clinical validation of percutaneous cochlear implant surgery: initial

The LaryngoscopeLippincott Williams & Wilkins© 2008 The American Laryngological,Rhinological and Otological Society, Inc.

Clinical Validation of Percutaneous CochlearImplant Surgery: Initial Report

Robert Frederick Labadie, MD, PhD; Jack H. Noble, BS; Benoit M. Dawant, PhD;Ramya Balachandran, PhD; Omid Majdani, MD, PhD; J. Michael Fitzpatrick, PhD

Objective: Percutaneous cochlear implant surgeryconsists of a single drill path from the lateral mastoidcortex to the cochlea via the facial recess. We sought toclinically validate this technique in patients undergoingtraditional cochlear implant surgery.

Study Design: Prospective clinical trial.Methods: After institutional regulatory board-

approved protocols, five ears were studied via thefollowing steps. 1) In the clinic under local anesthe-sia, bone-implanted anchors were placed surround-ing each mastoid. 2) Temporal-bone computed tomog-raphy (CT) scans were obtained. 3) On the CT scans,paths were planned from the lateral mastoid cortex,through the facial recess, to the basal turn of thecochlea both “manually” and “automatically” usingcomputer software. 4) Customized microstereotacticframes were rapid-prototyped to serve as drill guidesconstraining the drill to follow the appropriate path.5) During cochlear implant surgery, after drilling ofthe facial recess, drill guides were mounted on thebone-implanted anchors. 6) Accuracy of paths wasassessed via intraoperative photodocumentation.

Results: All surgical paths successfully tra-versed the facial recess and hit the basal turn of thecochlea. Distance in millimeters (average � SD) fromthe midpoint of the drill to the facial nerve was1.18 � 0.68 for the “manual” path and 1.24 � 0.44mm for the “automatic” path and for the chorda tym-pani 0.986 � 0.48 for the “manual” path and 1.22 �0.62 for the “automatic” path.

Conclusions: Percutaneous cochlear implant accessusing customized drill guides based on preoperative CTscans and image-guided surgery technology can be safelyaccomplished.

Key Words: Cochlear implantation, minimally inva-sive surgery, image-guided surgery, stereotactic frames,automated path planning.

Laryngoscope, 118:●●●–●●●, 2008

INTRODUCTIONWithin all fields of surgery, there has been a push

toward minimizing the invasiveness of procedures. Thishas occurred to limit comorbidity1 as well as to limitcost.2,3 Within the field of otolaryngology, perhaps themost notable shift to minimally invasive surgery is that offunctional endoscopic sinus surgery.4–6 Other areas thatare undergoing such transition include thyroid surgery7,8

and laryngeal surgery.9–11 In the continued effort to min-imize invasiveness, image-guided surgical systems (IGS)are being used more and more frequently. By using pre- orintraoperatively acquired radiographs, these systems al-low surgeons to determine the boundaries of the surgicalfield and positions of vital anatomic structures. Paralleldevelopment of IGS took place in America, Japan, andSwitzerland.12–14 Credit for introducing IGS to otolaryn-gology goes to Schlondorff et al.15,16

Within the subspecialty of otology/neurotology, anarea in which minimally invasive surgery has garneredattention is cochlear implant surgery. Early attempts tominimize cochlear implant surgery included the endo-meatal approach, in which a trough was drilled in theposterior external auditory canal to accommodate the con-necting wire from the internal receiver to the electrodearray. This approach was complicated by infection, wireextrusion into the external auditory canal, and cholestea-toma.17,18 More recent efforts have focused on smallerincision sizes,19,20 micromastoids,21 and single drilltroughs to the middle ear via the attic.22

The ultimate minimally invasive cochlear implant isone in which a single drill pass goes from the mastoid

From the Departments of Otolaryngology–Head and Neck Surgery(R.F.L., O.M.) and Electrical Engineering and Computer Science (J.H.N.,B.M.D., R.B., J.M.F.), Vanderbilt University, Nashville, Tennessee, U.S.A.

Editor’s Note: This Manuscript was accepted for publication January28, 2008.

Funded by the NIH (NIBIB R01 DC008408-01A1) and the Triologi-cal Society (Career Development Award).

Send correspondence to Robert F. Labadie, MD, Associate Professorof Otolaryngology–Head and Neck Surgery, 7209 Medical Center East,South Tower, Vanderbilt University Medical Center, Nashville, TN 37232-8605. E-mail: [email protected]

DOI: 10.1097/MLG.0b013e31816b309e

Laryngoscope 118: XX 2008 Labadie et al.: Percutaneous Cochlear Implant Surgery

1

cortex to the basal turn of the cochlea. This has beentermed “percutaneous cochlear implantation.” To achievethis radical approach, image guidance based on a preop-erative computed tomography (CT) scan is necessary. Ini-tial reports regarding this technique were made using IGSwith freehand drilling on cadaveric specimens.23 The au-thors of this study noted difficulty in keeping the drill onthe correct trajectory. The technique was then modified touse a microstereotactic frame as a drill guide constrainingthe drill to pass in the predetermined path.24 Both of thesereports were performed with cadaveric specimens.

