Publications by authors named "Aaron S Coyner"

12 Publications

  • Page 1 of 1

Single-Examination Risk Prediction of Severe Retinopathy of Prematurity.

Pediatrics 2021 Nov 23. Epub 2021 Nov 23.

Departments of Ophthalmology.

Background And Objectives: Retinopathy of prematurity (ROP) is a leading cause of childhood blindness. Screening and treatment reduces this risk, but requires multiple examinations of infants, most of whom will not develop severe disease. Previous work has suggested that artificial intelligence may be able to detect incident severe disease (treatment-requiring retinopathy of prematurity [TR-ROP]) before clinical diagnosis. We aimed to build a risk model that combined artificial intelligence with clinical demographics to reduce the number of examinations without missing cases of TR-ROP.

Methods: Infants undergoing routine ROP screening examinations (1579 total eyes, 190 with TR-ROP) were recruited from 8 North American study centers. A vascular severity score (VSS) was derived from retinal fundus images obtained at 32 to 33 weeks' postmenstrual age. Seven ElasticNet logistic regression models were trained on all combinations of birth weight, gestational age, and VSS. The area under the precision-recall curve was used to identify the highest-performing model.

Results: The gestational age + VSS model had the highest performance (mean ± SD area under the precision-recall curve: 0.35 ± 0.11). On 2 different test data sets (n = 444 and n = 132), sensitivity was 100% (positive predictive value: 28.1% and 22.6%) and specificity was 48.9% and 80.8% (negative predictive value: 100.0%).

Conclusions: Using a single examination, this model identified all infants who developed TR-ROP, on average, >1 month before diagnosis with moderate to high specificity. This approach could lead to earlier identification of incident severe ROP, reducing late diagnosis and treatment while simultaneously reducing the number of ROP examinations and unnecessary physiologic stress for low-risk infants.
View Article and Find Full Text PDF

Download full-text PDF

Source
http://dx.doi.org/10.1542/peds.2021-051772DOI Listing
November 2021

Diagnosability of Synthetic Retinal Fundus Images for Plus Disease Detection in Retinopathy of Prematurity.

AMIA Annu Symp Proc 2020 25;2020:329-337. Epub 2021 Jan 25.

Medical Informatics & Clinical Epidemiology.

Advances in generative adversarial networks have allowed for engineering of highly-realistic images. Many studies have applied these techniques to medical images. However, evaluation of generated medical images often relies upon image quality and reconstruction metrics, and subjective evaluation by laypersons. This is acceptable for generation of images depicting everyday objects, but not for medical images, where there may be subtle features experts rely upon for diagnosis. We implemented the pix2pix generative adversarial network for retinal fundus image generation, and evaluated the ability of experts to identify generated images as such and to form accurate diagnoses of plus disease in retinopathy of prematurity. We found that, while experts could discern between real and generated images, the diagnoses between image sets were similar. By directly evaluating and confirming physicians' abilities to diagnose generated retinal fundus images, this work supports conclusions that generated images may be viable for dataset augmentation and physician training.
View Article and Find Full Text PDF

Download full-text PDF

Source
http://www.ncbi.nlm.nih.gov/pmc/articles/PMC8075515PMC
June 2021

Deep Learning for the Diagnosis of Stage in Retinopathy of Prematurity: Accuracy and Generalizability across Populations and Cameras.

Ophthalmol Retina 2021 10 6;5(10):1027-1035. Epub 2021 Feb 6.

Department of Ophthalmology, Casey Eye Institute, Oregon Health & Science University, Portland, Oregon. Electronic address:

Purpose: Stage is an important feature to identify in retinal images of infants at risk of retinopathy of prematurity (ROP). The purpose of this study was to implement a convolutional neural network (CNN) for binary detection of stages 1, 2, and 3 in ROP and to evaluate its generalizability across different populations and camera systems.

Design: Diagnostic validation study of CNN for stage detection.

Participants: Retinal fundus images obtained from preterm infants during routine ROP screenings.

