Otology & Neurotology 36:1015Y1022 Ó 2015, Otology & Neurotology, Inc.
Manual Electrode Array Insertion Through a Robot-Assisted Minimal Invasive Cochleostomy: Feasibility and Comparison of Two Different Electrode Array Subtypes *†Frederic Venail, ‡Brett Bell, *Mohamed Akkari, ‡§Wilhelm Wimmer, ‡Tom Williamson, ‡Nicolas Gerber, ‡Kate Gavaghan, kFrancois Canovas, ‡Stefan Weber, ‡§Marco Caversaccio, and *†Alain Uziel *Otology and Neurotology Department, University Hospital of Montpellier, Montpellier, France; ÞInstitute for Neurosciences of Montpellier, INSERM U1051, Montpellier, France; þARTORG Center for Biomedical Engineering Research, University of Bern, Bern, Switzerland; §Department of ENT, Head and Neck Surgery, Inselspital, University of Bern, Bern, Switzerland; and kAnatomy Laboratory, Faculty of Medicine, University Hospital of Montpellier, Montpellier, France
Hypothesis: To evaluate the feasibility and the results of insertion of two types of electrode arrays in a robotically assisted surgical approach. Background: Recent publications demonstrated that robotassisted surgery allows the implantation of free-fitting electrode arrays through a cochleostomy drilled via a narrow bony tunnel (DCA). We investigated if electrode arrays from different manufacturers could be used with this approach. Methods: Cone-beam CT imaging was performed on fivecadaveric heads after placement of fiducial screws. Relevant anatomical structures were segmented and the DCA trajectory, including the position of the cochleostomy, was defined to target the center of the scala tympani while reducing the risk of lesions to the facial nerve. Med-El Flex 28 and Cochlear CI422 electrodes were implanted on both sides, and their position was verified by cone-beam CT. Finally, temporal bones were dissected
to assess the occurrence of damage to anatomical structures during DCA drilling. Results: The cochleostomy site was directed in the scala tympani in 9 of 10 cases. The insertion of electrode arrays was successful in 19 of 20 attempts. No facial nerve damage was observed. The average difference between the planned and the postoperative trajectory was 0.17 T 0.19 mm at the level of the facial nerve. The average depth of insertion was 305.5 T 55.2 and 243 T 32.1 degrees with Med-El and Cochlear arrays, respectively. Conclusions: Robot-assisted surgery is a reliable tool to allow cochlear implantation through a cochleostomy. Technical solutions must be developed to improve the electrode array insertion using this approach. Key Words: Cochlear implantationVCochleostomyVCone-beam CTVImage-guided surgeryVRobot-assisted surgery. Otol Neurotol 36:1015Y1022, 2015.
Advances in cochlear implant techniques emphasize the development of minimally invasive cochlear implantation methods. This concept implies the reduction of the surgical trauma not only for the electrode array insertion but also during the surgical approach through the facial recess. A ‘‘soft’’ electrode array insertion is required not only
for residual hearing preservation but also to maintain normal cochlear anatomy and avoid postoperative intracochlear fibrosis (1,2). Although some surgeons favor the round window for direct access into the scala tympani, some argue that insertion through the round window may cause cochlear damage. Conversely, an anteroinferior cochleostomy avoids these problems and facilitates an atraumatic insertion (3). However, anatomical variations can jeopardize the correct positioning of the cochleostomy and may lead to a scala vestibuli insertion (4). Recent work on image-guided robotic-assisted surgery has shown that a direct cochlear access (DCA) through a single bony tunnel can be drilled efficiently and safely in
Address correspondence and reprint requests to Frederic Venail, M.D., Ph.D., ENT Department,University Hospital Gui de Chauliac, 80 avenue Augustin Fliche, Montpellier, France; E-mail: [email protected] This work was financially supported by the Nano-Tera.ch initiative, FP7 HearEU, and the Swiss National Science Foundation. The authors report no conflicts of interest. Supplemental digital content is available in the text.
1015
Copyright © 2015 Otology & Neurotology, Inc. Unauthorized reproduction of this article is prohibited.
1016
F. VENAIL ET AL.
temporal bones using high-accuracy optical tracking camera while preserving sensitive structures like the facial nerve, ossicles, and the external ear canal (5). Recently, it was demonstrated that the drilling of a cochleostomy through a DCA tunnel is feasible and that insertion of free-fitting electrode arrays (Med-El 28 mm arrays) could be achieved without any intervention of the surgeon through the ear canal (6). A previous report described the successful use of different electrode arrays to perform image-guided cochlear implantation in a cadaver model using a stereotactic frame (7). In the present work, emphasis is placed on software modifications which allow the surgeon to plan the ideal insertion trajectory for drilling DCA tunnel and cochleostomies while evaluating the results achieved with freefitting electrode arrays from different manufacturers. MATERIALS AND METHODS
holding device which uses pressurized pads to stabilize the head. Next, a tracking reference was rigidly attached to the skull with a bone screw (2 5 mm). In a final step, the robot system was mounted to the OR table. The final plan was registered to the temporal bone using an optical tracking device to match the locations of the implanted fiducial screws. The robot was then moved, utilizing haptic control, into the field of view of the optical tracking camera before approaching the surface of the cortical mastoid bone. The DCA was then drilled by the robot using a custom ‘‘step’’drill having a proximal diameter of 2.5 mm with a length of 20 mm, and distal portion with a diameter of 1.8 mm and a length of 10 mm to the tip. The drill motor was started (5000 rpm) and the robot drilled with a feed rate of 0.5 mm/s using a ‘‘pecking’’motion until the middle ear cavity was reached. A cochleostomy was then drilled by the robot system using a 1-mm diamond burr. The drill speed was increased to 10,000 rpm, and the feed rate was reduced to 0.1 mm/s (Fig. 1). All robot movement was completed under the supervision of the surgeon using a three-stage enable switch, which allowed progression of the drilling or movement of the manipulator only when desired.
Specimen Preparation and Imaging The complete protocol (i.e., intervention planning, drilling, and array insertion) was performed in the experimental surgery laboratory of the University Hospital of Montpellier, France. Five human cadaver heads (n = 10 temporal bones) fixed with a 20% zinc-chloride intra-arterial injection (according to Sucquet [8]) were used in this study. Before the experiment, a retroauricular musculocutaneous flap was elevated to expose the mastoid cortical bone. Four titanium screws (M-5220.03; Medartis, Basel, Switzerland) were placed to be used as fiducial markers for patient-to-image registration. All imaging in this study (before and after surgery) was carried out using a cone-beam CT (NewTom 5G, Verona, Italy) and a high-resolution protocol (voxel size: 125 Km isotropic, 110 kVp, 19 mA). After the completion of the drilling and insertion experiments, all temporal bones underwent surgical microdissection to detect any damage to the facial nerve, the chorda tympani, or other middle ear structures during the DCA tunnel drilling.
Direct Cochlear Access and Cochleostomy Drilling A minimally invasive access to the tympanic cavity was drilled with a purpose-built robotic system developed earlier (9) as defined by the preoperative surgical planning (see Supplemental Digital Content for more details, http://links.lww.com/MAO/A294). The distances between the trajectory of the DCA tunnel and the facial nerve, the chorda tympani, the posterior wall of the external auditory canal, and the ossicles were calculated to minimize the risk of basilar membrane collision or tearing of the lateral wall. Deviation from the ideal trajectory towards the modiolus or the lateral wall is defined as the angle D, and deviation towards the basilar membrane or the bottom of the scala tympani is defined as the angle C (see Supplemental Data Figure, http://links.lww.com/MAO/A294). The distance of the cochleostomy from the round window was defined as the angular distance between 0 degrees (center of the round window) and 20 degrees (more anteriorly) by 2-degree angle steps (defined as 5C, see Supplemental Data Figure, http://links.lww.com/MAO/A294). At the beginning of the surgical procedure, the skin of the external auditory canal was removed and a tympanic flap was performed to introduce a 4-mm 30-degree endoscope for visual control of the cochleostomy drilling and the insertion of the electrode array. The DCA tunnel was drilled using the same protocol published previously (9). Briefly, the cadaver head was mounted in a noninvasive head
Insertion of Electrode Arrays and Evaluation of the Quality of Insertion Two types of electrodes arrays were inserted: Flex28 electrodes (28 mm length; Med-El company) and CI422 slim straight arrays (SSA, 20 mm length; Cochlear Corporation). Both electrode arrays are straight, free-fitting arrays. Before the insertion, the DCA tunnel was cleaned using saline solution irrigation and suction of bone dust. Flex28 electrodes arrays were inserted first (Fig. 1). Because these types of arrays are highly flexible and could bend during the mastoid part of the insertion, a custom insertion tool was designed to facilitate their guidance through the DCA tunnel. Insertion was performed with hyaluronic acid gel for lubrication under visual control with the endoscope (Fig. 1). Insertion was stopped at the point of first resistance. After completion of insertion, electrode arrays were fixed using sutures to prohibit movement during subsequent handling phases. Then, highresolution cone-beam CT scans were performed as previously described to evaluate the depth of insertion of the electrode arrays and their scalar positioning. Insertion depth was estimated by the insertion angle > (-) (Fig. 2, A and B) from the round window to the most apical electrode contact. Scalar localization of the electrode array was determined by a neuroradiologist experienced in cochlear implant imaging. Theoretical positions of scala tympani and scala vestibuli were determined based on previous studies (10,11). On midmodiolar reconstructions, scala tympani was located in the inferior part of a half-turn section of the cochlea and scala vestibuli in the superior part. The position of the electrode contact was inferred to be in the scala tympani when the contact was located in the lower half turn of the cochlear section. In the second part of the experiment, Flex28 arrays were gently removed and CI422 slim straight arrays were inserted in the same temporal bones to compare the insertion with both arrays. CI422 slim straight arrays are designed with a lateral fin for the handling of the array, making it too large to enter into the DCA tunnel. Therefore, the fin was removed from the electrode array before insertion. No insertion tool was used during the insertion of CI422 arrays which are larger and more rigid when compared to the Flex28 arrays. The rest of the procedure was performed as described for Flex28 arrays, and cone-beam CT scans were then
Otology & Neurotology, Vol. 36, No. 6, 2015
Copyright © 2015 Otology & Neurotology, Inc. Unauthorized reproduction of this article is prohibited.
