The volume and complexity of diagnostic imaging is increasing at a pace faster than the availability of human expertise to interpret it. Artificial intelligence has shown great promise in classifying ...two-dimensional photographs of some common diseases and typically relies on databases of millions of annotated images. Until now, the challenge of reaching the performance of expert clinicians in a real-world clinical pathway with three-dimensional diagnostic scans has remained unsolved. Here, we apply a novel deep learning architecture to a clinically heterogeneous set of three-dimensional optical coherence tomography scans from patients referred to a major eye hospital. We demonstrate performance in making a referral recommendation that reaches or exceeds that of experts on a range of sight-threatening retinal diseases after training on only 14,884 scans. Moreover, we demonstrate that the tissue segmentations produced by our architecture act as a device-independent representation; referral accuracy is maintained when using tissue segmentations from a different type of device. Our work removes previous barriers to wider clinical use without prohibitive training data requirements across multiple pathologies in a real-world setting.
With the increasing prevalence of diabetes, annual screening for diabetic retinopathy (DR) by expert human grading of retinal images is challenging. Automated DR image assessment systems (ARIAS) may ...provide clinically effective and cost-effective detection of retinopathy. We aimed to determine whether ARIAS can be safely introduced into DR screening pathways to replace human graders.
Observational measurement comparison study of human graders following a national screening program for DR versus ARIAS.
Retinal images from 20 258 consecutive patients attending routine annual diabetic eye screening between June 1, 2012, and November 4, 2013.
Retinal images were manually graded following a standard national protocol for DR screening and were processed by 3 ARIAS: iGradingM, Retmarker, and EyeArt. Discrepancies between manual grades and ARIAS results were sent to a reading center for arbitration.
Screening performance (sensitivity, false-positive rate) and diagnostic accuracy (95% confidence intervals of screening-performance measures) were determined. Economic analysis estimated the cost per appropriate screening outcome.
Sensitivity point estimates (95% confidence intervals) of the ARIAS were as follows: EyeArt 94.7% (94.2%-95.2%) for any retinopathy, 93.8% (92.9%-94.6%) for referable retinopathy (human graded as either ungradable, maculopathy, preproliferative, or proliferative), 99.6% (97.0%-99.9%) for proliferative retinopathy; Retmarker 73.0% (72.0 %-74.0%) for any retinopathy, 85.0% (83.6%-86.2%) for referable retinopathy, 97.9% (94.9%-99.1%) for proliferative retinopathy. iGradingM classified all images as either having disease or being ungradable. EyeArt and Retmarker saved costs compared with manual grading both as a replacement for initial human grading and as a filter prior to primary human grading, although the latter approach was less cost-effective.
Retmarker and EyeArt systems achieved acceptable sensitivity for referable retinopathy when compared with that of human graders and had sufficient specificity to make them cost-effective alternatives to manual grading alone. ARIAS have the potential to reduce costs in developed-world health care economies and to aid delivery of DR screening in developing or remote health care settings.
The aim of this study was to investigate the repeatability of manual measurements of choroidal thickness in healthy subjects imaged on spectral domain optical coherence tomography (OCT) using the ...enhanced depth imaging (EDI) technique.
Fifty consecutive, healthy, young, adult volunteers with no known eye disease were enrolled prospectively. Two good-quality horizontal and vertical line scans through the fovea were obtained for each eye. Using the manual calipers provided by the software of the proprietary device, two experienced OCT readers measured the subfoveal choroidal thickness (SFCT) of the horizontal and vertical line scans for all eyes. The readers were masked to each other's readings. Intraobserver, interobserver, and intrasession coefficients of repeatability (CRs) were calculated.
Mean (standard deviation SD) age of the study subjects was 38 (5) years (range, 30-49 years). Mean (SD) subfoveal choroidal thickness was 332 (90) μm (right eyes) and 332 (91) μm (left eyes). Intraobserver CR was approximately 23 (95% confidence interval CI, 19-26) μm, whereas interobserver and intrasession CRs were greater at 32 (95% CI, 30-34) and 34 (95% CI, 32-36) μm, respectively. There was no significant difference in SFCT between all pairs of SFCT measurements except for the two intrasession vertical line scans.
A change of >32 μm was likely to exceed interobserver variability in SFCT. Future studies are required to estimate the repeatability of SFCT measurements in patients with chorioretinal pathology.
To evaluate the application of an optical coherence tomography angiography (OCTA) system adapted for the assessment of anterior segment vasculature.
Cross-sectional, observational study.
Consecutive ...subjects with normal eyes on slit-lamp clinical examination and patients with abnormal corneal neovascularization.
All scans were performed using a commercially available AngioVue OCTA system (Optovue, Inc., Fremont, CA) using an anterior segment lens adapter and the split-spectrum amplitude decorrelation angiography algorithm. Each subject underwent scans from 4 quadrants (superior, inferior, nasal, and temporal) in each eye by 2 trained, independent operators.
