Temporal bone computed tomography findings associated with feasibility of endoscopic ear surgery

Preoperative prediction by artificial intelligence for mastoid extension in pars flaccida cholesteatoma using temporal bone high-resolution computed tomography: A retrospective study

Cholesteatoma is a progressive middle ear disease that can only be treated surgically but with a high recurrence rate. Depending on the extent of the disease, a surgical approach, such as microsurgery with a retroarticular incision or transcanal endoscopic surgery, is performed. However, the current examination cannot sufficiently predict the progression before surgery, and changes in approach may be made during the surgery. Large amounts of data are typically required to train deep neural network models; however, the prevalence of cholesteatomas is low (1-in-25, 000). Developing analysis methods that improve the accuracy with such a small number of samples is an important issue for medical artificial intelligence (AI) research. This paper presents an AI-based system to automatically detect mastoid extensions using CT. This retrospective study included 164 patients (80 with mastoid extension and 84 without mastoid extension) who underwent surgery. This study adopted a relatively lightweight neural network model called MobileNetV2 to learn and predict the CT images of 164 patients. The training was performed with eight divided groups for cross-validation and was performed 24 times with each of the eight groups to verify accuracy fluctuations caused by randomly augmented learning. An evaluation was performed by each of the 24 single-trained models, and 24 sets of ensemble predictions with 23 models for 100% original size images and 400% zoomed images. Fifteen otolaryngologists diagnosed the images and compared the results. The average accuracy of predicting 400% zoomed images using ensemble prediction model was 81.14% (sensitivity = 84.95%, specificity = 77.33%). The average accuracy of the otolaryngologists was 73.41% (sensitivity, 83.17%; specificity, 64.13%), which was not affected by their clinical experiences. Noteworthily, despite the small number of cases, we were able to create a highly accurate AI. These findings represent an important first step in the automatic diagnosis of the cholesteatoma extension.

Comparison of the Utility of High-Resolution CT-DWI and T2WI-DWI Fusion Images for the Localization of Cholesteatoma

BACKGROUND AND PURPOSE

Cholesteatoma is an aggressive disease that may lead to hearing impairment. This study aimed to compare the utility of high-resolution CT and TSE-DWI fusion images with that of T2WI and TSE-DWI fusion images in the localization of middle ear cholesteatoma.

MATERIALS AND METHODS

Seventy-one patients with middle ear cholesteatoma were retrospectively recruited. High-resolution CT, T2WI with fat suppression, and TSE-DWI scans were obtained, and image fusion was performed using a 3D reconstruction postprocessing workstation to form CT-DWI and T2WI-DWI fusion images. The quality of the 2 fused images was subjectively evaluated using a 5-point Likert scale with the horizontal semicircular canal transverse position as the reference. Receiver operating characteristic analysis was performed, and the diagnostic efficacies of CT-DWI and T2WI-DWI fusion images in localizing middle ear cholesteatoma were calculated.

RESULTS

The overall quality of T2WI-DWI fusion images was slightly higher than that of CT-DWI fusion images (P < .001), and the semicircular canal was slightly less clear on T2WI-DWI than on CT-DWI (P < .001). No statistical difference was found in the diagnostic confidence between them. In the localization of middle ear cholesteatoma, the accuracy, sensitivity, and specificity of T2WI-DWI fusion images and CT-DWI fusion images were equivalent for involvement of the attic, tympanic cavity, mastoid antrum, and mastoid process, with no statistically significant differences.

CONCLUSIONS

T2WI-DWI fusion images could replace CT-DWI in the preoperative selection of surgical options for middle ear cholesteatoma.

Comparative evaluation of temporal subtraction computed tomography and non-echoplanar diffusion-weighted imaging for the mastoid extension of middle ear cholesteatoma

OBJECTIVE

Preoperative imaging assessment influences the decision to perform mastoidectomy for the mastoid extension of middle ear cholesteatoma. This study compared the performance of temporal subtraction CT (TSCT) and non-echoplanar diffusion-weighted imaging (non-EP DWI) in evaluating such mastoid extensions.