MANUAL ELECTRODE ARRAY INSERTION
1017
FIG. 1. Visualization through the ear canal of the cochleostomy’s drilling and electrode array insertion into the cochlea. Step-by-step visualization of the drilling of the DCA with the cochleostomy and of the electrode array insertion of a Med-El Flex28 array in a left temporal bone. The round window (RW, arrow), the promontory (P), the long process of the incus (In), the stapes (St), and the pyramidal process (pyr) can be seen with the 30-degree endoscope. After drilling of the DCA (A), the diamond bur drills a 1-mm hole through the promontory of the cochlea (B). Then liquid and bone dust are removed from the cochleostomy (Co) (C), and the insertion tool (it) is inserted through the DCA tunnel close to the cochleostomy (D). A lubricated electrode array (ea) is pushed through the insertion tool into the cochlea until the first point of resistance (E). Finally, the insertion tool is gently removed from the DCA tunnel (F).
acquired for further evaluation. Additional cone-beam CT scans were performed after cochlear array removal to evaluate the position of the DCA tunnel for accuracy evaluation.
RESULTS Evaluation of Electrode Arrays’ Insertion Insertion of the array was possible in all cases using Flex28 electrode arrays and the insertion tool. Conversely, intracochlear insertion of the electrode array was not possible in one case with CI422 slim straight array. In this case, the electrode array could not be properly guided to the cochleostomy through the DCA tunnel in the absence of the insertion tool. The cochleostomy was successfully directed towards the scala tympani of the basal turn of the cochlea in 9 of 10 cases implanted with Flex28 and in eightof ninecases implanted with CI422 slim straight arrays (Table 1). In the last case, the cochleostomy ended in the scala vestibuli of the basal turn of the cochlea because of a large drilling error (Table 2). Full insertion of the electrode array was not achieved in any of the implanted cases. The average depth of insertion with Flex28 arrays was 305.5 T 55.2 degrees, with an average depth of 243 T 32.1 degrees with CI422 slim straight arrays. This resulted in an average of 2.3 T 0.9 and 3.9 T 1.4 electrodes outside the cochlea with Flex28 and CI422 arrays, respectively (Table 1, Fig. 2).
Accuracy of DCA Tunnel and Cochleostomy Drilling The accuracy at the cochleostomy target was measured at 0.29 T 0.23 mm with a range of 0.05 to 0.79 mm. In two cases, a target error larger than 0.5 mm was observed (Table 2). The target error was orientated anteriorly and posteriorly in specimen 1L and 1R, respectively. This caused penetration of the external auditory canal posterior bony wall without cutaneous damage and a cochleostomy ending in the scala vestibuli of the basal turn of the cochlea in specimen 1L (Fig. 3). A close passage of the facial nerve was observed in specimen 1R. As expected, the chorda tympani was damaged in specimens 2L and 2R (Fig. 4). During the planning, the distance (TSD) between the facial nerve and the DCA tunnel was estimated at 0.38 T 0.03 mm. Postoperatively, the distance between the DCA tunnel and the facial nerve was determined to be 0.39 T 0.26 mm (p = 0.82, Student’s t test for paired data)(see Table 2). The mean absolute difference between the planned and calculated distance to the facial nerve was 0.17 T 0.19 mm. According to postoperative CT image examination, the position of the drilled tunnel seemed to slightly touch the facial nerve in specimen 1R (distance of j0.081 mm, Table 2, Fig.3). However, surgical microdissection of the temporal bonerevealed that the drilled tunnel did not intersect with the facial nerve. Thus, the small negative value of the distance between the DCA tunnel and the Otology & Neurotology, Vol. 36, No. 6, 2015
Copyright © 2015 Otology & Neurotology, Inc. Unauthorized reproduction of this article is prohibited.
1018
F. VENAIL ET AL.
FIG. 2. Postoperative cone-beam CT of inserted Med-EL Flex28 and Cochlear CI422 arrays. Postoperative cone-beam CT of temporal bone 3R. Depth of insertion of the electrode array is evaluated at 259 degrees with CI422 electrode array (A) and at 342 degrees with Med-El Flex28 (B). Depth of insertion is measured as the angle > (-) between the cochleostomy (Co) and the most apical electrode in axial plane (A, B). Cross-sections perpendicular to the blue dashed line show that with CI422 electrode array (C) and Med-El Flex28 (D), the electrode array is entirely located in the scala tympani of the cochlea. Note that full insertion of electrode arrays was not achieved in both cases, leaving three and one electrodes outside the cochlea with CI422 and Med-El Flex28 arrays, respectively.
facial nerve was likely due to oversegmentation of the nerve during the planning stage. A standard mastoidectomy was subsequently performed on all specimens, confirming that the drilling of the DCA did not result in visible damage to the facial nerve (Fig. 4). The 5C angle chosen during the planning varied from 4 to 12 degrees (average of 9 T 3.68 degrees). This angle was compared to the type of cochleostomy observed through the endoscope during drilling (RW: round window, ERW: enlarged round window, CO: transpromontorial cochleostomy). Because of targeting errors, there was no clear correlation between the 5C angle defined on the surgical planning and the type of cochleostomy performed (Table 2).
DISCUSSION Electrode array insertion is a crucial step in cochlear implant surgery. This insertion can be performed through the round window membrane or through a transpromontorial cochleostomy. In the latter case, the surgeon must determine the position of the scala tympani of the basal turn of
the cochlea relying on the only constant landmark which is the round window itself. However, interindividual variation in round window and cochlear hook anatomy (12) can mislead the surgeon and cause to drill a cochleostomy ending in the scala vestibuli in approximately 20% of cases (4). Therefore, image-guided surgery was proposed to help in the positioning of the cochleostomy. Stelter et al. (13) showed that real-time image-guided navigation, based on conventional CT images (Brainlab), had an average deviation of 1.01 mm at the level of the round window. As shown in our study, a tool positioning errors below 1 mm at the level of the promontory could direct the cochleostomy to the scala vestibuli. Even if the accuracy of standard navigation techniques may be improved by using highresolution cone-beam CT images (14), it is likely that the accuracy will not be enough to be clinically implemented. Thus, improvements in cochleostomy positioning require additional technological solutions, including robotassisted approaches. Centerline computation (15) or active shape modeling (5,7,16) can be used to determine the position of the center of the scala tympani preoperatively
Otology & Neurotology, Vol. 36, No. 6, 2015
Copyright © 2015 Otology & Neurotology, Inc. Unauthorized reproduction of this article is prohibited.
