Quantitative Analysis of the Cochlea using Three-Dimensional Reconstruction based on Microcomputed Tomographic Images

Authors


Correspondence to: Ki-Seok Koh, Department of Anatomy, Research Institute of Medical Science, Konkuk University School of Medicine, 1 Hwayang-dong, Gwangjin-gu, Seoul 143–701, Republic of Korea. Fax: +82-2-2030–7845. E-mail: kskoh@kku.ac.kr

ABSTRACT

The aim of this study was to provide data on various dimensions of the normal cochlea using three-dimensional reconstruction based on high-resolution micro-CT images. The petrous parts of 39 temporal bones were scanned by micro-computed tomography (CT) with a slice thickness of 35 μm. The micro-CT images were used in reconstructing three-dimensional volumes of the bony labyrinth using computer software. The volumes were used to measure 12 dimensions of the cochlea, and statistical analysis was carried out. The dimensions of cochleae varied widely between different specimens. The mean height and length of the cochlea were 3.8 and 9.7 mm, respectively. The angle between the basal and middle turns was slightly larger in males than in females, while none of the other 11 dimensions differed significantly between males and females. The cochlear accessory canals were observed in about half of the cases (51.3%). Correlation analysis among measured items revealed positive correlations among several of the measured dimensions. The present study could investigate the detailed anatomy of the normal cochlea using high-resolution imaging technologies. The results of the present study could be helpful for the precise diagnosis of congenital cochlear malformations and for producing optimized cochlear implants. Anat Rec, 296:1083–1088, 2013. © 2013 Wiley Periodicals, Inc.

INTRODUCTION

The inner ear is a complex structure within the petrous part of the temporal bone that is involved in the hearing and balance senses. It includes the cochlea, vestibule, semicircular canals, and membranous labyrinth that is organized into membranous sacs and ducts interconnected within the bony labyrinth (Kirk and Gosselin-Ildari, 2009). Technological developments in imaging methods, such as high-resolution computed tomography (CT), have made it possible to characterize the inner ear structures (Sennaroglu and Saatci, 2002). Several related studies, including those of inner ear malformations of patients with congenital hearing loss, have suggested reference values for the inner ear structures by analyzing conventional CT images (Purcell et al., 2003; Krombach et al., 2005; Blaser et al., 2006; Chen et al., 2008; Lan et al., 2009; Giesemann et al., 2011). However, conventional CT images cannot provide complete information about the anatomical features of the inner ear due to limitations such as the spatial resolution, slice scanning thickness, and the three-dimensional (3D) complexity of the structures of the inner ear (Green et al., 1990; Mason et al., 2000; Krombach et al., 2005; Ma et al., 2008). Although conventional CT images can be used to detect severe inner ear malformations such as the Michel deformity and cochlear aplasia, their use in identifying subtle malformations such as cochlear hypoplasia is challenging (Jackler et al., 1987; Dimopoulos and Muren, 1990; Purcell et al., 2003). It was reported that one-third of these less-severe malformations are not detected on conventional CT images (Johnson and Lalwani, 2000). However, micro-CT images might be alternatives to overcome the resolution and slice-thickness limitations of conventional CT images. Additionally, various 3D reconstruction programs have been developed to improve the understanding of the tiny and complex inner ear (Mason et al., 2000; Verbist et al., 2009; Jang et al., 2011; Martinez-Monedero et al., 2011; Mukherjee et al., 2011). In the present study, the cochlea was reconstructed three dimensionally after scanning temporal bones with micro-CT, with the aim of providing data on various dimensions of the cochlea that could be used in the diagnosis of congenital cochlear malformations and helpful for designing optimized cochlear implants.

MATERIALS AND METHODS

The petrous parts of the 39 temporal bones included in this study (19 right and 20 left sides) were obtained from 23 cadavers fixed with formalin (12 males, 11 females; mean age 65.9 years, range 23–96 years). The specimens were wrapped to prevent drying. Serial micro-CT images of the specimens were acquired in the axial, coronal, and sagittal planes with a slice thickness of 35 µm using an in vivo X-ray microtomograph (model 1076, Skyscan, Belgium). All of the processes of micro-CT image manipulation and reconstruction of the 3D volume of the bony labyrinth performed using Lucion software (Cybermed, Seoul, Republic of Korea) were identical to those used in our previous study of the semicircular canal (Lee et al., 2013). The software was used to measure 12 dimensions of the cochlea; these measurements were made three times for each dimension in each image in order to reduce measurement errors, and the results were averaged. Several landmarks including the round window niche (RW niche) as described previously (Biedron et al., 2009; Erixon et al., 2009; Martinez-Monedero et al., 2011) were used to measure the dimensions of the cochlea. The midpoint of the RW niche was used as the starting point of the cochlea, and the starting line of the cochlea was defined as a line from the midpoint of the RW niche to the central axis of the cochlea (Fig. 1A).