Contained herein is an initial report on clinical trans-lation of the drill guide technique for percutaneous co-chlear implantation. This step consists of validation of asafe drill path. To accomplish this, patients underwentpreoperative imaging after having bone-implanted an-chors placed surrounding the ear. A customized drill guidewas created that constrained the path of a drill throughthe facial recess and to the basal turn of the cochlea. Then,after traditional cochlear implant surgery via mastoidec-tomy and facial recess but before performing the cochleos-tomy and implantation, the drill guide was affixed to theanchors to determine whether the percutaneous technique

would have been successful. The results for our first fourpatients (5 ears) are reported below.

METHODSInstitutional regulatory board (IRB) approval was obtained.

Each patient recruited underwent extensive preprocedural coun-seling detailing the specifics about the following procedures.

Anchor/Fiducial Marker PlacementAnchors with self-tapping threads were screwed into the

skull surrounding the temporal bone. These anchors serve twopurposes. First, they are used to register the patient’s anatomy tothe CT scan to allow referencing of the drill path (from lateralskull to cochlea) to the position of the anchors. Second, they havea hollow interior that is tapped to allow mounting of the drillguide at the time of surgery. Each anchor is made of titanium andis 4 mm wide and 8 mm long with 4 mm of this length screwedinto the skull. The shape and size can be seen in Figure 1. Threeanchors were placed surrounding the mastoid. Their locations are“mastoid,” “suprahelix,” and “posterior,” as shown in Figure 2,which is a view of a patient after anchor insertion and skinclosure and the corresponding radiograph.

Anchor placement was accomplished using local anesthesiain a procedure room in the clinic. The anchor sites were markedwith an indelible pen, cleaned with alcohol, and then injectedwith 1% lidocaine with epinephrine. The area was prepared anddraped without trimming of hair. At each site, a stab incision wasmade with a scalpel down to and through the periosteum. Hemo-stasis was achieved with direct pressure and, if necessary, bipolarelectrocautery. A nasal speculum was used for exposure. A peri-osteal elevator was used to expose bone. A modified surgicalscrewdriver (WayPoint driver, FHC, Inc., Bowdoin, ME) was usedin conjunction with an Osteomed electric driver (Model 450-0600,Adison, TX) to place anchors into the cranium to a depth of 4 mm,as shown in Figure 3. Incision sites were closed with subcutane-ous 3–0 Vicryl, 5–0 plain gut to skin and dressed with bacitracinzinc ointment.

CT ScanningAfter anchor placement, patients went directly to a CT

scanner where clinically applicable CT scans were obtained (he-lical scans, 0.8 mm slice thickness, with 0.4 mm overlap). Figure4 shows CT images of the three anchors. These CT scans wereloaded into proprietary software that was used to automaticallydetect the position of the anchors.

Path PlanningNext, an appropriate path from the lateral skull base to the

cochlea was planned either “manually” or “automatically.” Tomanually plan, the surgeon picked the target (the basal turn of

4 mm

4 mm

1.8 mm

4 mm

Fig. 1. Anchor. Left sample is shown from top. Right sample isshown from side. Anchor is titanium with self-tapping screw. Hole atupper end has female threads to which a bolt can be attached. Boltconnects drill guide rigidly to a set of these anchors. Smaller mark-ings on ruler are millimeters.

(c)

(b)

(a) Fig. 2. Anchor positioning. After anchor inser-tion and skin is closed. Locations: (a) mastoid,(b) suprahelix, (c) posterior.


2

the cochlea) as well as the midportion of the facial recess. Thistrajectory was then traced back to the surface of the lateral skull.Confirmation that this path did not violate any vital structureswas performed using a number of views including a “path offlight” view where the CT scans were aligned on the trajectory,allowing for subjective determination of distance to the facialnerve, chorda tympani, and external auditory canal.

The automated path planning was performed by means of acomputer program written in Matlab (The Mathworks, Inc.,Natick, MA). The program comprises the following three steps. 1)Identify and delimit select anatomy. Select anatomic landmarks(external auditory canal, the short process of the incus, and thebasal turn of the cochlea) were automatically localized in thepatient’s CT using an atlas-based technique.25 Temporal bone CTscans with normal anatomy (as identified and labeled by anexperienced otologist) were defined as atlases and used to createa master atlas. Then, for each patient plan, this atlas was regis-tered to the patient’s CT scan using affine registration, whichconsists of rotation, translation, skewing, and stretching of theatlas so that it fits the unknown scan.26–29 As a result of theregistration, the premapped anatomic landmarks in the atlasimage can be superimposed on the patient image, thus automat-ically identifying and delimiting those structures. It is importantto note that the final location of the basal turn of the cochlea isdependent on the manually identified location selected in theoriginal atlas. This location was defined as the intracochlear,fluid-containing space anterior and inferior to the round windowmembrane. 2) Refine the position of the facial nerve and chordatympani. After registration, specific segmentation of the facialnerve is necessary because atlas-based registration alone wasfound to be an insufficiently accurate predictor of its course. To

