RD, Carr JD, et al. Multiple regression andvector analyses of laser in situ keratomileusis for myopia andastigmatism. J Refract Surg 1999;15:538–49.

21. Pop M, Payette Y. Multipass versus single pass photorefrac-tive keratectomy for high myopia using a scanning laser. JRefract Surg 1999;15:444–50.

22. Thibos LN. Principles of Hartmann-Shack aberrometry. J Re-fract Surg 2000;16:S563–5.

23. Doane JF, Morris S, Border AD, et al. Wavefront analysis. In:Boyd BF, Agarwal S, Agarwal A, Agarwal A, eds. LASIK andbeyond LASIK: Wavefront Analysis and Customized Abla-tion. El Dorado, Panama: Highlights of Ophthalmology Inter-national; 2001.

24. Arbelaez MC. Super vision: dream or reality. J Refract Surg2001;17(suppl):S211–8.

1379