Abstract
To analyze the apparent diffusion coefficient (ADC) values of middle ear and mastoid lesions in Diffusion weighted Magnetic Resonance Imaging (DW-MRI) to arrive at a probable demarcating value to differentiate cholesteatoma from non-cholesteatomatous lesions. Accurate anatomic localization of the lesion was also done using High Resolution Computed Tomography (HRCT) temporal bone. The study cohort consisted of 30 patients who had undergone HRCT, DW-MRI and surgical intervention in clinically suspected cholesteatomatous lesions during the period August 2018 to August 2020.Sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV), and accuracy values of HRCT and MRI in relation to intraoperative findings and histopathological findings (gold standard) were calculated and compared using the 2-sided McNemar’s Chi Square test. Receiver operating characteristic (ROC) curve was used to predict the cut off value of ADC to differentiate between cholesteatoma and non cholesteatomatous lesions. Total patients were 30 out of which 15 were histopathologically proven cholesteatoma. MR DWI showed 100% sensitivity, 80% specificity, and 90% accuracy in diagnosing cholesteatoma compared to HPE. The probable cut off value of ADC in differentiating cholesteatoma from non-cholesteatomatous lesions was found to be < 1.226 × 10–3 mm2/s, statistically using ROC curve. HRCT showed 96.6% accuracy in identifying the location of the lesion. MR-DWI is a useful tool both individually and in combination with HRCT in the diagnosis of cholesteatomas with high accuracy. An ADC cut-off value could also significantly help increase the accuracy of diagnosis.
Keywords: Cholesteatoma, Diffusion weighted imaging, MRI, HRCT
Introduction
Cholesteatoma is a sac which contains keratin debris enveloped by keratinized squamous epithelium in a fibrous tissue matrix which frequently arise within pneumatized parts of the temporal bone. Cholesteatomas are classified as congenital and acquired. Acquired can be primary and secondary. Secondary acquired cholesteatomas are more common and are usually associated with squamous type of chronic otitis media (COM) having attic and marginal perforation [1]. Tos et al. found an annual incidence ranging from 9 to 12.6 cases per 100,000 adults for acquired cholesteatoma [2]. Kemppainen et al. gave an annual incidence ranging from 3 to 15 cases per 100,000 children [3].
Diagnosis of cholesteatoma is essentially clinical however the use of supporting imaging modalities increases the accuracy [4]. Computed tomography (CT) is the standard modality because it provides the ability to identify bony changes. Due to its excellent spatial resolution, it has a high sensitivity but low specificity in mass lesions as it cannot demarcate granulation tissue, cholesterol granulomas, or as other soft tissue growths from cholesteatoma [5, 6]. The imaging findings include a sharply marginated and expansile soft tissue lesion, scutum blunting, and erosion of the tympanic tegmen as well as ossicles, and retraction of the tympanic membrane [7]. Diffusion-weighted Magnetic Resonance Imaging (DW-MRI) is a valuable modality in the study of cholesteatoma as it eliminates the need for contrast injection [8]. DWI is a specialized technique in MRI that measures the molecular diffusion of water within the tissues, which can be quantified using Apparent Diffusion co-efficient (ADC). Restricted diffusion gives a low ADC value and vice versa. Unlike other middle ear pathologies, cholesteatoma causes diffusion restriction and is seen as a hyperintense focus on DW sequence [9]
The availability of published literature on the role of imaging in cholesteatoma especially from the Indian subcontinent was found to be inadeqaute. Hence this study was designed to delineate the relationship between ADC values of cholesteatoma and non-cholesteatomatous lesions and to arrive at a cut off value to distinguish between the two groups.
Materials and Methods
This prospective cross-sectional study was done in a tertiary care centre in South India on patients who were clinically diagnosed with cholesteatoma by the ENT surgeon. They had to be willing to undergo HRCT and MRI as pre-operative investigations, followed by surgery (modified radical mastoidectomy, MRM), and histopathological examination of the excised tissue. Clearance was obtained from institutional scientific review board and the institutional ethical committee prior to conduct of the study. Data set consisted of 30 patients who underwent HRCT and DW- MRI before surgery with subsequent histopathological analysis during the period of 2018–2020.
