Introduction

An underpneumatized mastoid bone has been shown to be related to atelectatic ear diseases, cholesteatoma and chronic otitis media with effusion [4, 12, 14]. However, it is not proven whether underdeveloped mastoid cells are a cause or consequence of otitis media.

In most previous studies, conventional X-ray (Schuller) films were used to show mastoid pneumatization. However, evaluation of the mastoid air cell system with computed tomography (CT) seems superior to conventional radiography [19]. Digital image processing programs can be used to measure mastoid pneumatization on CT examination.

A stereological technique based on Cavalieri’s principle has long been used in histological and pathological studies to obtain accurate and unbiased measures of a quantity in an anatomically defined volume, such as total number of particles as well as volume-weighted densities. Cavalieri’s principle is based on the theory that the volume of any structure can be estimated by cutting it into thin parallel slices, measuring the cross-sectional area of the structure in each slice, summing these areas, and multiplying the sum by the mean slice thickness [9].

In this study, we assessed mastoid pneumatization and variables of the sigmoid sinus in the healthy and diseased ears. This study differs from previous ones because we measured the size of the mastoid air cells rather than measuring the whole mastoid bone area.

Materials and methods

Thirty-eight patients with unilateral chronic otitis media (COM) were included in the study. Fifteen of them had unilateral COM with cholesteatoma (COM/+) and 23 had unilateral COM without cholesteatoma (COM/−). The contralateral ears of the patients were healthy. The diagnoses were made according to otoscopy, audiometry and CT. The patients had no history of previous surgical intervention for otological disease. In these patients, the ears with COM comprised the diseased mastoid group. Their contralateral normal ears comprised the dependent healthy mastoid group. Forty temporal bones of 20 subjects who had bilateral healthy ears and no history of previous chronic ear disease or surgery were also included in the study, and comprised the independent healthy mastoid group. In total there were 23 female and 35 male patients with ages ranging from 17 to 67 (mean 32.2) years.

CT scans of the temporal bones of the subjects were obtained with a Somatom Plus AR HP spiral scanner (Siemens Medical Systems, Erlangen, Germany). Contiguous sections 1 or 2 mm thick were obtained (120 kV, 200 mA) parallel to the orbitomeatal line, and the following assessments were made: (1) cross-sectional area of the air cells within the mastoid bone as shown in Fig. 1 (including the regions between the geniculate ganglion level, anterior border of the sigmoid sinus and mastoid apex); (2) the shape of the sigmoid sinus, i.e., whether it was of the protrusive, half-moon or saucer type as described previously [11] (Fig. 2); (3) the shortest distance between the spine of Henle (suprameatal spine) and the sigmoid sinus (Fig. 2a).

Fig 1
figure 1

Right (a) and left (b) axial CT sections of a 25-year-old woman with left chronic otitis media. The cross-sectional area of the mastoid air cells was estimated using a quantitative digital image processing program

Full size image
Fig. 2
figure 2

Types of the sigmoid sinus: a saucer type; b half-moon type; c protrusive type. In the CT images, the shortest distance (between the most prominent edge of the sinus and the spine of Henle) was selected between the spine of Henle (suprameatal spine) and sigmoid sinus in a

Full size image

Using a quantitative digital image processing program on temporal bone CT examination, the air cell area of the mastoid bone only was calculated; the area of the bony portion of the mastoid was not included in the calculation. The CT program was adjusted so that it could specifically measure the areas of air or soft tissue densities within the selected borders. Selection of the soft tissue areas was made in an effort to include the air cells that had been attacked by the inflammatory process so that the actual air cell area could be measured.

This technique allowed us to perform a quantitative analysis of the air cell area. For this purpose, a semi-automatic segmentation method was developed and threshold-based segmentation was used to isolate the mastoid air cell regions to be measured. We initially selected a mastoid air cell threshold value where the pixels above this value represented the bony structure. We used −1,000 to + 70 HU as an air or soft tissue threshold (upper value of attenuation defining the areas of bone: +70 to +1,000 HU). The issue of choosing between −1,000 and +70 HU was considered important for selecting both the air and soft tissue filled mastoid air cells.

In 44 temporal bone CT examinations of 22 patients (10 COM/+, 12 COM/−), the volume of the mastoid bone was estimated using the morphometric method of Cavalieri. This method is based on the principle developed by the Italian mathematician Bonaventura Cavalieri (1598–1647) in 1635, and subsequently updated [5]. This basic principle is a powerful and simple approach for volume estimation of sectioned material [6, 7]. The CT images of the temporal bones were projected on a screen for magnification, and the volume of mastoid bone was measured according to Cavalieri’s principle [5, 6, 7, 8, 13].

SPSS (Statistical Package for the Social Sciences) 8.0 for Windows was used for the statistical analyses. Paired-sample t-test and Spearman correlations were used to compare the results of the diseased mastoid group and dependent healthy mastoid group. A Mann-Whitney U-test or Spearman correlations were used to compare the results of the diseased mastoid group and independent healthy mastoid group.

