IE-Vnet: Deep Learning-Based Segmentation of the Inner Ear's Total Fluid Space

Abstract

BackgroundIn-vivo MR-based high-resolution volumetric quantification methods of the endolymphatic hydrops (ELH) are highly dependent on a reliable segmentation of the inner ear's total fluid space (TFS). This study aimed to develop a novel open-source inner ear TFS segmentation approach using a dedicated deep learning (DL) model.MethodsThe model was based on a V-Net architecture (IE-Vnet) and a multivariate (MR scans: T1, T2, FLAIR, SPACE) training dataset (D1, 179 consecutive patients with peripheral vestibulocochlear syndromes). Ground-truth TFS masks were generated in a semi-manual, atlas-assisted approach. IE-Vnet model segmentation performance, generalizability, and robustness to domain shift were evaluated on four heterogenous test datasets (D2-D5, n = 4 × 20 ears).ResultsThe IE-Vnet model predicted TFS masks with consistently high congruence to the ground-truth in all test datasets (Dice overlap coefficient: 0.9 ± 0.02, Hausdorff maximum surface distance: 0.93 ± 0.71 mm, mean surface distance: 0.022 ± 0.005 mm) without significant difference concerning side (two-sided Wilcoxon signed-rank test, p>0.05), or dataset (Kruskal-Wallis test, p>0.05; post-hoc Mann-Whitney U, FDR-corrected, all p>0.2). Prediction took 0.2 s, and was 2,000 times faster than a state-of-the-art atlas-based segmentation method.ConclusionIE-Vnet TFS segmentation demonstrated high accuracy, robustness toward domain shift, and rapid prediction times. Its output works seamlessly with a previously published open-source pipeline for automatic ELS segmentation. IE-Vnet could serve as a core tool for high-volume trans-institutional studies of the inner ear. Code and pre-trained models are available free and open-source under https://github.com/pydsgz/IEVNet.

Automatic segmentation of dura for quantitative analysis of lumbar stenosis: A deep learning study with 518 CT myelograms

“…Although multi‐atlas segmentation could potentially contour the dural sac on CTM, the use of deformable image registration in this method could significantly impair segmentation performance in cases complicated by pathological lesions (e.g., severe LSS or scoliosis) 42 . Numerous studies have confirmed the advantages of DL algorithms over the atlas‐based segmentation methods especially when segmenting cases with pathological lesions 43–45 . Thus, we have implemented DL algorithms to develop an automatic contouring tool for the lumbar dura on CTM images.…”

Endolymphatic space is age-dependent

“…3D- or volumetric quantification of the ELS ( v ELS) was achieved in two steps: First, segmentation of the total fluid space (TFS) was based on IE-Vnet [ 33 ], a recently proposed and pre-trained volumetric deep learning algorithm with V-net architecture. IE-Vnet was deployed via the TOMAAT module [ 34 ] into the 3D–Slicer toolbox (version 4.11, [ 35 ]).…”

Vestibular paroxysmia entails vestibular nerve function, microstructure and endolymphatic space changes linked to root-entry zone neurovascular compression

“…3D-quantification of the ELS consisted of three steps: first, segmentation of the total fluid space (TFS) was based on IE-Vnet [ 50 ], a recently proposed and pre-trained volumetric deep learning algorithm with V-net architecture that was deployed via the TOMAAT module [ 51 ] in a 3D–Slicer toolbox (version 4.11 [ 36 ]). Second, ELS and perilymphatic space (PS) were differentiated within the TFS using Vo lumetric L ocal T hresholding (VOLT; [ 52 ]) with ImageJ Fiji [ 53 ], the “Fuzzy and artificial neural networks image processing toolbox” [ 54 ], and the “MorphoLibJ Toolbox” [ 55 ].…”