segment the facial nerve, its beginning (most inferior 5%) and itsending (most anterior 5%) were identified from the aforemen-tioned atlas based on registration. These two points were thenconnected by a curved line to segment the track of the facial nerveby means of minimization of a cost function where cost is afunction of voxel intensity and shape of the track. This methodresults in excellent identification of the facial nerve. A similarprocedure was performed with the chorda tympani by picking abeginning 5 mm inferior to the atlas-determined branch pointand the ending being the superior medial corner of the annularring. As with the facial nerve, this method results in excellentidentification of the chorda tympani.30 3) Identify boundaries ofextended facial recess. Next, the extended facial recess is identi-fied using the location of the facial nerve, external auditory canal,and short process of the incus. The center of the recess is identi-fied, and a perpendicular line is drawn from the center of thefacial recess to the basal turn of the cochlea. This procedureallows a rough estimation of the plane of the facial recess. 4)Identify all paths that pass through the facial recess plane and hitthe target, the basal turn of the cochlea. Within the aforemen-tioned plane of the facial recess, a dense set of possible linearpaths to the basal turn of the cochlear are enumerated. 5) Cal-culation of safety. For each possible drill path, the path is sub-jected to thousands of random deviations of approximately thelevel of accuracy of the system. This creates a “cloud of uncer-tainty” for each possible drill path. The fraction of times the“cloud” hits any vital structure (facial nerve, external auditorycanal, chorda tympani, and short process of the incus) are calcu-lated. This fraction represents the probability, given the accuracyof the system, that a given drill path will violate one of thesestructures. Minimum acceptable safety probabilities were setsuch that there would be 99.9% certainty of avoiding the facialnerve, 95% certainty of avoiding the external auditory canal, 80%certainty of avoiding the incus, and 67% of avoiding the chordatympani. The path that achieves the highest probability for avoid-ance of each structure yet satisfies these minimum probabilitieswas chosen as the best path. An example of an automaticallychosen path is displayed in Figure 5. The entire path planningprogram takes approximately 35 minutes on a 3 GHz PC runningWindows XP.

Rapid Prototyping of Drill GuidesFor each patient, a drill guide was created that 1) mounted

to the anchors and 2) constrained the passage of a surgical drill toeither the “manual” or “automatic” path. Proprietary softwarewas used to create a digital file defining the drill guide thatsatisfies 1 and 2. This digital file was then sent to a manufactur-ing facility where the drill guide was built using rapid prototypetechnology (FHC, Inc., Bowdoin, ME). Manufactured drill guideswere sent via overnight shipping and were sterilized prior to use.Figure 6 shows both the graphic depiction of a drill guide as seenduring planning and the resulting guides as fabricated from theplan. Two examples are shown, one for a left ear and a second for

(b)(c)

(a)

Fig. 3. Anchor insertion. (a) Nasal speculum maintains exposure. (b)Waypoint driver passes through speculum to attach to anchor (notvisible). (c) Osteomed electric driver (model 450-0600) inserted intoWaypoint driver provides torque.

Fig. 4. Computed tomography images of threeanchors surrounding right ear. Positions: (a)mastoid, (b) suprahelix, (c) posterior.


3

a right ear. Figure 7 shows the physical arrangement of a drillguide and anchors. During surgery, a bolt is passed through eachfoot of the drill guide into the corresponding anchor.

Intraoperative ValidationStandard cochlear implant surgery was performed. The

mastoid and suprahelical anchors were exposed during initialpostauricular incision. After completion of mastoidectomy andextended facial recess, the cochlear promontory was visualized.Next, the posterior anchor was exposed with a stab incision. Todetermine whether the proposed drill path would safely traversethe facial recess and hit the basal turn of the cochlea, the cus-tomized drill guide was attached to the anchors, as shown inFigure 8, and sham drill bits (devoid of cutting flutes) of 1, 2, and3 mm were passed via the drill guide to the cochlear promontory.The largest size bit that safely traversed the facial recess wasnoted. The 1 mm sham drill bit was coated with ink from anoperative marking pen and used to mark where on the promon-tory the cochleostomy would have occurred. Photograph docu-mentation was performed using a 0-degree Hopkins rod that

traversed parallel to the drill path. Imaging software was used tomeasure the distance from the drill to vital anatomy.

