Localization of Damage in Progressive Hydroxychloroquine Retinopathy On and Off the Drug: Inner Versus Outer Retina, Parafovea Versus Peripheral Fovea

Abstract

Hydroxychloroquine retinopathy primarily involves outer retina (photoreceptors). Outer retinal thinning while using HCQ initially involves the parafovea, but becomes diffuse across the macula as damage progresses or after drug cessation. When HCQ is stopped at an early or moderate stage (before RPE damage), progression seems to be limited to the first year.

A New Objective Parameter in Hydroxychloroquine-Induced Retinal Toxicity Screening Test: Macular Retinal Ganglion Cell-Inner Plexiform Layer Thickness

“…19 Spectral-domain optical coherence tomography images show localized thinning of retinal layers in the parafoveal zone, and may reveal retinal toxicity at an earlier stage before visual field loss occurs. 6,11,[20][21][22] Majority of the previous studies were related to the pathologies associated with the outer retinal segment involving retinal pigment epithelium and photoreceptors. 13,21,23 However, some other studies also demonstrated HCQassociated damage in the inner retinal segment.…”

Automated intraretinal segmentation of SD-OCT images in normal and age-related macular degeneration eyes

“…A reliable automated segmentation approach, such as that presented herein, is crucial to enabling such work. Our work has also shown superior results to other state-of-the-art clinical software in eyes affected by hydroxychloroquine toxicity [15] and post-surgical macular hole eyes [6]. We plan to expand our segmentation method to SD-OCT scans acquired in the region surrounding the optic disc as well as to eyes presenting with other retinal pathologies, such as glaucoma and retinitis pigmentosa.…”

“…Our method is sought to address the aforementioned challenges of prior algorithms, producing accurate delineations of the intraretinal layered structure in scans acquired during clinical practice and maintaining a stable behavior in disease cases while also producing accurate outlines that follow the particular pathologies of the disease, not smoothing out important details. This method was used previously for the segmentation of the retina in hydroxychloroquine toxicity [15] and postsurgery in macular hole cases [6]. In these applications, the results of our methods were successful and superior to current commercial software by visual inspection, but details about the method and its evaluation have not yet been published.…”

“…Some of these commercial methods also require human interaction to initiate or refine the segmentation results [11][12][13], which can be timeconsuming. In addition, recent publications have reported substantial errors in their results [14][15][16][17]. A number of experimental algorithms have also been proposed by research groups, either in two-dimensional individual scans [18,19] or taking advantage of the threedimensional nature of the acquired data, considering a priori determined smoothness, continuity and layer interaction constraints.…”

“…A general graph-based approach for the segmentation of surfaces was proposed in [20], with applications to volumetric OCT data in [21][22][23][24] Although these and other automated or semi-automated segmentation methods have been applied to SD-OCT images [28][29][30][31][32][33][34][35], they have a number of limitations. First, most segmentation algorithms fail to produce acceptable results in complicated cases, such as severe retinal deformities [15][16][17]36] or in lower quality scans that are routinely found in clinical practice. Second, most methods were designed for use with SD-OCT images acquired by experimental systems or from a particular vendor.…”