Coregistration of preoperative computed tomography andintraoperative three-dimensional rotational X-ray imagesfor cochlear implant surgical evaluationPearlman, P.C.; Deurzen, van, M.H.W.; Pluim, J.P.W.; Grolman, W.

Published in:Otology & Neurotology

DOI:10.1097/MAO.0000000000000520

Published: 01/01/2014

Document VersionPublisher’s PDF, also known as Version of Record (includes final page, issue and volume numbers)

Please check the document version of this publication:

• A submitted manuscript is the author's version of the article upon submission and before peer-review. There can be important differencesbetween the submitted version and the official published version of record. People interested in the research are advised to contact theauthor for the final version of the publication, or visit the DOI to the publisher's website.• The final author version and the galley proof are versions of the publication after peer review.• The final published version features the final layout of the paper including the volume, issue and page numbers.

Link to publication

Citation for published version (APA):Pearlman, P. C., Deurzen, van, M. H. W., Pluim, J. P. W., & Grolman, W. (2014). Coregistration of preoperativecomputed tomography and intraoperative three-dimensional rotational X-ray images for cochlear implant surgicalevaluation. Otology & Neurotology, 35(10), 1759-1764. DOI: 10.1097/MAO.0000000000000520

General rightsCopyright and moral rights for the publications made accessible in the public portal are retained by the authors and/or other copyright ownersand it is a condition of accessing publications that users recognise and abide by the legal requirements associated with these rights.

• Users may download and print one copy of any publication from the public portal for the purpose of private study or research. • You may not further distribute the material or use it for any profit-making activity or commercial gain • You may freely distribute the URL identifying the publication in the public portal ?

Take down policyIf you believe that this document breaches copyright please contact us providing details, and we will remove access to the work immediatelyand investigate your claim.

Download date: 31. May. 2018

Coregistration of Preoperative Computed Tomographyand Intraoperative Three-Dimensional Rotational

X-Ray Images for Cochlear ImplantSurgical Evaluation

*Paul C. Pearlman, †‡Martinus H. W. van Deurzen, *Josien P. W. Pluim,and †‡Wilko Grolman

*Image Sciences Institute, ÞDepartment of Otorhinolaryngology and Head and Neck Surgery, and þBrainCenter Rudolf Magnus, University Medical Center Utrecht, Utrecht, The Netherlands

Objective: A registration procedure of intraoperative three-dimensional rotational x-ray (3DRX) imaging and preopera-tive computed tomography (CT) imaging so that intraoperativeCT quality imaging is available during cochlear implant surgery,providing detailed information concerning electrode position inthe cochlea and its relation to surrounding bony structures.Study Design: Retrospective case seriesSetting: Tertiary referral centerData: The imaging of five patients who had undergone cochlearimplant surgery is used to develop a semiautomatic registrationprocedure to integrate intraoperative 3DRX and preoperativeCT. The method is implemented in advanced medical imagingsoftware to compute the transformations. The electrode is seg-mented from the registered 3DRX images using a semiautomatedapproach. The segmented electrode is superimposed onto the CTdata. The methods are quantitatively validated based on expert-

labeled anatomical landmarks. These landmarks are identified inthe CT and 3DRX images by an expert.Main Outcome Measure: Mean error of the registration pro-cedure for five anatomical landmarks in millimeters.Results: Quantitative analysis showed a mean error of between0.5 and 1 mm for all anatomical landmarks, suggesting that theresults are trustworthy.Conclusion: We developed a reliable procedure for the regis-tration of intraoperative 3DRX imaging and preoperative CTimaging for cochlear implant surgery. This registration pro-cedure provides the ENT surgeon intraoperative high-qualityCT imaging during cochlear implant surgery. Key Words:3DRXVCochlear implantVDeafnessVHearing lossVIntraoperativeimagingVSurgery.

Otol Neurotol 35:1759Y1764, 2014.

At the time of the introduction of cochlear implanta-tion, the cochlear implants were only available for pa-tients afflicted with congenital deafness. Nowadays, theindication extends also to people with only high-frequency sensorineural hearing loss with residual hear-ing of the lower frequencies (1Y3). Implanting peoplewith residual hearing makes hearing preservation surgicaltechniques during cochlear implant surgery extremely

important. There are many factors influencing effectivehearing preservation during cochlear implant surgery.

