Deep learning based diagnostic quality assessment of choroidal OCT features with expert-evaluated explainability.

Various vision-threatening eye diseases including age-related macular degeneration (AMD) and central serous chorioretinopathy (CSCR) are caused due to the dysfunctions manifested in the highly vascular choroid layer of the posterior segment of the eye. In the current clinical practice, screening choroidal structural changes is widely based on optical coherence tomography (OCT) images. Accordingly, to assist clinicians, several automated choroidal biomarker detection methods using OCT images are developed. However, the performance of these algorithms is largely constrained by the quality of the OCT scan. Consequently, determining the quality of choroidal features in OCT scans is significant in building standardized quantification tools and hence constitutes our main objective. This study includes a dataset of 1593 good and 2581 bad quality Spectralis OCT images graded by an expert. Noting the efficacy of deep-learning (DL) in medical image analysis, we propose to train three state-of-the-art DL models: ResNet18, EfficientNet-B0 and EfficientNet-B3 to detect the quality of OCT images. The choice of these models was inspired by their ability to preserve the salient features across all the layers without information loss. To evaluate the attention of DL models on the choroid, we introduced color transparency maps (CTMs) based on GradCAM explanations. Further, we proposed two subjective grading scores: overall choroid coverage (OCC) and choroid coverage in the visible region(CCVR) based on CTMs to objectively correlate visual explanations vis-à-vis DL model attentions. We observed that the average accuracy and F-scores for the three DL models are greater than 96%. Further, the OCC and CCVR scores achieved for the three DL models under consideration substantiate that they mostly focus on the choroid layer in making the decision. In particular, of the three DL models, EfficientNet-B3 is in close agreement with the clinician's inference. The proposed DL-based framework demonstrated high detection accuracy as well as attention on the choroid layer, where EfficientNet-B3 reported superior performance. Our work assumes significance in bench-marking the automated choroid biomarker detection tools and facilitating high-throughput screening. Further, the methods proposed in this work can be adopted for evaluating the attention of DL-based approaches developed for other region-specific quality assessment tasks.

Volumetric quantification of choroid and Haller's sublayer using OCT scans: An accurate and unified approach based on stratified smoothing.

BACKGROUND AND OBJECTIVE: The choroid, a dense vascular structure in the posterior segment of the eye, maintains the health of the retina by supplying oxygen and nutrients, and assumes clinical significance in screening ocular diseases including age-related macular degeneration (AMD) and central serous chorioretinopathy (CSCR). As a technological assist, algorithmic estimation of choroidal biomarkers has been suggested based on sectional (B-scan) optical coherence tomography (OCT) images. However, most such 2D estimation techniques are compute-intensive, yet enjoy limited accuracy and have only been validated on OCT image datasets of healthy eyes. Not surprisingly, fine-scale analyses, including those involving Haller's sublayer, remain relatively rare and unsophisticated. Against this backdrop, we propose an efficient algorithm to quantify desired biomarkers with improved accuracy based on volume OCT scans. Specifically, we attempted an accurate, computationally light volumetric segmentation method involving stratified smoothing to detect choroid and Haller's sublayer. METHODS: For detecting the various boundaries of the choroid and the Haller's sublayer, we propose a common volumetric method that performs suitable exponential enhancement and maintains smooth spatial continuity across 2D B-scans. Further, we achieve suitable volumetric smoothing by primarily deploying light-duty linear regression, and sparingly using compute-intensive tensor voting, and hence significantly reduce overall complexity. The proposed methodology is tested on five health and five diseased OCT volumes considering various metrics including volumetric Dice coefficient and corresponding quotient measures to facilitate comparison vis-à-vis intra-observer repeatability. RESULTS: On five healthy and five diseased OCT volumes, respectively, the proposed method for choroid segmentation recorded volumetric Dice coefficients of 93.53 % and 93.30 %, which closely approximate the respective reference observer repeatability values of 95.60 % and 95.49 %. In terms of related quotient measures, our method achieved more than 50 % improvement over a recently reported method. In detecting Haller's sublayer as well, our algorithm records statistical performance closely matching that of reference manual method. CONCLUSION: Advancing the state-of-the-art, the proposed volumetric segmentation, tested on both healthy and diseased datasets, demonstrated close match with the manual reference. Our method assumes significance in accurate screening of chorioretinal diseases including AMD, CSCR and pachychoroid. Further, it enables generating accurate training data for developing deep learning models for improved detection of choroid and Haller's sublayer.