RESULTSFour patients agreed to partake in the study, with

one undergoing bilateral testing, for a total of five ears.Anchor placement was tolerated by all patients. The firstpatient took two hydrocodone 7.5 mg/acetaminophen 500mg pain tablets during 24 hours postprocedure. The sec-ond patient rated the pain as 1.5 on a scale of 10 and tooktwo pain tablets that evening. The third patient, who under-went bilateral testing, rated pain for the first five anchors as2 of 10 but, for the last anchor, had intense pain of 9.5 of 10.Notably, this patient did not verbalize pain during the pro-cedure. This patient took four pain tablets in the 48 hourspostprocedure. The last patient reported pain of 1 of 10during the procedure and 2 of 10 postprocedure because ofjaw discomfort. This patient took a total of three pain tablets.

Fig. 5. Example of automatically planned drillpath for right ear. (a) View from posterior toanterior. (b) View through facial recess. (c)Same view as panel a but with drill path (1 mmdiameter) shown. (d) Same view as panel b butwith drill path shown.

Fig. 6. Drill guide as designed and as fabri-cated device. (a) Left ear design. (b) Left eardevice. (c) Right ear design. (d) Right eardevice. Arrows point to pairs of parallel linesthat represent planned drill paths. Distancebetween lines of a pair are equal to planneddrill width (1 mm).


4

All patients reported transient tenderness and pain withmastication. Fourteen of 15 anchor sites healed nicely. Thelone difficulty was delayed healing without infection. Ofnote, this one site was the same site where the patientexperienced discomfort and was the only site where bipolarcautery was necessary to control bleeding.

Results of intraoperative validation testing areshown in Figure 9 and Table I. For each ear, the results forboth manual and automatic path planning are summa-rized in Table I, including 1) the largest sham drill bit thattraversed the facial nerve and 2) the distance from themidline of the trajectory to the facial nerve and 3) thedistance from the midline of the trajectory to the chordatympani.

Note that no “automatic” path is included for ear 1.This first patient had the posterior anchor extrude while achange was made from the manual to automatic path drillguide. In averaging the remaining four ears, distance tofacial nerve (�SD) was 1.24 (�0.44) mm for the auto-mated path planning and 1.18 (�0.68) mm for the manualpath planning. The chorda tympani was sacrificed in one“automatic” path. For the remaining “automatic” paths,

the average distance (�SD) from the midpoint of the drillpath to the chorda tympani was calculated to be 1.22(�0.62) mm. The chorda tympani was preserved in all“manual” paths, for which the average distance was cal-culated to be 0.986 (�0.48) mm.

Shown in Figure 10 are the positions of the cochleos-tomy for ears 2 to 5. Ear 1 is not included because thetechnique of photograph documentation was not in placefor this first case. The raw data—the purple dot where thesham drill bit hit the cochlear promontory—are high-lighted by a yellow circle. Also highlighted, by transparentgray annuli, are the round window niches. Inset withineach frame are the corresponding axial and coronal CTscan slices that show the bony overhang of the roundwindow. These insets are included because they describethe relationship between the relevant surface and subsur-face anatomy.

DISCUSSIONPresented herein is clinical validation of the concept

of percutaneous cochlear implantation. This concept,which has undergone extensive preclinical testing,23,24

consists of using image-guided surgical technology to cre-ate a customized drill guide, the path of which traversesthe facial recess (without injury to the facial nerve) andhits the cochlea anterior to the round window in the posi-tion loco typico for a cochleostomy. This technique hasbeen translated from neurosurgery where such devices areFood and Drug Administration approved for placement ofdeep brain stimulators in the subthalamic nucleus.

Although this is but a preliminary report, we areencouraged by these initial findings, especially 1) patienttolerance of the procedure and 2) accuracy of the device inbypassing the facial nerve and hitting the basal turn of thecochlea. In regard to point 1, the most arduous part of theproject for participants, placement of bone-implanted an-chors within the clinic using local anesthesia, was toler-ated quite well, with 14 of 15 anchors occurring with nomore than slight discomfort. The single anchor thatcaused significant discomfort appeared to be the result ofa technical error in the placement of the local anesthesia,and the patient voiced no complaints until after the pro-cedure was complete.

In regard to point 2, all trajectories bypassed thefacial nerve and hit the cochlea at the intended targetlocation. This result was not unexpected because a priorstudy established the accuracy of the microstereotacticframe as 0.44 mm at a depth of 125 mm from the surface,the depth typically used for deep brain implants.31 Thecochlea is more superficial than neurosurgical targets,and for cochlear implants, a depth of 75 mm is used.Assuming an approximately linear distribution of errorfrom the surface of the frame, we propose that the error forcochlear implant applications is 0.26 mm (0.44*75/125).This accuracy, afforded by the bone-implanted fiducialmarkers, which also serve as anchors for the frame, is wellbelow the 0.5 mm projected by Schipper et al.32 as neces-sary to achieve reliable access to the scala tympani.

The location of the cochleostomy proved difficult toassess. Our original concept of putting dye on the tip of asham drill bit produced the results shown in Figure 10.