MANUAL ELECTRODE ARRAY INSERTION TABLE 1. Comparison of depth of electrode insertion, number of electrodes outside the cochlea, and scalar position with Med-El Flex28 and CI422 electrode arrays No. Electrodes Outside the Cochlea
Angle of Insertion > (-) Temporal Bone Flex28 1L
CI422
264
Flex28 CI422 Flex28 3
317
2L
202
2R
295
3L
412
3R
342
4L
277
4R
323
5L
291
5R
332
Avg. TSD
305.5 55.2
5
SV/SV
2
ST/ST
NA
NA
NA
3
ST/ST
247
3
ST/ST
3
ST/ST
209
5
ST/ST
1
ST/ST
222
4
ST/ST
1
ST/ST
259
3
ST/ST
4
ST/ST
276
4
ST/ST
2
ST/ST
269
5
ST/ST
2
ST/ST
289
1
ST/ST
2 211 243.0 32.1
CI422
SV/SV
205 1R
Scalar Position of the Apical/ Basal Electrode
ST/ST
2.3 0.9
5 3.9 1.4
ST/ST
NA, non-applicableVthe electrode array could not be inserted; ST, scala tympani; SV, scala vestibuli.
from high-resolution cone-beam CT images. Both strategies have been utilized in temporal bone specimens (5,7,15) or in vivo (16) to determine suitable insertion sites. In these studies, electrode array insertion was performed either by the round window (5) or by a manually drilled cochleostomy (7,16), requiring an additional access to the middle ear through a tympanomeatal flap in both cases. Neither of these works performed the complete procedure, including TABLE 2.
cochleostomy and electrode insertion, through the same tunnel without the help of an additional middle ear approach as in the present study. In this study, a scala tympani cochleostomy was performed in 9 of 10 cases. In the last case, the cochleostomy ended in the scala vestibuli, but did not preclude cochlear array insertion. A drilling error of 0.79 mm at the level of the target (promontory) was responsible for this misalignment. The large deviation was later determined to be the caused by broken fiducial screws that were left in the specimen, causing smallest deviations in the computation of the fiducial’s center position. Such self-drilling screws were originally utilized to ease the preoperative workflow. However, their tendency to break in hard, dense cortical bone required us to identify different screws for future use with additional image-based checks implemented to ensure the safety of the robotic-assisted procedure. No damage to the facial nerve was observed resulting from the drilling of the DCA trajectory even in case of narrow facial recess (temporal bone 1L and 1R). The absolute difference calculated between the preoperative trajectory and the postoperative real trajectory was only 0.17 T 0.19 mm at the level of the facial nerve. Similar accuracy was also found in previous papers using active shape modeling of the cochlea (16). An erosion of the bony part of the external auditory canal, without skin lesion, was observed in specimen 1R. Insertion of CI422 slim straight arrays was possible through the DCA tunnel, after removal of the lateral fin that impairs a smooth progression of the array through the tunnel, in all cases but one. In this last case, the electrode array became stuck in the DCA tunnel and the tip was bent during the insertion. Thus, the geometry of the array is important for the success of the insertion through this minimally invasive approach. It may explain why, when using perimodiolar arrays with a lateral stylet, insertion through a tunnel was not possible in all cases as reported
Preoperative planning and postoperative data of direct cochlear access (DCA) trajectories and cochleostomies Preoperative Planning
Preoperative Planning
Distances (mm)
1L 1R 2L 2R 3L 3R 4L 4R 5L 5R Avg. TSD
1019
Angles (-)
Drilling
Distances (mm)
FN
ChT
EAC
In/Ma
St
C
D
5C
Opening
Target Error
0.44 0.37 0.37 0.32 0.37 0.43 0.38 0.38 0.39 0.36 0.38 0.03
0.12 0.00a G0.00a G0.00a 0.22 0.53 1.17 1.18 0.37 0.33 0.49 0.45
0.55 0.90 0.45 0.45 1.60 1.85 1.89 2.34 0.62 0.90 1.16 0.70
2.60 2.36 2.64 3.02 2.78 2.95 2.55 3.11 3.01 2.88 2.79 0.24
0.65 0.62 0.77 0.68 0.80 0.65 0.74 0.69 0.51 0.58 0.67 0.09
8 12 11 7 12 15 11 12 10 14 11.2 2.4
0 0 0 0 0 1 1 0 7 1 1.0 2.2
12 10 12 12 8 4 12 12 4 4 9.4 4.1
CO ERW ERW ERW CO ERW RW RW ERW ERW
0.79c 0.6c 0.07 0.14 0.05 0.33c 0.28 0.24 0.27c 0.22 0.29 0.23
a
FN
Absolute Differenceb
0.99c j0.081c 0.33 0.38 0.37 0.53c 0.58 0.28 0.40c 0.22 0.39 0.26
0.55c 0.45c 0.04 0.06 0.00 0.10c 0.20 0.10 0.01c 0.14 0.17 0.19
CO, transpromontorial cochleostomy; ERW, enlarged round window cochleostomy; RW, round window cochleostomy; FN, facial nerve; ChT, chorda tympani; EAC, external auditory canal; In/Ma, incus malleus. a At the level of the cochleostomy target. b Distance between the planned and the real trajectory at the level of the facial nerve. c Broken screw in image data. Otology & Neurotology, Vol. 36, No. 6, 2015
Copyright © 2015 Otology & Neurotology, Inc. Unauthorized reproduction of this article is prohibited.
1020
F. VENAIL ET AL.
FIG. 3. Postoperative cone-beam CT of specific cases. Cone-beam CT images of 1R (A, B) and 1L (C, D) temporal bones. Postoperative calculation of the distance between the DCA tunnel and the facial nerve (j0.081 mm) suggested a slight penetration in the facial nerve. As seen in (A) and (B), there is a close passage of the DCA tunnel without damage to the facial nerve. This fact was confirmed by the dissection of the temporal bone. This discrepancy was attributed to an oversegmentation of the facial nerve before 3D surface rendering to increase the safety of the surgical planning. In 1L, an error of registration 0.79 mm was responsible for shifting the cochleostomy (co) site above the basilar membrane (bm) plane toward the scala vestibuli of the basal turn (C). This resulted in a scala vestibuli insertion of the electrode array (D).
by Labadie et al. (16). This observation formed the basis for our comparison of this array to another straight but more flexible array (Flex28). When using the Flex28, a lubricated insertion tube was used to ensure the array would be aligned to the cochleostomy site without incorrectly bending in the middle ear space. The intracochlear insertion was performed only externally;the array was pushed manually through the DCA tunnel into the cochlea with no additional manipulation performed in the mastoid or through the ear canal. The tympanomeatal flap was only used to visualize the progression of the electrode insertion into the cochlea; insertion was stopped at first resistance. Unexpectedly, we could not achieve a full insertion of the electrode in the cochleae with either electrode arrays. The average depth of insertion with Flex28 arrays was 305 and 243 degrees with CI422 slim straight arrays. These insertion depths are far lower than what was observed in Wimmer et al. (645 degrees using Flex28 arrays [5]) or Skarzynski et al. (388 degrees using CI422 arrays [17]) who inserted the electrode arrays in temporal bones through the round window. Several factors may be involved to explain this discrepancy: the choice of the drilling trajectory during the surgical planning, the insertion
by the DCA tunnel, and finally the method for temporal bone fixation. The trajectory was determined by an experienced ENT surgeon to 1) avoid the risk of facial nerve lesions and 2) minimize the in-plane (D) and out-of-plane (C) angles to align to an ideal trajectory passing through the center of the basal turn of the cochlea. It is important to note that insertion trajectories tangential to the basal turn passed closely or intersected the facial nerve in all cases as previously described by Meshik et al. (15). Therefore, minimization of the angular deviation C of the planned trajectory was mainly restricted by the position of the facial nerve. In a previous insertion study using the same robotic system and straight electrode arrays, the RW membrane center was targeted for the drill trajectory which led to an average out-of-plane deviation of C = 17 degrees (5). In comparison, with the presented planning method, it was possible to reduce the deviation to an average value of C = 11 degrees. The difference in the surgical approach between the studies (i.e., RW insertion versus cochleostomy) does not allow for comparisons of the in-plane alignment angle D. Further, we tried to determine if there was a relationship between C and D angles and the depth of electrode array insertion. No correlation
Otology & Neurotology, Vol. 36, No. 6, 2015
Copyright © 2015 Otology & Neurotology, Inc. Unauthorized reproduction of this article is prohibited.