The position of the electrode is an important factor inthe functional outcome of cochlear implantation. Shep-herd et al. (4) determined significant threshold reductionswhen the electrode is located in the scala tympani, closeto the modiolus. Aschendorff et al. (5) reported betterspeech performance in scala tympani insertions. Speechperformance was poorer when dislocation of the elec-trode, from the scala tympani into the scala vestibule,occurs. The poorest speech performance occurs in scalavestibuli implantation. Skinner et al. (6) detected a negativecorrelation between scala vestibule insertions and wordtests. Finley et al. (7) reported a negative correlation be-tween scala vestibule insertions and word recognitionscores. Correct insertion of the electrode in the scala tym-pani close to the modiolus improves functional outcome.

Address correspondence and reprint requests to Martinus H. W. vanDeurzen, M.D., University Medical Centre Utrecht, G05.129, PO Box85500, 3508 GA Utrecht, The Netherlands; E-mail: ENT-research@umcutrecht.nlThe authors disclose no conflicts of interest.Wilko Grolman received unrestrictive research grants from Cochlear

Ltd., Med-El GmbH, and Advanced Bionics.

Otology & Neurotology35:1759Y1764 � 2014, Otology & Neurotology, Inc.

1759

Copyright © 2014 Otology & Neurotology, Inc. Unauthorized reproduction of this article is prohibited.

Others reported a correlation between insertion depth andspeech performance (7Y12).

Lehnhardt (13) described the soft surgery procedure toreduce intracochlear damage during insertion. Meshik et al.(14) determined that insertion inferior or anteroinferior to theround windowmay facilitate atraumatic insertion. Postelmanset al. (15) compared the suprameatal approach and a mas-toidectomy and posttympanic approach for cochlear im-plantation. There has been no significant reported differencein functional outcome between the methods. There is nodifference between a cochleostomy and a round window in-sertion regarding hearing preservation (16).

When intracochlear damage caused by electrode in-sertion occurs, it ranges from minor trauma to severedisruption of the organ of Corti. Several electrodes havebeen specifically developed to reduce insertion trauma.Briggs et al. (17) reported that narrow flexible straightarrays are the least traumatic. Curved or stiffer electrodes

are correlated with basilar membrane perforation. Rolandet al. (18) analyzed the effect of a modiolar huggingelectrode. They found cochlear damage caused by theelectrode. Briggs et al. (19) evaluated the modiolar re-search array, a very thin precurved perimodiolar proto-type electrode array that is held straight by a sheath. Thesheath is helpful during the insertion and does not causeadditional intracochlear damage. The sheath could besafely removed after insertion of the electrode.

To achieve hearing preservation during cochlear im-plantation, it is important for the surgeon to monitor theposition of the electrode. Imaging has become a criticaltool for monitoring electrode position. Carelsen et al. (20)described the value of intraoperative three-dimensionalrotational x-rays (3DRX) for monitoring the localizationof the electrode in cochlear implant surgery. With 3DRX,you can observe misdirection, or foldovers, of the elec-trode (21Y23). Intraoperative imaging is a valuable tool

FIG. 1. Overview of methods used in this study. Thumbnails of 2-D image sections are provided at each stage so that inputs and outputsare clear.

FIG. 2. Example slices from 3DRX (A) and CT (B), before registration. C, The ROI chosen from the CT before registration.

1760 P. C. PEARLMAN ET AL.

Otology & Neurotology, Vol. 35, No. 10, 2014

Copyright © 2014 Otology & Neurotology, Inc. Unauthorized reproduction of this article is prohibited.

especially in cases of postmeningitis or otosclerosis whereinsertion can be difficult (24,25). Intraoperative imaging isadvantageous because, in the case of suboptimal insertion,revisions can be made during the same procedure. Themain limitation of intraoperative 3DRX imaging is itsrelatively low resolution; scala tympani and scala vestibuleinsertions cannot be clearly resolved nor can the position ofthe electrode in relation to the modiolus be recovered. Theaim of this article is to register intraoperative 3DRX im-aging and preoperative computed tomography (CT) im-aging so that intraoperative CT quality imaging is availableduring cochlear implant surgery, providing detailed infor-mation concerning electrode position in the cochlea andthe position of surrounding bony structures.