(a)

(b)(c)

Fig. 7. Drill guide. Bolt (a) passes through hole in each foot (b),attaching foot to an anchor (c), as shown in inset at upper left.

(a)(b)

Fig. 8. Customized drill guide after attachment to three anchors inoperating room. (a) Entry for drill bit. (b) Entry for endoscope (Hopkinsrod). Images in Figures 9 and 10 were acquired with Hopkins rod.


5

Although these show cochleostomy locations acceptablefor access to the cochlea, it is difficult to interpret wherethe cochlea would be entered for two reasons. First, theendoscopic pictures are slightly skewed because they are atwo-dimensional representation of three-dimensional ob-jects and therefore dependent on angle. For example, theimage in the top left of Figure 10 is taken from below theproposed trajectory looking up to the cochlear promontory.From this perspective, the round window niche (which canbe seen in its entirety) appears much lower than theproposed cochleostomy. In learning from this experience,

subsequent images were taken in a much more uniformmanner, with photograph documentation along the axisof the trajectory producing the pictures in the remain-ing three panels. Second, the amount of bony overhangfrom the round window niche obscures the relationshipbetween the surface anatomy (the round window niche)and the subsurface anatomy (scalas vestibule, media,and tympani). Therefore, we have included the axialand coronal CT views of the round window to allowreaders to form their own interpretations about theproposed cochleostomy data. Recognizing the need to doc-

TABLE I.Results of Intraoperative Validation Testing

Ear Patient Planning MethodLargest Drill Bit

to Pass Recess (mm)Distance to Facial

Nerve (mm)Distance to Chorda

Tympani (mm)

1 1 Manual 3 2.23 1.13

1 1 Automatic NA NA NA

2 2 Manual 1 1.09 0.87

2 2 Automatic 2 1.22 1.57

3 3 Manual 1 0.69 1.73

3 3 Automatic 2 1 1.59

4 3 Manual 1 0.5 0.5

4 3 Automatic 1 0.87 Hit

5 4 Manual 2 1.4 0.7

5 4 Automatic 1 1.85 0.5

Ear 1 (Patient 1), Manual, 3 mm




Ear 5(Patient 4), Manual, 1mm

Ear 2 (Patient 2), Auto, 1 mm




Fig. 9. Intraoperative validation. Each imageshows sham drill bit (diameter noted undereach picture) that safely traversed facial recess.“Manual” means that a manual plan was used.“Automatic” means that plan based on com-puter algorithm was used. No automatic valida-tion was possible for ear 1.


6

ument precisely where in the cochlea the proposed trajectorylands, we have modified our future experimental protocol toinclude drilling away the round window overhang beforedocumentation of the proposed cochleostomy site.

The automated path planning worked as well if notbetter than the manual path planning. We attribute thisto the computer’s ability to use three-dimensional rela-tionships better than a surgeon, whose path planning isachieved with a collection of two-dimensional views of athree-dimensional structure. Although we have limiteddata, the automated path appeared to have more respectfor the facial nerve than the manual path planning per-formed by the surgeon. This may reflect the human ten-dency to desire a perfect solution (bypassing BOTH thefacial nerve and the chorda tympani), and, in cases inwhich the facial recess is extremely narrow, this may notbe technically achievable or advisable. Ear 4 provides justsuch an example for which the manual path successfullytraversed a very narrow facial recess (approximately 1mm), whereas the automated path resulted in sacrifice ofthe chorda tympani in keeping an acceptable margin ofsafety in avoidance of the facial nerve. Intraoperatively,the chorda tympani had to be sacrificed to allow visual-ization and instrumentation within the middle ear. Thisincidence of chorda tympani sacrifice is in line with theliterature, which reports an incidence of approximately20%.33

Perhaps a more fundamental question is, “Why pur-sue percutaneous cochlear implantation at all given anexisting surgical approach (mastoidectomy with facial re-cess) that has an excellent track record?” We argue thatpercutaneous cochlear implantation holds the prospect ofbenefit in several areas, including 1) standardization ofthe surgical approach, 2) decreased operative time andoverall cost, and 3) the possibility of unique implant op-portunities. Each of these will be discussed below.

Standardization of Surgical ApproachIn developing countries, where otolaryngology–head

and neck surgery training is highly variable, cochlearimplant surgery outcomes and complication rates are like-wise variable. In such settings, standardization of the sur-gical approach with automated path planning would offeradvantages, including reduced complications and improvedoutcomes. The technique proposed herein requires accessto a CT scanner, with the rest of the technology beingaccessed from afar via transfer of electronic files. Forapplication in developing countries, it is envisioned that aportable CT scanner (approximately $250,000 at thetime of this writing) would be centrally located in afacility where cochlear implant programming would beperformed.