MANUAL ELECTRODE ARRAY INSERTION
FIG. 4. Surgical dissection of implanted temporal bones. Examples of dissection of temporal bone 3R (A) and 2R (B). In 3R, the DCA tunnel passes about 0.5 mm away from the mastoid portion of the facial nerve (FN) (planned 0.43, calculated 0.52, A). In case 2R (B), a sacrifice of the chorda tympani (Cht) was planned to ensure a safe distance between the DCA and the facial nerve. After dissection, one can see that the chorda is interrupted (arrows) by the DCA tunnel. The incus (In), the lateral semicircular canal (LSCC), and the external ear canal (EAC) were not damaged in those specimens.
between the insertion trajectory deviation from the basal turn and the resulting insertion depth can be shown. Thus, the calculation of the trajectory cannot explain the underinsertion of the electrode arrays in the present study. In this study, electrode array insertion was performed through the DCA tunnel without any additional surgical handling of the array through the mastoid or the external canal as described in the Wimmer and Skarzynski studies (5,17). This study required the use of an insertion tool to guide the more flexible electrode arrays (Flex28) to the cochleostomy. The stiffer electrode arrays (CI422 slim straight) could be inserted directly through the DCA tunnel, but difficulties were experienced in one case in which the electrode array touched the edge of the tunnel, was bent, and was subsequently not suitable for implantation. The utilized insertion tool was not adapted to the insertion of the CI422 slim straight array, as this array does not have a cylindrical shape, and the presence of a lateral fin impairs its progression in the tool. In a clinical setting, the realization of a tympanomeatal flap to guide
1021
the electrode array insertion or the use of a backup implant may have fixed the problem, but this points out the fact that improvements have to be made to the electrode array to allow easier insertion through a minimally invasive approach. Labadie et al. (16) reported the clinical implementation of minimally invasive cochlear implantation using a microstereotactic frame to drill a minimally invasive approach to the cochlea. Although the cochleostomy was performed manually through the ear canal in this study, they experienced similar difficulties for electrode array insertion through their tunnel, especially for perimodiolar arrays (Cochlear Contour Advanced), which are larger and required additional manipulation to remove the stylet. In addition to improving the insertion tool, some adjustments will probably be necessary to use perimodiolar arrays using a robot-assisted minimally invasive approach. The use of an automated insertion tool (18Y21) could enhance the insertion array quality and safety by controlling the depth and the friction forces to avoid cochlear damages as scala vestibuli dislocation. The final point that may explain the incomplete insertion of the electrode array is the type of fixation used in this study. In the majority of previously published papers (5,17,22), temporal bones used were either freshly frozen or Thiel embalmed. Here, the fixation method consisted of zinc chloride intra-arterial injection (8). Although this method is very effective for soft tissue preservation, it may have altered the stiffness of the basilar membrane. Therefore, as the insertion of the electrode array was stopped at the first point of resistance, it may have led to systematic under-insertion of electrode arrays. A similar observation of under-insertion of electrode arrays was already reported by McRackan et al. (7) using Med-El standard (31.5 mm), Cochlear CI422, or Clarion HiFocus electrodes in fixed temporal bones. Looking at the postoperative position of the electrode arrays on cone-beam CT, we noticed no tip fold-over and no scalar dislocation (scala tympani to scala vestibuli and reciprocally). As most dislocation occurs at the level of the first turn of the cochlea (1,22), the utilization of a robot-assisted minimally invasive approach followed by the manual insertion of the electrode array was not responsible of any scalar dislocation in this study; a shallowest insertion depth of 205 degrees was observed. Altogether, our results suggest that the insertion of an electrode array through a robot-assisted cochleostomy is feasible, but many factors (i.e., shape and the flexibility of the electrode array, guiding tool into the DCA tunnel) in addition to the calculation of an optimized trajectory are crucial for complete electrode array insertion.
CONCLUSION This study shows that an image-guided robot-assisted cochleostomy can be performed with sufficient accuracy to allow cochlear implantation through a minimal invasive approach and to avoid damage to important anatomical structures such as the facial nerve. The drilling of Otology & Neurotology, Vol. 36, No. 6, 2015
Copyright © 2015 Otology & Neurotology, Inc. Unauthorized reproduction of this article is prohibited.
1022
F. VENAIL ET AL.
a DCA tunnel allows the insertion of several types of freefitting electrode arrays into the cochleostomy without additional intervention of the surgeon in the middle ear in most of the cases. Even if the positioning of the electrode array is appropriate, improvements have to be made to implement an effective insertion tool to control the depth and the forces of insertion of the electrode array to enhance the procedure of cochlear implantation. Acknowledgments: The authors would like to thank Jarek Tuszynski for the vectorized Matlab implementation of the ray/ triangle intersection algorithm; Prof. Guillaume Captier, Franck Meyer, and Hubert Taillades, University of Montpellier, for support during the experiments; and Prof. Alain Bonafe´ and Mr. Olivier Martin, University Hospital of Montpellier, for their help in acquiring and treating cone-beam CT images.
REFERENCES 1. Adunka O, Kiefer J, Unkelbach MH, et al. Evaluating cochlear implant trauma to the scala vestibuli. Clin Otolaryngol 2005;30: 121Y7. 2. Roland PS, Wright CG. Surgical aspects of cochlear implantation: mechanisms of insertional trauma. Adv Otorhinolaryngol 2006;64: 11Y30. 3. Adunka OF, Buchman CA. Scala tympani cochleostomy I: results of a survey. Laryngoscope 2007;117:2187Y94. 4. Adunka OF, Radeloff A, Gstoettner WK, et al. Scala tympani cochleostomy II: topography and histology. Laryngoscope 2007; 117:2195Y200. 5. Wimmer W, Bell B, Huth ME, et al. Cone beam and microcomputed tomography validation of manual array insertion for minimally invasive cochlear implantation. Audiol Neurootol 2014; 19:22Y30. 6. Wimmer W, Venail F, Williamson T, et al. Semiautomatic cochleostomy target and insertion trajectory planning for minimally invasive cochlear implantation. Biomed Res Int 2014;2014:596498. doi:10.1155/2014/596498. 7. McRackan TR, Balachandran R, Blachon GS, et al. Validation of minimally invasive, image-guided cochlear implantation using Advanced Bionics, Cochlear, and Medel electrodes in a cadaver model. Int J Comput Assist Radiol Surg 2013;8:989Y95.
8. Bryant CD. Handbook of Death &Dying. Thousand Oaks, Calif.: Sage Publications, 2003. 9. Bell B, Gerber N, Williamson T, et al. In vitro accuracy evaluation of image-guided robot system for direct cochlear access. Otol Neurotol 2013;34:1284Y90. 10. Lecerf P, Bakhos D, Cottier JP, et al. Midmodiolar reconstruction as a valuable tool to determine the exact position of the cochlear implant electrode array. Otol Neurotol 2011;32:1075Y81. 11. Marx M, Risi F, Escude B, et al. Reliability of cone beam computed tomography in scalar localization of the electrode array: a radio histological study. Eur Arch Otorhinolaryngol 2014;271:673Y9. 12. Pelosi S, Noble JH, Dawant BM, et al. Analysis of intersubject variations in intracochlear and middle ear surface anatomy for cochlear implantation. Otol Neurotol 2013;34:1675Y80. 13. Stelter K, Ledderose G, Hempel JM, et al. Image guided navigation by intraoperative CT scan for cochlear implantation. Comput Aided Surg 2012;17:153Y60. 14. Cushing SL, Daly MJ, Treaba CG, et al. High-resolution cone-beam computed tomography: a potential tool to improve atraumatic electrode design and position. Acta Otolaryngol 2012;132:361Y8. 15. Meshik X, Holden TA, Chole RA, et al. Optimal cochlear implant insertion vectors. Otol Neurotol 2010;31:58Y63. 16. Labadie RF, Balachandran R, Noble JH, et al. Minimally invasive image-guided cochlear implantation surgery: first report of clinical implementation. Laryngoscope 2014;124:1915Y22. 17. Skarzynski H, Podskarbi-Fayette R. A new cochlear implant electrode design for preservation of residual hearing: a temporal bone study. Acta Otolaryngol 2010;130:435Y42. 18. Majdani O, Schurzig D, Hussong A, et al. Force measurement of insertion of cochlear implant electrode arrays in vitro: comparison of surgeon to automated insertion tool. Acta Otolaryngol 2010; 130:31Y6. 19. Miroir M, Nguyen Y, Kazmitcheff G, et al. Friction force measurement during cochlear implant insertion: application to a force-controlled insertion tool design. Otol Neurotol 2012;33: 1092Y100. 20. Nguyen Y, Miroir M, Kazmitcheff G, et al. Cochlear implant insertion forces in microdissected human cochlea to evaluate a prototype array. Audiol Neurotol 2012;17:290Y8. 21. Zhang J, Wei W, Ding J, et al. Inroads toward robot-assisted cochlear implant surgery using steerable electrode arrays. Otol Neurotol 2010;31:1199Y206. 22. Marx M, Risi F, Escude´ B, et al. Reliability of cone beam computed tomography in scalar localization of the electrode array: a radio histological study. Eur Arch Otorhinolaryngol 2013;271:673Y9.
Otology & Neurotology, Vol. 36, No. 6, 2015
Copyright © 2015 Otology & Neurotology, Inc. Unauthorized reproduction of this article is prohibited.