MATERIALS AND METHODS

We present a registration pipeline designed to accurately registerintraoperative 3DRX to preoperative CT. Aside from a simple pre-processing stage, explained in the following sections, themethod isfully automated. Furthermore, we use a semiautomated method tosegment the implantedmultielectrode array from the 3DRX imageand then overlay the array in the CT image. Figure 1 provides anoverview of our entire approach.

Registration PipelineThe registration algorithm used in this work seeks to align the

images based on matching intensities in the regions containingbone in each image. The output of the algorithm is a transfor-mation that warps the 3DRX image into the CT image domain.

Before registration, the following preprocessing steps are performed:1. CT ROI Selection: A region of interest (ROI) containing

slightly more than all of the anatomical information in the3DRX image is manually delineated in the CT image.

2. Bone Segmentation: Simple intensity thresholding is usedto segment the bone from the CT and 3DRX images.

3. Masks: Segmentations are used as bone masks on originalimage pairs. Figure 2 contains example slices from the3DRX and CT, before registration, and the ROI is chosenfrom the CT. Figure 3 contains the bone masks determinedby thresholding these slices. Because the 3DRX imagecontains significantly more noise than the CT, its resultantbone mask is less accurate. Nonetheless, this mask is suit-able for registration. A coarse rotation, determined empiri-cally, is also applied to all 3DRX images as an initializationfor the registration. Each 3DRX image is acquired with areasonably similar field of view, and this rotation is largelyuniform for all data sets.

Registration proceeds in two stages. First, a similarity trans-form (scale, translation, and rotation) is computed to roughlyalign the images. An affine transform (scale, translation, rotation,and shearing) is then determined to obtain the final result. Bothstages seek to maximize, by means of registration, the normalizedcorrelation between the intensities in each image. This measure isideal for template-matching problems such as ours, where weexpect similar intensity distributions but not necessarily similarcontrast in the image pairs. Both stages compute a multiresolutionregistration, that is, the images are smoothed with a Gaussiankernel of decreasing width so that large structures are initiallymatched before fine details. The registration result is iterativelyrefined at each resolution. As an example, Figure 4 contains aregistered 3DRX image overlaid on the corresponding CT.

FIG. 3. Sections of bone masks for 3DRX (A) and CT (B). These masks are produced by thresholding.

FIG. 4. Section of registered 3DRX image is overlaid on the corresponding CT slice.

17613-D ROTATIONAL X-RAY IMAGES FOR COCHLEAR IMPLANT SURGERY

Otology & Neurotology, Vol. 35, No. 10, 2014

Copyright © 2014 Otology & Neurotology, Inc. Unauthorized reproduction of this article is prohibited.

FusionThe multielectrode array is segmented from the registered

3DRX using a simple semiautomated approach. As an example,Figure 5 contains a 3-D rendering of the 3DRX image with thesegmented array overlaid.A user sequentially clicks on the electrode locations in the

3DRX image, and a smooth curve is fit through these points. Thiscurve is then superimposed onto the CT data, where measure-ments can be made regarding its proximity to clinical landmarks.Renderings of the array in the CT are shown in Figure 6.

ExperimentsThe registration method is applied to five subjects who have

undergone surgery to implant a multielectrode array at theUniversity Medical Center Utrecht.

DataPreoperative CT images are reconstructed with a resolution of

0.21 � 0.21 � 0.3 mm3. Three-dimensional rotational x-rayimages of the implanted array and the surrounding structuresare reconstructed with a resolution of 0.71 � 0.71 � 0.71 mm3

using a motorized C-arm.Methods are implemented in advanced medical imaging

software, and the elastix toolbox is used to compute the trans-formations (26). Registrations take, on average, 14 minutes on a64-bit PC with eight 2.27-GHz processors and 16-GB RAM.

SubjectsFor this study, we used the imaging of five patients who

underwent cochlear implant surgery. All patients were operatedon using a suprameatal approach (27).

ValidationMethods are quantitatively validated using five expert-labeled

anatomical landmarks identified by an ENT resident. Theselandmarks are identified in both the CT and 3DRX images. Thetransformations computed by the registration are then used towarp the points in the 3DRX image into the CT domain, whereEuclidean distances are calculated by comparing these locationswith their corresponding locations in the CT image. Furthermore,because one patient required a postoperative CT, the corre-sponding intraoperative 3DRX is registered to this image toprovide additional validation. Electrode locations in the cochleaare identified in the warped 3DRX and the postoperative CT, andEuclidean distances are calculated between these points.

