No full-text available
To read the full-text of this research,
you can request a copy directly from the authors.
Glaucomatous Optic Neuropathy Evaluation Project Factors Associated With Underestimation of Glaucoma Likelihood
Abstract
Importance
Glaucoma is a significant health problem for which diagnosis remains suboptimal. Optic disc evaluation, which is fundamental to the diagnosis, is a difficult skill to acquire.Objectives
To determine the optic disc characteristics that most influence decision making in the assessment of glaucoma likelihood and to ascertain the optic disc features associated with overestimation and underestimation of glaucoma likelihood.Design, Setting, and Participants
This prospective, observational, Internet-based study with multinational participation included 197 ophthalmic clinicians (37 glaucoma subspecialists, 51 comprehensive ophthalmologists, and 109 ophthalmology trainees) from 22 countries who self-registered for the Glaucomatous Optic Neuropathy Evaluation (GONE) Project from December 1, 2008 through June 30, 2010.Interventions
A series of 42 monoscopic optic disc photographs of healthy and glaucomatous eyes were presented to clinicians using the GONE Project Program. Participants were asked to assess each disc according to 9 conventional topographic features and assign a presumptive grade for glaucoma likelihood.Main Outcomes and Measures
Agreement (κ and weighted κ) among participants for disc signs and glaucoma likelihood and contributions of disc-related factors to overestimation and underestimation of glaucoma likelihood.Results
Ophthalmology trainees and comprehensive ophthalmologists underestimated glaucoma likelihood in a mean (SD) of 22.1% (1.6%) and 23.8% (1.8%) of discs, respectively. Underestimation of vertical cup-disc ratio and failure to identify retinal nerve fiber layer loss, disc hemorrhage, or rim loss were most likely to lead to underestimation of glaucoma. When all 4 features were inaccurately assessed, underestimation of glaucoma likelihood increased to 43.0%. Ophthalmology trainees and comprehensive ophthalmologists overestimated glaucoma likelihood in a mean (SD) of 13.0% (1.2%) and 8.9% (1.3%) of discs, respectively. Overestimation of glaucoma likelihood was associated with overestimation of retinal nerve fiber layer loss, rim loss, vertical cup-disc ratio, disc hemorrhage, and incorrect assessment of disc tilt and was more likely in large discs.Conclusions and Relevance
Ophthalmology trainees and comprehensive ophthalmologists underestimated glaucoma likelihood in approximately 1 in 5 disc photographs and were twice as likely to underestimate as overestimate glaucoma likelihood. Underestimating the vertical cup-disc ratio and cup shape and missing retinal nerve fiber layer defects and disc hemorrhage were the key errors that led to underestimation. When all 4 parameters were incorrectly assessed, underestimation increased to almost 1 in 2.
... Furthermore, there is substantial variation in the visual interpretation of ONH features among clinicians [90]. This variation can be present between different clinicians (interobserver variation) and between assessments made at different times by the same clinician (intra-observer inter-session variation). ...
... This variation can be present between different clinicians (interobserver variation) and between assessments made at different times by the same clinician (intra-observer inter-session variation). A recent study [90] involving 197 ophthalmic clinicians from 22 countries showed substantial under diagnosis based on optic nerve head photos from patients with known glaucoma. Ophthalmology trainees (22%) and comprehensive ophthalmologists (24%) consistently underestimated the likelihood of glaucoma. ...
Glaucoma is the commonest cause of irreversible blindness worldwide, with over 70% of people affected remaining undiagnosed. Early detection is crucial for halting progressive visual impairment in glaucoma patients, as there is no cure available. This narrative review aims to: identify reasons for the significant under-diagnosis of glaucoma globally, particularly in Australia, elucidate the role of primary healthcare in glaucoma diagnosis using Australian healthcare as an example, and discuss how recent advances in artificial intelligence (AI) can be implemented to improve diagnostic outcomes. Glaucoma is a prevalent disease in ageing populations and can have improved visual outcomes through appropriate treatment, making it essential for general medical practice. In countries such as Australia, New Zealand, Canada, USA, and the UK, optometrists serve as the gatekeepers for primary eye care, and glaucoma detection often falls on their shoulders. However, there is significant variation in the capacity for glaucoma diagnosis among eye professionals. Automation with Artificial Intelligence (AI) analysis of optic nerve photos can help optometrists identify high-risk changes and mitigate the challenges of image interpretation rapidly and consistently. Despite its potential, there are significant barriers and challenges to address before AI can be deployed in primary healthcare settings, including external validation, high quality real-world implementation, protection of privacy and cybersecurity, and medico-legal implications. Overall, the incorporation of AI technology in primary healthcare has the potential to reduce the global prevalence of undiagnosed glaucoma cases by improving diagnostic accuracy and efficiency.
... The DLS outperformed non-physician graders in the classification of referable GON in this study. The variability in inter-assessor agreement among non-physician graders and ophthalmologists for the classification of ocular disorders is well known, especially glaucoma (37,38). The Glaucomatous optic neuropathy evaluation (GONE) project previously reported that ophthalmology trainees underestimated glaucoma likelihood in 22.1% of optic disks and overestimated 13.0% of included optic disks. ...
... The Glaucomatous optic neuropathy evaluation (GONE) project previously reported that ophthalmology trainees underestimated glaucoma likelihood in 22.1% of optic disks and overestimated 13.0% of included optic disks. This has been similar in our study where general ophthalmologists underestimated 23.8% and underestimated 8.9% of included optic disks (37). Furthermore, Breusegem et al. (38) reported that non-expert ophthalmologists had significantly lower accuracy compared with experts in the diagnosis of glaucoma. ...
Purpose:
The aim of this study was to prospectively quantify the level of agreement among the deep learning system, non-physician graders, and general ophthalmologists with different levels of clinical experience in detecting referable diabetic retinopathy, age-related macular degeneration, and glaucomatous optic neuropathy.