En-face analysis of short posterior ciliary arteries crossing the sclera to choroid using wide-field swept-source optical coherence tomography.

To study the topographic distribution of the short posterior ciliary arteries (SPCA) entry sites into the choroid in normal eyes using structural en-face swept source optical coherence tomography (SS-OCT). Retrospective analysis of SS-OCT scans (wide-field structural SS-OCT 12 × 12 mm) of 13 healthy subjects was performed. Cross-sectional swept-source OCT scans derived from a volume scan were represented as en-face image display following the Choroid-Scleral Interface to obtain en-face OCT. SPCAs in their last scleral location before choroidal entrance were identified manually, counted and localized by two masked observers. Correlations between two masked observers were analyzed using inter- and intra-class correlation. Accuracy for the choroidal inner and outer border segmentation was 95-99%. Eighteen eyes from 13 normal subjects were included for SPCA analysis. The mean number of arteries was 13.8 ± 3.5 per eye. Thirty-six percent were in the center of the posterior pole image; however, 21% were in the temporal part of the posterior pole. Median accuracy of the detection is 0.94. The correlation between the two observers was fair (0.54). Our algorithm allows visualization of the SPCA at the posterior pole of the eye using wide-field en-face SS-OCT. It can also help the clinicians to study the SPCAs in numerous ocular diseases, particularly its relationship with focal choroidal diseases.

Efficacy of navigated focal laser photocoagulation in diabetic macular edema planned with en face optical coherence tomography versus fluorescein angiography.

AIM: To analyze the efficacy of navigated focal laser photocoagulation (FLP) of microaneurysms in diabetic macular edema (DME) planned using en face optical coherence tomography (OCT) as against fluorescein angiography (FA). METHODS: Twenty-six eyes of 21 DME patients (12 males, 9 females, 69.5 ± 12.3 years) with mean BCVA of 0.52 ± 0.44 LogMAR were included. En face OCT images of deep capillary plexus slab and FA images were used to plan FLP targeting of leaky microaneurysms. The primary outcome measures were central retinal thickness (CRT) and macular volume. The secondary outcome measure was best-corrected visual acuity (BCVA). RESULTS: The difference in the change of CRT and macular volume between en face OCT and FA-planned FLP after 1 month and at the end of follow-up was not statistically significant (p > 0.05), except for a higher CRT reduction in the en face OCT-planning group (p = 0.007) at the end of mean follow-up of 2.6 ± 0.9 months. There was no difference in BCVA change between the two planning options (p = 0.42). CONCLUSION: En face OCT is a non-inferior alternative for FA in the planning of navigated FLP of microaneurysms in DME.

The role of tocilizumab in the treatment of inflammatory diseases of the eye and orbit: A useful alternative.

PURPOSE: With uveitis being one of the leading causes of blindness worldwide, biological therapies have arisen as an option for the treatment of refractory cases based on good results shown in clinical practice. The goal of this study is to provide a systematic review of current knowledge of the role and possible uses of tocilizumab in the field of ophthalmology. MATERIALS AND METHODS: We performed a search for records reporting the use of tocilizumab for various diseases in MEDLINE (PubMed and OVID). We conducted an analysis of several individual studies and their reported individual patient data (82 eyes of 45 patients) published from 2011 to 2017. CONCLUSIONS: Tocilizumab may prove to be an effective choice for the treatment of a variety of ocular conditions such as refractory uveitis, inflammatory macular edema, vitreo-retinal tumors and thyroid orbitopathy, leading to control of the inflammation in these patients. Further studies need to be conducted to establish its safety and efficacy.

