Application Research On Precise Localization And Invasion Prediction Of Cholesteatoma Based On MR/CT Plain Images

Part 1 The value of turbo spin-echo diffusion weighted imaging apparent diffusion coefficient in the diagnosis of middle ear cholesteatomaObjective:Cholesteatoma of the middle ear is a locally aggressive and osteolysis-enhancing disease that can lead to conductive hearing impairment,brain abscess,and facial palsy when the lesion expands and invades adjacent structures.Therefore,early diagnosis and surgical treatment of cholesteatoma are important to improve the quality of life of patients.The aim of this study was to explore the value of applying the apparent diffusion coefficient（ADC）to the quantitative diagnosis of middle ear cholesteatoma.Methods:Seventy-one patients clinically suspected of temporal bone cholesteatoma were prospectively enrolled.These patients underwent an ear magnetic resonance imaging（MRI）scan and turbo spin-echo diffusion weighted imaging（TSE-DWI）scan on a Philips Ingenia3.0T superconductive MRI scanner,and their ADCs were measured.Subsequently,all enrolled patients underwent surgery within 15 days of the MRI scan.Using Receiver-Operating-Characteristic（ROC）curve analysis the optimal threshold and diagnostic performance for diagnosing temporal bone cholesteatoma were determined.Logistic regression modeling was used to combine ADCs and T1-weighted image（T1WI）sequences and calculate the combined diagnostic performance.Results:Based on the pathology results,patients were categorized into the cholesteatoma group（CS group,n=43）and the non-cholesteatoma group（NCS group,n=28）.The non-cholesteatoma group included 7 cases of granulation tissues;2 cases of effusion;2 cases of papillary abscess;2 cases of neurilemmoma;3 cases of fibrous tissue;5 cases of fibrous tissue and calcification with old hemorrhage;3 cases of the tympanic polyp;2 cases of cholesterol crystals;and 2 cases of Aspergillus cenobium.In the CS group,ADCs were significantly lower(mean,0.87±0.31×10^-3mm²/s)than in the NCS group(mean,1.87±0.49×10^-3mm²/s)（p<0.001）;the area under the receiver operating characteristic curve（AUC）was 0.915.Based on the optimal threshold of ADC,≤1.23×10^-3mm²/s,the diagnostic performance was high;the sensitivity and specificity for diagnosing cholesteatoma were 95.35%and 85.71%,respectively.By combining the ADC and T1WI sequence,AUC increased to 0.953,and the sensitivity and specificity were 90.70%and96.43%,respectively.Conclusions:ADC quantitative analysis has a high value in the preoperative diagnosis of cholesteatoma,and the combination of ADC and T1WI sequences can improve the diagnostic accuracy and specificity and thereby facilitate clinically effective diagnosis.Part 2 Comparison of the utility of HRCT-DWI and T2WI-DWI fusion images for accomplishing the localization of cholesteatomaObjective:Cholesteatoma is an aggressive disease that may lead to hearing impairment.Surgical resection is the only way to treat cholesteatoma.The choice of surgical approach varies depending on the localization and extent of cholesteatoma involvement,so precise preoperative localization is important.This study aimed to compare the utility of high-resolution CT（HRCT）and turbo spin-echo diffusion weighted imaging（TSE-DWI）fusion images,and T2 Weighted Images（T2WI）and TSE-DWI fusion images in the localization of middle ear cholesteatoma.Methods:Seventy-one patients with middle ear cholesteatoma were retrospectively recruited,and HRCT scans of the ear were performed with a Philip 256-row CT.T2WI fat suppression sequences and TSE-DWI scans were performed with a Philips Ingenia 3.0T superconducting scanner,and image fusion was performed using a 3D reconstruction post-processing workstation（Netherlands）.The fusion process was performed by experienced otolaryngologists and head and neck radiologists.The HRCT and T2WI fat suppression,and TSE-DWI were changed to the same random number code,and the HRCT and T2WI fat suppression were reconstructed to have the same field of view and layer thickness as TSE-DWI images.The fusion images are then automatically generated and manually fine-tuned to form CT-DWI and T2WI-DWI fusion according to the structure of the internal auditory canal and temporal bone.The quality of the two fused images（overall quality of the fusion image,semicircular canal display,clarity of the lesion,and diagnostic confidence）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 efficacy of CT-DWI and T2WI-DWI fusion images in localizing middle ear cholesteatoma was calculated.Results:The overall quality of T2WI-DWI fusion images was slightly higher than that of CT-DWI fusion images（P<0.001）,and the semicircular canal was slightly less clear than that of CT-DWI（P<0.001）;no statistical difference was found in the diagnostic confidence between them.In the localization of middle ear cholesteatoma,the diagnostic accuracy,sensitivity,and specificity of T2WI-DWI fusion images and CT-DWI fusion images were higher for the attic,tympanic cavity,mastoid antrum,and mastoid process,with no statistically significant difference.Conclusions:The overall quality of T2WI-DWI fusion images can be comparable to that of CT-DWI fusion images,and can clearly assess the anatomical location and extent of involvement of cholesteatoma on the attic,tympanic cavity,mastoid sinus,and mastoid process of concern to clinicians,which can replace CT-DWI fusion images to assist clinicians in preoperative evaluation,selection of the best surgical plan and planning of the optimal surgical approach.Part 3 Application value of radiomic nomogram based on preoperative HRCT in predicting fallopian canal involvement in cholesteatomaObjective:Cholesteatoma growth pattern predicts frequent involvement of the tympanic segment of the fallopian canal,and fallopian canal dehiscence is a major risk factor for facial nerve injury during middle ear mastoid surgery.The purpose of this study was to establish and validate a non-contrast HRCT based radiomic nomogram for predicting fallopian canal involvement in patients with cholesteatoma.Methods:214 consecutive patients（157 in the training set and 57 in the test set）who underwent surgery and had pathologically confirmed cholesteatoma were retrospectively included.All of whom had microscopic surgical records as the gold standard for the integrity or involvement of the fallopian canal,and completed ear scan HRCT within one month before surgery.Open-source radiomics software（Ma Zda,version 4.6;Institute of Electronics,Technical University of Lodz）was used to select the largest slice of the fallopian canal segment and extract features on the non-contrast HRCT.A radiomics signature model was constructed to fallopian canal involvement using the least absolute shrinkage and selection operator（LASSO）.Radiomic features,clinically independent risk factors were incorporated into multivariate logistic regression analysis and presented in the formation of a nomogram for predicting fallopian canal involvement in both the training and test sets,which were evaluated by its discrimination,calibration,and clinical usefulness.Results:A radiomics model consisting of seven radiomic features with a good prognostic ability in the training set with the area under the curve（AUC）of 0.85（95%CI,0.78-0.90）,which was also confirmed in the test set with an AUC of 0.81（95%CI,0.68-0.90）.Multivariate logistic regression identifies low-lying middle cranial fossa and stapes destruction as independent risk factors for fallopian canal involvement in cholesteatoma.The nomogram consisting of the radiomics scores and low-lying middle cranial fossa and stapes destruction showed good discrimination ability in the training and test sets with an AUC of 0.91（95%CI,0.85-0.95）and an AUC of 0.83（95%CI,0.71-0.92）,respectively.Decision curve analysis showed that the radiomic nomogram was clinically useful.Conclusions:Radiomic nomogram based on the non-contrast HRCT is a useful tool for the prediction of fallopian canal involvement in patients with cholesteatoma,which can assist clinicians for optimized risk stratification of patients preoperatively and take effective measures to prevent intraoperative facial nerve injury.