Methods:
Deep learning systems for diabetic retinopathy, age-related macular degeneration, and glaucomatous optic neuropathy classification, with accuracy proven through internal and external validation, were established using 210,473 fundus photographs. Five trained non-physician graders and 47 general ophthalmologists from China were chosen randomly and included in the analysis. A test set of 300 fundus photographs were randomly identified from an independent dataset of 42,388 gradable images. The grading outcomes of five retinal and five glaucoma specialists were used as the reference standard that was considered achieved when ≥50% of gradings were consistent among the included specialists. The area under receiver operator characteristic curve of different groups in relation to the reference standard was used to compare agreement for referable diabetic retinopathy, age-related macular degeneration, and glaucomatous optic neuropathy.
Results:
The test set included 45 images (15.0%) with referable diabetic retinopathy, 46 (15.3%) with age-related macular degeneration, 46 (15.3%) with glaucomatous optic neuropathy, and 163 (55.4%) without these diseases. The area under receiver operator characteristic curve for non-physician graders, ophthalmologists with 3-5 years of clinical practice, ophthalmologists with 5-10 years of clinical practice, ophthalmologists with >10 years of clinical practice, and the deep learning system for referable diabetic retinopathy were 0.984, 0.964, 0.965, 0.954, and 0.990 (p = 0.415), respectively. The results for referable age-related macular degeneration were 0.912, 0.933, 0.946, 0.958, and 0.945, respectively, (p = 0.145), and 0.675, 0.862, 0.894, 0.976, and 0.994 for referable glaucomatous optic neuropathy, respectively (p < 0.001).
Conclusion:
The findings of this study suggest that the accuracy of this deep learning system is comparable to that of trained non-physician graders and general ophthalmologists for referable diabetic retinopathy and age-related macular degeneration, but the deep learning system performance is better than that of trained non-physician graders for the detection of referable glaucomatous optic neuropathy.
... 11 O'Neill et al. showed that ophthalmology residents under training and even comprehensive ophthalmologists could underestimate the presence of glaucoma in 1 of 5 patients. 12 Rim loss, disc haemorrhage, ovoid disc and error in the evaluation of retinal nerve fibre layer thickness, as well as vertical cup disc ratio, according to their research, were some factors that are likely to be underestimated, causing underdiagnosis of glaucoma. Other studies, have shown similar variability in the assessment of optic discs, even amongst glaucoma subspecialists. ...
Objective: To determine the frequency of high cup-to-disc ratio amongst the healthy young population of the Central Punjab province of Pakistan. Study Design: Cross-sectional study. Place and Duration of Study: Department of Eye, Combined Military Hospital Lahore, from Jun 2022 to Feb 2023. Methodology: One hundred fifteen candidates who presented for medical fitness examinations were enrolled; they all belonged to the Central Punjab province of Pakistan. All the participants underwent thorough ocular examination, including visual acuity testing, slit lamp bio-microscopy, fundoscopy and tonometry. Candidates with a cup-to-disc ratio >0.6 were further investigated to rule out glaucoma using pachymetry, visual field (VF) testing and spectral-domain optical coherence tomography.Results: The mean age of participants of the study was 18.97±2.60 years. There were 72(62.6%) males and 43(37.4%) females.The mean intraocular pressure was 13.08±2.40 mmHg. The mean cup-to-disc ratio was calculated to be 0.34±0.11. It was found that 7 out of 115(6.1%) participants have a cup-to-disc ratio of more than 0.6, whereas the remaining 108(93.9%) participants had less than 0.6 cup-to-disc ratio. Conclusion: The screened population revealed a high cup-to-disc ratio in 6.1% of the candidates.
... Despite the existence of multiple tests for detecting glaucoma, diagnosis is still largely subjective with great variability between clinicians. Based on findings of the Glaucoma Optic Neuropathy Evaluation Project [12], underestimating the vertical cup-to-disc ratio and cup shape and missing both retinal nerve fiber layer (RNFL) defects including presence of disc hemorrhage, were key errors that led to glaucoma underestimation. Similar challenges can also lead to glaucoma overestimation. ...
The purpose of this study was to evaluate the capabilities of large language models such as Chat Generative Pretrained Transformer (ChatGPT) to diagnose glaucoma based on specific clinical case descriptions with comparison to the performance of senior ophthalmology resident trainees.
We selected 11 cases with primary and secondary glaucoma from a publicly accessible online database of case reports. A total of four cases had primary glaucoma including open-angle, juvenile, normal-tension, and angle-closure glaucoma, while seven cases had secondary glaucoma including pseudo-exfoliation, pigment dispersion glaucoma, glaucomatocyclitic crisis, aphakic, neovascular, aqueous misdirection, and inflammatory glaucoma. We input the text of each case detail into ChatGPT and asked for provisional and differential diagnoses. We then presented the details of 11 cases to three senior ophthalmology residents and recorded their provisional and differential diagnoses. We finally evaluated the responses based on the correct diagnoses and evaluated agreements.
The provisional diagnosis based on ChatGPT was correct in eight out of 11 (72.7%) cases and three ophthalmology residents were correct in six (54.5%), eight (72.7%), and eight (72.7%) cases, respectively. The agreement between ChatGPT and the first, second, and third ophthalmology residents were 9, 7, and 7, respectively.
The accuracy of ChatGPT in diagnosing patients with primary and secondary glaucoma, using specific case examples, was similar or better than senior ophthalmology residents. With further development, ChatGPT may have the potential to be used in clinical care settings, such as primary care offices, for triaging and in eye care clinical practices to provide objective and quick diagnoses of patients with glaucoma.
... CDR can be measured horizontally or vertically, giving the horizontal cup-to-disc ratio (HCDR) and vertical cup-to-disc ratio (VCDR), respectively. VCDR is usually used as an indication for glaucoma screening [1,[4][5][6]. An illustration of the ONH and VCDR is shown in Fig 1. VCDR equal to or larger than 0.8, or a difference in VCDR between eyes by more than 0.2, is pathological [7]. ...
Purpose
Accurate evaluation of the cup-to-disc ratio is crucial for optometrists and may be influenced by their clinical experience. The study’s objective was to compare the thresholds for discriminating vertical cup-to-disc ratio (VCDR) between years 2, 3, and 4 clinical optometry trainees.