Figure 1.

Measurement items of the cochlea. Landmarks and three turns: basal turn (red line), middle turn (blue line), and apical turn (green line) (A). (1) Height of the cochlea (HC), (2) height of basal turn (HBT), (3) height of apical turn (HAT) (B). (4) Height of middle turn (HMT) (C). (5) Angle between basal turn and middle turn (ABM) (D). (6) Angle between middle turn and apical turn (AMA) (E). (7) Length of the cochlea (LC), (8) width of basal turn (WBT), (9) width of middle turn (WMT) (F). (10) Length of cochlear coiling (LCC) (G). (11) Number of cochlear turn (NCT) (H). (12) Angle of accessory canal coiling (ACC) (I). RW niche, round window niche.

The following dimensions were measured (Fig. 1):

  1. Height of the cochlea (HC): on a line perpendicular to the basal turn along the midline of the cochlea.
  2. Height of the basal turn (HBT): on the same line as for HC.
  3. Height of the apical turn (HAT): on the same line as for HC.
  4. Height of the middle turn (HMT): on the same line as for HC.
  5. Angle between the basal and middle turns (ABM): between a line parallel to the axis of the basal turn and another line parallel to the axis of the middle turn.
  6. Angle between the middle and apical turns (AMA): between a line parallel to the axis of the middle turn and another line parallel to the axis of the apical turn.
  7. Length of the cochlea (LC): on a line drawn from the midpoint of the RW niche to the polar-opposite point through the central axis of the cochlea.
  8. Width of the basal turn (WBT): on a line perpendicular to LC through the central axis of the cochlea.
  9. Width of the middle turn (WMT): on the same line as for WBT.
  10. Length of cochlear coiling (LCC): a curved line drawn from the starting line of the cochlea to the terminal point of the apical turn through the middle of the three turns.
  11. Number of cochlear turns (NCT): angle between the starting line of the cochlea and another line from the central axis of the cochlea to the terminal point of apical turn.
  12. Angle of accessory canal coiling (ACC): between the starting line of the cochlea and another line from the central axis of the cochlea to the terminal point of the cochlear accessory canal.

To standardize the same position among specimens during measurements, the anteromedial view (Fig. 1B,C) and the inferior view (Fig. 1D,E) of the cochlea were defined as when the axis of the basal turn was aligned with a horizontal line on the software and when HC and ABM were largest on the views, respectively. The anterolateral view (Fig. 1F,G,H,I) was defined as when both LC and WBT were largest. Measurements were performed using a 3D ruler function of the software to avoid errors due to changes in the viewing angle. Independent-samples t-tests between both genders and sides, and correlation analysis among measured dimensions were performed using SPSS Statistics (ver. 19, SPSS, NY).

RESULTS

Large variations were observed in the measured dimensions. HC was 3.8 mm and LC was 9.7 mm; other dimensions related to height, width, and LCC are listed in Table 1. HBT, HMT, and HAT were 1.9, 1.8, and 0.7 mm, respectively (corresponding to 49.5, 47.9, and 19.4% of HC) (Table 2). The mean NCT was 2.54 turns, and ranged from 2.36 to 2.80 turns (mean 916.2 degrees; range 850.7–1007.7 degrees). Statistically, all of the measurements conformed to a normal distribution. There was no significant difference between both sides, and all dimensions except for ABM were not different between both genders. ABM was slightly larger for males than for females (P < 0.05) (Table 3).

Table 1. Height, width, length of the cochlea, and length of cochlear coiling (mm)
 Total (N = 39)Male (N = 22)Female (N = 17)
MeanSDMeanSDMeanSD
  1. HC, height of the cochlea; HBT, height of basal turn; HMT, height of middle turn; HAT, height of apical turn; LC, length of the cochlea; WBT, width of basal turn; WMT, width of middle turn; LCC, length of cochlear coiling.

