Otology & Neurotology 36:631Y637 Ó 2015, Otology & Neurotology, Inc.
Study of the Feasible Size of a Bone Conduction Implant Transducer in the Temporal Bone ¨ stli, *Bo Håkansson, *Hamidreza Taghavi, *Sabine Reinfeldt, *Per O †Måns Eeg-Olofsson, and †Joacim Stalfors *Department of Signals and Systems, Chalmers University of Technology, Gothenburg, Sweden; and ÞEar, Nose, and Throat Department, Sahlgrenska University Hospital, Gothenburg, Sweden
Hypothesis: The aim was to assess the temporal bone volume to determine the suitable size and position of a bone conduction implant (BCI) transducer. Background: A BCI transducer needs to be sufficiently small to fit in the mastoid portion of the temporal bone for a majority of patients. The anatomical geometry limits both the dimension of an implanted transducer and its positions in the temporal bone to provide a safe and simple surgery. Methods: Computed tomography (CT) scans of temporal bones from 22 subjects were virtually reconstructed. With an algorithm in MATLAB, the maximum transducer diameter as function of the maximum transducer depth in the temporal bone, and the most suitable position were calculated in all subjects. Results: An implanted transducer diameter of 16 mm inserted at a depth of 4 mm statistically fitted 95% of the subjects.
If changing the transducer diameter to 12 mm, a depth of 6 mm would fit in 95% of the subjects. The most suitable position was found to be around 20 mm behind the ear canal. Conclusion: The present BCI transducer casing, used in ongoing clinical trials, was designed from the results in this study, demonstrating that the present BCI transducer casing (largest diameter [diagonal]: 15.5 mm, height: 6.4 mm) will statistically fit more than 95% of the subjects. Hence, the present BCI transducer is concluded to be sufficiently small to fit most normal-sized temporal bones and should be placed approximately 20 mm behind the ear canal. Key Words: Bone conductionVimplantVComputed tomographyVMaximum sizeV Temporal boneVTransducer. Otol Neurotol 36:631Y637, 2015.
Subjects with conductive or mixed hearing loss can improve their hearing by using bone conduction devices (BCD) (1,2). In the present bone-anchored hearing device (BAHA), the sound processor is attached to a percutaneous titanium screw anchored in the temporal bone, transmitting sound vibrations to the cochlea without skin attenuation. The BAHA has been a global success with over 150,000 patients (3,4). The system gives excellent audiological outcome (5,6), but complications resulting from implant losses and skin reactions, sometimes demanding revision surgery, are reported (7Y9). Transcutaneous bone conduction devices have been developed to minimize complications related to percutaneous implants by keeping the skin intact. Two active transcutaneous BCDs are the bone conduction implant (BCI), which is developed by research groups at Chalmers
University of Technology and Sahlgrenska University Hospital, Gothenburg, Sweden, and is now in clinical trials (10Y13), and the Bonebridge by MED-EL (14), which is commercially available. Active transcutaneous BCDs consist of external and internal units, and the signal is transmitted from the external audio processor via an inductive link through the skin to the implanted transducer which is directly attached and osseointegrated to the bone. The BCI system is depicted in Figure 1 with the audio processor, the inductive link, and the implanted unit, which is called the bridging bone conductor (BBC). Figure 2 shows the BBC, which includes the internal retention magnet, the receiver coil, the demodulator, and the BCI transducer. The size of the transducer with casing must be sufficiently small to fit in the mastoid portion of the temporal bone. For an electromagnetic transducer, the size of the transducer is directly related to its performance, especially at low frequencies where the resonance frequency is inversely proportional to the square root of the counterweight mass (15). Also, the biasing magnetic flux must be sufficiently high (16) requiring a certain size of permanent magnets and soft iron material to avoid saturation.
Address correspondence and reprint requests to Sabine Reinfeldt, Ph.D., Department of Signals and Systems, Chalmers University of Technology, SE-412 96 Gothenburg, Sweden. E-mail: [email protected] All authors disclose no conflict of interest, except Bo Håkansson, who holds several patents related to the BCI device. This study was supported by Vinnova (Sweden’s Innovation Agency).
631
Copyright © 2015 Otology & Neurotology, Inc. Unauthorized reproduction of this article is prohibited.
632
S. REINFELDT ET AL. Aim of Study The primary aim of this study is to assess the maximum size of the BCI transducer to fit the mastoid portion of the temporal bone of most patients. The secondary aim is to find a suitable position for the BCI transducer. MATERIALS AND METHODS
FIG. 1. The BCI audio processor receives the sound and transmits the signal via an inductive link through the intact skin to the implanted transducer that is positioned in the temporal bone. Not in scale.
In the present BAHA, the transducer including casing is too big to fit in most temporal bones. Therefore, a new transducer principle has been developed for the BCI, called balanced electromagnetic separation transducer (BEST). This transducer principle allows for a considerably smaller counterweight mass for a given resonance frequency as compared with the transducers used in the BAHA systems and it gives less distortion (16). The dimensions of the transducer casing are shown in Figure 3. The Bonebridge transducer is of floating mass type with 15.8 mm in diameter and a depth of 8.7 mm underneath the retention arms. As a flat seat for these retention arms are made, the whole casing is recessed in the temporal bone. The distance between the anchor holes for the screws in the retention arms is 23.8 mm. The optimal position of a transducer in the temporal bone must be based on several findings. Earlier studies have suggested that a transducer positioned closer to the cochlea gives higher sensitivity (11,17Y19). The challenge is to maintain safe and uncomplicated surgery while avoiding damage on vital anatomical structures. Previous investigations have shown that the temporal bone has an average volume of 8.2 ml (20) with large individual variations (20,21). Furthermore, definite anatomical landmarks, such as the dura of the middle and posterior cranial fossa, the sigmoid sinus, the ear canal, and the labyrinth, provide natural limitations for the positioning. Knowledge of the maximum size of the transducer casing that can be fitted in most patients is essential before the definite size of the casing is decided.