HRCT was performed on a 256-row multi-detector CT scanner (Brilliance-iCT, Philips Healthcare, Cleveland, OH) using an axial volume scan (section thickness of 0.67 mm, an increment of 0.67 mm, 120 kV) with coronal and sagittal reformations (0.67 mm) parallel and perpendicular to the lateral semi-circular canal). The raw data were reconstructed by using an ultrasharp bone algorithm to provide optimum visualization of the bony anatomy of the temporal bone from the level of superior border of petrous temporal bone till the level of mastoid tip. HRCT findings of tympanic location of the lesion were compared with the intraoperative findings.
All MRI examinations were performed on a 3 T scanner (GE Discovery MR 750 W, Milwaukee, Wisconsin, USA) using a standard head coil. All patients were imaged using the institutional cochlear implant protocol which includes the following sequences: Axial T2 weighted FSE slices, COR T2 weighted FSE, Axial CUBE T2 IAC, Axial T2 FLAIR FS, Axial diffusion weighted propeller images (B 0 and B 1000), Axial T1 3D SPGR, Axial 3D FIESTA-C. The imaging sections covered the temporal bone from the level of superior border of petrous temporal bone till the level of mastoid tip. Propeller DWI is a type of non-echo planar fast spin-echo-based DWI sequence which is performed in axial plane. MRI findings of diffusion restriction and ADC values were compared with the histopathological diagnosis.
The DWI MRI data were transferred to a commercial workstation (Advantage Windows workstation, GE Healthcare, Milwaukee, WI, USA) and then processed using proprietary diffusion analysis software (Functool, GE Medical Systems, Milwaukee, WI, USA). A region of interest (ROI) was manually drawn based on the area with diffusion restriction to include the entire lesion on a slice with the largest axial dimension, and the conventional sequence images were used as an anatomic reference. All the patients underwent surgery within 1 month following MR study.
Statistical Analysis
The statistical analysis was done using the IBM SPSS version 20.0 software (IBM SPSS, USA). To test the statistical significance of the difference in the proportion of 2 different methods (CT findings with Intraoperative findings and MR diffusion restriction with HPE) 2 sided Mc Nemar Chi-Square test was used. ROC curve was used to predict the cut off value of ADC to differentiate between cholesteatoma and non-cholesteatomatous lesions. Diagnostic parameters such as sensitivity, specificity, accuracy, PPV and NPV values of HRCT and MRI were calculated. P value less than 0.05 was considered as significant.
Results
There were a total of 30 patients in the study. Their mean age was 33 ± 16.8 years (Male: Female: 14:16). Among 15 patients with HPE diagnosis of cholesteatoma, all (100%) were reported as having diffusion restriction present in MRI. Among 15 patients with HPE diagnosis other than cholesteatoma (suppurative, granulation tissue, cholesterol granuloma), 3 cases showed diffusion restriction, of these 2 were mastoid abscess and 1 was granulation tissue. DW MRI compared to histopathology for the detection of cholesteatoma showed a sensitivity 100%, specificity 80%, PPV 83.3%, NPV 100%, accuracy 90% and a P value of 0.250 where HPE finding was considered as gold standard (Table 1). ROC curve for ADC value to differentiate cholesteatoma and non-cholesteatoma lesions gave an optimal cut off value of ≤ 0.0012265 mm2/s which was seen to offer the best parameter for diagnosing cholesteatoma using ADC VALUE, with a sensitivity of 100%, specificity of 93.3% and this was found to be statistically significant (Fig. 1).
Fig.2.
Cholesteatoma (A) Axial CT showing soft tissue in middle ear and mastoid with ossicular erosion. (B) Axial DWI showing bright signal in the lesion (C) Axial ADC image showing dark signal in the lesion (D) ROI drawn showing ADC value of 0.84 × 10−3 mm2/s
Table 1.
Comparison of MR diffusion restriction with HPE for cholesteatoma
| MR Diffusion restriction | HPE cholesteatoma | Sensitivity % | Specificity % | Accuracy % | P value | |
|---|---|---|---|---|---|---|
| Yes | No | |||||
| n (%) | n (%) | |||||
| Yes | 15 (100) | 3 (20) | 100 | 80 | 90 | 0.250 |
| No | 0 (0) | 12 (80) | ||||
Fig. 1.