Methods
One hundred fundus photos with various VCDR sizes were selected from a clinic database. The median VCDR (0.43) photo was assigned as the standard, while the other 99 were assigned as the test photos. The participant’s task was to discriminate using a 2-alternate-forced-choice paradigm whether the test photos’ VCDR were larger or smaller than the standard VCDR. Data were fit with a Weibull function, and three discrimination thresholds were determined: the point of subjective equality (PSE), the range of VCDR uncertainty, and the ability to judge VCDR that was 0.1 unit larger than the standard photo.
Results
Year 4 trainees had better VCDR discrimination thresholds. However, the difference between the three participant groups was not statistically different for all measurements (PSE: F(2,27) = 0.43, p = 0.657; VCDR uncertainty range: F(2,27) = 0.12, p = 0.887), and thresholds for correctly discriminating VCDR 0.1 larger than the standard photo’s VCDR: F(2,27) = 0.69, p = 0.512).
Conclusion
Although Year 4 optometry trainees performed slightly better at estimating VCDR than their Year 3 and Year 2 peers, the number of years of clinical experiences did not significantly affect their VCDR discrimination thresholds when 2-dimensional fundus photos were used as stimuli.
... It has been noted that trainees and comprehensive ophthalmologists tend to underestimate glaucoma likelihood in 22.1% and 23.8% discs in an internet-based assessment. It was further noted that underestimating the vertical CDR was one of the key errors [37]. However, it has also been shown that with practice and training even non-physician graders can perform reliably well and with greater agreement [38,39]. ...
Optic nerve head (ONH) evaluation is the cornerstone of the basic eye exam. A spectrum of conditions can affect the ONH ranging from optic disc swelling in life threatening intracranial space occupying lesions to minor morphological and congenital variations. However before one can identify the abnormal ONH, it is important to characterise what constitutes a normal ONH appearance. This chapter delves into the range of normal ONH findings and the key features that can help classify eyes which need further evaluation using other diagnostic modalities like visual field testing and ONH imaging. Prudent evaluation of the ONH on the slit lamp, using a systematic approach can thus conserve valuable resources and reduce unnecessary patient testing by other imaging and functional tests.
... However, subjective evaluation of fundus photographs to identify signs of glaucomatous optic neuropathy (GON) is challenging with significant disagreement between glaucoma specialists. Results from the Glaucomatous Optic Neuropathy Evaluation Project [2] suggests that ophthalmology trainees and comprehensive ophthalmologists underestimated glaucoma likelihood in one out of five disc photographs and were twice as likely to underestimate or overestimate glaucoma likelihood due to factors including variability in cup-to-disc ratio, peripapillary RNFL atrophy, and presence of disc hemorrhage. Therefore, methods for quantifying and classifying structural data captured by emerging technologies such as OCT can aid clinicians in glaucoma assessment. ...
... All of the interventions and more than half of the referrals for additional examinations took place in participants who, if ever, last visited their ophthalmologist more than one year ago. is highlights the importance of the interval of ophthalmologist visits, but at the same time confirms that the detection threshold in the setting of study screening is different from that in clinical practice. is has been investigated by O'Neill et al. who showed that ophthalmology trainees as well as comprehensive ophthalmologists underestimated glaucoma likelihood in approximately one out of five-disc photographs and that they were twice as likely to underestimate as overestimate glaucoma likelihood, compared to glaucoma specialists [50]. e health system would not be able to cope with the burden of overreferral of "glaucoma suspects" with current management strategies. ...
Background. Age-related conditions such as glaucoma, age-related macular degeneration, diabetic retinopathy, and cataract have become the major cause of visual impairment and blindness in high-income countries. The aim of the current study is to investigate the prevalence of these eye diseases in a cohort of self-proclaimed healthy elderly and thus get a rough estimation of the prevalence of undiagnosed age-related eye conditions in the Belgian population. Methods. Individuals aged 55 and older without ophthalmological complaints were asked to fill in a general medical questionnaire and underwent an ophthalmological examination, which included a biomicroscopic examination, intraocular pressure measurement, axial length measurement, and acquisition of fundus pictures and optical coherence tomography scans. Information regarding follow-up was collected in those who received the advice of referral to an ophthalmologist or the advice to have more frequent follow-up visits, based on their study evaluation. Results. The cohort included 102 people and comprised 46% men (median age 70 years, range 57–85 years). Referral for additional examinations was made in 26 participants (25%). The advice to have more regular follow-up ophthalmologist visits was given to nine additional participants (9%). No significant correlations between baseline characteristics and the need for referral could be identified. Follow-up information was available for 25 out of 26 referred volunteers. Out of these, four underwent a therapeutic intervention based on study referral, up until 18 months after study participation. All four interventions took place in the age group 65–74 years. Conclusions. This study shows that, even in an elderly population with self-proclaimed healthy eyes and good general health, a significant proportion of subjects showed ocular findings that need regular follow-up and/or intervention. The frequency of prior ophthalmological examinations does not seem to be relevant to this proportion, meaning that everyone above 55 years old needs a routine ophthalmological evaluation.
1. Introduction
Age-related ocular diseases are the major cause of visual impairment and blindness in high-income countries and carry a major socioeconomic burden. Western Europe roughly counts one million (0.6%) blind and three to ten million (1.7–5.6%) visually impaired persons older than 40 years [1]. Age-related conditions such as cataract, age-related macular degeneration (AMD), glaucoma, and diabetic retinopathy (DRP) have become the mainstay of visual decline in the Western world and account for more than half of cases of blindness in those aged 50 or older and for 35% of cases of visual impairment in the same age group. As a single cause of visual impairment, uncorrected refractive error continues to take the lead in all age groups, worldwide as well as in high-income countries [2].
Projections made by Finger and Scholl estimate that visual impairment will affect 5–25% of an elderly population in a high-income region over a 5- to 15-year period, with age being the most significant risk factor [1]. Visual impairment in the elderly negatively affects quality of life and increases the need for care because of increased fall risk, loss of independence, depression, and increased all-cause mortality [3–8]. A large longitudinal observational study in American adults concluded that regular eye examinations for those aged 65 or older are a protective factor for the development of decline in both vision and functional status [9]. This link between receipt of care and visual and functional outcome reinforces the current professional guidelines by the American Academy of Ophthalmology (AAO), which advocate a complete eye exam with an ophthalmologist every year or two after the age of 65 [10]. When it comes to screening for impaired visual acuity in elderly, however, current evidence appears to be insufficient to assess the balance of benefits and harms, as concluded by the US Preventive Services Task Force. The reason for these findings is the lack of well-designed studies demonstrating conclusive benefits of universal eye screening in the elderly [11].