METHODS

We retrospectively evaluated 239 consecutive patients with surgically proven middle ear cholesteatoma between April 2016 and April 2021. The diagnostic performance of TSCT, wherein the presence of black color indicated progressive bone erosion, and non-EP DWI, wherein high signal intensity in the mastoid region suggested mastoid extension, was compared using Fisher's exact test.

RESULTS

In 34 patients with evaluable TSCT images, black color was significantly more common in patients with mastoid extension than in those without; the sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV), and accuracy of TSCT were 1.00, 0.95, 0.94, 1.00, and 0.97, respectively. In 90 patients with evaluable non-EP DWI, high signal intensity was significantly more common in patients with mastoid extension than in those without; the sensitivity, specificity, PPV, NPV, and accuracy of non-EP DWI were 0.88, 0.85, 0.91, 0.81, and 0.87, respectively. In 16 patients with both evaluable TSCT and non-EP DWI, the diagnostic performance of the TSCT was slightly superior to that of the non-EP DWI for predicting mastoid extension, although the difference was not significant.

CONCLUSIONS

TSCT images generated using consecutively acquired preoperative high-resolution CT images are useful for predicting mastoid extension of middle ear cholesteatoma, and the diagnostic performance of TSCT is non-inferior to that of non-EP DWI.

Preoperative prediction for mastoid extension of middle ear cholesteatoma using temporal subtraction serial HRCT studies

OBJECTIVES

This study investigated the utility of temporal subtraction computed tomography (TSCT) obtained with temporal bone high-resolution computed tomography (HRCT) for the preoperative prediction of mastoid extension of middle ear cholesteatomas.

METHODS

Twenty-eight consecutive patients with surgically proven middle ear cholesteatomas were retrospectively evaluated. The presence of black color in the mastoid region on TSCT suggested progressive changes caused by bone erosion. Enlarged width of the anterior part of mastoid on HRCT was interpreted as suggestive of mastoid extension. Fisher's exact test was used to compare the widths and black color on TSCT for cases with and without mastoid extension. The diagnostic accuracy of TSCT and HRCT for detecting mastoid extension and interobserver agreement during the evaluation of black color on TSCT were calculated.

RESULTS

There were 15 cases of surgically proven mastoid extension and 13 cases without mastoid extension. Patients with black color on TSCT were significantly more likely to have a mastoid extension (p < 0.001). The sensitivity and specificity of TSCT were 0.93 and 1.00, respectively. Patients in whom the width of the anterior part of the mastoid was enlarged were significantly more likely to have a mastoid extension (p = 0.007). The sensitivity and specificity of HRCT to detect the width of the anterior part of the mastoid were 0.80 and 0.77, respectively. Interobserver agreement during the evaluation of TSCT findings was good (k = 0.71).

CONCLUSIONS

This novel TSCT technique and preoperative evaluations are useful for assessing mastoid extension of middle ear cholesteatomas and making treatment decisions.

KEY POINTS

•TSCT shows a clear black color in the mastoid region when the middle ear cholesteatoma is accompanied by mastoid extension. •TSCT obtained with preoperative serial HRCT of the temporal bone is useful for assessing mastoid extension of middle ear cholesteatomas.

Preoperative predictive criteria for mastoid extension in pars flaccida cholesteatoma in assessments using temporal bone high-resolution computed tomography

OBJECTIVES

We aimed to clarify the usefulness of high-resolution computed tomography (HRCT) and establish HRCT criteria for presurgical assessment of the mastoid extension in pars flaccida cholesteatomas of the middle ear.

METHODS

Retrospective observational study. Patients who underwent primary surgery for pars flaccida cholesteatoma and those who underwent temporal bone HRCT for pretreatment evaluation were reviewed. The distance in the anterior-most portion of the mastoid sinus on HRCT was measured, and the presence of surgically verified mastoid extension of cholesteatoma was evaluated. All cases were subclassified as M+ (surgically mastoid extension-positive) or M- (surgically mastoid extension-negative).

