Purpose This study investigated the feasibility of a new image analysis technique (radiomics) on conventional MRI for the computer-aided diagnosis of Meniere's disease. Materials and methods A... Show morePurpose This study investigated the feasibility of a new image analysis technique (radiomics) on conventional MRI for the computer-aided diagnosis of Meniere's disease. Materials and methods A retrospective, multicentric diagnostic case-control study was performed. This study included 120 patients with unilateral or bilateral Meniere's disease and 140 controls from four centers in the Netherlands and Belgium. Multiple radiomic features were extracted from conventional MRI scans and used to train a machine learning-based, multi-layer perceptron classification model to distinguish patients with Meniere's disease from controls. The primary outcomes were accuracy, sensitivity, specificity, positive predictive value, and negative predictive value of the classification model. Results The classification accuracy of the machine learning model on the test set was 82%, with a sensitivity of 83%, and a specificity of 82%. The positive and negative predictive values were 71%, and 90%, respectively. Conclusion The multi-layer perceptron classification model yielded a precise, high-diagnostic performance in identifying patients with Meniere's disease based on radiomic features extracted from conventional T2-weighted MRI scans. In the future, radiomics might serve as a fast and noninvasive decision support system, next to clinical evaluation in the diagnosis of Meniere's disease. Show less
Vaidyanathan, A.; Lubbe, M.F.J.A. van der; Leijenaar, R.T.H.; Hoof, M. van; Zerka, F.; Miraglio, B.; ... ; Lambin, P. 2021
Segmentation of anatomical structures is valuable in a variety of tasks, including 3D visualization, surgical planning, and quantitative image analysis. Manual segmentation is time-consuming and... Show moreSegmentation of anatomical structures is valuable in a variety of tasks, including 3D visualization, surgical planning, and quantitative image analysis. Manual segmentation is time-consuming and deals with intra and inter-observer variability. To develop a deep-learning approach for the fully automated segmentation of the inner ear in MRI, a 3D U-net was trained on 944 MRI scans with manually segmented inner ears as reference standard. The model was validated on an independent, multicentric dataset consisting of 177 MRI scans from three different centers. The model was also evaluated on a clinical validation set containing eight MRI scans with severe changes in the morphology of the labyrinth. The 3D U-net model showed precise Dice Similarity Coefficient scores (mean DSC-0.8790) with a high True Positive Rate (91.5%) and low False Discovery Rate and False Negative Rates (14.8% and 8.49% respectively) across images from three different centers. The model proved to perform well with a DSC of 0.8768 on the clinical validation dataset. The proposed auto-segmentation model is equivalent to human readers and is a reliable, consistent, and efficient method for inner ear segmentation, which can be used in a variety of clinical applications such as surgical planning and quantitative image analysis. Show less
Lubbe, M.F.J.A. van der; Vaidyanathan, A.; Rompaey, V. van; Postma, A.A.; Bruintjes, T.D.; Kimenai, D.M.; ... ; Berg, R. van de 2020
Background Classifying and diagnosing peripheral vestibular disorders based on their symptoms is challenging due to possible symptom overlap or atypical clinical presentation. To improve the... Show moreBackground Classifying and diagnosing peripheral vestibular disorders based on their symptoms is challenging due to possible symptom overlap or atypical clinical presentation. To improve the diagnostic trajectory, gadolinium-based contrast-enhanced magnetic resonance imaging of the inner ear is nowadays frequently used for the in vivo confirmation of endolymphatic hydrops in humans. However, hydrops is visualized in both healthy subjects and patients with vestibular disorders, which might make the clinical value of hydrops detection on MRI questionable. Objective To investigate the diagnostic value of clinical and radiological features, including the in vivo visualization of endolymphatic hydrops, for the classification and diagnosis of vestibular disorders. Methods A literature search was performed in February and March 2019 to estimate the prevalence of various features in healthy subjects and in common vestibular disorders to make a graphical comparison between healthy and abnormal. Results Of the features studied, hydrops was found to be a highly prevalent feature in Meniere's disease (99.4%). Though, hydrops has also a relatively high prevalence in patients with vestibular schwannoma (48.2%) and in healthy temporal bones (12.5%) as well. In patients diagnosed with (definite or probable) Meniere's disease, hydrops is less frequently diagnosed by magnetic resonance imaging compared to the histological confirmation (82.3% versus 99.4%). The mean prevalence of radiologically diagnosed hydrops was 31% in healthy subjects, 28.1% in patients with vestibular migraine, and 25.9% in patients with vestibular schwannoma. An interesting finding was an absolute difference in hydrops prevalence between the two diagnostic techniques (histology and radiology) of 25.2% in patients with Meniere's disease and 29% in patients with vestibular schwannoma. Conclusions Although the visualization of hydrops has a high diagnostic value in patients with definite Meniere's disease, it is important to appreciate the relatively high prevalence of hydrops in healthy populations and other vestibular disorders. Endolymphatic hydrops is not a pathognomic phenomenon, and detecting hydrops should not directly indicate a diagnosis of Meniere's disease. Both symptom-driven and hydrops-based classification systems have disadvantages. Therefore, it might be worth to explore features "beyond" hydrops. New analysis techniques, such as Radiomics, might play an essential role in (re)classifying vestibular disorders in the future. Show less
Aghajani, M.; Veer, I.; Hoof, M. van; Rombouts, S.; Wee, N. van der; Vermeiren, R. 2015