HC3.80.23.90.23.80.1
HBT1.90.11.90.11.90.1
HMT1.80.21.80.21.90.2
HAT0.70.10.70.10.80.1
LC9.70.39.80.39.70.3
WBT7.00.37.00.37.00.3
WMT3.90.23.90.23.80.2
LCC30.01.630.11.329.91.9
Table 2. The ratio among several dimensions of the cochlea (%: a/b × 100)
 Total (N = 39)Male (N = 22)Female (N = 17)
MeanSDMeanSDMeanSD
  1. HC, height of the cochlea; LC, length of the cochlea; HBT, height of basal turn; HMT, height of middle turn; HAT, height of apical turn.

HC/LC39.62.139.62.439.61.6
HBT/HC49.52.549.92.749.02.1
HMT/HC47.94.247.44.448.53.9
HAT/HC19.42.419.02.520.02.3
HMT/HBT96.68.394.77.499.19.1
HAT/HBT39.66.538.46.841.15.8
HAT/HMT41.37.141.08.341.85.4
Table 3. The number of cochlear turn and angles between cochlear turns (unit of NCT: turn; unit of ABM; and AMA: degree)
 Total (N = 39)Male (N = 22)Female (N = 17)
MeanSDMeanSDMeanSD
  1. NCT, the number of cochlear turn; ABM, angle between basal turn and middle turn; AMA, angle between middle turn and apical turn.

  2. a

    Statistical difference between male and female (P < 0.05).

NCT2.540.092.530.102.570.07
ABMa9.02.69.92.87.91.7
AMA2.40.82.50.92.40.7

The cochlear accessory canal was observed in 20 of the 39 specimens (51.3%). The mean ACC was 114.7 degrees, and ranged from 71.4 to 137.7 degrees (Table 4). Additionally, the structure was observed in seven of the 12 males (58.3%) and in seven of the 11 females (63.6%), and occurred bilaterally in two of the males (28.6%) and four of the females (57.1%).

Table 4. The angle of the accessory canal coiling (degree)
 Total (N = 20)Male (N = 9)Female (N = 11)
MeanSDMeanSDMeanSD
  1. ACC, angle of the accessory canal coiling.

ACC114.717.8106.617.3121.316.0

Correlation analysis of the measured dimensions revealed that LC was positively correlated with WBT, WMT, and LCC (P < 0.05), and HBT was positively correlated with ABM (P < 0.05).

DISCUSSION

Imaging technologies such as CT and magnetic resonance imaging have previously been used to identify detailed anatomical structures of the temporal bone (Naganawa et al., 2002, 2003). However, the complex anatomy of the cochlea may not always be understood on conventional CT images (Erixon et al., 2009). Therefore, the present study used micro-CT and a 3D reconstruction program to obtain exact reference values for 12 cochlea dimensions.

The actual starting point of the cochlea which is located within the hook region near the round window could not be shown on the outer surface of the cochlea (Erixon et al., 2009). The RW niche lies posteroinferior to the promontory overlying the basal turn of the cochlea, and the round window membrane is positioned within the RW niche (Gulya, 2007). Therefore, using the midpoint of the RW niche (which is a bony structure) as a reference point could be optimal because it is the closest clear landmark to the actual starting point of the cochlea on the anterolateral view (Fig. 1A). Furthermore, this reference point has the advantages of representing a suitable criterion for measuring various dimensions such as LC, WBT, LCC, NCT, and ACC (Fig. 1), and for allowing comparisons with previous studies that used the same point (Biedron et al., 2009; Erixon et al., 2009; Martinez-Monedero et al., 2011).

Previous studies have measured the height of the cochlea on conventional CT images (Purcell et al., 2003; Krombach et al., 2005; Blaser et al., 2006; Giesemann et al., 2011) or on images of a corrosion casting model (Erixon et al., 2009). However, those approaches had limitations for measuring some of the cochlear dimensions, such as HMT and HAT, since the border between the middle and apical turns was indistinct. Accurate information about the height of the three turns is very important in the design of cochlear implants, and the actual height of three turns in the present study was first measured as far as we know using the 3D reconstructed volume (Fig. 1B,C). Surprisingly, the actual HMT was almost identical to HBT, and the height decreased sharply where the apical turn was coiled (Tables 1 and 2; Fig. 1C).