Methods: Two datasets were used: 5943 fundus images obtained by RetCam camera (Natus Medical, Pleasanton, CA) from 9 North American institutions and 5049 images obtained by 3nethra camera (Forus Health Incorporated, Bengaluru, India) from 4 hospitals in Nepal. Images were labeled based on the presence of stage by 1 to 3 expert graders. Three CNN models were trained using 5-fold cross-validation on datasets from North America alone, Nepal alone, and a combined dataset and were evaluated on 2 held-out test sets consisting of 708 and 247 images from the Nepali and North American datasets, respectively.

Main Outcome Measures: Convolutional neural network performance was evaluated using area under the receiver operating characteristic curve (AUROC), area under the precision-recall curve (AUPRC), sensitivity, and specificity.

Results: Both the North American- and Nepali-trained models demonstrated high performance on a test set from the same population: AUROC, 0.99; AUPRC, 0.98; sensitivity, 94%; and AUROC, 0.97; AUPRC, 0.91; and sensitivity, 73%; respectively. However, the performance of each model decreased to AUROC of 0.96 and AUPRC of 0.88 (sensitivity, 52%) and AUROC of 0.62 and AUPRC of 0.36 (sensitivity, 44%) when evaluated on a test set from the other population. Compared with the models trained on individual datasets, the model trained on a combined dataset achieved improved performance on each respective test set: sensitivity improved from 94% to 98% on the North American test set and from 73% to 82% on the Nepali test set.

Conclusions: A CNN can identify accurately the presence of ROP stage in retinal images, but performance depends on the similarity between training and testing populations. We demonstrated that internal and external performance can be improved by increasing the heterogeneity of the training dataset features of the training dataset, in this case by combining images from different populations and cameras.
View Article and Find Full Text PDF

Download full-text PDF

Source
http://dx.doi.org/10.1016/j.oret.2020.12.013DOI Listing
http://www.ncbi.nlm.nih.gov/pmc/articles/PMC8364291PMC
October 2021

Introduction to Machine Learning, Neural Networks, and Deep Learning.

Transl Vis Sci Technol 2020 02 27;9(2):14. Epub 2020 Feb 27.

Department of Ophthalmology, Casey Eye Institute, Oregon Health & Science University (OHSU), Portland, Oregon, United States.

Purpose: To present an overview of current machine learning methods and their use in medical research, focusing on select machine learning techniques, best practices, and deep learning.

Methods: A systematic literature search in PubMed was performed for articles pertinent to the topic of artificial intelligence methods used in medicine with an emphasis on ophthalmology.

Results: A review of machine learning and deep learning methodology for the audience without an extensive technical computer programming background.

Conclusions: Artificial intelligence has a promising future in medicine; however, many challenges remain.

Translational Relevance: The aim of this review article is to provide the nontechnical readers a layman's explanation of the machine learning methods being used in medicine today. The goal is to provide the reader a better understanding of the potential and challenges of artificial intelligence within the field of medicine.
View Article and Find Full Text PDF

Download full-text PDF

Source
http://dx.doi.org/10.1167/tvst.9.2.14DOI Listing
http://www.ncbi.nlm.nih.gov/pmc/articles/PMC7347027PMC
February 2020

Automated Fundus Image Quality Assessment in Retinopathy of Prematurity Using Deep Convolutional Neural Networks.

Ophthalmol Retina 2019 05 31;3(5):444-450. Epub 2019 Jan 31.

Department of Medical Informatics and Clinical Epidemiology, Oregon Health and Science University, Portland, Oregon; Department of Ophthalmology, Casey Eye Institute, Oregon Health and Science University, Portland, Oregon. Electronic address:

Purpose: Accurate image-based ophthalmic diagnosis relies on fundus image clarity. This has important implications for the quality of ophthalmic diagnoses and for emerging methods such as telemedicine and computer-based image analysis. The purpose of this study was to implement a deep convolutional neural network (CNN) for automated assessment of fundus image quality in retinopathy of prematurity (ROP).