Furthermore, glaucoma, and to a lesser extent AMD and DRP, are characterized by irreversible damage and vision loss, emphasizing the need for early diagnosis and treatment to delay the development of significant visual impairment. More than half of glaucoma cases remain undiagnosed, even in developed regions [12–15], despite widely available eye care facilities [16–18]. Glaucoma screening remains controversial because of the lack of data, economic evaluations, and accurate screening test algorithms [19–21], and diagnosis is mostly made by routine opportunistic case finding as there is no evidence for a useful screening tool to date. Nevertheless, the debate is ongoing and more evidence could argue for targeted screening or even for mass screening for glaucoma if more effective diagnostic tools become available. On the other end of the spectrum, glaucoma overdiagnosis and overtreatment is a relevant health issue, as pointed out by various authors and reviewed by González-Martín-Moro and Zarallo-Gallardo [12, 17, 22–25].
This study investigates the prevalence of age-related eye diseases in a cohort of self-proclaimed healthy elderly to get a rough estimation of the prevalence of undiagnosed age-related eye conditions in the Belgian population. The results underline the potential benefit of screening for a subset of prevalent sight-threatening age-related eye diseases in an elderly population, preferentially before the onset of impactful visual decline.
2. Materials and Methods
2.1. Study Design
This single-center cross-sectional study took place at University Hospitals UZ Leuven, Department of Ophthalmology, Leuven, Belgium, during April 2017. Elderly individuals free of known ophthalmological diseases were recruited from the members of the Third Age University Leuven, a KU Leuven initiative that offers a continued education program to the over-55-year-olds.
Those who received the advice of referral to an ophthalmologist or the advice to have more frequent follow-up visits, based on their participation in this study, received a questionnaire regarding their follow-up status in November 2018 (18 months after study participation).
Individuals with a known ophthalmological condition, besides refractive error or pseudophakia, and those with subjectively suboptimal visual acuity were excluded.
2.2. Study Population and Research Methods
The cohort included 102 people and comprised 46% men (median age 70 years, range 57–85 years). Inclusion criteria were members of the Third Age University Leuven (thus aged ≥55 years), with self-proclaimed healthy eyes and good general health. Written informed consent was obtained from each volunteer prior to inclusion in the study in compliance with relevant regulation on clinical trials. Individual results were discussed with each participant. In the case of an abnormal result, the participant was referred for further diagnostic work-up or regular follow-up.
2.3. Ophthalmological Examination
Subjects were asked to fill in a questionnaire on their personal and familial general and ocular history. A basic ophthalmological examination of both eyes of each participant was performed, including biomicroscopy, intraocular pressure (IOP) measurement by rebound tonometry using an iCare® TA01i tonometer (Tiolat Oy, Helsinki, Finland), axial length (AXL) measurement by IOL Master 700 (Carl Zeiss Meditec AG, Jena, Germany), dilated fundoscopy, and stereoscopic optic disc photography as well as macula-centered fundus photography using the Visucam PRO NM (Carl Zeiss Meditec AG, Jena, Germany) and optical coherence tomography (OCT) using the glaucoma module of the OCT Spectralis (Heidelberg Engineering, Heidelberg, Germany). The subset of volunteers that received advice to have further exams or more frequent follow-up received a questionnaire concerning their follow-up status where they were asked to answer questions about ophthalmologist visits since participating in this study and ocular interventions. All data were anonymized prior to analysis.
2.4. Statistical Analysis
Statistical analyses were performed using IBM SPSS® 25.0 for Windows (IBM, Armonk, New York, USA). Continuous variables were tested for normality using the Shapiro–Wilk test. Continuous variables are presented as the median, minimum, and maximum because they were not normally distributed . Binary variables are presented as numbers and percentages. Nominal variables are presented as numbers with percentages per category. To statistically compare variables between groups, the Mann–Whitney U test was used for continuous variables and the chi-square test was used for dichotomous and nominal variables. Pairwise correlation was additionally assessed using Spearman’s rank correlation coefficients. The influence of age, gender, education, smoking status, diabetes, arterial hypertension, neurological pathology, autoimmune pathology, intraocular lens status, familial history of glaucoma, familial history of AMD, intake of vitamin preparations for the prevention of AMD, the presence of a corrected refractive error, and previous ophthalmologist visits on the need for referral to an ophthalmologist was studied. Statistical significance was accepted based on two-sided values of <0.05.
2.5. Compliance
The study was conducted in compliance with the principles of the European Union Directive on Clinical Trials (2001/20/EC) and all local/regional requirements required to conform with the provisions of the Declaration of Helsinki (World Medical Association, Edinburgh, 2000). Approval was issued by the Ethics Committee of the University Hospitals Leuven before the study commenced.
3. Results
3.1. Patients’ Characteristics
The cohort included 102 people, who were all included for further analysis. Thus, 102 subjects and 203 eyes (one visitor was monophthalmic due to trauma) remained in the study group. Detailed baseline characteristics are listed in Table 1. The age ranged from 57 to 85 years (median 69.50 years) with 83% of the study population being aged ≥65 years. Overall, the subjects included were highly educated, with significantly more men reporting more than three years of higher education . All subjects that had been diagnosed with systemic pathology stated that this pathology was well controlled and under follow-up. In accordance with the inclusion criteria, the ocular status was deemed healthy by all 102 volunteers, with the absence of subjective visual impairment.