[Optical coherence tomography in diagnostics and treatment of central serous chorioretinopathy]. / Opticheskaia kogerentnaia tomografiia v diagnostike i lechenii tsentral'noi seroznoi khorioretinopatii.

PURPOSE: To identify the signs of leakage points in central serous chorioretinopathy (CSC) patients using optical coherence tomography (OCT) allowing focal laser photocoagulation (FLP) to be performed without the need for fluorescein angiography (FA) in their treatment; to evaluate the effectiveness and safety of the approach. MATERIAL AND METHODS: The retrospective part of the study included 54 patients (54 eyes) with distinctive signs of CSC further confirmed by the presence of a leakage point as could be seen with FA. The number of pigment epithelial detachments (PED), the thickness of the photoreceptor outer segments (PROS) layer, maximum horizontal length of the photoreceptor thinning area, and morphological changes in the outer retinal layers were evaluated with OCT. Using FA for each leakage point, its localization and coincidence with PED were evaluated in retro-mode confocal scanning laser ophthalmoscopy and OCT images, respectively. The prospective part of the study included 11 CSC patients (11 eyes) who underwent en face OCT-guided FLP of the leakage point with NAVILAS 532 system. RESULTS: The leakage points were found to coincide with PED in 50 eyes (92.6%). Mean distance from the superior border of the retinal detachment area to the leakage point was 28.8±14.1% of the retinal detachment height. In 90.3% of cases the PROS layer in the area above the PED was thinner (23.3±7.8 µm) compared to the area outside the leakage point (64.0±9.4 µm) (p<0.001). FA-free OCT-guided FLP required in average 3.3±0.5 laser burns per leakage point. Complete resolution of subretinal fluid occurred in all treated eyes within 7.5±1.0 weeks after the surgery, and no adverse events were observed. CONCLUSION: Coincidence of PED in the superior half of the area of retinal detachment with thinning of the PROS layer as seen on OCT can be considered a reliable sign of the leakage point. The FA-free FLP based on OCT appears to be an effective and safe procedure for the treatment of CSC.

Actualización en técnicas de imagen coroidea: pasado, presente y futuro / Update of choroidal imaging techniques: past, present and future

La coroides es la capa media del ojo, un tejido muy vascularizado y pigmentado, cuyo papel ya ha sido probado en numerosas enfermedades oftálmicas. Pero no ha sido hasta los últimos años cuando hemos sido capaces de cuantificar y medir de una forma fiable, precisa y reproducible su grosor y su forma. Las técnicas de imagen basadas en ultrasonidos y la angiografía con verde de indocianina fueron las primeras en ser utilizadas para el estudio de la coroides, y aún mantienen su uso y sus indicaciones clínicas para el diagnóstico y manejo de algunas enfermedades oculares. Pero fue la llegada de la tomografía de coherencia óptica la que supuso el mayor avance en cuanto a la visualización de la coroides se refiere. En esta revisión se discutirán modalidades de imagen coroidea pasadas, presentes y futuras, con una especial atención sobre la tomografía de coherencia óptica y su más reciente desarrollo

The choroid is the middle layer of the eye, a very vascular and pigmented tissue, with its role in several ophthalmological pathologies already having been clearly established. But it was not until the last few years that we have been able to reliably and precisely measure and quantify its shape and thickness. Ultrasound technology and indocyanine green angiography were the first techniques used for the study of the choroid, and they still maintain their use and clinical indications for the diagnosis and management of several pathologies. But it was the advent of optical coherence tomography that was the greatest breakthrough in choroidal imaging. In this chapter, the past, current and future image modalities for the study of the choroid will be discussed, with special focus on optical coherence tomography and its latest developments

Update of choroidal imaging techniques: Past, present and future. / Actualización en técnicas de imagen coroidea: pasado, presente y futuro.

The choroid is the middle layer of the eye, a very vascular and pigmented tissue, with its role in several ophthalmological pathologies already having been clearly established. But it was not until the last few years that we have been able to reliably and precisely measure and quantify its shape and thickness. Ultrasound technology and indocyanine green angiography were the first techniques used for the study of the choroid, and they still maintain their use and clinical indications for the diagnosis and management of several pathologies. But it was the advent of optical coherence tomography that was the greatest breakthrough in choroidal imaging. In this chapter, the past, current and future image modalities for the study of the choroid will be discussed, with special focus on optical coherence tomography and its latest developments.