RESULTS

Note in the overlay images in Figure 4 that the 3DRXoverlay visually corresponds with the CT image. There is noobvious misalignment. Nonetheless, the results are limitedby the resolution of the images, and the regions registered inthis project contain small anatomies of clinical interest. Forthis reason, the quantitative figures are more illuminating.

Quantitative AnalysisTable 1 contains the target registration errors for the

five anatomical landmarks identified by an expert ob-server. The errors for each landmark along with the meanerror are less than a voxel, suggesting that the results aretrustworthy. It should be noted that the registration erroris larger for some targets than others, suggesting that theregistration results are poorer for these locations. Alter-natively, these locations may have been more difficult toidentify manually in the images. Furthermore, the large

FIG. 5. Three-dimensional rendering of the 3DRX image with thesegmented array overlaid.

FIG. 6. Three-dimensional rendering of CT data including seg-mented multielectrode array from 3DRX.

FIG. 7. Multielectrode array from CT image segmented andoverlaid on 3DRX image.

1762 P. C. PEARLMAN ET AL.

Otology & Neurotology, Vol. 35, No. 10, 2014

Copyright © 2014 Otology & Neurotology, Inc. Unauthorized reproduction of this article is prohibited.

variance associated with several of the measures can beattributed to large differences in the field of view between3DRX scans, making landmark identification in someimages more difficult than others.

Figure 7 contains the multielectrode array from thepostoperative CT image segmented and overlaid on 3DRXimage. Note that the overlay corresponds with the arraylocations in the 3DRX image. The first five electrode lo-cations are quantitatively compared in Table 2.

DISCUSSION

We described the registration of intraoperative 3DRXimaging to preoperative CT imaging. With this registration,we can place the electrode array in the preoperative CT scan,which allows us to relate the position of the electrode to itssurrounding bony cochlea with the advantage of a betterquality and higher-resolution CT scan images.

Hearing preservation becomes more important duringcochlear implant surgery, so do the position of the elec-trode. The electrode should be placed in the scala tympaniclose to the modiolus. With this registration tool, it’s afirst step to detailed high-quality intraoperative imagingduring cochlear implant surgery.

The registration is still time-consuming, it takes approx-imately 10 minutes to make the 3DRX imaging and it takesanother 15 minutes for the registration. But in patients withresidual hearing, postmeningitis, or otosclerosis, we areconvinced that it is justified to do the registration to monitorthe exact location. Registration offers the surgeon a chanceto improve the position or to do an electrode array rein-sertion during the same procedure.

In the future, for other indications, intraoperative3DRX imaging may prove useful. For example, in case ofmiddle ear implantations like a sound bridge of a bonebridge and even for a mastoidectomy or radicalization,these registration tools allow for intraoperative monitor-ing of surgery.

Schipper et al. (28) described the use of computer-assisted cochlear implant surgery using a navigationsystem. They reported a deviation of 1.6 mm around thecochleostomy site. Matsumoto et al. (29) developed anoninvasive image guidance for cochlear implant surgeryusing the STAMP method and navigation software. Theyreported an average registration error of 2.4 mm. It givesthe surgeon feedback of his performance during cochlearimplant surgery. Stelter et al. (30) used an intraoperativeCT scan for image-guided cochlear implantation. It givesa detailed view of anatomical landmarks in the mastoid.The average deviation was between 0.9 and 1.01 mm andis quite precise. Drawbacks of this navigation are thecosts and the time needed for preparing the navigation.Therefore, interdisciplinary use by neurosurgeons, max-illofacial surgeons, and orthopedic surgeons is required tomake it cost-effective.

Compared with these navigation tools, our registrationis more accurate with a standard error between 0.5 and 1mm. Our method is more economic. It cannot help uswhile we are performing the insertion, but we have acheck immediately after insertion in the same session.

In the future, minimally invasive cochlear implantsurgery will grow in popularity. Majdani et al. (31,32)and Reda et al. (31,32) described a registration procedurefor surgical navigation using intraoperative flat-panelvolume computed tomography. Their registration proce-dure is reported to have an accuracy of less than 0.2 mm.Their registration procedure allows the ENT surgeon todrill a straight tunnel to the cochlea and perform acochleostomy using a minimally invasive technique.Their method is more precise than ours, but their patientswere exposed to relative high doses of x-rays by usingflat-panel volume computed tomography compared withusing 3DRX and preoperative CT scan.