Measurements of the maximum size of transducer casing were performed from preoperative CT scans of patients due for cochlear implantation. The subjects were identified consecutively with exclusion of children and subjects with a history of chronic ear disease or ear surgery. Also, incomplete scans or malformed temporal bones were excluded. Only examinations with a spatial resolution of 0.5 mm or less and without artifacts were included. Primarily, the left temporal bone was used. Out of CT scans from 26 subjects, 22 were included in this study. The study cohort had a median age of 68 years (range 38Y89 yr) and both genders were represented equally. The studied part of the temporal bone was the space where a simple canal wall up mastoidectomy is normally performed. The bony ear canal is intact and the space is limited medially by the labyrinth, laterally by the cortical bone, superiorly by the dura of the middle cranial fossa, posteriorly by the dura of the posterior cranial fossa and the sigmoid sinus, and inferiorly by the tip of the mastoid process.
Segmentation of the Temporal Bone CT scans (2D) in DICOM format were imported to MATLAB and a 3D reconstruction was created for each subject. A graphical user interface was created in MATLAB to view the reconstructions in three projections: transverse, sagittal, and coronal. The reconstructions were cropped to only contain the temporal bones used, and if the right side was chosen, the reconstruction was mirrored in the sagittal plane. The ear canal opening in the sagittal plane was set by visual inspection. The temporal bone contains both compact bone and air cells, and the compact bone relevant in this study was mainly located at the lateral border of the temporal bone. To identify the anatomical limitations of the temporal bone in the reconstructions, the following segmentation method was automated by a MATLAB script. Air, bone, and tissue values were estimated by histograms from each 3D reconstruction. First, all voxels (volumetric pixels) in a reconstruction were assigned to bone, except at the boundary. The boundary outside of the body was set to air, while the other boundaries were set to tissue. The voxels at the transitions between bone, tissue, and air were then classified.
FIG. 2. The bridging bone conductor (BBC) unit including the transducer to the right.
Otology & Neurotology, Vol. 36, No. 4, 2015
Copyright © 2015 Otology & Neurotology, Inc. Unauthorized reproduction of this article is prohibited.
FEASIBLE SIZE OF BCI TRANSDUCER IN TEMPORAL BONE
FIG. 3.
633
Dimensions of the BCI transducer casing. A, Top view; B, side view.
This was repeated to make the areas with tissue and air grow inwards. For a voxel to be classified as a certain type (bone, tissue, or air), the demand was to have five neighboring voxels of the same type and the voxel value should be within this type’s interval. Finally, a fine tuning was performed to remove noisy areas by increasing the demand to nine instead of five neighboring voxels of the same type. Figure 4A shows an image from a CT reconstruction after the segmentation procedure.
Implant Sizes and Positions A virtual implant with cylindrical shape was created and inserted in the segmented MATLAB 3D reconstruction. The maximum diameter and insertion depth, as well as the position, were determined by the boundaries of the mastoid portion of the temporal bone. The tests were automatized by a MATLAB script also including localization constraints according to the following procedure. First, a starting point was defined 10 mm posteriorly and 10 mm superiorly to the center of the ear canal. The virtual implant, with a diameter of 7.9 mm and height of
10 mm, was then positioned at the starting point. If the implant could fit within the temporal bone reconstruction, the diameter was increased in steps of 0.1 mm. If not fitted, the implant was rotated and its position was randomly changed to find a better position. Then the implant height was decreased by 0.1 mm and the procedure was repeated. For each reconstruction, at least 2,500 tests for most suitable fitting were performed. In the script, the implant must be partly or fully within one area (area 1) and completely outside of one other area (area 2) (see Fig. 4, B and C). The virtual implant was inserted into the temporal bone with its surface parallel to the bone surface with residual bone wall set to 1 mm (see Fig. 5).
Calculations The maximum implant diameter as a function of maximum implant insertion depth in the temporal bone was calculated for each CT scan. The average, quartile, and 95% confidence interval was calculated for all CT scans.
FIG. 4. A, An image from a CT reconstruction of a temporal bone after the segmentation. The implant must be (B) partly or fully within area 1 and (C) completely outside area 2. Otology & Neurotology, Vol. 36, No. 4, 2015
Copyright © 2015 Otology & Neurotology, Inc. Unauthorized reproduction of this article is prohibited.
634
S. REINFELDT ET AL.
FIG. 5. The virtual implant (gray) is placed with its surface parallel to the bone (yellow in online version) surface and the residual bone wall (blue in online version) is 1 mm.
To investigate possible gender differences, a two-tailed paired t test was performed with the null hypothesis that there were no differences between the genders.