Manual Electrode Array Insertion Through a Robot-Assisted Minimal Invasive Cochleostomy: Feasibility and Comparison of Two Different Electrode Array Subtypes *†Frederic Venail, ‡Brett Bell, *Mohamed Akkari, ‡§Wilhelm Wimmer, ‡Tom Williamson, ‡Nicolas Gerber, ‡Kate Gavaghan, kFrancois Canovas, ‡Stefan Weber, ‡§Marco Caversaccio, and *†Alain Uziel *Otology and Neurotology Department, University Hospital of Montpellier, Montpellier, France; ÞInstitute for Neurosciences of Montpellier, INSERM U1051, Montpellier, France; þARTORG Center for Biomedical Engineering Research, University of Bern, Bern, Switzerland; §Department of ENT, Head and Neck Surgery, Inselspital, University of Bern, Bern, Switzerland; and kAnatomy Laboratory, Faculty of Medicine, University Hospital of Montpellier, Montpellier, France
Hypothesis: To evaluate the feasibility and the results of insertion of two types of electrode arrays in a robotically assisted surgical approach. Background: Recent publications demonstrated that robotassisted surgery allows the implantation of free-fitting electrode arrays through a cochleostomy drilled via a narrow bony tunnel (DCA). We investigated if electrode arrays from different manufacturers could be used with this approach. Methods: Cone-beam CT imaging was performed on fivecadaveric heads after placement of fiducial screws. Relevant anatomical structures were segmented and the DCA trajectory, including the position of the cochleostomy, was defined to target the center of the scala tympani while reducing the risk of lesions to the facial nerve. Med-El Flex 28 and Cochlear CI422 electrodes were implanted on both sides, and their position was verified by cone-beam CT. Finally, temporal bones were dissected
to assess the occurrence of damage to anatomical structures during DCA drilling. Results: The cochleostomy site was directed in the scala tympani in 9 of 10 cases. The insertion of electrode arrays was successful in 19 of 20 attempts. No facial nerve damage was observed. The average difference between the planned and the postoperative trajectory was 0.17 T 0.19 mm at the level of the facial nerve. The average depth of insertion was 305.5 T 55.2 and 243 T 32.1 degrees with Med-El and Cochlear arrays, respectively. Conclusions: Robot-assisted surgery is a reliable tool to allow cochlear implantation through a cochleostomy. Technical solutions must be developed to improve the electrode array insertion using this approach. Key Words: Cochlear implantationVCochleostomyVCone-beam CTVImage-guided surgeryVRobot-assisted surgery. Otol Neurotol 36:1015Y1022, 2015.
Advances in cochlear implant techniques emphasize the development of minimally invasive cochlear implantation methods. This concept implies the reduction of the surgical trauma not only for the electrode array insertion but also during the surgical approach through the facial recess. A ‘‘soft’’ electrode array insertion is required not only
for residual hearing preservation but also to maintain normal cochlear anatomy and avoid postoperative intracochlear fibrosis (1,2). Although some surgeons favor the round window for direct access into the scala tympani, some argue that insertion through the round window may cause cochlear damage. Conversely, an anteroinferior cochleostomy avoids these problems and facilitates an atraumatic insertion (3). However, anatomical variations can jeopardize the correct positioning of the cochleostomy and may lead to a scala vestibuli insertion (4). Recent work on image-guided robotic-assisted surgery has shown that a direct cochlear access (DCA) through a single bony tunnel can be drilled efficiently and safely in
Address correspondence and reprint requests to Frederic Venail, M.D., Ph.D., ENT Department,University Hospital Gui de Chauliac, 80 avenue Augustin Fliche, Montpellier, France; E-mail: [email protected] This work was financially supported by the Nano-Tera.ch initiative, FP7 HearEU, and the Swiss National Science Foundation. The authors report no conflicts of interest. Supplemental digital content is available in the text.
1015
Copyright © 2015 Otology & Neurotology, Inc. Unauthorized reproduction of this article is prohibited.
1016
F. VENAIL ET AL.
temporal bones using high-accuracy optical tracking camera while preserving sensitive structures like the facial nerve, ossicles, and the external ear canal (5). Recently, it was demonstrated that the drilling of a cochleostomy through a DCA tunnel is feasible and that insertion of free-fitting electrode arrays (Med-El 28 mm arrays) could be achieved without any intervention of the surgeon through the ear canal (6). A previous report described the successful use of different electrode arrays to perform image-guided cochlear implantation in a cadaver model using a stereotactic frame (7). In the present work, emphasis is placed on software modifications which allow the surgeon to plan the ideal insertion trajectory for drilling DCA tunnel and cochleostomies while evaluating the results achieved with freefitting electrode arrays from different manufacturers. MATERIALS AND METHODS
holding device which uses pressurized pads to stabilize the head. Next, a tracking reference was rigidly attached to the skull with a bone screw (2 5 mm). In a final step, the robot system was mounted to the OR table. The final plan was registered to the temporal bone using an optical tracking device to match the locations of the implanted fiducial screws. The robot was then moved, utilizing haptic control, into the field of view of the optical tracking camera before approaching the surface of the cortical mastoid bone. The DCA was then drilled by the robot using a custom ‘‘step’’drill having a proximal diameter of 2.5 mm with a length of 20 mm, and distal portion with a diameter of 1.8 mm and a length of 10 mm to the tip. The drill motor was started (5000 rpm) and the robot drilled with a feed rate of 0.5 mm/s using a ‘‘pecking’’motion until the middle ear cavity was reached. A cochleostomy was then drilled by the robot system using a 1-mm diamond burr. The drill speed was increased to 10,000 rpm, and the feed rate was reduced to 0.1 mm/s (Fig. 1). All robot movement was completed under the supervision of the surgeon using a three-stage enable switch, which allowed progression of the drilling or movement of the manipulator only when desired.
Specimen Preparation and Imaging The complete protocol (i.e., intervention planning, drilling, and array insertion) was performed in the experimental surgery laboratory of the University Hospital of Montpellier, France. Five human cadaver heads (n = 10 temporal bones) fixed with a 20% zinc-chloride intra-arterial injection (according to Sucquet [8]) were used in this study. Before the experiment, a retroauricular musculocutaneous flap was elevated to expose the mastoid cortical bone. Four titanium screws (M-5220.03; Medartis, Basel, Switzerland) were placed to be used as fiducial markers for patient-to-image registration. All imaging in this study (before and after surgery) was carried out using a cone-beam CT (NewTom 5G, Verona, Italy) and a high-resolution protocol (voxel size: 125 Km isotropic, 110 kVp, 19 mA). After the completion of the drilling and insertion experiments, all temporal bones underwent surgical microdissection to detect any damage to the facial nerve, the chorda tympani, or other middle ear structures during the DCA tunnel drilling.
Direct Cochlear Access and Cochleostomy Drilling A minimally invasive access to the tympanic cavity was drilled with a purpose-built robotic system developed earlier (9) as defined by the preoperative surgical planning (see Supplemental Digital Content for more details, http://links.lww.com/MAO/A294). The distances between the trajectory of the DCA tunnel and the facial nerve, the chorda tympani, the posterior wall of the external auditory canal, and the ossicles were calculated to minimize the risk of basilar membrane collision or tearing of the lateral wall. Deviation from the ideal trajectory towards the modiolus or the lateral wall is defined as the angle D, and deviation towards the basilar membrane or the bottom of the scala tympani is defined as the angle C (see Supplemental Data Figure, http://links.lww.com/MAO/A294). The distance of the cochleostomy from the round window was defined as the angular distance between 0 degrees (center of the round window) and 20 degrees (more anteriorly) by 2-degree angle steps (defined as 5C, see Supplemental Data Figure, http://links.lww.com/MAO/A294). At the beginning of the surgical procedure, the skin of the external auditory canal was removed and a tympanic flap was performed to introduce a 4-mm 30-degree endoscope for visual control of the cochleostomy drilling and the insertion of the electrode array. The DCA tunnel was drilled using the same protocol published previously (9). Briefly, the cadaver head was mounted in a noninvasive head
Insertion of Electrode Arrays and Evaluation of the Quality of Insertion Two types of electrodes arrays were inserted: Flex28 electrodes (28 mm length; Med-El company) and CI422 slim straight arrays (SSA, 20 mm length; Cochlear Corporation). Both electrode arrays are straight, free-fitting arrays. Before the insertion, the DCA tunnel was cleaned using saline solution irrigation and suction of bone dust. Flex28 electrodes arrays were inserted first (Fig. 1). Because these types of arrays are highly flexible and could bend during the mastoid part of the insertion, a custom insertion tool was designed to facilitate their guidance through the DCA tunnel. Insertion was performed with hyaluronic acid gel for lubrication under visual control with the endoscope (Fig. 1). Insertion was stopped at the point of first resistance. After completion of insertion, electrode arrays were fixed using sutures to prohibit movement during subsequent handling phases. Then, highresolution cone-beam CT scans were performed as previously described to evaluate the depth of insertion of the electrode arrays and their scalar positioning. Insertion depth was estimated by the insertion angle > (-) (Fig. 2, A and B) from the round window to the most apical electrode contact. Scalar localization of the electrode array was determined by a neuroradiologist experienced in cochlear implant imaging. Theoretical positions of scala tympani and scala vestibuli were determined based on previous studies (10,11). On midmodiolar reconstructions, scala tympani was located in the inferior part of a half-turn section of the cochlea and scala vestibuli in the superior part. The position of the electrode contact was inferred to be in the scala tympani when the contact was located in the lower half turn of the cochlear section. In the second part of the experiment, Flex28 arrays were gently removed and CI422 slim straight arrays were inserted in the same temporal bones to compare the insertion with both arrays. CI422 slim straight arrays are designed with a lateral fin for the handling of the array, making it too large to enter into the DCA tunnel. Therefore, the fin was removed from the electrode array before insertion. No insertion tool was used during the insertion of CI422 arrays which are larger and more rigid when compared to the Flex28 arrays. The rest of the procedure was performed as described for Flex28 arrays, and cone-beam CT scans were then
Otology & Neurotology, Vol. 36, No. 6, 2015
Copyright © 2015 Otology & Neurotology, Inc. Unauthorized reproduction of this article is prohibited.