Results

Air cell area and sigmoid to Henle distance

In COM/+ patients, the distance and mastoid area were greater in the dependent healthy mastoids than in the diseased mastoids (p=0.002 and p=0.0005, respectively). In COM/− patients, the distance and air cell area were greater in the dependent healthy mastoids than the diseased mastoids (p=0.009 and p=0.0003, respectively). The distance and area were significantly greater in the independent healthy mastoids than in the dependent healthy mastoids (p=0.004 and p=0.02, respectively) and diseased mastoids (p=0.0001 and p=0.0003) (Table 1).

Table 1 Mastoid size related parameters in the patients and controls (H-SS spine of Henle to sigmoid sinus, COM/+ chronic otitis media with cholesteatoma, COM/− chronic otitis media without cholesteatoma, DM diseased mastoid of the patients with COM, DHM dependent healthy mastoid or contralateral healthy mastoid of the patients with COM, IDHM independent healthy mastoid of the control subjects)
Full size table

Air cell volume

In both COM/+ and COM/− patients, the mastoid volume was greater in dependent healthy mastoids than in the diseased mastoids (t=−3.559, p<0.01, and t=−6.534, p<0.01, respectively). In COM/+ and COM/− patients, the volumes of diseased mastoids were not significantly different (p>0.05). The healthy mastoid volumes of COM/+ and COM/− patients were not significantly different either (p>0.05). In dependent healthy mastoids, there was a significant correlation between the sigmoid to Henle distance and mastoid volume (r=0.592, p=0.04). In the diseased mastoids, there was no correlation between the sigmoid to Henle distance and mastoid volume (r=0.173, p=0.4).

Sigmoid shape

The half-moon, protrusive and saucer types of sigmoid shape were encountered in 73 (total 62%; right 32%, left 30%), 26 (total 22%; right 11%, left 11%) and 19 (total 16%; right 7%, left 9%) temporal bones, respectively (Table 2). Whatever the sinus shape, the area of the mastoid air cells in the diseased mastoids was smaller than in the independent healthy mastoids (p<0.001) (Table 3).

Table 2 Sigmoid sinus shapes in the patients and controls (COM/+ chronic otitis media with cholesteatoma, COM/− chronic otitis media without cholesteatoma, IDHM independent healthy mastoid of the control subjects)
Full size table
Table 3 Sigmoid shape versus mastoid air cell size in the patients and healthy controls
Full size table

Discussion

Mastoid air cells act as a buffer to maintain the middle ear partial gas pressure [16]. In this context, it would be convenient to measure the mastoid air cell size instead of measuring the size of the whole mastoid bone. For this reason, the overall size of the individual mastoid air cells was measured in this study, excluding the compact bony parts of the mastoid bone.

Quantitative CT techniques have been used for other purposes, such as in emphysema patients who are candidates for lung volume reduction surgery by resection of relatively non-functioning areas of the pulmonary tissue [10]. Using a modification of this CT technique it was possible to estimate the individual air cell area in the mastoid by excluding the compact areas of the bone. The CT program selected the areas of air and soft tissue densities. Selection of the soft tissue areas facilitated including the air cells that had been attacked by an inflammatory process. Choosing the threshold values between −1,000 and +70 HU was considered important when selection of the air and soft tissue filled mastoid air cells was attempted. Thus, it was possible to measure the area of the air cells that were obliterated by inflammatory or soft tissue.

The sigmoid sinus develops in fetal life, but pneumatization of the mastoid continues after birth until puberty. We hypothesize that the shape of the sigmoid or its distance from the external ear canal is one of the factors that could determine to what extent the mastoid bone and its air cells develop after birth. However, middle ear infections in childhood that block the mastoid air cells through inflammation may further contribute to the arrest of mastoid pneumatization [1]. Inflammatory changes in the middle ear mucosa have been shown to affect mastoid pneumatization [2]. Therefore, both genetic and environmental factors appear to be associated with COM.

It was previously shown that cholesteatoma is associated with poor mastoid pneumatization [14, 17]. In this study, the distance of the spine of Henle to the sigmoid was significantly shorter in the diseased mastoids than in the dependent healthy mastoids, and in both these groups this distance was shorter than it was in the independent healthy mastoids. The same was also true regarding mastoid air cell size. Compared with independent healthy mastoids, there was a bilaterally diminished air cell area in both diseased mastoids and dependent healthy mastoids. Compared with dependent healthy mastoids, the poor pneumatization of the diseased mastoids suggests that ear infections after birth may further contribute to arrest of mastoid pneumatization.

Previous studies have shown a relationship between mastoid pneumatization and sigmoid sinus position [3, 18, 20]. In this study, there were also several sorts of relationships between the sigmoid sinus and mastoid air cell area. Half-moon was the most common sigmoid sinus shape in all groups. This was followed by protrusive and saucer types in all groups except for the independent healthy mastoid group, in which the saucer type was more frequent than the protrusive. Our results showed a relationship between air cell size and sigmoid shape, that is, as the sigmoid became more protrusive the air cell size diminished.

Conclusion

A digital image processing CT program allowed us to estimate the individual area of the air and soft tissue filled mastoid air cells. The mastoid size in both intact and diseased ears of COM patients was smaller than in the healthy controls. Mastoid size may be determined genetically. However, environmental factors such as infection may also affect mastoid size. Therefore, both genetic and environmental factors may be related to chronic otitis media as far as mastoid size is concerned.