has implications on the transducer size because a too large transducer will not fit in all patients. This study had its starting point in finding the largest size for a BCI transducer given that it should fit most temporal bones. The study group was chosen from patients eligible for cochlear implantation, but with an otherwise healthy ear. Indications for bone conduction hearing amplification can be found in patients with a deficit in the middle ear but healthy mastoid, as in this study group. However, patients with previous surgery in the mastoid or atretic ear patients (often also with an atretic temporal bone) are eligible for bone conduction devices as well, and this group of patients has not been subject to this study. Likely, the disposable anatomical limitations in these patients will be further restricting. If the choice would have been to investigate 22 subjects with middle ear disease and who have different histories of surgical interventions, most likely the high degree of variability would have made it impossible to make a general conclusion. Such patients were therefore excluded here. This study shows that the BCI can be installed in most patients with otherwise normal temporal bones, and that
RESULTS Fitting of the implant in MATLAB was performed in CT reconstructions from 22 individuals. Maximum implant diameter for a statistical fit of 95% of the subjects was determined by the algorithm to be 16 mm with an implant depth in the temporal bone of 4 mm. If the implant diameter was reduced to 12 mm, the insertion depth could be increased to 6 mm to statistically fit 95% of the subjects. The statistical measure of ‘‘95%’’ of the subjects corresponds to 21 of 22 subjects in our study. The individual maximum implant diameters are shown as function of the maximum insertion depth into the bone in Figure 6A. The maximum diameter was found to decrease almost linearly with the depth for most subjects. The average of the maximum implant diameter for all subjects is shown together with the quartile and the 5 and 95% confidence intervals in Figure 6B. The data indicated a slight gender difference with maximum size of the implant larger for men than for women. However, the difference was not statistically significant according to a two-tailed paired t test. The position of the implant center in the sagittal plane is shown in Figure 7 for 6 mm depth. The spread of the positions were found to be maximum 35 mm, but on average the implant center was positioned within 20 mm behind the ear canal entrance and spread T10 mm in vertical direction. The position of the implant center in the sagittal plane for 4 mm depth was similar as for 6 mm depth. DISCUSSION This study shows anatomical variation in the size of normal temporal bones, and that a transducer casing with a diameter of 16 mm and a height of 4 mm would statistically fit in 95% of the subjects. The result in this study
FIG. 6. Maximum implant diameter as a function of maximum implant depth in the temporal bone: (A) individual results and (B) average (blue solid line in online version), quartile (gray area), 95% confidence interval (dashed green line in online version), and 5% confidence interval (dashed red line in online version).
Otology & Neurotology, Vol. 36, No. 4, 2015
Copyright © 2015 Otology & Neurotology, Inc. Unauthorized reproduction of this article is prohibited.
FEASIBLE SIZE OF BCI TRANSDUCER IN TEMPORAL BONE
FIG. 7.
Position of the implant center in the sagittal plane at 6 mm depth.
the proposed positioning is around 20 mm behind the ear canal (see Fig. 8). A CT scan for preoperative planning is recommended in patients with a mastoid cavity after ear surgery, where the BCI transducer will be placed posterior to the cavity, or in patients with malformations of the temporal bone. The BCI transducer is placed against a bony flat surface 20 mm behind the ear canal and a thin titanium wire on top of the transducer applies a static pressure against the bone. The transducer commonly protrudes 1Y3 mm, but it can, if necessary, protrude even more. Extra protrusion of the implantable part is possible and seen in the implant coil part of, for example, cochlear implants and the Bonebridge, with a thickness of around 3Y5 mm (the implant coil part of Bonebridge has a height of 4.5 mm (14)). To illustrate on how many patients a BCI transducer would fit, Figure 6A was reconstructed to show maximum implant radius on the y axis and to include a BCI transducer with diameter of 13.7 mm, which is the longest distance of the transducer if disregarding the diagonal. The size of the temporal bone in the posterior
FIG. 8.
635
Proposed position of the transducer casing.
direction from the ear canal is typically what limits the virtual implant diameter. Therefore, the BCI transducer is designed to have its shortest dimension in that direction. So it would be more realistic to show a BCI transducer of 11.7 mm diameter, but worst case was chosen instead.
FIG. 9. Fitting of BCI transducer with a diameter of 13.7 mm and protrusion of 2 mm in graph of individual maximum radius versus maximum depth in the temporal bone. The BCI transducer fits all patients as no curve overlaps the transducer casing. Otology & Neurotology, Vol. 36, No. 4, 2015
Copyright © 2015 Otology & Neurotology, Inc. Unauthorized reproduction of this article is prohibited.
636
S. REINFELDT ET AL.
Figure 9 shows a BCI transducer with a radius of 6.9 mm (diameter 13.7 mm) in the graph describing the relation between maximum insertion depth and maximum radius. The BCI transducer shown here protrudes 2 mm. If the transducer would have ended up overlapping any of the individual curves, the transducer would not have fitted the patient who this curve belonged to. The figure shows that the BCI transducer fits all patients of this study with a good margin and therefore most patients with normal temporal bones. There is also some margin for diseased or malformed temporal bones. Furthermore, protrusion more than 1Y3 mm is a possible solution for small temporal bones. If, for example, a transducer protrudes over the bone surface by 3 mm and the diameter is 16 mm, the transducer height could be 7 mm. This maximum size was preliminary known from a pilot study when the final size of the BCI transducer was decided (22). From a design point of view, it would be difficult to decrease the size of the transducer much further without reducing low-frequency sensitivity, and furthermore, the BCI transducer does not need to be made smaller to fit more patients. The present BCI transducer was designed with a maximum diameter of 15.5 mm and a height of 6.4 mm as shown in Figure 3, and would hence statistically fit more than 95% of the patients with good margin. The Bonebridge by MED-EL (14) requires a larger space in the temporal bone (diameter 15.8 mm, depth 8.7 mm). If disregarding the screws, the implant would fit around 40% of the subjects included in this study (see Fig. 6B). The screws must also be taken into consideration to provide enough space on opposite sides of the transducer to fit them. The segmentation method and the method of finding maximum implant diameter and insertion depth was regarded as reliable, however with the following considerations: 1) The CT exams were performed at different machines with different settings and different quality of the images. In the highly processed images, the edges were enhanced, which might have caused tissue next to compact bone to appear as air or bone. 2) The bone next to dura was at some places too thin to be visible in the CT images; hence, the limits between tissue and bone were seen as limits between tissue and air. The thin bone in the air cell system was sometimes interpreted as the surrounding tissue, which complicated the segmentation at some places. 3) The middle ear had no visible limit towards the air cells in the temporal bone; it was possible for the segmentation algorithm to interpret the most medial air cells in the temporal bone as middle ear. The first consideration could affect the maximum diameter both ways, and therefore, the errors were expected to be evened out. The second and third points might have given smaller maximum diameter of the transducer, but these effects were considered small and negligible. Therefore, these limitations are not likely to affect the conclusions of this study. Furthermore, visual inspection of the results of all subjects was conducted to control all measures accomplished with the automated method.