ROC curve for ADC cut off value to differentiate cholesteatoma and non-cholesteatomatous lesions. Area under the curve: 0.964 (95% CI: 0.893 to 1) (p < 0.001)
HRCT showed middle ear soft tissue in tympanic (hypotympanum/mesotympanum) location in 4 (13.3%) patients, attic location in 6 (20%) patients and atticotympanic location in 20 (66.7%) patients. The intraoperative (IOP) findings showed middle ear soft tissue in tympanic location in 3 (10%) patients, attic location in 7 (23.3%) patients, and atticotympanic location in 20 (66.7%) patients. There was no statistically significant difference between the HRCT and IOP findings for the detection of tympanic location of middle ear soft tissue (p = 0.317) with an accuracy of 96.6%.
Discussion
DW MRI is highly useful in accurately giving a diagnosis of cholesteatoma as cholesteatomatous lesions tend to produce diffusion restriction. T1 and T2 sequences were not found to be particularly useful in differentiating cholesteatoma from non-cholesteatomatous lesions.
In our study, MR-DWI was accurately able to diagnose cholesteatoma in more than 90% cases with a high sensitivity (100%) and specificity (80%). There were two false positive cases of mastoid abscess with diffusion restriction. Abscess may be differentiated from cholesteatoma due to its peripheral enhancement on post contrast images while cholesteatomas do not take up contrast media. One out of the 12 HPE proved granulation tissue cases was bright on diffusion weighted imaging, which reduced our specificity. However quantitative ADC values were higher in this case (1.56 × 10–3 mm2/s) suggesting the possibility of “T2 shine through” than true diffusion restriction. This recommends the need for ADC calculation in differentiating cholesteatoma from non-cholesteatoma lesions. Barath et al. reported that cholesteatomas have a high signal intensity, attributable to restricted water diffusion due to an oily consistency of the retained fluid on MR DWI [5]. The sensitivity and specificity of DWI in our study were comparable to the ones reported by Vercruysse (81%and 100% respectively) [9].
Looking at the apparent diffusion coefficient of MR in this study, the mean ADC value (0.688 ± 0.217 × 10−3mm2/s) in the cholesteatoma cases were lower compared to non cholesteatomatous cases (2.001 ± 0.514 × 10−3mm2/s) (p = < 0.001).ROC cut off ADC value of ≤ 1.226 × 10 −3 mm2/s could differentiate cholesteatoma cases from non cholesteatomatous cases with an optimal sensitivity (100%) and specificity (93.3%). The slightly lower ADC value of 0.966 × 10−3mm2/s can also be considered, since the specificity (93.3%) is remaining same with only mild reduction sensitivity of 93.3%. Russo et al. derived a mean ADC value of 0.859 ± 0.276 × 10–3 mm2/s for cholesteatoma and proposed an interval ADC value 0.318 × 10–3 to 1.265 × 10–3 as an appropriate benchmark range for the accurate differentiation of cholesteatoma from non-cholesteatomatous inflammatory lesions including granulation tissue [10].
Finally coming to the HRCT localization of middle ear soft tissue lesion, all the intraoperatively detected atticotympanic and tympanic location of the middle ear soft tissue were correctly identified by HRCT. HRCT correctly localized 6 of the 7 intraoperatively detected attic location. One case of attic soft tissue was misinterpreted as tympanic by HRCT. This was attributed due to the caudally placed attic lesion. Overall accuracy of HRCT in localizing the middle ear soft tissue abnormality was 96.6%. Locketz et al. also reported an accuracy 90% for HRCT which is comparable to our study [11].
HRCT combined with MR DWI can be used as a guiding tool in pre-operative localization of middle ear pathology, so that it can aid in the surgical planning. MR DWI along with ADC values helps in differentiating cholesteatoma from granulation tissue. This is important in the clinical management as recurrent cholesteatoma needs surgical intervention whereas granulation tissue is usually managed conservatively.
Limitations
The sample size was small given the time frame of the study. This could have caused the study to be under-powered to achieve the objective. Lastly, there could have been a selection bias in selecting participants for the study as those patients who opted for non-surgical management could have been the ones with smaller or less aggressive cholesteatoma and hence this could have confounded the results.
