Intraretinal Layer Segmentation Using Cascaded Compressed U-Nets

Abstract

Reliable biomarkers quantifying neurodegeneration and neuroinflammation in central nervous system disorders such as Multiple Sclerosis, Alzheimer’s dementia or Parkinson’s disease are an unmet clinical need. Intraretinal layer thicknesses on macular optical coherence tomography (OCT) images are promising noninvasive biomarkers querying neuroretinal structures with near cellular resolution. However, changes are typically subtle, while tissue gradients can be weak, making intraretinal segmentation a challenging task. A robust and efficient method that requires no or minimal manual correction is an unmet need to foster reliable and reproducible research as well as clinical application. Here, we propose and validate a cascaded two-stage network for intraretinal layer segmentation, with both networks being compressed versions of U-Net (CCU-INSEG). The first network is responsible for retinal tissue segmentation from OCT B-scans. The second network segments eight intraretinal layers with high fidelity. At the post-processing stage, we introduce Laplacian-based outlier detection with layer surface hole filling by adaptive non-linear interpolation. Additionally, we propose a weighted version of focal loss to minimize the foreground–background pixel imbalance in the training data. We train our method using 17,458 B-scans from patients with autoimmune optic neuropathies, i.e., multiple sclerosis, and healthy controls. Voxel-wise comparison against manual segmentation produces a mean absolute error of 2.3 μm, outperforming current state-of-the-art methods on the same data set. Voxel-wise comparison against external glaucoma data leads to a mean absolute error of 2.6 μm when using the same gold standard segmentation approach, and 3.7 μm mean absolute error in an externally segmented data set. In scans from patients with severe optic atrophy, 3.5% of B-scan segmentation results were rejected by an experienced grader, whereas this was the case in 41.4% of B-scans segmented with a graph-based reference method. The validation results suggest that the proposed method can robustly segment macular scans from eyes with even severe neuroretinal changes.

The Acute Optic Neuritis Network (ACON): Study protocol of a non-interventional prospective multicenter study on diagnosis and treatment of acute optic neuritis

“…Documented OCT measurements include the average pRNFL thickness, GCL or GCC as collected by the different platforms (e.g., Spectralis measuring GCL, Zeiss measuring GCC). The OCT images will subsequently be analyzed using post-hoc analysis with semi-automatic, device-independent algorithms (70).…”

Automatic Segmentation of Retinal Layers in Multiple Neurodegenerative Disorder Scenarios

“…The method was evaluated on a dataset containing MS patients and controls, reporting a drop in performance on the MS patients when compared with controls. In [47], the authors proposed the use of a cascaded two-stage network, with each stage being composed of a compressed version of U-Net. A post-processing stage was incorporated in which a Laplacian-based outlier detection is applied, followed by an adaptive non-linear interpolation with the intention of filling any inconsistent holes in the segmentation.…”

“…The contour surrounding every layer was extracted both from each ground truth and automatically segmented output image. Following other related retinal layer segmentation works (for reference, [41], [42], [47]), the Mean Absolute Error (MAE) was computed between the pixel heights for every image column as a measurement of contour error (MAE C ), along with the MAE for the thickness of each layer (MAE T ). In order to provide a robust measurement of these metrics, and following other similar approaches in the literature [65], [66], bootstrapping with 1000 repetitions was applied at the image level to all tests.…”

Hybrid deep learning and optimal graph search method for optical coherence tomography layer segmentation in diseases affecting the optic nerve

“…He et al [ 26 ] trained a residual U-Net with two output branches for two probability maps of the retinal layers and junction surfaces and merged probability maps in an iterative surface topology module to estimate the retinal surface locations. Furthermore, Yadav et al [ 27 ] proposed a cascaded two-stage design, in which the first U-Net identified the retina in 2D OCT B-scan and the following U-Net focused on identifying each retinal layer based on the pre-segmented retina. Most recently, He et al [ 28 ] aimed to segment longitudinal OCT scans by proposing a long-short-term-memory U-Net (LSTM-UNet) to incorporate adjacent 2D B-scans information and convert the segmented surfaces as longitudinal priors.…”