CONCLUSION It is found in this study that a cylindrical transducer casing with maximum diameter of 16 mm can have a maximum insertion depth of 4 mm to statistically fit more than 95% of normal temporal bones. If the transducer is protruding the bone surface 3 mm, the height of the same transducer can be 7 mm. The present BCI transducer casing is designed to comply with the maximum size found in this study, that is, it will be sufficiently small to fit in most healthy adult temporal bones, and it should be placed with its center 20 mm behind the ear canal. Acknowledgments: The authors thank the master thesis students Alireza Arjomand and Siavash Esmaeili Fashtakeh who performed the pilot study.
REFERENCES 1. von Be´kesy G. Experiments in Hearing. Toronto: McGraw-Hill Book Company, Incorporated, 1960. 2. Stenfelt S, Goode R. Bone conducted sound: physiological and clinical aspects. Otol Neurotol 2005;26:1245Y61. 3. News and research in bone conduction hearing - craniofacial osseointegration [press release]. Mo¨lnlycke: Cochlear Bone Anchored Solutions AB; 2011 [June 11] Available from: http:// www.cochlear.com/wps/wcm/connect/0a7aa3b7-8ad8-40ab-b28820054c04e3e2/en-Baha-Ospdf?MOD=AJPERES&CONVERT_ TO=url&CACHEID=0a7aa3b7-8ad8-40ab-b288-20054c04e3e2. Accessed December 4, 2014. 4. Oticon Medical. 2014. Oticon MedicalVtaking ideas further. http://www. oticonmedical.com/~asset/cache.ashx?id=7229&type=14&format=web. Accessed March 4, 2015. 5. Tjellstro¨m A, Håkansson B, Granstro¨m G. Bone-anchored hearing aids: current status in adults and children. Otolaryngol Clin North Am 2001;34:337Y64. 6. Snik A, Mylanus E, Proops D, et al. Consensus statements on the BAHA system: where do we stand at present? Ann Otol Rhinol Laryngol Suppl 2005;195:2Y12. 7. Wallberg E, Granstro¨m G, Tjellstro¨m A, Stalfors J. Implant survival rate in bone-anchored hearing aid users: long-term results. J Laryngol Otol 2011;125:1131Y5. 8. Dun CAJ, Faber HTF, de Wolf MJF, Mylanus EAM, Cremers CWRJ, Hol MKS. Assessment of more than 1,000 implanted percutaneous bone conduction devices: skin reactions and implant survival. Otol Neurotol 2012;33:192Y8. 9. Hobson JC, Roper AJ, Andrew R, Rothera MP, Hill P, Green KM. Complications of bone-anchored hearing aid implantation. J Laryngol Otol 2010;124:132Y6. 10. Håkansson B, Eeg-Olofsson M, Reinfeldt S, Stenfelt S, Granstro¨m G. Percutaneous vs transcutaneous bone conduction implant system: a feasibility study on a cadaver head. Otol Neurotol 2008;29:1132Y9. 11. Håkansson B, Reinfeldt S, Eeg-Olofsson M, et al. A novel bone conduction implant (BCI): engineering aspects and pre-clinical studies. Int J Audiol 2010;49:203Y15. 12. Taghavi H, Håkansson B, Reinfeldt S. A Novel Bone Conduction Implant SystemVAnalog Radio Frequency Data and Power Link Design. Proceedings of the 9th IASTED International Conference on Biomedical Engineering, Innsbruck, Austria, 2012a:327Y35. 13. Taghavi H, Håkansson B, Reinfeldt S. Analysis and design of RF power and data link using amplitude modulation of class-E for a novel bone conduction implant. IEEE Trans Biomed Eng 2012b;59:3050Y9. 14. MedEl 2014. BONEBRIDGEiVInformation for Surgeons (with BCI lifts). Available at http://s3.medel.com/pdf/INT/information_surgeons_ bci-lifts.pdf. Accessed March 4, 2015.
Otology & Neurotology, Vol. 36, No. 4, 2015
Copyright © 2015 Otology & Neurotology, Inc. Unauthorized reproduction of this article is prohibited.
FEASIBLE SIZE OF BCI TRANSDUCER IN TEMPORAL BONE 15. Håkansson B, Brandt A, Carlsson P, Tjellstro¨m A. Resonance frequencies of the human skull in vivo. J Acoust Soc Am 1994;95: 1474Y81. 16. Håkansson B. The balanced electromagnetic separation transducer: a new bone conduction transducer. J Acoust Soc Am 2003;113: 818Y25. 17. Eeg-Olofsson M, Stenfelt S, Tjellstro¨m A, Granstro¨m G. Transmission of bone-conducted sound in the human skull measured by cochlear vibrations. Int J Audiol 2008;47:761Y9. 18. Reinfeldt S, Håkansson B, Taghavi H, Eeg-Olofsson M. Bone conduction hearing sensitivity in normal hearing subjectsVtranscutaneous stimulation at BAHA versus BCI Position. Int J Audiol 2014;53: 360Y9.
637
19. Eeg-Olofsson M, Håkansson B, Reinfeldt S, et al. The bone conduction implantVfirst implantation, surgical and audiological aspects. Otol Neurotol 2014;35:679Y85. 20. Cinamon U. The growth rate and size of the mastoid air cell system and mastoid bone: a review and reference. Eur Arch Otorhinolaryngol 2009;266:781Y6. 21. Swarts JD, Doyle BM, Doyle WJ. Relationship between surface area and volume of the mastoid air cell system in adult humans. J Laryngol Otol 2011;125:580Y4. 22. Arjomand A, Fashtakeh S. Geometrical Analysis of Temporal Bones for Bone Conduction Implants (BCI). Master of Science Thesis. Go¨teborg, Sweden: Chalmers University of Technology, 2009. EX084/2009.
Otology & Neurotology, Vol. 36, No. 4, 2015
Copyright © 2015 Otology & Neurotology, Inc. Unauthorized reproduction of this article is prohibited.