RESULTS

A total of 107 patients with pars flaccida cholesteatoma were included. The distance in the M+ cases was significantly longer than that in the M- cases, and the cutoff value was 3.6 mm. The difference between the ipsilateral/diseased-side distance and the contralateral/evaluable side (difference value) in M+ cases was larger than that in M- cases, with a cutoff value of 0.6 mm. The inter-rater reliability of this distance measurement was excellent, regardless of imaging experience.

CONCLUSIONS

The cutoff values of the distance and the difference value can be used for pretreatment HRCT evaluation of mastoid extension in middle ear cholesteatoma with relatively high accuracy, regardless of the experience and skill levels of the evaluator.

Novel Radiologic Approaches for Cholesteatoma Detection: Implications for Endoscopic Ear Surgery

Technological advancement in computed tomography (CT) and MRI has improved cholesteatoma detection rates considerably in the past decade. Accurately predicting disease location and extension is essential for staging, planning, and preoperative counseling, in particular in the newer approach of endoscopic ear surgery. Improved sensitivity and specificity of these radiological methods may allow the surgeon to confidently monitor patients, therefore avoiding unnecessary surgery. This article outlines recent advances in CT and MRI technology and advantages and disadvantages of the newer techniques. Emphasis on improving the feedback loop between the radiologist and surgeon will increase the accuracy of these new technologies.

Management of squamous cell carcinomas of the skull-base

PURPOSE

The purpose of this review is to assess the recent evidence regarding the management of squamous cell carcinoma of the skull-base and to discuss the implications of these findings on clinical practice.

METHOD

Free text Medline and MeSH term search of publications relating to Squamous Cell Carcinoma & Skull-base and Skull base, Neoplasm respectively. Multidisciplinary clinical guidelines were also reviewed.

RESULTS

The primary search yielded a total of 271 papers which following initial review was reduced to 28. Secondary search yielded 56 papers. There were no randomised controlled trials relating to squamous cell carcinoma of the skull-base and as such this review is based on cohort studies, case series and expert opinion.

CONCLUSION

Squamous cell carcinoma (SCC) is the most common cancer occurring in the Head and Neck. Squamous cell carcinoma is also the most common cancer arising within the nose and sinuses of which skull-base squamous cell carcinoma is a rare subgroup. Evidence relating to the management and survival of skull-base SCC is based on expert opinion and. retrospective analyses Clinical examination and biopsy, imaging and a broad multidisciplinary team are key to the management of skull-base SCC. The information gathered should be used to guide informed discussion by suitably trained experts with patients regarding surgical approach, post-operative recovery and adjuvant or neoadjuvant treatments. The standard of care is currently to perform skull base resection with or without additional craniotomy, pedicled or free flap reconstruction in multiple layers and post-operative radiation (usually photons or protons). Open approaches have traditionally been the mainstay, however in certain cases endoscopic approaches can yield equivalent results and offer many advantages. Despite advances in care survival remains poor with a nearly one in five risk of nodal recurrence within two years.

The Antrum-Malleus-Tegmen Score: A Pilot Study Assessing Preoperative Radiographic Predictors for Transcanal Endoscopic Cholesteatoma Dissection

OBJECTIVE

Identify features on computed tomography (CT) that predict mastoidectomy conversion (MC) during transcanal endoscopic ear surgery (TEES).

STUDY DESIGN

Retrospective case-control.

SETTING

University otology practice.

PATIENTS

Consecutive patients with cholesteatoma.

INTERVENTION

TEES cholesteatoma dissection versus those requiring MC.

MAIN OUTCOME MEASURES

Antrum opacification, depth of scutum involvement, and erosion of the mastoid trabeculae, ossicular chain, and tegmen were evaluated. Univariable and multivariable regression analysis was performed. The Antrum-Malleus-Tegmen (AMT) score was created using receiver operating characteristic curves to assess feasibility of performing TEES for cholesteatoma dissection.

RESULTS

There were 39 TEES and 19 MC cases. Groups had similar age (median 28.5 yr), gender, laterality, and revision surgery status. Median surgical time for MC cases was longer than TEES (231 min vs. 171 min, p < 0.001). Radiographic predictors of MC by multivariable regression included antrum opacification (p = 0.036), malleus erosion (p = 0.044), and tegmen erosion (p = 0.023). The AMT score predicted the feasibility of TEES without MC with a sensitivity of 90% and specificity of 75% when ≥2 of the following conditions were met on preoperative CT: intact tegmen, intact malleus, and absence of antrum opacification.