Analysis of signal strength, image quality, and reproducibility of corneal vascular measurements was performed.
In our study of 20 normal subjects (10 men, 10 women; mean age, 25.3±7.8 years), we found good repeatability (κ coefficient, 0.76) for image quality score and good interobserver agreement for vasculature measurements (intraclass coefficient, 0.94). After optimization of the angiography scan protocol, vascular measurements within the regions of interest were compared in the superior versus inferior quadrants (mean vascular loops, 3.34±1.16 vs. 3.12 ± 0.90 P = 0.768; segment-to-loop ratio, 4.18±0.71 vs. 4.32±0.87 P = 0.129; fractal dimension Df value, 1.78±0.06 vs. 1.78±0.06 P = 0.94; vascular loop area, 25.9±14.5 vs. 25.9±10.7 × 10(-3) mm(2) P = 0.21) and nasal versus temporal quadrant (mean vascular loops, 2.89±0.98 vs. 3.57±0.99 P < 0.001; segment-to-loop ratio, 3.94±0.69 vs. 4.55±0.78 P = 0.897; Df value, 1.78±0.06 vs. 1.77±0.06 P = 0.14; vascular loop area, 29.7±15.7 vs. 22.1±7.1 × 10(-3) mm(2) P = 0.38. We then used the established OCTA scanning protocol to visualize abnormal vasculature successfully in 5 patients with various corneal pathologic features, including graft-associated neovascularization, postherpetic keratitis scarring, lipid keratopathy, and limbal stem cell deficiency.
This preliminary study describes a method for acquiring OCTA images of the cornea and limbal vasculature with substantial consistency. This technique may be useful for the objective evaluation of corneal neovascularization in the future.
Age-related macular degeneration (AMD) remains a major cause of blindness, with dysfunction and loss of retinal pigment epithelium (RPE) central to disease progression. We engineered an RPE patch ...comprising a fully differentiated, human embryonic stem cell (hESC)-derived RPE monolayer on a coated, synthetic basement membrane. We delivered the patch, using a purpose-designed microsurgical tool, into the subretinal space of one eye in each of two patients with severe exudative AMD. Primary endpoints were incidence and severity of adverse events and proportion of subjects with improved best-corrected visual acuity of 15 letters or more. We report successful delivery and survival of the RPE patch by biomicroscopy and optical coherence tomography, and a visual acuity gain of 29 and 21 letters in the two patients, respectively, over 12 months. Only local immunosuppression was used long-term. We also present the preclinical surgical, cell safety and tumorigenicity studies leading to trial approval. This work supports the feasibility and safety of hESC-RPE patch transplantation as a regenerative strategy for AMD.
Purpose To evaluate the efficacy of spectral-domain optical coherence tomography (SD OCT) compared to indocyanine green angiography (ICGA) in detecting idiopathic polypoidal choroidal vasculopathy ...(PCV) and in differentiating between PCV and occult choroidal neovascularization (CNV). Design Retrospective observational case-control study. Methods SD OCTs of 51 eyes of 44 consecutive patients who presented with 1 or more pigment epithelial detachments (PEDs) attributable to either PCV or occult CNV were retrospectively reviewed by a grader masked to the final diagnosis. A qualitative analysis based on the following tomographic findings was performed: sharp PED peak, PED notch, hyporeflective lumen within hyperreflective lesions adherent to retinal pigment epithelium. The diagnosis based on SD OCT alone was compared with the final diagnosis made using ICGA and fluorescein angiography. Sensitivity and specificity were calculated. Patients with classic CNV and central serous chorioretinopathy were excluded. Results Among 51 eyes of 44 patients, 37 had an ICGA-confirmed diagnosis of PCV and 14 had occult CNV. SD OCT based on the features above detected 35 of 37 true-positive PCV lesions but missed 2 ICGA-confirmed lesions (false negatives). SD OCT correctly excluded 13 of 14 non-PCV lesions but misidentified 1 PCV lesion (false positive). These data showed a sensitivity of 94.6% and a specificity of 92.9% for the above SD OCT features in identifying PCV lesions. Conclusions SD OCT based on the features above allowed for good detection of PCV and differentiation between PCV and occult CNV in this selected clinic population. A careful qualitative analysis of the tomographic findings in patients presenting with PEDs may allow ophthalmologists to distinguish between PCV and occult CNV, decreasing the need for ICGA and the risks related to this procedure.