Design: Experimental study.

Participants: Retinal fundus images were collected from preterm infants during routine ROP screenings.

Methods: Six thousand one hundred thirty-nine retinal fundus images were collected from 9 academic institutions. Each image was graded for quality (acceptable quality [AQ], possibly acceptable quality [PAQ], or not acceptable quality [NAQ]) by 3 independent experts. Quality was defined as the ability to assess an image confidently for the presence of ROP. Of the 6139 images, NAQ, PAQ, and AQ images represented 5.6%, 43.6%, and 50.8% of the image set, respectively. Because of low representation of NAQ images in the data set, images labeled NAQ were grouped into the PAQ category, and a binary CNN classifier was trained using 5-fold cross-validation on 4000 images. A test set of 2109 images was held out for final model evaluation. Additionally, 30 images were ranked from worst to best quality by 6 experts via pairwise comparisons, and the CNN's ability to rank quality, regardless of quality classification, was assessed.

Main Outcome Measures: The CNN performance was evaluated using area under the receiver operating characteristic curve (AUC). A Spearman's rank correlation was calculated to evaluate the overall ability of the CNN to rank images from worst to best quality as compared with experts.

Results: The mean AUC for 5-fold cross-validation was 0.958 (standard deviation, 0.005) for the diagnosis of AQ versus PAQ images. The AUC was 0.965 for the test set. The Spearman's rank correlation coefficient on the set of 30 images was 0.90 as compared with the overall expert consensus ranking.

Conclusions: This model accurately assessed retinal fundus image quality in a comparable manner with that of experts. This fully automated model has potential for application in clinical settings, telemedicine, and computer-based image analysis in ROP and for generalizability to other ophthalmic diseases.
View Article and Find Full Text PDF

Download full-text PDF

Source
http://dx.doi.org/10.1016/j.oret.2019.01.015DOI Listing
http://www.ncbi.nlm.nih.gov/pmc/articles/PMC6501831PMC
May 2019

Demystifying the Jargon: The Bridge between Ophthalmology and Artificial Intelligence.

Ophthalmol Retina 2019 04;3(4):291-293

Department of Medical Informatics and Clinical Epidemiology, Oregon Health and Science University, Portland, Oregon.; Department of Ophthalmology, Casey Eye Institute, Oregon Health and Science University, Portland, Oregon.. Electronic address:

View Article and Find Full Text PDF

Download full-text PDF

Source
http://dx.doi.org/10.1016/j.oret.2018.12.008DOI Listing
http://www.ncbi.nlm.nih.gov/pmc/articles/PMC7874933PMC
April 2019

Deep Learning for Image Quality Assessment of Fundus Images in Retinopathy of Prematurity.

AMIA Annu Symp Proc 2018 5;2018:1224-1232. Epub 2018 Dec 5.

Medical Informatics & Clinical Epidemiology, and.

Accurate image-based medical diagnosis relies upon adequate image quality and clarity. This has important implications for clinical diagnosis, and for emerging methods such as telemedicine and computer-based image analysis. In this study, we trained a convolutional neural network (CNN) to automatically assess the quality of retinal fundus images in a representative ophthalmic disease, retinopathy of prematurity (ROP). 6,043 wide-angle fundus images were collected from preterm infants during routine ROP screening examinations. Images were assessed by clinical experts for quality regarding ability to diagnose ROP accurately, and were labeled "acceptable" or "not acceptable." The CNN training, validation and test sets consisted of 2,770 images, 200 images, and 3,073 images, respectively. Test set accuracy was 89.1%, with area under the receiver operating curve equal to 0.964, and area under the precision-recall curve equal to 0.966. Taken together, our CNN shows promise as a useful prescreening method for telemedicine and computer-based image analysis applications. We feel this methodology is generalizable to all clinical domains involving image-based diagnosis.
View Article and Find Full Text PDF

Download full-text PDF

Source
http://www.ncbi.nlm.nih.gov/pmc/articles/PMC6371336PMC
December 2019

The Role of ERK1/2 Activation in Sarpogrelate-Mediated Neuroprotection.

