Abstract
Purpose
This study was performed to evaluate the usefulness of Hounsfield unit (HU) to better distinguish cholesteatoma from other inflammatory conditions in the mastoid ad antrum before primary mastoid surgery.
Materials and methods
We enrolled 82 patients who underwent tympanomastoidectomy for treatment of chronic otitis. Forty-one patients were pathologically diagnosed with cholesteatoma, whereas the others were diagnosed with inflammatory granulation. These lesions were confirmed, and HU was measured in preoperative computed tomography. The difference in HU between cholesteatoma and non-cholesteatoma tissues was analyzed, and the improvement in the diagnosis of cholesteatoma after inclusion of HU data was calculated.
Results
The HU was calculated as 42.68 ± 24.42 in the cholesteatoma group and 86.07 ± 26.50 in the non-cholesteatoma group. The differences between the 2 groups were statistically different (Student t test, P < .01). By applying the HU, the sensitivity (51.2%–80.5%), specificity (80.5%–87.8%), positive predictive value (72.4%–86.8%), and negative predictive value (62.3%–81.8%) to diagnose cholesteatoma improved.
Conclusions
The HU density was found to be statistically different between cholesteatoma and inflammatory granulation tissue in mastoid antrum. An improved diagnosis of cholesteatoma was achieved after adjusting for the HU.
1
Introduction
Preoperative identification of cholesteatoma is important because it can determine the necessity of surgery and the surgical method. In non–contrast-enhanced temporal bone computed tomography (CT), which is the most commonly used method for the preoperative evaluation, soft tissue lesion combined with bony erosion is diagnosed with cholesteatoma . The cholesteatoma may erode the scutum, ossicles, semicircular canals, and the bony facial canal. In combination with the erosion of the ossicles, it occurs in approximately 90% of cases of pars tensa cholesteatoma and in approximately 75% of patients diagnosed with pars flaccida cholesteatoma . However, non–contrast-enhanced temporal bone CT has its limitation with respect to differential diagnosis in early cholesteatoma, which emerges as a soft tissue lesion and does not show characteristic bony destruction patterns.
The Hounsfield unit (HU) density is a basic characteristic of the CT imaging technique. HU is the linear transition product of the linear attenuation coefficient and is often used for biological tissues. The HU of pure water is 0, that of air is −1000, and that of dense bone is +1000. Every biological tissue has its unique HU, which ranges from −1000 to +1000 . Thus, in cases of 2 lesions with similar gross findings in CT, a measurement of the HU of these lesions would afford a more precise differential diagnosis.
This study was performed to evaluate HU differences between cholesteatoma and non-cholesteatoma tissues in non–contrast-enhanced temporal bone CT and to prove the usefulness of HU in distinguishing cholesteatoma from inflammatory granulation tissue in the mastoid ad antrum.
2
Materials and methods
Patients who underwent tympanomastoidectomy for treatment of chronic otitis media were enrolled between December 2006 and February 2009. The cases who received revision surgery were excluded. Tissues of the mastoid antrum were obtained, and all participants with a confirmed pathologic diagnosis of either inflammatory granulation tissue or cholesteatoma were included. We enrolled 41 patients with a pathologic diagnosis of cholesteatoma. As control group, we included 41 patients whose pathologic diagnosis revealed inflammatory granulation tissue. The medical records and CT scans of all enrolled cases were reviewed retrospectively. The male-to-female ratio was 37:45, and the average age of the patients was 48.95 (±13.98) years.
Non–contrast-enhanced, high-resolution, multislice CT examination (GE Medical System, Milwaukee, WI) of the temporal bone was performed using the following parameters of acquisition: collimation width 0.6 mm, 120 kV (peak), and 340 mA. The HU was measured at the mastoid antrum. The size of the region of interest (ROI) was 2.49 mm 2 . In axial slices, we placed the ROI circle at the same level of the lateral semicircular canal in the mastoid antrum ( Fig. 1 ). All measurements were performed by 2 different authors who were blinded to the patients’ data. In addition, the HU was measured in 3 adjacent axial slices (1 slice that contained the suspected lesion, and an adjacent upper and lower slice). The lowest measured HU was used for analysis.
The cutoff point of HU between cholesteatoma and inflammatory granulation tissue was calculated by using the median value from the average HU in the cholesteatoma group and the average HU in the group with inflammatory granulation tissue; it was also the average value of the HU in all patients. The initial CT scan reading was performed by radiologist independently from this study. We revised the patients’ temporal bone CT reading by adjusting for the individual HU score, and the change of diagnosis of cholesteatoma was analyzed.
We performed Student t test and Fisher exact test using SPSS version 13.0 (SPSS Inc, Chicago, IL) for statistical analysis. A P value of less than .05 was considered significant.
This study was approved by the institutional review board of the Seoul National University Boramae Medical Center.
2
Materials and methods
Patients who underwent tympanomastoidectomy for treatment of chronic otitis media were enrolled between December 2006 and February 2009. The cases who received revision surgery were excluded. Tissues of the mastoid antrum were obtained, and all participants with a confirmed pathologic diagnosis of either inflammatory granulation tissue or cholesteatoma were included. We enrolled 41 patients with a pathologic diagnosis of cholesteatoma. As control group, we included 41 patients whose pathologic diagnosis revealed inflammatory granulation tissue. The medical records and CT scans of all enrolled cases were reviewed retrospectively. The male-to-female ratio was 37:45, and the average age of the patients was 48.95 (±13.98) years.
Non–contrast-enhanced, high-resolution, multislice CT examination (GE Medical System, Milwaukee, WI) of the temporal bone was performed using the following parameters of acquisition: collimation width 0.6 mm, 120 kV (peak), and 340 mA. The HU was measured at the mastoid antrum. The size of the region of interest (ROI) was 2.49 mm 2 . In axial slices, we placed the ROI circle at the same level of the lateral semicircular canal in the mastoid antrum ( Fig. 1 ). All measurements were performed by 2 different authors who were blinded to the patients’ data. In addition, the HU was measured in 3 adjacent axial slices (1 slice that contained the suspected lesion, and an adjacent upper and lower slice). The lowest measured HU was used for analysis.
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