Even in countries where otolaryngological surgicaltraining is highly regimented (e.g., United States, Europe)and complications rates are low (injury to the facial nerveestimated at 1.5%),34–36 disagreement about such funda-mental questions such as proper position of the cochleos-tomy persists.37–39 Automated path planning holds out theprospect of optimizing electrode placement, potentiallyimproving outcome by decreasing power consumption.

Decreased Operative Time and Overall CostSurgical times for cochlear implantation are fre-

quently discussed but rarely published. Work from adecade ago estimated operative time at 3 hours forinexperienced (i.e., resident) surgeons.40 Estimates forexperienced surgeons range from 50 to 75 minutes, withoutliers occurring routinely. Retrospectively reviewingdata from two experienced institutions with multiple sur-geons showed the following results: Institute A: n � 457,average surgical time 209 min, minimum 40 min, maxi-mum 310 min; Institute B: n � 108: average surgical time147 min, minimum 84 min, maximum 280 min (exclusion

Fig. 10. Positions of cochleostomy for eachpath are shown as purple spots on the lateralwall of each cochlea. Yellow circles have beenadded to highlight this position while transpar-ent gray annuli overlie round window niches.Inset within each panel are corresponding axialand coronal scans showing bony overhang ofround window. Note, no data presented for firstear because technique was developed afterfirst surgery.


7

criteria: bilateral implantation, revision cases, anomalies,and patients involved in intraoperative research studies;note that reported operative times include intraoperativeaudiologic testing, which can take up to 30 min for eachpatient). With use of this information and the extremelylimited data regarding time for implantation, a very con-servative estimate is that an experienced surgeon cansafely perform a routine cochlear implantation in 70 min-utes, and an inexperienced surgeon can do such in 150minutes. Conservatively estimating the individual compo-nents of a percutaneous cochlear implant, we predict thatpercutaneous cochlear implantation would take approxi-mately 50 minutes (exposure of anchors 10 min, attach-ment of drill guide 5 min, drilling from lateral cortex tomiddle ear 5 min, creation of pocket for internal receiver10 min, cochleostomy and insertion of electrode 5 min,removal of drill guide 2.5 min, removal of anchors 2.5 min,closure 10 min). Although this time differential is notdramatic for an experienced surgeon, it could have a dra-matic impact on surgical times for inexperiencedsurgeons.

In regard to actual surgical costs, as with surgicaltimes, limited data are available. A report from Germanyestimates cost to vary between $46,000 and $57,000 de-pending on the discount on device.41 In the United King-dom, costs are approximately $44,000.42 Direct sugicalcosts in the United States are a minimum of $24,475.43 Incomparing cost of traditional cochlear implant surgerywith percutaneous cochlear implant surgery, we proposethat a more appropriate comparison is to equate operativetime with operative cost. At the home institution of thelead author, the first 1/2 hour facility fee is $1,581, andsubsequent 1/2 hour fees are $1,265. In also taking intoaccount the additional cost of the drill guide (approxi-mately $1,200), the facility fee cost for experienced versusinexperienced surgeons is estimated below in Table II. Aswith time, although the cost savings for experienced sur-geons may be minimal, the cost savings for inexperiencedsurgeons potentially could be substantial.

Unique Implant OpportunitiesThe cost estimate noted above takes into account only

the time for procedural intervention. Significant cost sav-ings could also be realized if the intervention could bemoved from the operating room to a procedure room underlocal anesthesia. Local anesthesia has been used in selectcases for cochlear implant surgery44,45 and is the preferredmode of anesthesia for placement of deep brain stimula-tors using techniques similar to those described in this

paper. Given that the procedure could be performed underlocal anesthesia, the concept of same-day service for co-chlear implantation may be achievable. This would in-clude audiologic testing, CT acquisition, automated pathplanning, custom drill guide creation, implantation, andactivation. Such a radical paradigm shift may allow per-cutaneous cochlear implantation to become analogous toLASIX surgery.

The preceding four paragraphs provide a glimpse atthe possibilities that percutaneous cochlear implant sur-gery may allow. At present, we have demonstrated thefeasibility of the technique in patients undergoing tradi-tional cochlear implantation. This continuing study,funded by the National Institute on Deafness and OtherCommunication Disorders, will now start a multisite com-ponent during which the findings contained within thisinitial report will be tested by other surgeons. Pendingconfirmation of the success of the technique and afteraddressing other technical issues (e.g., electrode insertiontools and drill strategies currently under development inour temporal bone laboratory), the method will be used toperform cochlear implantation in the near future.