Invest Ophthalmol Vis Sci 2018 01;59(1):462-471

Casey Eye Institute, Oregon Health & Science University, Portland, Oregon, United States.

Purpose: To characterize the mediators of 5-HT2A serotonin receptor-driven retinal neuroprotection.

Methods: Albino mice were treated intraperitoneally with saline or sarpogrelate, a 5-HT2A antagonist, immediately before light exposure (LE). Following LE, retinas were harvested for a high-throughput phosphorylation microarray to quantify activated phosphorylated proteins in G protein-coupled receptor (GPCR) signaling. To confirm microarray results and define temporal changes, Western blots of select GPCR signaling proteins were performed. Since both methodologies implicated MAPK/ERK activation, the functional significance of sarpogrelate-mediated ERK1/2 activation was examined by inhibition of ERK1/2 phosphorylation via pretreatment with the MEK inhibitor (MEKi) PD0325901. The degree of neuroprotection was evaluated with spectral-domain optical coherence tomography (SD-OCT) and electroretinography (ERG). To determine the effects of sarpogrelate on gene expression, a qPCR array measuring the expression of 84 genes involved in oxidative stress and cell death was performed 48 hours post LE.

Results: Sarpogrelate led to an activation of the MAPK/ERK pathway. Temporal analysis further demonstrated a transient activation of ERK1/2, starting with an early inhibition 20 minutes into LE, a maximum activation at 3 hours post LE, and a return to baseline at 7 hours post LE. Inhibition of ERK1/2 with MEKi pretreatment led to attenuation of sarpogrelate-mediated neuroprotection. LE caused significant changes in the expression of genes involved in iron metabolism, oxidative stress, and apoptosis. These changes were prevented by sarpogrelate treatment.

Conclusions: Sarpogrelate-mediated retinal protection involves a transient activation of the MAPK/ERK pathway, although this pathway alone does not account for the full effect of neuroprotection.
View Article and Find Full Text PDF

Download full-text PDF

Source
http://dx.doi.org/10.1167/iovs.17-23159DOI Listing
http://www.ncbi.nlm.nih.gov/pmc/articles/PMC5786286PMC
January 2018

SCLERAL PITS IN CHOROIDEREMIA: Implications for Retinal Gene Therapy.

Retina 2018 09;38(9):1725-1730

Casey Eye Institute, Oregon Health & Science University, Portland, Oregon.

Purpose: We report a novel finding on spectral domain optical coherence tomography in patients with choroideremia, which we describe as scleral pits (SCPs).

Methods: Cross-sectional observational case series of 36 patients with choroideremia, who underwent ophthalmic examination and multimodal imaging, including optical coherence tomography of the macula. Optical coherence tomography images were reviewed for SCP, which were defined as discrete tracts of hyporeflectivity that traverse the sclera with or without the involvement of Bruch membrane, retinal pigment epithelium, and retina. Unpaired two-tailed t-test with Welch correction was used for statistical analysis.

Results: Of the 36 patients, 19 had SCP in at least one eye. Scleral pits were confined to areas of advanced chorioretinal degeneration and never involved the foveola. Type 1 SCP affected only the sclera, whereas Type 2 SCP also involved the Bruch membrane and the retinal pigment epithelium. Type 3 SCP additionally had a full-thickness retinal defect. Patients with SCP were significantly older (51 ± 2 vs. 33 ± 4 years; P < 0.05) and had lower best-corrected visual acuity (20/160 vs. 20/30 or 0.9 ± 0.2 vs. 0.2 ± 0.07 logarithm of the minimum angle of resolution; P < 0.05) than patients without SCP. Patients with SCP had a greater myopic refractive error compared with patients without SCP (-2.6 ± 0.5 vs. -0.3 ± 0.5D; P < 0.05), but there was no significant correlation between the number of SCPs with refraction. Short posterior ciliary arteries were observed to enter the eye through one Type 3 SCP.