Purpose To define a minimum set of outcome measures for tracking, comparing, and improving macular degeneration care. Design Recommendations from working-group of international experts in macular ...degeneration outcomes registry development and patient advocates, facilitated by the International Consortium for Health Outcomes Measurement (ICHOM). Methods Modified Delphi technique, supported by structured teleconferences, followed by online-surveys to drive consensus decisions. Potential outcomes were identified through literature review of outcomes collected in existing registries and reported in major clinical trials. Outcomes were refined by the working-group and selected based upon impact on patients, relationship to good clinical care and feasibility of measurement in routine clinical practice. Results Standardized measurement of the following outcomes is recommended: visual functioning and quality of life (distance visual acuity, mobility and independence, emotional well-being, reading and accessing information); number of treatments; complications of treatment; and disease-control. Proposed data-collection sources include administrative, clinical data during routine clinical visits and patient-reported sources annually. Recording the following clinical characteristics is recommended to enable risk-adjustment: age; gender; ethnicity; smoking status; baseline visual acuity in both eyes; type of macular degeneration; presence of geographic atrophy, subretinal fibrosis or pigment epithelial detachment; previous macular degeneration treatment; ocular co-morbidities. Conclusions The recommended minimum outcomes and pragmatic reporting standards should enable standardized, meaningful assessments and comparisons of macular degeneration treatment outcomes. Adoption could accelerate global improvements in standardized data-gathering and reporting of patient-centered outcomes. This can facilitate informed decisions by patients and health care providers, plus allow long-term monitoring of aggregate data, ultimately improving understanding of disease progression and treatment responses.
Despite advances in artificial intelligence (AI), its application in medical imaging has been burdened and limited by expert-generated labels. We used images from optical coherence tomography ...angiography (OCTA), a relatively new imaging modality that measures retinal blood flow, to train an AI algorithm to generate flow maps from standard optical coherence tomography (OCT) images, exceeding the ability and bypassing the need for expert labeling. Deep learning was able to infer flow from single structural OCT images with similar fidelity to OCTA and significantly better than expert clinicians (P < 0.00001). Our model allows generating flow maps from large volumes of previously collected OCT data in existing clinical trials and clinical practice. This finding demonstrates a novel application of AI to medical imaging, whereby subtle regularities between different modalities are used to image the same body part and AI is used to generate detailed inferences of tissue function from structure imaging.
Purpose:
Dexamethasone intravitreal implant and intravitreal ranibizumab are indicated for the treatment of macular edema secondary to retinal vein occlusion. This non-inferiority study compared ...dexamethasone with ranibizumab in patients with branch retinal vein occlusion.
Methods:
In this randomized, 12-month head-to-head comparison, subjects with branch retinal vein occlusion were assigned to dexamethasone 0.7 mg at day 1 and month 5 with the option of retreatment at month 10 or 11, or ranibizumab 0.5 mg at day 1 and monthly through month 5 with subsequent as-needed injections at month 6–month 11. The primary efficacy outcome was the mean change from baseline in best-corrected visual acuity at month 12; secondary outcomes included average change in best-corrected visual acuity, proportion of eyes with ≥10- and ≥15-letter gain/loss, change in central retinal thickness, and change in Vision Functioning Questionnaire-25 score.
Results:
In all, 307 of a planned 400 patients were enrolled in the study and received (mean) 2.5 dexamethasone injections (n = 154) and 8.0 ranibizumab injections (n = 153) over 12 months. The mean change from baseline in best-corrected visual acuity at month 12 was 7.4 letters for dexamethasone versus 17.4 letters for ranibizumab (least-squares mean difference (dexamethasone minus ranibizumab), −10.1 letters; 95% confidence interval, −12.9, −7.2; p = 0.0006).
Conclusion:
Dexamethasone and ranibizumab improved best-corrected visual acuity and anatomical outcomes; however, dexamethasone did not show non-inferiority to ranibizumab in this under-powered study. Dexamethasone was associated with an increased risk of intraocular pressure elevation and cataract progression, but a lower injection burden, compared to ranibizumab.
A pilot cross sectional study was conducted to investigate the role of red blood cells (RBC) deformability in type 2 diabetes mellitus (T2DM) without and with diabetic retinopathy (DR) using a dual ...optical tweezers stretching technique. A dual optical tweezers was made by splitting and recombining a single Nd:YAG laser beam. RBCs were trapped directly (i.e., without microbead handles) in the dual optical tweezers where they were observed to adopt a "side-on" orientation. RBC initial and final lengths after stretching were measured by digital video microscopy, and a Deformability index (DI) calculated. Blood from 8 healthy controls, 5 T2DM and 7 DR patients with respective mean age of 52.4 yrs, 51.6 yrs and 52 yrs was analysed. Initial average length of RBCs for control group was 8.45 ± 0.25 μm, 8.68 ± 0.49 μm for DM RBCs and 8.82 ± 0.32 μm for DR RBCs (p < 0.001). The DI for control group was 0.0698 ± 0.0224, and that for DM RBCs was 0.0645 ± 0.03 and 0.0635 ± 0.028 (p < 0.001) for DR group. DI was inversely related to basal length of RBCs (p = .02). DI of RBC from DM and DR patients was significantly lower in comparison with normal healthy controls. A dual optical tweezers method can hence be reliably used to assess RBC deformability.