MANUAL ELECTRODE ARRAY INSERTION
1017
FIG. 1. Visualization through the ear canal of the cochleostomy’s drilling and electrode array insertion into the cochlea. Step-by-step visualization of the drilling of the DCA with the cochleostomy and of the electrode array insertion of a Med-El Flex28 array in a left temporal bone. The round window (RW, arrow), the promontory (P), the long process of the incus (In), the stapes (St), and the pyramidal process (pyr) can be seen with the 30-degree endoscope. After drilling of the DCA (A), the diamond bur drills a 1-mm hole through the promontory of the cochlea (B). Then liquid and bone dust are removed from the cochleostomy (Co) (C), and the insertion tool (it) is inserted through the DCA tunnel close to the cochleostomy (D). A lubricated electrode array (ea) is pushed through the insertion tool into the cochlea until the first point of resistance (E). Finally, the insertion tool is gently removed from the DCA tunnel (F).
acquired for further evaluation. Additional cone-beam CT scans were performed after cochlear array removal to evaluate the position of the DCA tunnel for accuracy evaluation.
RESULTS Evaluation of Electrode Arrays’ Insertion Insertion of the array was possible in all cases using Flex28 electrode arrays and the insertion tool. Conversely, intracochlear insertion of the electrode array was not possible in one case with CI422 slim straight array. In this case, the electrode array could not be properly guided to the cochleostomy through the DCA tunnel in the absence of the insertion tool. The cochleostomy was successfully directed towards the scala tympani of the basal turn of the cochlea in 9 of 10 cases implanted with Flex28 and in eightof ninecases implanted with CI422 slim straight arrays (Table 1). In the last case, the cochleostomy ended in the scala vestibuli of the basal turn of the cochlea because of a large drilling error (Table 2). Full insertion of the electrode array was not achieved in any of the implanted cases. The average depth of insertion with Flex28 arrays was 305.5 T 55.2 degrees, with an average depth of 243 T 32.1 degrees with CI422 slim straight arrays. This resulted in an average of 2.3 T 0.9 and 3.9 T 1.4 electrodes outside the cochlea with Flex28 and CI422 arrays, respectively (Table 1, Fig. 2).
Accuracy of DCA Tunnel and Cochleostomy Drilling The accuracy at the cochleostomy target was measured at 0.29 T 0.23 mm with a range of 0.05 to 0.79 mm. In two cases, a target error larger than 0.5 mm was observed (Table 2). The target error was orientated anteriorly and posteriorly in specimen 1L and 1R, respectively. This caused penetration of the external auditory canal posterior bony wall without cutaneous damage and a cochleostomy ending in the scala vestibuli of the basal turn of the cochlea in specimen 1L (Fig. 3). A close passage of the facial nerve was observed in specimen 1R. As expected, the chorda tympani was damaged in specimens 2L and 2R (Fig. 4). During the planning, the distance (TSD) between the facial nerve and the DCA tunnel was estimated at 0.38 T 0.03 mm. Postoperatively, the distance between the DCA tunnel and the facial nerve was determined to be 0.39 T 0.26 mm (p = 0.82, Student’s t test for paired data)(see Table 2). The mean absolute difference between the planned and calculated distance to the facial nerve was 0.17 T 0.19 mm. According to postoperative CT image examination, the position of the drilled tunnel seemed to slightly touch the facial nerve in specimen 1R (distance of j0.081 mm, Table 2, Fig.3). However, surgical microdissection of the temporal bonerevealed that the drilled tunnel did not intersect with the facial nerve. Thus, the small negative value of the distance between the DCA tunnel and the Otology & Neurotology, Vol. 36, No. 6, 2015
Copyright © 2015 Otology & Neurotology, Inc. Unauthorized reproduction of this article is prohibited.
1018
F. VENAIL ET AL.
FIG. 2. Postoperative cone-beam CT of inserted Med-EL Flex28 and Cochlear CI422 arrays. Postoperative cone-beam CT of temporal bone 3R. Depth of insertion of the electrode array is evaluated at 259 degrees with CI422 electrode array (A) and at 342 degrees with Med-El Flex28 (B). Depth of insertion is measured as the angle > (-) between the cochleostomy (Co) and the most apical electrode in axial plane (A, B). Cross-sections perpendicular to the blue dashed line show that with CI422 electrode array (C) and Med-El Flex28 (D), the electrode array is entirely located in the scala tympani of the cochlea. Note that full insertion of electrode arrays was not achieved in both cases, leaving three and one electrodes outside the cochlea with CI422 and Med-El Flex28 arrays, respectively.
facial nerve was likely due to oversegmentation of the nerve during the planning stage. A standard mastoidectomy was subsequently performed on all specimens, confirming that the drilling of the DCA did not result in visible damage to the facial nerve (Fig. 4). The 5C angle chosen during the planning varied from 4 to 12 degrees (average of 9 T 3.68 degrees). This angle was compared to the type of cochleostomy observed through the endoscope during drilling (RW: round window, ERW: enlarged round window, CO: transpromontorial cochleostomy). Because of targeting errors, there was no clear correlation between the 5C angle defined on the surgical planning and the type of cochleostomy performed (Table 2).
DISCUSSION Electrode array insertion is a crucial step in cochlear implant surgery. This insertion can be performed through the round window membrane or through a transpromontorial cochleostomy. In the latter case, the surgeon must determine the position of the scala tympani of the basal turn of
the cochlea relying on the only constant landmark which is the round window itself. However, interindividual variation in round window and cochlear hook anatomy (12) can mislead the surgeon and cause to drill a cochleostomy ending in the scala vestibuli in approximately 20% of cases (4). Therefore, image-guided surgery was proposed to help in the positioning of the cochleostomy. Stelter et al. (13) showed that real-time image-guided navigation, based on conventional CT images (Brainlab), had an average deviation of 1.01 mm at the level of the round window. As shown in our study, a tool positioning errors below 1 mm at the level of the promontory could direct the cochleostomy to the scala vestibuli. Even if the accuracy of standard navigation techniques may be improved by using highresolution cone-beam CT images (14), it is likely that the accuracy will not be enough to be clinically implemented. Thus, improvements in cochleostomy positioning require additional technological solutions, including robotassisted approaches. Centerline computation (15) or active shape modeling (5,7,16) can be used to determine the position of the center of the scala tympani preoperatively
Otology & Neurotology, Vol. 36, No. 6, 2015
Copyright © 2015 Otology & Neurotology, Inc. Unauthorized reproduction of this article is prohibited.