Study of the Feasible Size of a Bone Conduction Implant Transducer in the Temporal Bone ¨ stli, *Bo Håkansson, *Hamidreza Taghavi, *Sabine Reinfeldt, *Per O †Måns Eeg-Olofsson, and †Joacim Stalfors *Department of Signals and Systems, Chalmers University of Technology, Gothenburg, Sweden; and ÞEar, Nose, and Throat Department, Sahlgrenska University Hospital, Gothenburg, Sweden
Hypothesis: The aim was to assess the temporal bone volume to determine the suitable size and position of a bone conduction implant (BCI) transducer. Background: A BCI transducer needs to be sufficiently small to fit in the mastoid portion of the temporal bone for a majority of patients. The anatomical geometry limits both the dimension of an implanted transducer and its positions in the temporal bone to provide a safe and simple surgery. Methods: Computed tomography (CT) scans of temporal bones from 22 subjects were virtually reconstructed. With an algorithm in MATLAB, the maximum transducer diameter as function of the maximum transducer depth in the temporal bone, and the most suitable position were calculated in all subjects. Results: An implanted transducer diameter of 16 mm inserted at a depth of 4 mm statistically fitted 95% of the subjects.
If changing the transducer diameter to 12 mm, a depth of 6 mm would fit in 95% of the subjects. The most suitable position was found to be around 20 mm behind the ear canal. Conclusion: The present BCI transducer casing, used in ongoing clinical trials, was designed from the results in this study, demonstrating that the present BCI transducer casing (largest diameter [diagonal]: 15.5 mm, height: 6.4 mm) will statistically fit more than 95% of the subjects. Hence, the present BCI transducer is concluded to be sufficiently small to fit most normal-sized temporal bones and should be placed approximately 20 mm behind the ear canal. Key Words: Bone conductionVimplantVComputed tomographyVMaximum sizeV Temporal boneVTransducer. Otol Neurotol 36:631Y637, 2015.
Subjects with conductive or mixed hearing loss can improve their hearing by using bone conduction devices (BCD) (1,2). In the present bone-anchored hearing device (BAHA), the sound processor is attached to a percutaneous titanium screw anchored in the temporal bone, transmitting sound vibrations to the cochlea without skin attenuation. The BAHA has been a global success with over 150,000 patients (3,4). The system gives excellent audiological outcome (5,6), but complications resulting from implant losses and skin reactions, sometimes demanding revision surgery, are reported (7Y9). Transcutaneous bone conduction devices have been developed to minimize complications related to percutaneous implants by keeping the skin intact. Two active transcutaneous BCDs are the bone conduction implant (BCI), which is developed by research groups at Chalmers
University of Technology and Sahlgrenska University Hospital, Gothenburg, Sweden, and is now in clinical trials (10Y13), and the Bonebridge by MED-EL (14), which is commercially available. Active transcutaneous BCDs consist of external and internal units, and the signal is transmitted from the external audio processor via an inductive link through the skin to the implanted transducer which is directly attached and osseointegrated to the bone. The BCI system is depicted in Figure 1 with the audio processor, the inductive link, and the implanted unit, which is called the bridging bone conductor (BBC). Figure 2 shows the BBC, which includes the internal retention magnet, the receiver coil, the demodulator, and the BCI transducer. The size of the transducer with casing must be sufficiently small to fit in the mastoid portion of the temporal bone. For an electromagnetic transducer, the size of the transducer is directly related to its performance, especially at low frequencies where the resonance frequency is inversely proportional to the square root of the counterweight mass (15). Also, the biasing magnetic flux must be sufficiently high (16) requiring a certain size of permanent magnets and soft iron material to avoid saturation.
Address correspondence and reprint requests to Sabine Reinfeldt, Ph.D., Department of Signals and Systems, Chalmers University of Technology, SE-412 96 Gothenburg, Sweden. E-mail: [email protected] All authors disclose no conflict of interest, except Bo Håkansson, who holds several patents related to the BCI device. This study was supported by Vinnova (Sweden’s Innovation Agency).
631
Copyright © 2015 Otology & Neurotology, Inc. Unauthorized reproduction of this article is prohibited.
632
S. REINFELDT ET AL. Aim of Study The primary aim of this study is to assess the maximum size of the BCI transducer to fit the mastoid portion of the temporal bone of most patients. The secondary aim is to find a suitable position for the BCI transducer. MATERIALS AND METHODS
FIG. 1. The BCI audio processor receives the sound and transmits the signal via an inductive link through the intact skin to the implanted transducer that is positioned in the temporal bone. Not in scale.
In the present BAHA, the transducer including casing is too big to fit in most temporal bones. Therefore, a new transducer principle has been developed for the BCI, called balanced electromagnetic separation transducer (BEST). This transducer principle allows for a considerably smaller counterweight mass for a given resonance frequency as compared with the transducers used in the BAHA systems and it gives less distortion (16). The dimensions of the transducer casing are shown in Figure 3. The Bonebridge transducer is of floating mass type with 15.8 mm in diameter and a depth of 8.7 mm underneath the retention arms. As a flat seat for these retention arms are made, the whole casing is recessed in the temporal bone. The distance between the anchor holes for the screws in the retention arms is 23.8 mm. The optimal position of a transducer in the temporal bone must be based on several findings. Earlier studies have suggested that a transducer positioned closer to the cochlea gives higher sensitivity (11,17Y19). The challenge is to maintain safe and uncomplicated surgery while avoiding damage on vital anatomical structures. Previous investigations have shown that the temporal bone has an average volume of 8.2 ml (20) with large individual variations (20,21). Furthermore, definite anatomical landmarks, such as the dura of the middle and posterior cranial fossa, the sigmoid sinus, the ear canal, and the labyrinth, provide natural limitations for the positioning. Knowledge of the maximum size of the transducer casing that can be fitted in most patients is essential before the definite size of the casing is decided.