Intravitreal dexamethasone implant for recurrent cystoid macular edema due to Irvine-Gass syndrome: a prospective case series.

PurposeTo determine the preliminary efficacy and safety of off-label dexamethasone implant for treatment of recurrent cystoid macular edema (CME) secondary to Irvine-Gass syndrome (IGS).Patients and methodsThis study was set in Raghudeep Eye Clinic, Ahmedabad and LV Prasad Eye Institute, Hyderabad (India). It is a Prospective Case Series. Prospective case series comprising of patients with uncomplicated pseudophakia and CME due to IGS who recurred after one course of topical steroids with NSAIDS and a sub-Tenon corticosteroid injection. A complete ocular and systemic exam, fluorescein angiography, and central subfield thickness (CST) on optical coherence tomography scans were performed. Follow-up visits were on days 1, 15, and 30 and then monthly for a year. Appropriate statistical analysis was done. The primary outcome measure was the change in CDVA at months 1, 6, and 12. Secondary outcome measures were recurrence of CME and complications if any as noted at months 1, 2, 6, and 12.ResultsAbout 27 patients (27 eyes) with 16 males were included. Median age: 63.24±5.62 years. At 1 month, the CDVA improved to 0.04±0.02 (20/25) logMAR from 0.52±0.12 logMAR (20/70) (P=0.001) with a reduction in CST from 454.2±45.3 to 218.32±38.15 microns(P=0.013). The CDVA was 0.04±0.03 logMAR(P<0.001) at month 6 and 0.05±0.02 logMAR(P<0.001) at month 12. The CST was 221±35.2 microns (P=0.013) at month 6 and 214±43.34 microns (P=0.0124) at month 12. All improvements were maintained for a year. Only one patient required a second injection. No complications were noted.ConclusionThe implant is safe and effective for the treatment of recurrent CME due to IGS.

Subfoveal choroidal thickness as a predictor of central serous chorioretinopathy.

PurposeThe purpose of the study was to evaluate the role of subfoveal choroidal thickness (CT) measurements as a predictor for the course of central serous chorioretinopathy (CSC).Patients and methodsThirty-eight eyes of 33 patients with CSC were analyzed retrospectively. Key inclusion criteria were naive (acute or recurrent) CSC; a minimum of 12 months of follow up; and the availability of good quality enhanced depth imaging SD-OCT images at each visit. Eyes with changes suggestive of chronic CSC and history of any treatment were excluded. Collected data included demographic profile, visual acuity, subfoveal CT, central macular thickness (CMT), and treatment details. Univariate and multivariate analyses for association of baseline features with need for treatment were performed. On the basis of results of this analysis, the 38 eyes were divided into two groups: Group A (subfoveal CT≤356 µm) and Group B (subfoveal CT>356 µm).ResultsOn univariate analysis, the need for treatment had significantly positive correlation with duration of symptoms (P=0.02), negative correlation with the baseline CT (P<0.01), and no significant correlation with CMT (P=0.13). On multivariate regression analysis, only baseline CT had a statistically significant association with the need for treatment (odds ratio (OR), 0.989; CI, 0.979-0.999; P=0.048). Group A eyes were more likely to require treatment (Group A: 54.55%, 12 of 22 eyes; Group B: 18.75%, 3 of 16 eyes; OR: 5.2, 95% CI, 1.15-23.54; P=0.04). There was a significant decrease in subfoveal CT in Group B (-105.62±108.91 µm; P=0.002).ConclusionCSC with a subfoveal CT≤356 µm is of chronic nature and is more likely to require treatment rather than observation.

Surgical challenges and outcomes of rhegmatogenous retinal detachment in albinism.