Characteristics, n = 102
Age (years), range (median)
57–85 (69.50)
Age (years in range), n (%)
<65
17 (17)
65–74
64 (63)
75–84
20 (19)
≥85
1 (1)
Male sex, n (%)
47 (46)
Education, n (%)
High school
12 (12)
≤3 years of higher education
66 (65)
>3 years of higher education
24 (23)
Smoking status, n (%), pack years, range (median)
Current smoker
3 (3)–4.4–50.0 (28.0)
Former smoker
47 (46)–0.1–39.0 (8.5)
Never smoked
52 (51)
Diabetes, n (%)
8 (8)
Arterial hypertension, n (%)
39 (38)
Neurological disorder, n (%)
16 (16)
Autoimmune disorder, n (%)
14 (14)
Family history of glaucoma, n (%)
17 (17)
Family history of AMD, n (%)
6 (6)
Intake of vitamin preparations for the prevention of AMD, n (%)
3 (3)
Pseudophakia/aphakia in at least one eye, n (%)
9 (9)
Correcting spectacles, n (%)
Distant sight
67 (66)
Reading
90 (88)
Time since last eye doctor visit, n (%)
<1 year
34 (33)
1–4 years
48 (47)
>4 years
13 (13)
Never
7 (7)
AMD = age-related macular degeneration.
... A number of imaging modalities have been used for the diagnosis of eye diseases; however, the interpretation of these images is highly dependent on the skill and experience of physicians, and this process is often subjective, with substantial interobserver variation. [4][5][6] Evidence supporting this includes a study wherein even senior glaucoma specialists could only achieve a "substantial" level of agreement (k ¼ 0.63) for the classification of glaucoma based on optic disc photography. This agreement could be even poorer among general ophthalmologists (k ¼ 0.51) and trainees (k ¼ 0.50). ...
Artificial intelligence has rapidly evolved from the experimental phase to the implementation phase in many image-driven clinical disciplines, including ophthalmology. A combination of the increasing availability of large datasets and computing power with revolutionary progress in deep learning has created unprecedented opportunities for major breakthrough improvements in the performance and accuracy of automated diagnoses that primarily focus on image recognition and feature detection. Such an automated disease classification would significantly improve the accessibility, efficiency, and cost-effectiveness of eye care systems where it is less dependent on human input, potentially enabling diagnosis to be cheaper, quicker, and more consistent. Although this technology will have a profound impact on clinical flow and practice patterns sooner or later, translating such a technology into clinical practice is challenging and requires similar levels of accountability and effectiveness as any new medication or medical device due to the potential problems of bias, and ethical, medical, and legal issues that might arise. The objective of this review is to summarize the opportunities and challenges of this transition and to facilitate the integration of artificial intelligence (AI) into routine clinical practice based on our best understanding and experience in this area.
... Disc photography reviewed by experienced specialists can achieve similar diagnostic performance as that of advanced imaging technology such as optical coherence tomography (OCT) [10], and combining other modalities with fundus images did not necessarily increase the diagnostic power in teleglaucoma [7]. However, evaluation of the optic disc is subjective, and likely to encounter interobserver discordance even among glaucoma specialists [8,11]. Qualified interpretation of fundus images at limited cost through automated glaucoma detection using artificial intelligence may make fundus photography-based glaucoma screening feasible. ...
Purpose
To evaluate ways to improve the generalizability of a deep learning algorithm for identifying glaucomatous optic neuropathy (GON) using a limited number of fundus photographs, as well as the key features being used for classification.
Methods
A total of 944 fundus images from Taipei Veterans General Hospital (TVGH) were retrospectively collected. Clinical and demographic characteristics, including structural and functional measurements of the images with GON, were recorded. Transfer learning based on VGGNet was used to construct a convolutional neural network (CNN) to identify GON. To avoid missing cases with advanced GON, an ensemble model was adopted in which a support vector machine classifier would make final classification based on cup-to-disc ratio if the CNN classifier had low-confidence score. The CNN classifier was first established using TVGH dataset, and then fine-tuned by combining the training images of TVGH and Drishti-GS datasets. Class activation map (CAM) was used to identify key features used for CNN classification. Performance of each classifier was determined through area under receiver operating characteristic curve (AUC) and compared with the ensemble model by diagnostic accuracy.
Results
In 187 TVGH test images, the accuracy, sensitivity, and specificity of the CNN classifier were 95.0%, 95.7%, and 94.2%, respectively, and the AUC was 0.992 compared to the 92.8% accuracy rate of the ensemble model. For the Drishti-GS test images, the accuracy of the CNN, the fine-tuned CNN and ensemble model was 33.3%, 80.3%, and 80.3%, respectively. The CNN classifier did not misclassify images with moderate to severe diseases. Class-discriminative regions revealed by CAM co-localized with known characteristics of GON.
Conclusions
The ensemble model or a fine-tuned CNN classifier may be potential designs to build a generalizable deep learning model for glaucoma detection when large image databases are not available.
... 7 Eyes with large physiologic cups, for example, are frequently diagnosed as having glaucoma, whereas eyes with small optic discs but showing significant rim loss may go undetected. 7,33 A fundamental aspect of the development of a deep learning network is VOL. -5 HUMAN VS MACHINE FOR GLAUCOMA DIAGNOSIS establishing a reliable ''ground-truth'' or reference label that can be used to train the algorithm. ...
Purpose:
To compare the diagnostic performance of human gradings versus predictions provided by a machine-to-machine (M2M) deep learning (DL) algorithm trained to quantify retinal nerve fiber layer (RNFL) damage on fundus photographs.
Design:
Evaluation of a machine learning algorithm.
Methods:
A M2M DL algorithm trained with RNFL thickness parameters from spectral-domain optical coherence tomography was applied to a subset of 490 fundus photos of 490 eyes of 370 subjects graded by two glaucoma specialists for the probability of glaucomatous optical neuropathy (GON), and estimates of cup-to-disc (C/D) ratios. Spearman correlations with standard automated perimetry (SAP) global indices were compared between the human gradings versus the M2M DL-predicted RNFL thickness values. The area under the receiver operating characteristic curves (AUC) and partial AUC for the region of clinically meaningful specificity (85-100%) were used to compare the ability of each output to discriminate eyes with repeatable glaucomatous SAP defects versus eyes with normal fields.
Results:
The M2M DL-predicted RNFL thickness had a significantly stronger absolute correlation with SAP mean deviation (rho=0.54) than the probability of GON given by human graders (rho=0.48; P<0.001). The partial AUC for the M2M DL algorithm was significantly higher than that for the probability of GON by human graders (partial AUC = 0.529 vs. 0.411, respectively; P=0.016).