Measurements of the maximum size of transducer casing were performed from preoperative CT scans of patients due for cochlear implantation. The subjects were identified consecutively with exclusion of children and subjects with a history of chronic ear disease or ear surgery. Also, incomplete scans or malformed temporal bones were excluded. Only examinations with a spatial resolution of 0.5 mm or less and without artifacts were included. Primarily, the left temporal bone was used. Out of CT scans from 26 subjects, 22 were included in this study. The study cohort had a median age of 68 years (range 38Y89 yr) and both genders were represented equally. The studied part of the temporal bone was the space where a simple canal wall up mastoidectomy is normally performed. The bony ear canal is intact and the space is limited medially by the labyrinth, laterally by the cortical bone, superiorly by the dura of the middle cranial fossa, posteriorly by the dura of the posterior cranial fossa and the sigmoid sinus, and inferiorly by the tip of the mastoid process.
Segmentation of the Temporal Bone CT scans (2D) in DICOM format were imported to MATLAB and a 3D reconstruction was created for each subject. A graphical user interface was created in MATLAB to view the reconstructions in three projections: transverse, sagittal, and coronal. The reconstructions were cropped to only contain the temporal bones used, and if the right side was chosen, the reconstruction was mirrored in the sagittal plane. The ear canal opening in the sagittal plane was set by visual inspection. The temporal bone contains both compact bone and air cells, and the compact bone relevant in this study was mainly located at the lateral border of the temporal bone. To identify the anatomical limitations of the temporal bone in the reconstructions, the following segmentation method was automated by a MATLAB script. Air, bone, and tissue values were estimated by histograms from each 3D reconstruction. First, all voxels (volumetric pixels) in a reconstruction were assigned to bone, except at the boundary. The boundary outside of the body was set to air, while the other boundaries were set to tissue. The voxels at the transitions between bone, tissue, and air were then classified.
FIG. 2. The bridging bone conductor (BBC) unit including the transducer to the right.
Otology & Neurotology, Vol. 36, No. 4, 2015
Copyright © 2015 Otology & Neurotology, Inc. Unauthorized reproduction of this article is prohibited.
FEASIBLE SIZE OF BCI TRANSDUCER IN TEMPORAL BONE
FIG. 3.
633
Dimensions of the BCI transducer casing. A, Top view; B, side view.
This was repeated to make the areas with tissue and air grow inwards. For a voxel to be classified as a certain type (bone, tissue, or air), the demand was to have five neighboring voxels of the same type and the voxel value should be within this type’s interval. Finally, a fine tuning was performed to remove noisy areas by increasing the demand to nine instead of five neighboring voxels of the same type. Figure 4A shows an image from a CT reconstruction after the segmentation procedure.
Implant Sizes and Positions A virtual implant with cylindrical shape was created and inserted in the segmented MATLAB 3D reconstruction. The maximum diameter and insertion depth, as well as the position, were determined by the boundaries of the mastoid portion of the temporal bone. The tests were automatized by a MATLAB script also including localization constraints according to the following procedure. First, a starting point was defined 10 mm posteriorly and 10 mm superiorly to the center of the ear canal. The virtual implant, with a diameter of 7.9 mm and height of
10 mm, was then positioned at the starting point. If the implant could fit within the temporal bone reconstruction, the diameter was increased in steps of 0.1 mm. If not fitted, the implant was rotated and its position was randomly changed to find a better position. Then the implant height was decreased by 0.1 mm and the procedure was repeated. For each reconstruction, at least 2,500 tests for most suitable fitting were performed. In the script, the implant must be partly or fully within one area (area 1) and completely outside of one other area (area 2) (see Fig. 4, B and C). The virtual implant was inserted into the temporal bone with its surface parallel to the bone surface with residual bone wall set to 1 mm (see Fig. 5).
Calculations The maximum implant diameter as a function of maximum implant insertion depth in the temporal bone was calculated for each CT scan. The average, quartile, and 95% confidence interval was calculated for all CT scans.
FIG. 4. A, An image from a CT reconstruction of a temporal bone after the segmentation. The implant must be (B) partly or fully within area 1 and (C) completely outside area 2. Otology & Neurotology, Vol. 36, No. 4, 2015
Copyright © 2015 Otology & Neurotology, Inc. Unauthorized reproduction of this article is prohibited.
634
S. REINFELDT ET AL.
FIG. 5. The virtual implant (gray) is placed with its surface parallel to the bone (yellow in online version) surface and the residual bone wall (blue in online version) is 1 mm.
To investigate possible gender differences, a two-tailed paired t test was performed with the null hypothesis that there were no differences between the genders.
