Automated measurement of hydrops ratio from MRI in patients with Ménière’s disease using CNN-based segmentation

Abstract

Ménière's Disease (MD) is difficult to diagnose and evaluate objectively over the course of treatment. Recently, several studies have reported MD diagnoses by MRI-based endolymphatic hydrops (EH) analysis. However, this method is time-consuming and complicated. Therefore, a fast, objective, and accurate evaluation tool is necessary. The purpose of this study was to develop an algorithm that can accurately analyze EH on intravenous (IV) gadolinium (Gd)-enhanced inner-ear MRI using artificial intelligence (AI) with deep learning. In this study, we developed a convolutional neural network (CNN)based deep-learning model named INHEARIT (INner ear Hydrops Estimation via ARtificial InTelligence) for the automatic segmentation of the cochlea and vestibule, and calculation of the EH ratio in the segmented region. Measurement of the EH ratio was performed manually by a neuro-otologist and neuro-radiologist and by estimation with the INHEARIT model and were highly consistent (intraclass correlation coefficient = 0.971). This is the first study to demonstrate that automated EH ratio measurements are possible, which is important in the current clinical context where the usefulness of IV-Gd inner-ear MRI for MD diagnosis is increasing. Ménière's disease (MD) is a multifactorial disorder with typical symptoms of recurrent spontaneous attacks of vertigo, fluctuating hearing loss, tinnitus, and sensations of ear fullness. Endolymphatic hydrops (EH) is a pathological finding where the endolymphatic spaces are distended by enlargements of endolymphatic volume, a histologic hallmark of MD 1-3. According to a 1995 consensus statement from the Committee on Hearing and Equilibrium of the American Association of Otolaryngology-Head and Neck Surgery (AAO-HNS), "certain" MD cases can only be confirmed by the histological demonstration of EH in postmortem temporal bone specimens 4. Therefore in 2015, a committee of the Bárány Society revised the diagnostic criteria to remove the concept of "certain MD" 5. Thus far, the diagnostic criteria have been changed due to the lack of tools to objectively find EH during life. However, with the advancement of imaging technology, MRI can be used to identify endolymphatic hydrops in MD patients as an objective marker. In 2004, Duan et al. succeeded in visualizing EH in vivo for the first time in a guinea pig using 4.7 T MRI 6. Nakashima et al. succeeded in confirming EH after injecting contrast media through intratympanic (IT) and intravenous (IV) injections into MD patients using 3 T MRI 7,8. Recently, many reports have been published regarding the use of MRI to assess EH. In particular, IV gadolinium (Gd)-enhanced inner-ear MRI has shown good results 9,10. We have also proven through previous studies that IV-Gd inner-ear MRI is very useful for diagnosing MD by demonstrating the correlation of hydrops with

Endolymphatic Hydrops in Fluctuating Hearing Loss and Recurrent Vertigo

“…Although volumetric evaluation of EH has been reported (53), it is not normally used in routine clinical practice, mainly because there is no dedicated software that has been developed. However, recent advances in the use of artificial intelligence with deep learning (still to be circulated) represent a major step forward for the purpose of providing an objective measure (54). Nonetheless, the correlation obtained between semiquantitative scales and volumetry is very good (55).…”

Advanced Imaging of the Vestibular Endolymphatic Space in Ménière's Disease

“…In order to eliminate the subjective (examiner's) bias, new automated images segmentation processing algorithms have been introduced ( 43 , 66 ), also associated with deep-learning models based on Artificial Intelligence ( 45 ). By adding the quantification of perilymphatic enhancement to the grading of EH, van Steekelenburg et al ( 46 ) recently reported to improve the positive predictive value of Gd-enhanced MRI from 0.92 to 0.97 in the confirmation of definite MD.…”

Application of artificial intelligence in the otoneurology diagnostic process – an overview

“…Zautomatyzowane pomiary współczynnika przerostu worka endolimfatycznego wykazują się wysoką celnością. Ponadto, ze względu na coraz większe wykorzystanie badania MRI z kontrastem w praktyce medycznej, z kontrastem, mogą być dodatkową wartością w kolejnych badaniach klinicznych [19].…”