Conclusion:
A M2M DL algorithm performed as well as, if not better than, human graders at detecting eyes with repeatable glaucomatous visual field loss. This DL algorithm could potentially replace human graders in population screening efforts for glaucoma.
... However, VCDR assessment is inherently subjective and inter-grader variability is high, even among ophthalmologists. [9][10][11][12][13][14] We test in the present study whether our software made assessment of the VCDR more reliable between graders. If so, this would support the idea of non-expert graders performing VCDR assessment at the point of care, without the need for a reading center. ...
Purpose:
Glaucoma screening can be performed by assessing the vertical-cup-to-disk ratio (VCDR) of the optic nerve head from fundus photography, but VCDR grading is inherently subjective. This study investigated whether computer software could improve the accuracy and repeatability of VCDR assessment.
Methods:
In this cross-sectional diagnostic accuracy study, 5 ophthalmologists independently assessed the VCDR from a set of 200 optic disk images, with the median grade used as the reference standard for subsequent analyses. Eight non-ophthalmologists graded each image by two different methods: by visual inspection and with assistance from a custom-made publicly available software program. Agreement with the reference standard grade was assessed for each method by calculating the intraclass correlation coefficient (ICC), and the sensitivity and specificity determined relative to a median ophthalmologist grade of ≥0.7.
Results:
VCDR grades ranged from 0.1 to 0.9 for visual assessment and from 0.1 to 1.0 for software-assisted grading, with a median grade of 0.4 for each. Agreement between each of the 8 graders and the reference standard was higher for visual inspection (median ICC 0.65, interquartile range 0.57 to 0.82) than for software-assisted grading (median ICC 0.59, IQR 0.44 to 0.71); P = 0.02, Wilcoxon signed-rank test). Visual inspection and software assistance had similar sensitivity and specificity for detecting glaucomatous cupping.
Conclusion:
The computer software used in this study did not improve the reproducibility or validity of VCDR grading from fundus photographs compared with simple visual inspection. More clinical experience was correlated with higher agreement with the ophthalmologist VCDR reference standard.
... Incorrect assessment of glaucoma likelihood can arise from a failure to identify RNFL loss, optic disc hemorrhages, neuroretinal rim loss, and incorrect assessment of the VCDR. 34 For both orthoptic groups in the current study, the characteristics with the poorest level of agreement were the assessment of cup-to-disc ratio, cup shape (neuroretinal rim loss), and RNFL. It appears likely that poor assessment of these characteristics may have led to the fair to moderate agreement for glaucoma likelihood for orthoptists. ...
Introduction:
The aim of this study was to determine the accuracy of orthoptists when examining the optic disc for signs of glaucoma, and to explore the impact of targeted clinical education on accuracy.
Methods:
In this randomized controlled trial, 42 monoscopic color optic disc images were presented to 46 orthoptists who assessed the likelihood of glaucoma as well as optic disc size, shape, tilting, vertical cup-to-disc ratio, cup shape, depth, presence of hemorrhage, peripapillary atrophy, and retinal nerve fiber layer. The level of agreement with specialist ophthalmologists was assessed. Participants were then randomly assigned to an experimental group (targeted postgraduate education on optic disc assessment) or to no intervention. The educational program was designed to increase knowledge of the characteristic features associated with glaucomatous optic neuropathy. All participants re-examined the included optic disc images after a period of 6 to 8 weeks. The primary outcome measure was a change in agreement between attempts.
Results:
The education group showed significant improvements between attempts for identifying hemorrhages (P = .013), retinal nerve fiber layer defects (0.035), disc size (P = .001), peripapillary atrophy (P = .030), and glaucoma likelihood (P = .023). The control group did not show any statistically significant improvement. The intervention group showed significantly more improvement when identifying hemorrhages (P = .013), disc size (P = .001), disc shape (P = .033), and cup shape (P = .020) compared with the control group.
Discussion:
Orthoptists who received additional postgraduate online education based on principles of adult learning were more accurate at assessing the optic disc for glaucoma. These results highlight the value of continuing education to optimize clinical practice in allied health professionals.
... The results have been published, and they concluded that ophthalmology trainees and comprehensive ophthalmologists underestimated glaucoma likelihood in approximately one in five disc photographs and were twice as likely to underestimate as overestimate glaucoma likelihood. Underestimating the vertical cup-disc ratio and cup shape and missing retinal nerve fibre layer defects and disc haemorrhage were the key errors that led to underestimation [11]. A partner application has also been developed for the iPad, which allows the GONE project to be used in an applet form. ...
Glaucoma being a disease that requires lifelong medical management and strict compliance is one of the best candidates for app development. Mobile devices have been used for intraocular pressure (IOP) assessment and recording to screening of disc photographs [1]. With the sudden boom in mobile augmented imaging of the optic disc for screening of patients to rapid assessment of visual field, applications of mobile computing to diagnosis and management of glaucoma are bound to increase with time [2, 3]. We shall review the best available current applications on Android and iOS platforms that can be easily adapted into the daily workflow of your glaucoma practice.
... A recently published article studied nine topographic features of the disk and RNFL in monoscopic optic disk photographs and found that moderate or large peripapillary atrophy, ovoid horizontal disk shape, and error in the assessment of the vertical C:D ratio, RNFL, cup shape (rim loss), or disk hemorrhage lead to underestimation. 9 Our results agreed with the findings of this published study in terms of absolute percentages, but had lower κ coefficients. This was probably due to sample heterogeneity (κ paradox) 10 and the specific glaucomatous ONH changes investigated in our study, which allowed for a better analysis of interobserver variability. ...
Objectives
To assess agreement among ophthalmology residents and a glaucoma expert in the evaluation of cross-sectional glaucomatous optic nerve head characteristics using stereoscopic photographs.
Methods
Twenty stereo photographs were analyzed by ophthalmology residents just after completion of their first (First-Year Group) or third (Third-Year Group) year of residency and by a glaucoma expert. The agreement was assessed using the kappa statistic (κ) and limits of agreement.