CONCLUSIONSContained herein are the initial validation studies

documenting translation of the concept of “percutaneouscochlear access” from the laboratory to the operatingroom. To accomplish this, under IRB-approved protocols,the following procedures were performed: 1) patients hadthree bone-implanted markers placed surrounding themastoid; 2) patients underwent clinically applicable CTscanning; 3) with use of proprietary computer softwareand each patient’s CT scan, trajectories from the lateralcortex of the mastoid, through the facial recess, and to thebasal turn of the cochlear were planned; 4) customized drillguides to achieve these trajectories were rapid prototyped; 5)intraoperatively, after traditional cochlear implant surgery,drill guides were mounted on the markers, and accuracy ofthe proposed trajectory was assessed. All planned trajecto-ries were accurate, with five of five traversing the facialrecess without injury to the facial nerve. Four of five trajec-tories also bypassed the chorda tympani, with the lonechorda tympani injury planned a priori on the pre-intervention CT scan. Although additional work is necessarybefore full clinical implementation, these results support thefeasibility of this novel surgical approach.

AcknowledgmentsWe are appreciative of the efforts of Wendy Lips-

comb, RN, our research nurse, and Jason Mitchell, MS,our Machinist.

BIBLIOGRAPHY1. Attwood SE, Hill AD, Mealy K. A prospective comparison of

laparoscopic versus open cholecystectomy. Ann R Coll SurgEngl 1992;74:397–400.

2. Soper NJ, Barteau JA, Clayman RV, Ashley SW. Comparisonof early postoperative results for laparoscopic versus stan-dard open cholecystectomy. Surg Gynecol Obstet 1992;174:114–118.

3. Vierra M. Minimally invasive surgery. Annu Rev Med 1995;46:147–158.

TABLE II.Time and Cost Estimates for Cochlear Implant (CI) Surgery

Traditional CISurgery

Percutaneous CISurgery

TIME of experienced surgeon 70 minutes 50 minutes

COST of experienced surgeon $4,111 $4,046

TIME of inexperienced surgeon 150 minutes 50 minutes

COST of inexperiencedsurgeon

$6,641 $4,046


8

4. Messerklinger W. Endoscopy of the Nose. Baltimore: Urban &Schwarzenberg, 1978.

5. Draf W. Therapeutic endoscopy of the paranasal sinuses.Endoscopy 1978;10:247–254.

6. Kennedy DW, Zinreich SJ, Rosenbaum AE. Functional endo-scopic sinus surgery. Theory and diagnostic evaluation.Arch Otolaryngol 1985;111:576–582.

7. Choe JH, Kim SW, Chung KW, et al. Endoscopic thyroidec-tomy using a new bilateral axillo-breast approach. WorldJ Surg 2007;31:601–606.

8. Terris DJ, Gourin CG. Minimally invasive thyroidectomy:basic and advanced techniques. Laryngoscope 2006;116:350–356.

9. Iro H, Waldfahrer F, tendorf-Hofmann A, Weidenbecher M,Sauer R. Transoral laser surgery of supraglottic cancer:follow-up of 141 patients. Arch Otolaryngol Head NeckSurg 1998;124:1245–1250.

10. Strong MS. Laser surgery in the larynx. Early clinical expe-rience with continuous CO2 laser. Ann Otol Rhinol Laryn-gol 1972;81:791–798.

11. Vaughan CW. Transoral laryngeal surgery using the CO2laser: laboratory experiments and clinical experience.Laryngoscope 1978;88:1399–1420.

12. Watanabe E, Watanabe T, Manaka S, Mayanagi Y. Three-dimensional digitizer (neuronavigator): new equipment forcomputed tomography-guided stereotaxic surgery. SurgNeurol 1987;27:543–547.

13. Reinhardt H, Meyer H. A computer-assisted device for theintraoperative CT-correlated localization of brain tumors.Eur Surg Res 1988;20:51–58.

14. Roberts DW, Strohbehn JW, Hatch JF, Murray W. A framelessstereotaxic integration of computerized tomographic imagingand the operating microscope. J Neurosurg 1986;65:545–549.

15. Mosges R. A new imaging method for intraoperative therapycontrol in skull-base surgery. Neurosurg Rev 11:245–247,1988.

16. Schlondorff G, Mosges R, Meyer-Ebrecht D, Krybus W. CAS(computer assisted surgery). A new procedure in head andneck surgery. HNO 1989;37:187–190.

17. Banfai P, Kubik S. Our extra-scalar operating method ofcochlear implantation. Experience with 46 cases. Acta Oto-laryngol Suppl 1984;411:9–12.

18. Chouard CH. Implantation of multiple intracochlear elec-trodes for rehabilitation of total deafness: preliminary re-port. Laryngoscope 1976;86:1743–1751.

19. Roberson JB, Stidham KR, Scott KM. Cochlear implantation:minimal hair removal technique. Otolaryngol Head NeckSurg 2000;122:625–629.

20. O’Donoghue GM. Minimal access surgery for pediatric co-chlear implantation. Otol Neurotol 2002;23:891–894.

21. Kiratzidis T. ‘Veria operation’: cochlear implantation without amastoidectomy and a posterior tympanotomy. A new surgicaltechnique. Adv Otorhinolaryngol 2000;57:127–130.

22. Kronenberg J, Migirov L. Suprameatal approach: new surgi-cal approach for cochlear implantation. J Laryngol Otol2001;115:283–285.