CONCLUSION

We developed a reliable semiautomatic procedure for theregistration of intraoperative 3DRX imaging and preopera-tive CT imaging for cochlear implant surgery. This regis-tration procedure provides the ENT surgeon intraoperativehigh-quality CT imaging with the intraoperatively made3DRX images fused with it. The intraoperative 3DRXscanning causes low-dose radiation exposure. The next stepto help improve the clinically relevant images available tothe surgeon is by fusing intraoperative acquired images withpreoperative high-resolution magnetic resonance imagesVthe obvious advantage being a more detailed overview ofthe soft tissues surrounding the electrode array. This regis-tration procedure can be a next helpful step to a new method

TABLE 1. Target registration errors

Vestibulocochlear nerve Facial nerve Vestibular aqueduct Sigmoid sinus Horizontal canal Mean

0.592 T 0.283 0.983 T 0.172 0.720 T 0.141 0.499 T 0.140 0.887 T 0.338 0.736 T 0.201

Five landmarks are identified in each image by a trained expert, and distance after registration is calculated between the registration target and thetransformed point in the three-dimensional rotational x-rays. Values are given in millimeters.

TABLE 2. Quantitative comparison between transformedelectrode array in three-dimensional rotational x-rays with

array in postoperative computed tomography

Electrode Distance

1 0.2082 0.2953 0.4664 0.4995 0.742Mean 0.442 T 0.206

Values are given in millimeters.

17633-D ROTATIONAL X-RAY IMAGES FOR COCHLEAR IMPLANT SURGERY

Otology & Neurotology, Vol. 35, No. 10, 2014

Copyright © 2014 Otology & Neurotology, Inc. Unauthorized reproduction of this article is prohibited.

for intraoperative navigation during minimally invasivecochlear implant surgery.

REFERENCES

1. Fraysse B, Dillier N, Klenzner T, et al. Cochlear implants for adultsobtaining marginal benefit from acoustic amplification: a Europeanstudy. Am J Otol 1998;19:591Y7.

2. Gantz BJ, Turner C. Combining acoustic and electrical speech processing:Iowa/Nucleus hybrid implant. Acta Otolaryngol 2004;124:344Y7.

3. von Ilberg C, Kiefer J, Tillein J, et al. Electric-acoustic stimulationof the auditory system. New technology for severe hearing loss.ORL J Otorhinolaryngol Relat Spec 1999;61:334Y40.

4. Shepherd RK, Hatsushika S, Clark GM. Electrical stimulation of theauditory nerve: the effect of electrode position on neural excitation.Hear Res 1993;66:108Y20.

5. Aschendorff A, Kromeier J, Klenzner T, Laszig R. Quality controlafter insertion of the nucleus contour and contour advance electrodein adults. Ear Hear 2007;28:75sY9s.

6. Skinner MW, Holden TA, Whiting BR, et al. In vivo estimates ofthe position of advanced bionics electrode arrays in the humancochlea. Ann Otol Rhinol Laryngol Suppl 2007;197:2Y24.

7. Finley CC, Holden TA, Holden LK, et al. Role of electrodeplacement as a contributor to variability in cochlear implant out-comes. Otol Neurotol 2008;29:920Y8.

8. Skinner MW, Ketten DR, Holden LK, et al. CT-derived estimationof cochlear morphology and electrode array position in relation toword recognition in Nucleus-22 recipients. J Assoc Res Otolaryngol2002;3:332Y50.

9. Hamzavi J, Baumgartner WD, Pok SM, Franz P, Gstoettner W.Variables affecting speech perception in postlingually deaf adultsfollowing cochlear implantation. Acta Otolaryngol 2003;123:493Y8.

10. Hamzavi J, Arnoldner C. Effect of deep insertion of the cochlearimplant electrode array on pitch estimation and speech perception.Acta Otolaryngol 2006;126:1182Y7.

11. Boyd PJ. Potential benefits from deeply inserted cochlear implantelectrodes. Ear Hear 2011;32:411Y27.

12. Lazard DS, Vincent C, Venail F, et al. Pre-, per- and postoperativefactors affecting performance of postlinguistically deaf adults usingcochlear implants: a new conceptual model over time. PloS ONE2012;7:e48739.