Results
Agreement among resident groups and the expert ranged from poor to moderate. Agreement between Third Years and the expert seems to be better than that between First Years and the expert, especially in the evaluation of “nasal cupping”, “barring circumlinear vessel,” “notching”, and “retinal nerve fiber layer defect” criteria. However, no improvement was seen in the agreement with the expert regarding glaucomatous optic neuropathy, which was 64% (κ=0.19) for First Years and 63% (κ=0.20) for Third Years.
Conclusion
Agreement between residents and the expert was poor to moderate and similar when comparing both groups. This may suggest that the residents learn how to identify glaucoma signals during the first year of training, and the results of this study may facilitate the creation of targeted teaching tools in residency training.
... They then explored the factors associated with underdiagnosis of glaucoma in trainees using monoscopic disc photograph. The most common factor was inability to identify RNFL defect, followed by unseen disc hemorrhage, neural rim loss, and vertical cup-to-disc ratio [17]. These results were more pronounced in ophthalmology trainees than in fully trained ophthalmologists. ...
Background
Assessment of color disc photograph (C-DP) is affected by image quality, which decreases the ability to detect glaucoma. High-dynamic-range (HDR) imaging provides a greater range of luminosity. Therefore, the objective of this study was to evaluate the capability of ophthalmology residents to detect glaucoma using HDR-concept disc photography (HDR-DP) compared to C-DP.
Design
Cross-sectional study.
Methods
Twenty subjects were classified by 3 glaucoma specialists as either glaucoma, glaucoma suspect, or control. All C-DPs were converted to HDR-DPs and randomly presented and assessed by 10 first-year ophthalmology residents. Sensitivity and specificity of glaucoma detection were compared.
Results
The mean ± SD of averaged retinal nerve fiber layer (RNFL) thickness was 74.0 ± 6.1 μm, 100.2 ± 9.6 μm, and 105.8 ± 17.2 μm for glaucoma, glaucoma suspect, and controls, respectively. The diagnostic sensitivity of HDR-DP was higher than that of C-DP (87% versus 68%, mean difference: 19.0, 95% CI: 4.91 to 33.1; p = 0.014). Regarding diagnostic specificity, HDR-DP and C-DP yielded 46% and 75% (mean difference: 29.0, 95% CI: 13.4 to 44.6; p = 0.002).
Conclusions
HDR-DP statistically increased diagnostic sensitivity but not specificity. HDR-DP may be a screening tool for nonexpert ophthalmologists.
... These studies, therefore, suggest that failure of surveillance or referral may be more important. Supporting the alternative hypothesis, in the Glaucoma Optic Neuropathy Evaluation Project, 23 ophthalmology trainees and comprehensive ophthalmologists underestimated glaucoma likelihood in approximately one in five disc photographs and were twice as likely to underestimate as overestimate glaucoma likelihood. Underestimating the vertical cup-disc ratio and cup shape and missing retinal nerve fiber layer (RNFL) defects and disc hemorrhage were the key errors that led to underestimation. ...
Drug Induced Glaucoma (Chapter 11) is a unique type of Secondary Glaucoma that can be brought on by a variety of regularly prescribed medications. This chapter gives clinical insight to approach such cases and its management.
Purpose:
To evaluate the longitudinal change of cup to disc ratio (CDR) in premature infants with enlarged CDR and normal intraocular pressure (IOP).
Methods:
This retrospective, observational study included 283 eyes of 283 premature infants at single center. Infants were divided into enlarged CDR and control groups. Data on demographics, gestational age (GA), birth weight (BW), vertical cup to disc ratio (vCDR), IOP, and corneal diameter were analyzed.
Results:
Of the 283 patients, 38 (13.4%) and 245 (86.6%) were in the enlarged CDR and control groups, respectively. In the enlarged CDR group, the vCDR and baseline IOP was 0.63 ± 0.12 and 12.7 ± 2.2 mmHg, respectively. In the control group, the vCDR and baseline IOP was 0.18 ± 0.05 and 10.9 ± 1.7 mmHg, respectively. The IOP at a GA of 40 weeks and at 1 and 2 years of age were significantly lower than that at baseline. The differences in vCDR between baseline and each follow-up visit were not significant. vCDR was negatively correlated with GA and BW; however, these correlations were not significant.
Conclusions:
The vCDR did not significantly change up to age of 2 years of age in premature infants with enlarged CDR and normal IOP. However, close follow-up is needed until other reliable glaucoma examinations, such as optical coherence tomography and visual fields, can be possible.
Purpose
To identify patterns of visual field (VF) loss based on unsupervised machine learning and to identify patterns that are associated with rapid progression.
Design
Cross-sectional and longitudinal study.
Participants
A total of 2231 abnormal VFs from 205 eyes of 176 OHTS participants followed over approximately 16 years.
Methods
VFs were assessed by an unsupervised deep archetypal analysis algorithm as well as an OHTS certified VF reader to identify prevalent patterns of VF loss. Machine-identified patterns of glaucoma damage were compared against those patterns previously identified (expert-identified) in the OHTS in 2003. Based on the longitudinal VFs of each eye, VF loss patterns that were strongly associated with rapid glaucoma progression were identified.
Main Outcome Measures
Machine-expert correspondence and type of patterns of VF loss associated with rapid progression.
Results
The average VF mean deviation (MD) at conversion to glaucoma was -2.7 dB (Standard Deviation (SD) = 2.4 dB) while the average MD of the eyes at the last visit was -5.2 dB (SD = 5.5 dB). Fifty out of 205 eyes had MD rate of -1 dB/year or worse and were considered rapid progressors. Eighteen machine-identified patterns of VF loss were compared with expert-identified patterns in which 13 patterns of VF loss were similar. The most prevalent expert-identified patterns included partial arcuate, paracentral, and nasal step defects, and the most prevalent machine-identified patterns included temporal wedge, partial arcuate, nasal step, and paracentral VF defects. One of the machine-identified patterns of VF loss predicted future rapid VF progression after adjustment for age, sex, and initial MD.
Conclusions
An automated machine learning system can identify patterns of VF loss and could provide objective, and reproducible nomenclature for characterizing early signs of visual defects and rapid progression in patients with glaucoma.