23. Labadie RF, Chodhury P, Cetinkaya E, et al. Minimally in-vasive, image-guided, facial-recess approach to the middleear: demonstration of the concept of percutaneous cochlearaccess in vitro. Otol Neurotol 2005;26:557–562.

24. Warren F, Balachandran R, Fitzpatrick J, Labadie RF. Percuta-neous cochlear access using bone-mounted, customized drillguides: demonstration of concept in vitro. Otol Neurotol 2007;28:325–329.

25. Al-Marzouqi H, Noble J, Warren F, et al. Planning a safedrilling path for cochlear implantation surgery using im-age registration techniques. From the Proceedings of theSPIE Conference on Medical Implants, February 2007,volume 8, no. 28.

26. Maes F, Collignon A, Vandermeulen D, Marchal G, SuetensP. Multimodality image registration by maximization ofmutual information. IEEE Trans Med Imaging 1997;16:187–198.

27. Wells WM III, Viola P, Atsumi H, Nakajima S, Kikinis R.Multi-modal volume registration by maximization of mu-tual information. Med Image Anal 1996;1:35–51.

28. Collignon A, Maes F, Delaere D, Vandermeulen D, Suetens P,Marchal G. Automated multimodality medical image reg-istration using information theory. In: Bizais Y, Barillot C,Di Paola R, eds. Proceedings of the 14th International Con-ference of Information Processing in Medical Imaging(IPMI’95), vol. 3, Computational imaging and vision. Ile deBerder, France: 1995:263–274.

29. Gustavo KR, Aldroubi A, Dawant BM. The adaptive basesalgorithm for intensity-based nonrigid image registration.IEEE Trans Med Imaging 2003;22:1470–1479.

30. Noble JH, Warren FH, Labadie RF, Dawant BA. Automaticsegmentation of the facial nerve and chorda tympani usingimage registration and statistical priors. Proceedings ofSPIE, Medical Imaging 2008: Image Processing, San Diego,CA, 17–21 February 2008. 2008;6914: 69140P.

31. Balachandran R, Mitchell J, Dawant B, Fitzpatrick JM. Eval-uation of targeting frames for deep-brain stimulation usingvirtual targets. From the Proceedings of the IEEE Inter-national Symposium on Biomedical Imaging: From Nanoto Macro, April 12–15, 2007; Washington, DC.

32. Schipper J, Klenzner T, Aschendorff A, Arapakis I, RidderGJ. Navigation-controlled cochleostomy. Is an improve-ment in the quality of results for cochlear implant surgerypossible? HNO 2004;52:329–335.

33. Bhatia K, Gibbin KP, Nikolopoulos TP. Surgical complica-tions and their management in a series of 300 consecutivepediatric cochlear implantations. Otol Neurotol 2004;25:730–739.

34. Fayad JN, Wanna GB, Micheletto JN. Facial nerve paralysisfollowing cochlear implant surgery. Laryngoscope 2003;113:1344–1346.

35. Cohen NL. Complications of cochlear implant surgery inadults and children. Ann Otol Rhinol Laryngol 1991;100:708–711.

36. Hoffman RA. Complications of cochlear implant surgery. AnnOtol Rhinol Laryngol Suppl 1995;166:420–422.

37. Adunka OF, Radeloff A, Gstoettner WK, Pillsbury HC,Buchman CA. Scala tympani cochleostomy II: topographyand histology. Laryngoscope 2007;2195–2200.

38. Adunka OF, Buchman CA. Scala tympani cochleostomy I:results of a survey. Laryngoscope 2007;2187–2194.

39. Roland PS, Wright GG, Issacson B. Cochlear implant elec-trode insertion: the round window revisited. Laryngoscope2007;117:1397–1402.

40. Labadie RF, Carrasco VN, Gilmer CH, Pillsbury HC. Co-chlear implant performance in older adults. OtolaryngolHead Neck Surg 2000:123:419–424.

41. Schulze-Gattermann H, Illg A, Schoenermark M, Lenarz SA.Cost-benefit analysis of pediatric cochlear implantation:German experience. Otol Neurotol 2002;23:674–681.

42. O’Neill C, O’Donoghue GM, Archbold SM, Normand C. Acost-utility analysis of pediatric cochlear implantation. La-ryngoscope 110:156–160.

43. Cheng AK, Rubin HR, Powe NR, Mellon NK, Francis HW.Cost-utility analysis of the cochlear implant in children.JAMA 2000;284:850–856.

44. Toner JG, John G. Cochlear implantation under local anes-thesia, the Belfast experience. J Laryngol Otol 1998;112:533–536.

45. Djalilian HR, Roy S, Benson AG, Regala C, McDonald TB.Transcanal cochlear implantation under monitored anes-thesia care. Otol Neurotol 2005;26:674–677.


9

clinical validation of percutaneous cochlear implant surgery: initial

Documents