has implications on the transducer size because a too large transducer will not fit in all patients. This study had its starting point in finding the largest size for a BCI transducer given that it should fit most temporal bones. The study group was chosen from patients eligible for cochlear implantation, but with an otherwise healthy ear. Indications for bone conduction hearing amplification can be found in patients with a deficit in the middle ear but healthy mastoid, as in this study group. However, patients with previous surgery in the mastoid or atretic ear patients (often also with an atretic temporal bone) are eligible for bone conduction devices as well, and this group of patients has not been subject to this study. Likely, the disposable anatomical limitations in these patients will be further restricting. If the choice would have been to investigate 22 subjects with middle ear disease and who have different histories of surgical interventions, most likely the high degree of variability would have made it impossible to make a general conclusion. Such patients were therefore excluded here. This study shows that the BCI can be installed in most patients with otherwise normal temporal bones, and that
RESULTS Fitting of the implant in MATLAB was performed in CT reconstructions from 22 individuals. Maximum implant diameter for a statistical fit of 95% of the subjects was determined by the algorithm to be 16 mm with an implant depth in the temporal bone of 4 mm. If the implant diameter was reduced to 12 mm, the insertion depth could be increased to 6 mm to statistically fit 95% of the subjects. The statistical measure of ‘‘95%’’ of the subjects corresponds to 21 of 22 subjects in our study. The individual maximum implant diameters are shown as function of the maximum insertion depth into the bone in Figure 6A. The maximum diameter was found to decrease almost linearly with the depth for most subjects. The average of the maximum implant diameter for all subjects is shown together with the quartile and the 5 and 95% confidence intervals in Figure 6B. The data indicated a slight gender difference with maximum size of the implant larger for men than for women. However, the difference was not statistically significant according to a two-tailed paired t test. The position of the implant center in the sagittal plane is shown in Figure 7 for 6 mm depth. The spread of the positions were found to be maximum 35 mm, but on average the implant center was positioned within 20 mm behind the ear canal entrance and spread T10 mm in vertical direction. The position of the implant center in the sagittal plane for 4 mm depth was similar as for 6 mm depth. DISCUSSION This study shows anatomical variation in the size of normal temporal bones, and that a transducer casing with a diameter of 16 mm and a height of 4 mm would statistically fit in 95% of the subjects. The result in this study
FIG. 6. Maximum implant diameter as a function of maximum implant depth in the temporal bone: (A) individual results and (B) average (blue solid line in online version), quartile (gray area), 95% confidence interval (dashed green line in online version), and 5% confidence interval (dashed red line in online version).
Otology & Neurotology, Vol. 36, No. 4, 2015
Copyright © 2015 Otology & Neurotology, Inc. Unauthorized reproduction of this article is prohibited.
FEASIBLE SIZE OF BCI TRANSDUCER IN TEMPORAL BONE
FIG. 7.
Position of the implant center in the sagittal plane at 6 mm depth.
the proposed positioning is around 20 mm behind the ear canal (see Fig. 8). A CT scan for preoperative planning is recommended in patients with a mastoid cavity after ear surgery, where the BCI transducer will be placed posterior to the cavity, or in patients with malformations of the temporal bone. The BCI transducer is placed against a bony flat surface 20 mm behind the ear canal and a thin titanium wire on top of the transducer applies a static pressure against the bone. The transducer commonly protrudes 1Y3 mm, but it can, if necessary, protrude even more. Extra protrusion of the implantable part is possible and seen in the implant coil part of, for example, cochlear implants and the Bonebridge, with a thickness of around 3Y5 mm (the implant coil part of Bonebridge has a height of 4.5 mm (14)). To illustrate on how many patients a BCI transducer would fit, Figure 6A was reconstructed to show maximum implant radius on the y axis and to include a BCI transducer with diameter of 13.7 mm, which is the longest distance of the transducer if disregarding the diagonal. The size of the temporal bone in the posterior
FIG. 8.
635
Proposed position of the transducer casing.
direction from the ear canal is typically what limits the virtual implant diameter. Therefore, the BCI transducer is designed to have its shortest dimension in that direction. So it would be more realistic to show a BCI transducer of 11.7 mm diameter, but worst case was chosen instead.
FIG. 9. Fitting of BCI transducer with a diameter of 13.7 mm and protrusion of 2 mm in graph of individual maximum radius versus maximum depth in the temporal bone. The BCI transducer fits all patients as no curve overlaps the transducer casing. Otology & Neurotology, Vol. 36, No. 4, 2015
Copyright © 2015 Otology & Neurotology, Inc. Unauthorized reproduction of this article is prohibited.
636
S. REINFELDT ET AL.
Figure 9 shows a BCI transducer with a radius of 6.9 mm (diameter 13.7 mm) in the graph describing the relation between maximum insertion depth and maximum radius. The BCI transducer shown here protrudes 2 mm. If the transducer would have ended up overlapping any of the individual curves, the transducer would not have fitted the patient who this curve belonged to. The figure shows that the BCI transducer fits all patients of this study with a good margin and therefore most patients with normal temporal bones. There is also some margin for diseased or malformed temporal bones. Furthermore, protrusion more than 1Y3 mm is a possible solution for small temporal bones. If, for example, a transducer protrudes over the bone surface by 3 mm and the diameter is 16 mm, the transducer height could be 7 mm. This maximum size was preliminary known from a pilot study when the final size of the BCI transducer was decided (22). From a design point of view, it would be difficult to decrease the size of the transducer much further without reducing low-frequency sensitivity, and furthermore, the BCI transducer does not need to be made smaller to fit more patients. The present BCI transducer was designed with a maximum diameter of 15.5 mm and a height of 6.4 mm as shown in Figure 3, and would hence statistically fit more than 95% of the patients with good margin. The Bonebridge by MED-EL (14) requires a larger space in the temporal bone (diameter 15.8 mm, depth 8.7 mm). If disregarding the screws, the implant would fit around 40% of the subjects included in this study (see Fig. 6B). The screws must also be taken into consideration to provide enough space on opposite sides of the transducer to fit them. The segmentation method and the method of finding maximum implant diameter and insertion depth was regarded as reliable, however with the following considerations: 1) The CT exams were performed at different machines with different settings and different quality of the images. In the highly processed images, the edges were enhanced, which might have caused tissue next to compact bone to appear as air or bone. 2) The bone next to dura was at some places too thin to be visible in the CT images; hence, the limits between tissue and bone were seen as limits between tissue and air. The thin bone in the air cell system was sometimes interpreted as the surrounding tissue, which complicated the segmentation at some places. 3) The middle ear had no visible limit towards the air cells in the temporal bone; it was possible for the segmentation algorithm to interpret the most medial air cells in the temporal bone as middle ear. The first consideration could affect the maximum diameter both ways, and therefore, the errors were expected to be evened out. The second and third points might have given smaller maximum diameter of the transducer, but these effects were considered small and negligible. Therefore, these limitations are not likely to affect the conclusions of this study. Furthermore, visual inspection of the results of all subjects was conducted to control all measures accomplished with the automated method.