13. Lehnhardt E. [Intracochlear placement of cochlear implant elec-trodes in soft surgery technique]. HNO 1993;41:356Y9.

14. Meshik X, Holden TA, Chole RA, Hullar TE. Optimal cochlearimplant insertion vectors. Otol Neurotol 2010;31:58Y63.

15. Postelmans JT, Stokroos RJ, van Spronsen E, et al. Comparison oftwo cochlear implantation techniques and their effects on thepreservation of residual hearing. Is the surgical approach of anyimportance? Eur Arch Otorhinolaryngol 2014;271:997Y1005.

16. Havenith S, Lammers MJ, Tange RA, et al. Hearing preservationsurgery: cochleostomy or round window approach? A systematicreview. Otol Neurotol 2013;34:667Y74.

17. Briggs RJ, Tykocinski M, Saunders E, et al. Surgical implicationsof perimodiolar cochlear implant electrode design: avoidingintracochlear damage and scala vestibuli insertion. Cochlear Im-plants Int 2001;2:135Y49.

18. Roland JT Jr, Fishman AJ, Alexiades G, Cohen NL. Electrode tomodiolus proximity: a fluoroscopic and histologic analysis. Am JOtol 2000;21:218Y25.

19. Briggs RJ, Tykocinski M, Lazsig R, et al. Development andevaluation of the modiolar research arrayVmulticentre collabo-rative study in human temporal bones. Cochlear Implants Int 2011;12:129Y39.

20. Carelsen B, Grolman W, Tange R, et al. Cochlear implant electrodearray insertion monitoring with intra-operative 3D rotational X-ray.Clin Otolaryngol 2007;32:46Y50.

21. Tange RA, Grolman W, Maat A. Intracochlear misdirected im-plantation of a cochlear implant. Acta Otolaryngol 2006;126:650Y2.

22. Grolman W, Maat A, Verdam F, et al. Spread of excitation mea-surements for the detection of electrode array foldovers: a pro-spective study comparing 3-dimensional rotational x-ray andintraoperative spread of excitation measurements. Otol Neurotol2009;30:27Y33.

23. Viccaro M, De Seta E, Covelli E, Marvaso V, Filipo R. Anotherreason for intra-operative imaging during cochlear implantation. JLaryngol Otol 2008;122:e12.

24. Proops DW, Stoddart RL, Donaldson I. Medical, surgical and au-diological complications of the first 100 adult cochlear implantpatients in Birmingham. J Laryngol Otol Suppl 1999;24:14Y7.

25. Rotteveel LJ, Proops DW, Ramsden RT, Saeed SR, van Olphen AF,Mylanus EA. Cochlear implantation in 53 patients with otosclero-sis: demographics, computed tomographic scanning, surgery, andcomplications. Otol Neurotol 2004;25:943Y52.

26. Klein S, Staring M, Murphy K, Viergever MA, Pluim JP. Elastix: atoolbox for intensity-based medical image registration. IEEE TransMed Imaging 2010;29:196Y205.

27. Tange RA, Grolman W. Een mastoidsparende benadering voor hetinbrengen van een cochleair implantaat. Ned Tijdschr KNO-heelk03-07-2004 2004;10:143Y7.

28. Schipper J, Aschendorff A, Arapakis I, et al. Navigation as a qualitymanagement tool in cochlear implant surgery. J Laryngol Otol2004;118:764Y70.

29. Matsumoto N, Oka M, Cho B, et al. Cochlear implantation assistedby noninvasive image guidance. Otol Neurotol 2012;33:1333Y8.

30. Stelter K, Ledderose G, Hempel JM, et al. Image guided navigationby intraoperative CT scan for cochlear implantation. Comput AidedSurg 2012;17:153Y60.

31. Majdani O, Bartling SH, Leinung M, et al. A true minimally in-vasive approach for cochlear implantation: high accuracy in cranialbase navigation through flat-panel-based volume computed to-mography. Otol Neurotol 2008;29:120Y3.

32. Reda FA, Noble JH, Labadie RF, Dawant BM. Automatic pre- tointra-operative CT registration for image-guided cochlear implantsurgery. IEEE Trans Biomed Eng 2012;59:3070Y7.

1764 P. C. PEARLMAN ET AL.

Otology & Neurotology, Vol. 35, No. 10, 2014

Copyright © 2014 Otology & Neurotology, Inc. Unauthorized reproduction of this article is prohibited.