Conclusion: Scleral pits are, to the best of our knowledge, a novel optical coherence tomography finding in advanced choroideremia that likely represents the abnormal juxtaposition of penetrating short posterior ciliary arteries with the retina.
View Article and Find Full Text PDF

Download full-text PDF

Source
http://dx.doi.org/10.1097/IAE.0000000000001957DOI Listing
http://www.ncbi.nlm.nih.gov/pmc/articles/PMC5955782PMC
September 2018

Long-term Characterization of Retinal Degeneration in Royal College of Surgeons Rats Using Spectral-Domain Optical Coherence Tomography.

Invest Ophthalmol Vis Sci 2017 03;58(3):1378-1386

Casey Eye Institute, Oregon Health & Science University, Portland, Oregon, United States 3Department of Neuroscience, Oregon National Primate Research Center, Oregon Health & Science University, Beaverton, Oregon, United States.

Purpose: Prospective treatments for age-related macular degeneration and inherited retinal degenerations are commonly evaluated in the Royal College of Surgeons (RCS) rat before translation into clinical application. Historically, retinal thickness obtained through postmortem anatomic assessments has been a key outcome measure; however, utility of this measurement is limited because it precludes the ability to perform longitudinal studies. To overcome this limitation, the present study was designed to provide a baseline longitudinal quantification of retinal thickness in the RCS rat by using spectral-domain optical coherence tomography (SD-OCT).

Methods: Horizontal and vertical linear SD-OCT scans centered on the optic nerve were captured from Long-Evans control rats at P30, P60, P90 and from RCS rats between P17 and P90. Total retina (TR), outer nuclear layer+ (ONL+), inner nuclear layer (INL), and retinal pigment epithelium (RPE) thicknesses were quantified. Histologic sections of RCS retina obtained from P21 to P60 were compared to SD-OCT images.

Results: In RCS rats, TR and ONL+ thickness decreased significantly as compared to Long-Evans controls. Changes in INL and RPE thickness were not significantly different between control and RCS retinas. From P30 to P90 a subretinal hyperreflective layer (HRL) was observed and quantified in RCS rats. After correlation with histology, the HRL was identified as disorganized outer segments and the location of accumulated debris.

Conclusions: Retinal layer thickness can be quantified longitudinally throughout the course of retinal degeneration in the RCS rat by using SD-OCT. Thickness measurements obtained with SD-OCT were consistent with previous anatomic thickness assessments. This study provides baseline data for future longitudinal assessment of therapeutic agents in the RCS rat.
View Article and Find Full Text PDF

Download full-text PDF

Source
http://dx.doi.org/10.1167/iovs.16-20363DOI Listing
http://www.ncbi.nlm.nih.gov/pmc/articles/PMC5361458PMC
March 2017

Retinal Neuroprotective Effects of Flibanserin, an FDA-Approved Dual Serotonin Receptor Agonist-Antagonist.

PLoS One 2016 22;11(7):e0159776. Epub 2016 Jul 22.

Casey Eye Institute, Oregon Health & Science University, Portland, Oregon, United States of America.

Purpose: To assess the neuroprotective effects of flibanserin (formerly BIMT-17), a dual 5-HT1A agonist and 5-HT2A antagonist, in a light-induced retinopathy model.

Methods: Albino BALB/c mice were injected intraperitoneally with either vehicle or increasing doses of flibanserin ranging from 0.75 to 15 mg/kg flibanserin. To assess 5-HT1A-mediated effects, BALB/c mice were injected with 10 mg/kg WAY 100635, a 5-HT1A antagonist, prior to 6 mg/kg flibanserin and 5-HT1A knockout mice were injected with 6 mg/kg flibanserin. Injections were administered once immediately prior to light exposure or over the course of five days. Light exposure lasted for one hour at an intensity of 10,000 lux. Retinal structure was assessed using spectral domain optical coherence tomography and retinal function was assessed using electroretinography. To investigate the mechanisms of flibanserin-mediated neuroprotection, gene expression, measured by RT-qPCR, was assessed following five days of daily 15 mg/kg flibanserin injections.