MANUAL ELECTRODE ARRAY INSERTION TABLE 1. Comparison of depth of electrode insertion, number of electrodes outside the cochlea, and scalar position with Med-El Flex28 and CI422 electrode arrays No. Electrodes Outside the Cochlea
Angle of Insertion > (-) Temporal Bone Flex28 1L
CI422
264
Flex28 CI422 Flex28 3
317
2L
202
2R
295
3L
412
3R
342
4L
277
4R
323
5L
291
5R
332
Avg. TSD
305.5 55.2
5
SV/SV
2
ST/ST
NA
NA
NA
3
ST/ST
247
3
ST/ST
3
ST/ST
209
5
ST/ST
1
ST/ST
222
4
ST/ST
1
ST/ST
259
3
ST/ST
4
ST/ST
276
4
ST/ST
2
ST/ST
269
5
ST/ST
2
ST/ST
289
1
ST/ST
2 211 243.0 32.1
CI422
SV/SV
205 1R
Scalar Position of the Apical/ Basal Electrode
ST/ST
2.3 0.9
5 3.9 1.4
ST/ST
NA, non-applicableVthe electrode array could not be inserted; ST, scala tympani; SV, scala vestibuli.
from high-resolution cone-beam CT images. Both strategies have been utilized in temporal bone specimens (5,7,15) or in vivo (16) to determine suitable insertion sites. In these studies, electrode array insertion was performed either by the round window (5) or by a manually drilled cochleostomy (7,16), requiring an additional access to the middle ear through a tympanomeatal flap in both cases. Neither of these works performed the complete procedure, including TABLE 2.
cochleostomy and electrode insertion, through the same tunnel without the help of an additional middle ear approach as in the present study. In this study, a scala tympani cochleostomy was performed in 9 of 10 cases. In the last case, the cochleostomy ended in the scala vestibuli, but did not preclude cochlear array insertion. A drilling error of 0.79 mm at the level of the target (promontory) was responsible for this misalignment. The large deviation was later determined to be the caused by broken fiducial screws that were left in the specimen, causing smallest deviations in the computation of the fiducial’s center position. Such self-drilling screws were originally utilized to ease the preoperative workflow. However, their tendency to break in hard, dense cortical bone required us to identify different screws for future use with additional image-based checks implemented to ensure the safety of the robotic-assisted procedure. No damage to the facial nerve was observed resulting from the drilling of the DCA trajectory even in case of narrow facial recess (temporal bone 1L and 1R). The absolute difference calculated between the preoperative trajectory and the postoperative real trajectory was only 0.17 T 0.19 mm at the level of the facial nerve. Similar accuracy was also found in previous papers using active shape modeling of the cochlea (16). An erosion of the bony part of the external auditory canal, without skin lesion, was observed in specimen 1R. Insertion of CI422 slim straight arrays was possible through the DCA tunnel, after removal of the lateral fin that impairs a smooth progression of the array through the tunnel, in all cases but one. In this last case, the electrode array became stuck in the DCA tunnel and the tip was bent during the insertion. Thus, the geometry of the array is important for the success of the insertion through this minimally invasive approach. It may explain why, when using perimodiolar arrays with a lateral stylet, insertion through a tunnel was not possible in all cases as reported
Preoperative planning and postoperative data of direct cochlear access (DCA) trajectories and cochleostomies Preoperative Planning
Preoperative Planning
Distances (mm)
1L 1R 2L 2R 3L 3R 4L 4R 5L 5R Avg. TSD
1019
Angles (-)
Drilling
Distances (mm)
FN
ChT
EAC
In/Ma
St
C
D
5C
Opening
Target Error
0.44 0.37 0.37 0.32 0.37 0.43 0.38 0.38 0.39 0.36 0.38 0.03
0.12 0.00a G0.00a G0.00a 0.22 0.53 1.17 1.18 0.37 0.33 0.49 0.45
0.55 0.90 0.45 0.45 1.60 1.85 1.89 2.34 0.62 0.90 1.16 0.70
2.60 2.36 2.64 3.02 2.78 2.95 2.55 3.11 3.01 2.88 2.79 0.24
0.65 0.62 0.77 0.68 0.80 0.65 0.74 0.69 0.51 0.58 0.67 0.09
8 12 11 7 12 15 11 12 10 14 11.2 2.4
0 0 0 0 0 1 1 0 7 1 1.0 2.2
12 10 12 12 8 4 12 12 4 4 9.4 4.1
CO ERW ERW ERW CO ERW RW RW ERW ERW
0.79c 0.6c 0.07 0.14 0.05 0.33c 0.28 0.24 0.27c 0.22 0.29 0.23
a
FN
Absolute Differenceb
0.99c j0.081c 0.33 0.38 0.37 0.53c 0.58 0.28 0.40c 0.22 0.39 0.26
0.55c 0.45c 0.04 0.06 0.00 0.10c 0.20 0.10 0.01c 0.14 0.17 0.19
CO, transpromontorial cochleostomy; ERW, enlarged round window cochleostomy; RW, round window cochleostomy; FN, facial nerve; ChT, chorda tympani; EAC, external auditory canal; In/Ma, incus malleus. a At the level of the cochleostomy target. b Distance between the planned and the real trajectory at the level of the facial nerve. c Broken screw in image data. Otology & Neurotology, Vol. 36, No. 6, 2015
Copyright © 2015 Otology & Neurotology, Inc. Unauthorized reproduction of this article is prohibited.
1020
F. VENAIL ET AL.
FIG. 3. Postoperative cone-beam CT of specific cases. Cone-beam CT images of 1R (A, B) and 1L (C, D) temporal bones. Postoperative calculation of the distance between the DCA tunnel and the facial nerve (j0.081 mm) suggested a slight penetration in the facial nerve. As seen in (A) and (B), there is a close passage of the DCA tunnel without damage to the facial nerve. This fact was confirmed by the dissection of the temporal bone. This discrepancy was attributed to an oversegmentation of the facial nerve before 3D surface rendering to increase the safety of the surgical planning. In 1L, an error of registration 0.79 mm was responsible for shifting the cochleostomy (co) site above the basilar membrane (bm) plane toward the scala vestibuli of the basal turn (C). This resulted in a scala vestibuli insertion of the electrode array (D).
by Labadie et al. (16). This observation formed the basis for our comparison of this array to another straight but more flexible array (Flex28). When using the Flex28, a lubricated insertion tube was used to ensure the array would be aligned to the cochleostomy site without incorrectly bending in the middle ear space. The intracochlear insertion was performed only externally;the array was pushed manually through the DCA tunnel into the cochlea with no additional manipulation performed in the mastoid or through the ear canal. The tympanomeatal flap was only used to visualize the progression of the electrode insertion into the cochlea; insertion was stopped at first resistance. Unexpectedly, we could not achieve a full insertion of the electrode in the cochleae with either electrode arrays. The average depth of insertion with Flex28 arrays was 305 and 243 degrees with CI422 slim straight arrays. These insertion depths are far lower than what was observed in Wimmer et al. (645 degrees using Flex28 arrays [5]) or Skarzynski et al. (388 degrees using CI422 arrays [17]) who inserted the electrode arrays in temporal bones through the round window. Several factors may be involved to explain this discrepancy: the choice of the drilling trajectory during the surgical planning, the insertion
by the DCA tunnel, and finally the method for temporal bone fixation. The trajectory was determined by an experienced ENT surgeon to 1) avoid the risk of facial nerve lesions and 2) minimize the in-plane (D) and out-of-plane (C) angles to align to an ideal trajectory passing through the center of the basal turn of the cochlea. It is important to note that insertion trajectories tangential to the basal turn passed closely or intersected the facial nerve in all cases as previously described by Meshik et al. (15). Therefore, minimization of the angular deviation C of the planned trajectory was mainly restricted by the position of the facial nerve. In a previous insertion study using the same robotic system and straight electrode arrays, the RW membrane center was targeted for the drill trajectory which led to an average out-of-plane deviation of C = 17 degrees (5). In comparison, with the presented planning method, it was possible to reduce the deviation to an average value of C = 11 degrees. The difference in the surgical approach between the studies (i.e., RW insertion versus cochleostomy) does not allow for comparisons of the in-plane alignment angle D. Further, we tried to determine if there was a relationship between C and D angles and the depth of electrode array insertion. No correlation
Otology & Neurotology, Vol. 36, No. 6, 2015
Copyright © 2015 Otology & Neurotology, Inc. Unauthorized reproduction of this article is prohibited.