PURPOSE: To report the outcomes and surgical difficulties during rhegmatogenous retinal detachment (RRD) repair in patients with albinism. METHODS: Retrospective analysis of 10 eyes of 9 patients with albinism that underwent RRD repair was performed. Collected data included demographic details, preoperative examination details, surgical procedure, surgical difficulties, anatomical, and visual outcomes. Outcome measures were retinal reattachment and visual acuity at the last follow-up. RESULTS: Mean preoperative best-corrected visual acuity (BCVA) was logMAR (Logarithm of the Minimum Angle of Resolution) 2.15 (range 0.9-3.0) with preoperative localization of causative break in six eyes. One eye had proliferative vitreoretinopathy grade C1 preoperatively. Four eyes underwent scleral buckling (SB) and six underwent 20G pars plana vitrectomy (PPV) with silicone oil injection. Intraoperative complication as iatrogenic retinal break occurred in four eyes. For retinopexy during vitrectomy, endolaser delivery was possible in three out of six eyes, whereas three eyes had cryopexy. The mean follow-up was 12 months in SB group (range 1-12; median 12 months) and 5.33 months (range 1-12; median 3 months) in PPV group. Among vitrectomized eye, two eyes had recurrence at 3 months with oil in situ. Rest of the eyes had attached retina at last follow-up. Mean BCVA at last follow-up was logMAR -1.46 (range 0.7-2.0) with mean improvement of -0.57 logMAR. CONCLUSIONS: Identification of break, induction of posterior vitreous detachment, and endolaser delivery may be difficult during RRD repair in patients with albinism. The incidence of PVR appeared less in these eyes. Both SB and PPV were efficacious and appear to be good surgical techniques for use in this patient population.

Dexamethasone implant in diabetic macular edema in real-life situations.

PURPOSE: To report outcome of eyes with recalcitrant and naive eyes with diabetic macular edema (DME) treated with intravitreal dexamethasone implants (Ozurdex) injection. METHODS: Retrospective multicenter data analysis of eyes with DME treated with Ozurdex implant and with minimum follow-up of at least one year after the first implant. Data collected included demographic details, history of presenting illness, past treatment history, clinical examination details including visual acuity at presentation, and follow-up with imaging and treatment details. Paired sample t-test was used to measure mean differences between pre- and post-implant values obtained at baseline and last follow-up. RESULTS: A total of 79 eyes (62 subjects) were included. Sixty-four eyes had been previously treated; 15 eyes were naive. Among the previously treated eyes, mean interval between first Ozurdex injection and any previous treatment was 7.69±8.2 months. In naive eyes, the visual acuity improved from baseline 0.58±0.25 to 0.44±0.33 logMAR at last follow-up (P=0.05). In eyes that had been previously treated, the improvement was from 0.65±0.34 at baseline to 0.48±0.35 logMAR (P=0.01). Mean treatment-free interval was 6.5±4.5 months. Nine eyes were steroid responder with controlled intraocular pressure (IOP), none showed any spike in IOP during the follow-up period. CONCLUSIONS: Ozurdex implant could be a good alternative for recalcitrant as well as naive eyes with DME. The visual gain after initial implant injection was fairly maintained, with additional treatment usually after 6 months in naive eyes. Ozurdex appeared safe even in steroid responders with good control of IOP with antiglaucoma medications.

Fluorescein angiography and optical coherence tomography in myopic choroidal neovascularization.