Results: A five-day treatment regimen of 3 to 15 mg/kg of flibanserin significantly preserved outer retinal structure and function in a dose-dependent manner. Additionally, a single-day treatment regimen of 6 to 15 mg/kg of flibanserin still provided significant protection. The action of flibanserin was hindered by the 5-HT1A antagonist, WAY 100635, and was not effective in 5-HT1A knockout mice. Creb, c-Jun, c-Fos, Bcl-2, Cast1, Nqo1, Sod1, and Cat were significantly increased in flibanserin-injected mice versus vehicle-injected mice.

Conclusions: Intraperitoneal delivery of flibanserin in a light-induced retinopathy mouse model provides retinal neuroprotection. Mechanistic data suggests that this effect is mediated through 5-HT1A receptors and that flibanserin augments the expression of genes capable of reducing mitochondrial dysfunction and oxidative stress. Since flibanserin is already FDA-approved for other indications, the potential to repurpose this drug for treating retinal degenerations merits further investigation.
View Article and Find Full Text PDF

Download full-text PDF

Source
http://journals.plos.org/plosone/article?id=10.1371/journal.pone.0159776PLOS
http://www.ncbi.nlm.nih.gov/pmc/articles/PMC4957778PMC
July 2017

Sarpogrelate, a 5-HT2A Receptor Antagonist, Protects the Retina From Light-Induced Retinopathy.

Invest Ophthalmol Vis Sci 2015 Jul;56(8):4560-9

Casey Eye Institute Oregon Health & Science University, Portland, Oregon, United States.

Purpose: To determine if sarpogrelate, a selective 5-HT2A receptor antagonist, is protective against light-induced retinopathy in BALB/c mice.

Methods: BALB/c mice were dosed intraperitoneally with 5, 15, 30, 40, or 50 mg/kg sarpogrelate 48, 24, and 0 hours prior to bright light exposure (10,000 lux) as well as 24 and 48 hours after exposure. Additionally, a single injection regimen was evaluated by injecting mice with 50 mg/kg sarpogrelate once immediately prior to light exposure. To investigate the potential for additive effects of serotonin receptor agents, a combination therapy consisting of sarpogrelate (15 mg/kg) and 8-OH-DPAT (1 mg/kg) was evaluated with the 5-day treatment regimen. Neuroprotection was characterized by the preservation of retinal thickness and function, measured by spectral-domain optical coherence tomography (SD-OCT) and electroretinography (ERG), respectively.

Results: Mice that were light damaged and injected with saline had significantly reduced outer retinal thickness, total retinal thickness, and ERG amplitudes compared with naïve mice. A 5-day administration of 15, 30, or 40 mg/kg of sarpogrelate was able to partially protect retinal morphology and full protection of retinal morphology was achieved with a 50 mg/kg dose. Both 15 and 30 mg/kg doses of sarpogrelate partially preserved retinal function measured by ERG, whereas 40 and 50 mg/kg doses fully preserved retinal function. Additionally, a single administration of 50 mg/kg sarpogrelate was able to fully preserve both retinal morphology and function. Administration of 15 mg/kg of sarpogrelate and 1 mg/kg of 8-OH-DPAT together demonstrated an additive effect and fully preserved retinal morphology.

Conclusions: A 5- or 1-day treatment with 50 mg/kg sarpogrelate can completely protect the retina of BALB/c mice from light-induced retinopathy. Partial protection can be achieved with lower doses starting at 15 mg/kg and protection increases in a dose-dependent manner. Treatment with low doses of sarpogrelate and 8-OH-DPAT elicits an additive effect that results in full protection of retinal morphology.
View Article and Find Full Text PDF

Download full-text PDF

Source
http://dx.doi.org/10.1167/iovs.15-16378DOI Listing
http://www.ncbi.nlm.nih.gov/pmc/articles/PMC4515947PMC
July 2015
-->