Conclusion
MR DWI has good accuracy (90%) in diagnosing cholesteatoma with mastoid abscess being the major confounding factor. ADC cut off value of 1.2 × 10–3 mm2/s can differentiate cholesteatoma and non-cholesteatoma with good sensitivity (100%) and specificity (93.3%).MR DWI with ADC value is a useful tool both individually and in combination with HRCT in the diagnosis of cholesteatomas with a high accuracy.
Funding
No external sources of funding involved in this article.
Declarations
Conflict of interest
Authors have no conflict of interest to declare.
Ethical Standards
All procedures performed in studies involving human participants were in accordance with the ethical standards of the institutional and/or national research committee and with the 1964 Helsinki declaration and its later amendments or comparable ethical standards.
Informed Consent
Informed consent was obtained by all individuals participating in the study.
Footnotes
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Contributor Information
Nazreen Abbass Ayyaril, Email: nazreenniyaz@gmail.com.
Sandya Chirukandath Jayasankaran, Email: sandya_cj@yahoo.com.
Unnikrishnan Menon, Email: unnikrishnanmenon@aims.amrita.edu.
Srikanth Moorthy, Email: smoorthy@aims.amrita.edu.
References
- 1.Rout MR, Mohanty D, Vijaylaxmi Y, Kamalesh B, Chakradhar M. Prevalence of cholesteatoma in chronic suppurative otitis media with central perforation. Indian J Otol. 2012;18(1):7. doi: 10.4103/0971-7749.98280. [DOI] [Google Scholar]
- 2.Tos M (1988) Incidence, etiology and pathogenesis of cholesteatoma in children. Pediatric Otology . Karger Publishers 40: 110-117 [DOI] [PubMed]
- 3.Kemppainen HO, Puhakka HJ, Laippala PJ, Sipilä MM, Manninen MP, Karma PH. Epidemiology and aetiology of middle ear cholesteatoma. Actaoto-laryngologica. 1999;119(5):568–572. doi: 10.1080/00016489950180801. [DOI] [PubMed] [Google Scholar]
- 4.Pusalkar AG. Cholesteatoma and its management. Indian J Otolaryngol Head Neck Surg. 2015;67(3):201–204. doi: 10.1007/s12070-015-0891-y. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Baráth K, Huber AM, Stämpfli P, Varga Z, Kollias S. Neuroradiology of cholesteatomas. Am J Neuroradiol. 2011;32(2):221–229. doi: 10.3174/ajnr.A2052. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Venail F, Bonafé A, Poirrier V, Mondain M, Uziel A. Comparison of echo-planar diffusion-weighted imaging and delayed postcontrast T1-weighted MR imaging for the detection of residual cholesteatoma. Am J Neuroradiol. 2008;29(7):1363–1368. doi: 10.3174/ajnr.A1100. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Corrales CE, Blevins NH. Imaging for evaluation of cholesteatoma: current concepts and future directions. Curr Opin Otolaryngol Head Neck Surg. 2013;21(5):461–467. doi: 10.1097/MOO.0b013e328364b473. [DOI] [PubMed] [Google Scholar]
- 8.Henninger B, Kremser C. Diffusion weighted imaging for the detection and evaluation of cholesteatoma. World J Radiol. 2017;9(5):217–222. doi: 10.4329/wjr.v9.i5.217. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Vercruysse JP, De Foer B, Pouillon M, Somers T, Casselman J, Offeciers E. The value of diffusion-weighted MR imaging in the diagnosis of primary acquired and residual cholesteatoma: a surgical verified study of 100 patients. Eur Radiol. 2006;16(7):1461–1467. doi: 10.1007/s00330-006-0160-2. [DOI] [PubMed] [Google Scholar]
- 10.Russo C, Elefante A, Di Lullo AM, Carotenuto B, D’Amico A, Cavaliere M & Brunetti A (2018) ADC benchmark range for correct diagnosis of primary and recurrent middle ear cholesteatoma. BioMed research international, 2018. [DOI] [PMC free article] [PubMed]
- 11.Locketz GD, Li PM, Fischbein NJ, Holdsworth SJ, Blevins NH. Fusion of computed tomography and PROPELLER diffusion-weighted magnetic resonance imaging for the detection and localization of middle ear cholesteatoma. JAMA Otolaryngol-Head Neck Surg. 2016;142(10):947–953. doi: 10.1001/jamaoto.2016.1663. [DOI] [PubMed] [Google Scholar]