PURPOSE: To assess intra/inter-observer agreement, and diagnostic capabilities of a color fundus photograph, fundus fluorescein angiography (FFA), and spectral domain optical coherence tomography (SD-OCT) in making a diagnosis of myopic choroidal neovascularization (CNV). PATIENTS AND METHODS: Two masked observers evaluated FFA and SD-OCT images to identify the presence of myopic CNV in 80 high-myopic eyes of 57 patients. A third masked observer identified CNV on a color fundus photo. Presence of myopic CNV on a fundus photo was defined as presence of subretinal hemorrhage, thickening of the retina and/or visible membrane at the macula. Presence of myopic CNV on FFA was defined as hyperfluorescence in the early phase with increase in intensity and size in the late phase; presence of a large irregular lesion; and hypofluorsescence due to subretinal hemorrhage. Myopic CNV on SD-OCT was defined as the hyper-reflective lesion with or without intraretinal fluid or subretinal fluid with retinal thickening. RESULTS: Intraobserver repeatability on FFA and SD-OCT was 0.54 and 0.44, respectively. Agreement (kappa) between FFA and SD-OCT was 0.38 and 0.3, respectively. Among 34 eyes, which had the presence of CNV on a color fundus photo, CNV was diagnosed in 18 (53%) eyes on FFA and in 20 (58.8%) eyes on SD-OCT. Sensitivity and specificity of FFA was 47 and 80.4%, respectively, and that of SD-OCT was 58.8 and 86.9%, respectively. CONCLUSION: Repeatability and reproducibility for diagnosis of myopic CNV was better with FFA compared with SD-OCT; however, agreement is very poor between FFA and SD-OCT. SD-OCT is comparatively a better tool to rule out presence of myopic CNV.

Macular toxicity following brilliant blue G-assisted macular hole surgery - a report of three cases.

INTRODUCTION: Brilliant blue G is a new dye used for staining the internal limiting membrane to ease its peeling in cases like a macular hole. CASES: Three patients presented with full-thickness idiopathic macular hole. They underwent pars planavitrectomy, Brilliant Blue G (BBG) stained internal limiting membrane peeling and fluid gas exchange. OBSERVATIONS: Postoperatively,the macular hole closed but foveal thinning and perifoveal hyperpigmentation presumably due to BBG toxicity were observed in all three patients. All of them had a subnormal final best corrected visual acuity. CONCLUSION: This case series highlights the unusual occurrence of macular toxicity following brillantblue G-assisted macular hole surgery.

Visual function assessment in simulated real-life situations in patients with age-related macular degeneration compared to normal subjects.

PURPOSE: To evaluate visual function variations in eyes with age-related macular degeneration (AMD) compared to normal eyes under different light/contrast conditions using a time-dependent visual acuity testing instrument, the Central Vision Analyzer (CVA). METHODS: Overall, 37 AMD eyes and 35 normal eyes were consecutively tested with the CVA after assessing best-corrected visual acuity (BCVA) using ETDRS charts. The CVA established visual thresholds for three mesopic environments (M1 (high contrast), M2 (medium contrast), and M3 (low contrast)) and three backlight-glare environments (G1 (high contrast, equivalent to ETDRS), G2 (medium contrast), and G3 (low contrast)) under timed conditions. Vision drop across environments was calculated, and repeatability of visual scores was determined. RESULTS: BCVA significantly reduced with decreasing contrast in all eyes. M1 scores for BCVA were greater than M2 and M3 (P<0.001); G1 scores were greater than G2 and G3 (P<0.01). BCVA dropped more in AMD eyes than in normal eyes between M1 and M2 (P=0.002) and between M1 and M3 (P=0.003). In AMD eyes, BCVA was better using ETDRS charts compared to G1 (P<0.001). The drop in visual function between ETDRS and G1 was greater in AMD eyes compared to normal eyes (P=0.004). Standard deviations of test-retest ranged from 0.100 to 0.139 logMAR. CONCLUSION: The CVA allowed analysis of the visual complaints that AMD patients experience with different lighting/contrast time-dependent conditions. BCVA changed significantly under different lighting/contrast conditions in all eyes, however, AMD eyes were more affected by contrast reduction than normal eyes. In AMD eyes, timed conditions using the CVA led to worse BCVA compared to non-timed ETDRS charts.

Automated detection of choroid boundary and vessels in optical coherence tomography images.

Structural changes in the choroid, a layer located between the retina and sclera, could indicate various vision impairments. Consequently, ophthalmologists inspect optical coherence tomography (OCT) scans of the posterior section of the eye towards making diagnosis. With a view to assist diagnosis, we propose an automated technique for segmentation of the choroid layer. Specifically, we detect the upper and lower boundaries of the choroid using structural similarity and adaptive Hessian analysis. Subsequently, we detect choroid vessels within those boundaries using a level set method. Experimental results are presented using spectral domain (SD) OCT images.