MANUAL ELECTRODE ARRAY INSERTION
FIG. 4. Surgical dissection of implanted temporal bones. Examples of dissection of temporal bone 3R (A) and 2R (B). In 3R, the DCA tunnel passes about 0.5 mm away from the mastoid portion of the facial nerve (FN) (planned 0.43, calculated 0.52, A). In case 2R (B), a sacrifice of the chorda tympani (Cht) was planned to ensure a safe distance between the DCA and the facial nerve. After dissection, one can see that the chorda is interrupted (arrows) by the DCA tunnel. The incus (In), the lateral semicircular canal (LSCC), and the external ear canal (EAC) were not damaged in those specimens.
between the insertion trajectory deviation from the basal turn and the resulting insertion depth can be shown. Thus, the calculation of the trajectory cannot explain the underinsertion of the electrode arrays in the present study. In this study, electrode array insertion was performed through the DCA tunnel without any additional surgical handling of the array through the mastoid or the external canal as described in the Wimmer and Skarzynski studies (5,17). This study required the use of an insertion tool to guide the more flexible electrode arrays (Flex28) to the cochleostomy. The stiffer electrode arrays (CI422 slim straight) could be inserted directly through the DCA tunnel, but difficulties were experienced in one case in which the electrode array touched the edge of the tunnel, was bent, and was subsequently not suitable for implantation. The utilized insertion tool was not adapted to the insertion of the CI422 slim straight array, as this array does not have a cylindrical shape, and the presence of a lateral fin impairs its progression in the tool. In a clinical setting, the realization of a tympanomeatal flap to guide
1021
the electrode array insertion or the use of a backup implant may have fixed the problem, but this points out the fact that improvements have to be made to the electrode array to allow easier insertion through a minimally invasive approach. Labadie et al. (16) reported the clinical implementation of minimally invasive cochlear implantation using a microstereotactic frame to drill a minimally invasive approach to the cochlea. Although the cochleostomy was performed manually through the ear canal in this study, they experienced similar difficulties for electrode array insertion through their tunnel, especially for perimodiolar arrays (Cochlear Contour Advanced), which are larger and required additional manipulation to remove the stylet. In addition to improving the insertion tool, some adjustments will probably be necessary to use perimodiolar arrays using a robot-assisted minimally invasive approach. The use of an automated insertion tool (18Y21) could enhance the insertion array quality and safety by controlling the depth and the friction forces to avoid cochlear damages as scala vestibuli dislocation. The final point that may explain the incomplete insertion of the electrode array is the type of fixation used in this study. In the majority of previously published papers (5,17,22), temporal bones used were either freshly frozen or Thiel embalmed. Here, the fixation method consisted of zinc chloride intra-arterial injection (8). Although this method is very effective for soft tissue preservation, it may have altered the stiffness of the basilar membrane. Therefore, as the insertion of the electrode array was stopped at the first point of resistance, it may have led to systematic under-insertion of electrode arrays. A similar observation of under-insertion of electrode arrays was already reported by McRackan et al. (7) using Med-El standard (31.5 mm), Cochlear CI422, or Clarion HiFocus electrodes in fixed temporal bones. Looking at the postoperative position of the electrode arrays on cone-beam CT, we noticed no tip fold-over and no scalar dislocation (scala tympani to scala vestibuli and reciprocally). As most dislocation occurs at the level of the first turn of the cochlea (1,22), the utilization of a robot-assisted minimally invasive approach followed by the manual insertion of the electrode array was not responsible of any scalar dislocation in this study; a shallowest insertion depth of 205 degrees was observed. Altogether, our results suggest that the insertion of an electrode array through a robot-assisted cochleostomy is feasible, but many factors (i.e., shape and the flexibility of the electrode array, guiding tool into the DCA tunnel) in addition to the calculation of an optimized trajectory are crucial for complete electrode array insertion.
CONCLUSION This study shows that an image-guided robot-assisted cochleostomy can be performed with sufficient accuracy to allow cochlear implantation through a minimal invasive approach and to avoid damage to important anatomical structures such as the facial nerve. The drilling of Otology & Neurotology, Vol. 36, No. 6, 2015
Copyright © 2015 Otology & Neurotology, Inc. Unauthorized reproduction of this article is prohibited.
1022
F. VENAIL ET AL.
a DCA tunnel allows the insertion of several types of freefitting electrode arrays into the cochleostomy without additional intervention of the surgeon in the middle ear in most of the cases. Even if the positioning of the electrode array is appropriate, improvements have to be made to implement an effective insertion tool to control the depth and the forces of insertion of the electrode array to enhance the procedure of cochlear implantation. Acknowledgments: The authors would like to thank Jarek Tuszynski for the vectorized Matlab implementation of the ray/ triangle intersection algorithm; Prof. Guillaume Captier, Franck Meyer, and Hubert Taillades, University of Montpellier, for support during the experiments; and Prof. Alain Bonafe´ and Mr. Olivier Martin, University Hospital of Montpellier, for their help in acquiring and treating cone-beam CT images.
REFERENCES 1. Adunka O, Kiefer J, Unkelbach MH, et al. Evaluating cochlear implant trauma to the scala vestibuli. Clin Otolaryngol 2005;30: 121Y7. 2. Roland PS, Wright CG. Surgical aspects of cochlear implantation: mechanisms of insertional trauma. Adv Otorhinolaryngol 2006;64: 11Y30. 3. Adunka OF, Buchman CA. Scala tympani cochleostomy I: results of a survey. Laryngoscope 2007;117:2187Y94. 4. Adunka OF, Radeloff A, Gstoettner WK, et al. Scala tympani cochleostomy II: topography and histology. Laryngoscope 2007; 117:2195Y200. 5. Wimmer W, Bell B, Huth ME, et al. Cone beam and microcomputed tomography validation of manual array insertion for minimally invasive cochlear implantation. Audiol Neurootol 2014; 19:22Y30. 6. Wimmer W, Venail F, Williamson T, et al. Semiautomatic cochleostomy target and insertion trajectory planning for minimally invasive cochlear implantation. Biomed Res Int 2014;2014:596498. doi:10.1155/2014/596498. 7. McRackan TR, Balachandran R, Blachon GS, et al. Validation of minimally invasive, image-guided cochlear implantation using Advanced Bionics, Cochlear, and Medel electrodes in a cadaver model. Int J Comput Assist Radiol Surg 2013;8:989Y95.
8. Bryant CD. Handbook of Death &Dying. Thousand Oaks, Calif.: Sage Publications, 2003. 9. Bell B, Gerber N, Williamson T, et al. In vitro accuracy evaluation of image-guided robot system for direct cochlear access. Otol Neurotol 2013;34:1284Y90. 10. Lecerf P, Bakhos D, Cottier JP, et al. Midmodiolar reconstruction as a valuable tool to determine the exact position of the cochlear implant electrode array. Otol Neurotol 2011;32:1075Y81. 11. Marx M, Risi F, Escude B, et al. Reliability of cone beam computed tomography in scalar localization of the electrode array: a radio histological study. Eur Arch Otorhinolaryngol 2014;271:673Y9. 12. Pelosi S, Noble JH, Dawant BM, et al. Analysis of intersubject variations in intracochlear and middle ear surface anatomy for cochlear implantation. Otol Neurotol 2013;34:1675Y80. 13. Stelter K, Ledderose G, Hempel JM, et al. Image guided navigation by intraoperative CT scan for cochlear implantation. Comput Aided Surg 2012;17:153Y60. 14. Cushing SL, Daly MJ, Treaba CG, et al. High-resolution cone-beam computed tomography: a potential tool to improve atraumatic electrode design and position. Acta Otolaryngol 2012;132:361Y8. 15. Meshik X, Holden TA, Chole RA, et al. Optimal cochlear implant insertion vectors. Otol Neurotol 2010;31:58Y63. 16. Labadie RF, Balachandran R, Noble JH, et al. Minimally invasive image-guided cochlear implantation surgery: first report of clinical implementation. Laryngoscope 2014;124:1915Y22. 17. Skarzynski H, Podskarbi-Fayette R. A new cochlear implant electrode design for preservation of residual hearing: a temporal bone study. Acta Otolaryngol 2010;130:435Y42. 18. Majdani O, Schurzig D, Hussong A, et al. Force measurement of insertion of cochlear implant electrode arrays in vitro: comparison of surgeon to automated insertion tool. Acta Otolaryngol 2010; 130:31Y6. 19. Miroir M, Nguyen Y, Kazmitcheff G, et al. Friction force measurement during cochlear implant insertion: application to a force-controlled insertion tool design. Otol Neurotol 2012;33: 1092Y100. 20. Nguyen Y, Miroir M, Kazmitcheff G, et al. Cochlear implant insertion forces in microdissected human cochlea to evaluate a prototype array. Audiol Neurotol 2012;17:290Y8. 21. Zhang J, Wei W, Ding J, et al. Inroads toward robot-assisted cochlear implant surgery using steerable electrode arrays. Otol Neurotol 2010;31:1199Y206. 22. Marx M, Risi F, Escude´ B, et al. Reliability of cone beam computed tomography in scalar localization of the electrode array: a radio histological study. Eur Arch Otorhinolaryngol 2013;271:673Y9.
Otology & Neurotology, Vol. 36, No. 6, 2015
Copyright © 2015 Otology & Neurotology, Inc. Unauthorized reproduction of this article is prohibited.