HC was slightly shorter (3.8 mm) in the present study than previously reported values of 3.93 mm (Dimopoulos and Muren, 1990) and 3.9 mm (Erixon et al., 2009). LC was slightly larger than that found by Giesemann et al. (2011), while NCT was in accordance with the measurements of Hardy (1938), Erixon et al. (2009), and Biedron et al. (2009). We consider that such variations between cochlear studies might be due to differences in methodology. Many previous studies that used conventional CT images found HC to be as large as 4–5 mm. However, these results may be inaccurate due to the relatively low resolution and indistinct boundaries of the cochlea on a CT image.

Correlation analysis of the measured dimensions revealed that LC had strong positive correlations with WBT, WMT, and LCC. Dimopoulos and Muren (1990) also described a positive correlation between LC and WBT. Combining these results could explain the coiling characteristic of the cochlea. The cochlea with a longer LC is less compact because the basal turn is distally coiled from the cochlea starting point, which would result in larger WBT, WMT, and LCC (Fig. 2). Erixon et al. (2009) differentiated the coiling characteristic into “distal coiling” and “proximal coiling” based only on the LC variability. However, we consider that LCC also influences the coiling characteristic due to the positive correlation between LC and LCC, with a cochlea with a longer LCC having the distal coiling characteristic.

Figure 2.

The coiling patterns of the cochlea. LC, length of the cochlea; WBT, width of basal turn; WMT, width of middle turn; LCC, length of cochlear coiling. The cochlea with relatively longer LC represented less compact coiling by the distal coiling of basal turn (A), but shorter LC showed more compact coiling form by the proximal coiling of basal turn (B).

The positive correlations between HBT and ABM could explain the large variation in the angles between the turns. ABM is larger for a cochlea with a larger HBT, which in turn indicates that the thickness of the basal turn increases, and the donut-like space formed by the basal turn is relatively narrow. This results in the axis of the middle turn becoming more slanted due to spatial limitations when the middle turn is coiled (Fig. 3).

Figure 3.

Two types of ABM. HBT, height of basal turn; ABM, angle between basal turn and middle turn. As HBT increase the thickness of basal turn is thickened, and the donut-like space which is formed by basal turn will be narrowed (A). It is thought that the cochlea with relatively larger HBT had larger ABM (B) than in the cochlea with relatively lesser HBT (C) due to spatial limitation where middle turn will be placed.

The accessory canal was observed in both the 3D volume and the micro-CT image. We named this structure the “cochlear accessory canal,” which has never been shown in previous studies. In all cases, this canal started from the vestibule, but its terminal point differed among the specimens. Moreover, the structure started from the posteromedial side of the basal turn near to the vestibule and gradually ran toward the anterior of the basal turn and was a clear anatomical unit separated by a bony septum (Fig. 4). Previous studies might not have observed the accessory canal due to either its very small size or its border with the cochlea being obscured on conventional CT images. In contrast, we found that the structure could be clearly identified using micro-CT and 3D reconstruction in the present study. However, our histological investigation revealed that the accessory canal had no specific structure such as a membranous labyrinth. Because the accessory canal was thought to be a developmental remnant and was observed in seven of the 12 males (58.3%) and in seven of the 11 females (63.6%), it could be regarded as a nonfunctional structure and a normal variation, rather than as a pathological malformation.

Figure 4.

The detection of the cochlear accessory canal. The accessory canal which could be both observed in three-dimensional volume (A) and micro-CT image (B) of the cochlea. The traveling course of the accessory canal is shown (red arrows). The accessory canal is an anatomical structure obviously distinct from basal turn (white arrow).

The precise diagnosis of cochlear malformations using CT imaging is very important to achieving a successful surgical outcome because variations in their morphological characteristics greatly influence the optimal surgical strategy. The present study has yielded data on accurate dimensions of the normal cochlea and has identified detailed anatomical variations using micro-CT and 3D reconstruction. Although it is impossible to apply micro-CT to patients, the data obtained in the present study could be used to determine quantitative criteria for defining the normal cochlea, which would be useful for otology surgeons diagnosing cochlear malformations and for designers of cochlear implants. In addition to clinical applications, the reported data on accurate dimensions of the cochlea will facilitate future morphological and functional studies of the inner ear structure.

Ancillary