CONCLUSIONS

An aerated antrum, intact malleus, and intact tegmen suggest that TEES cholesteatoma dissection without the need for mastoidectomy is highly likely. A score of ≥2 on the AMT score predicts this with a positive predictive value of 88% and negative predictive value of 78%.

Improved Assessment of Middle Ear Recurrent Cholesteatomas Using a Fusion of Conventional CT and Non-EPI-DWI MRI

BACKGROUND AND PURPOSE

Recurrent middle ear cholesteatomas are commonly preoperatively assessed using MR imaging (non-EPI-DWI) and CT. Both modalities are used with the aim of distinguishing scar tissue from cholesteatoma and determining the extent of bone erosions. Inflammation and scar tissue associated with the lesions might hamper a proper delineation of the corresponding extensions on CT images. Using surgical findings as the criterion standard, we assessed the recurrent middle ear cholesteatoma extent using either uncoregistered or fused CT-MR imaging datasets and determined the corresponding accuracy and repeatability.

MATERIALS AND METHODS

Twenty consecutive patients with suspected recurrent middle ear cholesteatoma and preoperative CT-MR imaging datasets were prospectively included. A double-blind assessment and coregistration of the recurrent middle ear cholesteatoma extent and manual delineation of 18 presumed recurrent middle ear cholesteatomas were performed by 2 radiologists and compared with the criterion standard. "Reliability score" was defined to qualify radiologists' confidence. For each volume, segmentation repeatability was assessed on the basis of intraclass correlation coefficient and overlap indices.

RESULTS

For the whole set of patients, recurrent middle ear cholesteatoma was further supported by surgical results. Two lesions were excluded from the analysis, given that MR imaging did not show a restricted diffusion. Lesions were accurately localized using the fused datasets, whereas significantly fewer lesions (85%) were correctly localized using uncoregistered images. Reliability scores were larger for fused datasets. Segmentation repeatability showed an almost perfect intraclass correlation coefficient regarding volumes, while overlaps were significantly lower in uncoregistered (52%) compared with fused (60%, P < .001) datasets.

CONCLUSIONS

The use of coregistered CT-MR images significantly improved the assessment of recurrent middle ear cholesteatoma with a greater accuracy and better reliability and repeatability.

Applications of Augmented Reality in Otolaryngology: A Systematic Review

OBJECTIVE

Augmented reality (AR) is a rapidly developing technology. The aim of this systematic review was to (1) identify and evaluate applications of AR in otolaryngology and (2) examine trends in publication over time.

DATA SOURCES

PubMed and EMBASE.

REVIEW METHODS

A systematic review was performed according to PRISMA guidelines without temporal limits. Studies were included if they reported otolaryngology-related applications of AR. Exclusion criteria included non-English articles, abstracts, letters/commentaries, and reviews. A linear regression model was used to compare publication trends over time.

RESULTS

Twenty-three articles representing 18 AR platforms were included. Publications increased between 1997 and 2018 ( P < .05). Twelve studies were level 5 evidence; 9 studies, level 4; 1 study, level 2; and 1 study, level 1. There was no trend toward increased level of evidence over time. The most common subspecialties represented were rhinology (52.2%), head and neck (30.4%), and neurotology (26%). The most common purpose of AR was intraoperative guidance (54.5%), followed by surgical planning (24.2%) and procedural simulations (9.1%). The most common source of visual inputs was endoscopes (50%), followed by eyewear (22.2%) and microscopes (4.5%). Computed tomography was the most common virtual input (83.3%). Optical trackers and fiducial markers were the most common forms of tracking and registration, respectively (38.9% and 44.4%). Mean registration error was 2.48 mm.

CONCLUSION

AR holds promise in simulation, surgical planning, and perioperative navigation. Although level of evidence remains modest, the role of AR in otolaryngology has grown rapidly and continues to expand.