CONCLUSION It is found in this study that a cylindrical transducer casing with maximum diameter of 16 mm can have a maximum insertion depth of 4 mm to statistically fit more than 95% of normal temporal bones. If the transducer is protruding the bone surface 3 mm, the height of the same transducer can be 7 mm. The present BCI transducer casing is designed to comply with the maximum size found in this study, that is, it will be sufficiently small to fit in most healthy adult temporal bones, and it should be placed with its center 20 mm behind the ear canal. Acknowledgments: The authors thank the master thesis students Alireza Arjomand and Siavash Esmaeili Fashtakeh who performed the pilot study.
REFERENCES 1. von Be´kesy G. Experiments in Hearing. Toronto: McGraw-Hill Book Company, Incorporated, 1960. 2. Stenfelt S, Goode R. Bone conducted sound: physiological and clinical aspects. Otol Neurotol 2005;26:1245Y61. 3. News and research in bone conduction hearing - craniofacial osseointegration [press release]. Mo¨lnlycke: Cochlear Bone Anchored Solutions AB; 2011 [June 11] Available from: http:// www.cochlear.com/wps/wcm/connect/0a7aa3b7-8ad8-40ab-b28820054c04e3e2/en-Baha-Ospdf?MOD=AJPERES&CONVERT_ TO=url&CACHEID=0a7aa3b7-8ad8-40ab-b288-20054c04e3e2. Accessed December 4, 2014. 4. Oticon Medical. 2014. Oticon MedicalVtaking ideas further. http://www. oticonmedical.com/~asset/cache.ashx?id=7229&type=14&format=web. Accessed March 4, 2015. 5. Tjellstro¨m A, Håkansson B, Granstro¨m G. Bone-anchored hearing aids: current status in adults and children. Otolaryngol Clin North Am 2001;34:337Y64. 6. Snik A, Mylanus E, Proops D, et al. Consensus statements on the BAHA system: where do we stand at present? Ann Otol Rhinol Laryngol Suppl 2005;195:2Y12. 7. Wallberg E, Granstro¨m G, Tjellstro¨m A, Stalfors J. Implant survival rate in bone-anchored hearing aid users: long-term results. J Laryngol Otol 2011;125:1131Y5. 8. Dun CAJ, Faber HTF, de Wolf MJF, Mylanus EAM, Cremers CWRJ, Hol MKS. Assessment of more than 1,000 implanted percutaneous bone conduction devices: skin reactions and implant survival. Otol Neurotol 2012;33:192Y8. 9. Hobson JC, Roper AJ, Andrew R, Rothera MP, Hill P, Green KM. Complications of bone-anchored hearing aid implantation. J Laryngol Otol 2010;124:132Y6. 10. Håkansson B, Eeg-Olofsson M, Reinfeldt S, Stenfelt S, Granstro¨m G. Percutaneous vs transcutaneous bone conduction implant system: a feasibility study on a cadaver head. Otol Neurotol 2008;29:1132Y9. 11. Håkansson B, Reinfeldt S, Eeg-Olofsson M, et al. A novel bone conduction implant (BCI): engineering aspects and pre-clinical studies. Int J Audiol 2010;49:203Y15. 12. Taghavi H, Håkansson B, Reinfeldt S. A Novel Bone Conduction Implant SystemVAnalog Radio Frequency Data and Power Link Design. Proceedings of the 9th IASTED International Conference on Biomedical Engineering, Innsbruck, Austria, 2012a:327Y35. 13. Taghavi H, Håkansson B, Reinfeldt S. Analysis and design of RF power and data link using amplitude modulation of class-E for a novel bone conduction implant. IEEE Trans Biomed Eng 2012b;59:3050Y9. 14. MedEl 2014. BONEBRIDGEiVInformation for Surgeons (with BCI lifts). Available at http://s3.medel.com/pdf/INT/information_surgeons_ bci-lifts.pdf. Accessed March 4, 2015.
Otology & Neurotology, Vol. 36, No. 4, 2015
Copyright © 2015 Otology & Neurotology, Inc. Unauthorized reproduction of this article is prohibited.
FEASIBLE SIZE OF BCI TRANSDUCER IN TEMPORAL BONE 15. Håkansson B, Brandt A, Carlsson P, Tjellstro¨m A. Resonance frequencies of the human skull in vivo. J Acoust Soc Am 1994;95: 1474Y81. 16. Håkansson B. The balanced electromagnetic separation transducer: a new bone conduction transducer. J Acoust Soc Am 2003;113: 818Y25. 17. Eeg-Olofsson M, Stenfelt S, Tjellstro¨m A, Granstro¨m G. Transmission of bone-conducted sound in the human skull measured by cochlear vibrations. Int J Audiol 2008;47:761Y9. 18. Reinfeldt S, Håkansson B, Taghavi H, Eeg-Olofsson M. Bone conduction hearing sensitivity in normal hearing subjectsVtranscutaneous stimulation at BAHA versus BCI Position. Int J Audiol 2014;53: 360Y9.
637
19. Eeg-Olofsson M, Håkansson B, Reinfeldt S, et al. The bone conduction implantVfirst implantation, surgical and audiological aspects. Otol Neurotol 2014;35:679Y85. 20. Cinamon U. The growth rate and size of the mastoid air cell system and mastoid bone: a review and reference. Eur Arch Otorhinolaryngol 2009;266:781Y6. 21. Swarts JD, Doyle BM, Doyle WJ. Relationship between surface area and volume of the mastoid air cell system in adult humans. J Laryngol Otol 2011;125:580Y4. 22. Arjomand A, Fashtakeh S. Geometrical Analysis of Temporal Bones for Bone Conduction Implants (BCI). Master of Science Thesis. Go¨teborg, Sweden: Chalmers University of Technology, 2009. EX084/2009.
Otology & Neurotology, Vol. 36, No. 4, 2015
Copyright © 2015 Otology & Neurotology, Inc. Unauthorized reproduction of this article is prohibited.