Ophthalmology is the branch of medicine that encompasses diseases and treatments of the eye. Its technical, clinical and public health features have enabled it to be an ideal field for the development and deployment of artificial intelligence (AI). Indeed, it is within ophthalmology that many of healthcare’s most promising AI applications have emerged – from early-stage tools in development, to regulatory-approved and commercialised platforms in real-world clinical use. The field’s technical, clinical and public health contexts are described within this chapter and are further illustrated through case studies in diabetic retinopathy (DR), glaucoma and retinopathy of prematurity (ROP).
Purpose
To investigate the prevalence and factors associated with optic disc tilt in the eyes of African Americans with glaucoma.
Design
Cross-sectional.
Participants
Subjects with glaucoma participating in the Primary Open-Angle African American Glaucoma Genetics (POAAGG) study.
Methods
Stereo pairs of optic disc images were assessed independently by POAAGG-certified non-physician graders for quantitative features including maximum and minimum linear disc diameters, and qualitative features including gradeability of images, shape of the cup, rim plane position, beta-peripapillary atrophy, sloping region adjacent to the outer disc margin, and rim pallor. Discrepancies were adjudicated by an ophthalmologist. Descriptive statistics and p-values were generated for associations of tilt with demographic and ocular characteristics. Stepwise multivariable analysis was performed with logistic regression using Generalized Estimating Equations (GEE) to account for inter-eye correlation within subjects.
Main Outcome Measures
Tilt defined by Ovality Index (TOI) of >1.3 and Stereoscopically Identified optic disc Tilt (SIT).
Results
Among 1251 subjects with data on both eyes, 104 (8.3%) had TOI. Subjects with TOI were less likely to be male (Adjusted Odds Ratio [aOR] 0.46, 95% CI 0.29-0.74, p<.001), Eyes with TOI were less likely to have large Cup Disc ratios (aOR 0.18, 95% CI 0.06-0.53, p<0.001) and less likely to have cylinder shaped cups compared to conical shaped cups (aOR 0.31, 95% CI 0.19-0.49, p<0.001). Among 1007 subjects with data on both eyes, 254 (25.2%) had SIT. Subjects with SIT were younger (aOR 0.95, 95% CI 0.93-0.96, p<0.001) and eyes with SIT were more likely to have oval shaped discs compared to round discs (aOR 1.82, 95% CI 1.32-2.52, p<0.001), more likely to have a sloping region adjacent to the outer disc margin instead of being flat (aOR 3.26, 95% CI 2.32-4.59, p<0.001) and less likely to have cylinder shaped cups compared to conical shaped cups (aOR 0.59, 95% CI 0.41-0.85, p<0.001). Both TOI and SIT were not associated with myopia.
Conclusions
There are substantial numbers of tilted optic discs in glaucoma patients with African American ancestry. They occur more frequently in females and younger subjects and are associated with several ocular features but not with myopia.
Optic nerve head appearance can vary significantly in glaucomatous as well as non-glaucomatous eyes. This creates difficulties in correct classification of the discs from glaucoma point of view, more so where the underlying structure is unexpected or not clearly identified. This chapter addresses some basic aspects that pose difficulty in interpretation and some aberrant discs that pose a diagnostic dilemma.
Diagnosis of glaucoma requires more than the optic nerve head (ONH) evaluation findings alone, but ONH assessment is crucial to detection of glaucoma, assessment of severity, and monitoring for disease progression. Objective assessment of the ONH must be done independent of other clinical features. This chapter explores the process of examining the disc for glaucoma, where errors may occur, and strategies to avoid them.
The use of monoscopic cameras for glaucoma screening is increasing due to their portability, lower cost, and non-mydriatic capabilities. However, it is important to compare the accuracy of such devices with stereoscopic cameras that are used clinically and are considered the gold standard in optic disc assessment. The aim of this study is to compare vertical cup-to-disc ratio (VCDR) estimates obtained using images taken with a monoscopic and stereoscopic camera.
Participants were selected from the Tema Eye Survey. Eligible subjects had images of at least one eye taken with two cameras. They were classified as meeting the glaucoma threshold if an eye had a VCDR estimate >97.5th percentile, corresponding to >0.725 for this population. Hence, we used 0.725 as the cutoff to group eyes into two categories: positive and negative. We calculated sensitivity, specificity, and predictive values of VCDR assessed by expert readers at a reading center for monoscopic photos using stereoscopic photos as the gold standard.
Three hundred and seventy-nine eyes of 206 participants were included in the study. Most participants were female (60.2%) and the most common age group was 50–59 years (36.4%). Sixteen eyes met the glaucoma threshold (VCDR > 0.725). Of these, the VCDR estimates of 14 eyes (87.5%) disagreed on the glaucoma threshold from the two cameras. The sensitivity to detect glaucoma with the monoscopic camera was 14.3% (95% CI: 4.0, 40.3).
The low sensitivity of monoscopic photos suggests that stereoscopic photos are more useful in the diagnosis of glaucoma.
Purpose of review:
The study was undertaken to assess the current importance of clinical examination of the optic nerve head (ONH) in glaucoma management.
Recent findings:
ONH changes in glaucoma are easily assessed with good specificity on examination, whereas imaging has not been able to identify pallor of the neuroretinal rim (NRR), disc hemorrhages, or vascular signs of acquired cupping. Glaucomatous neuropathy in primary open angle glaucoma (POAG) has extensive databases on imaging machines, however, other glaucomas such as primary angle closure glaucoma (PACG), juvenile open angle glaucoma (JOAG), congenital, and secondary glaucomas have been reported to have a significantly different ONH morphology. Clinical evaluation with knowledge of the underlying cause of glaucoma allows easy diagnosis of any abnormality, whereas the absence of an imaging database in such eyes makes diagnosis difficult on optical coherence tomography (OCT) and so forth. In eyes with congenital anomalies and dysplastic discs, clinical examination and history provide better identification of glaucomatous damage. Staging of glaucoma on ONH examination when perimetry is unreliable provides a means of determining the amount of damage and appropriate reduction of intraocular pressure (IOP).
Summary:
Clinical examination of the ONH is very relevant for screening, diagnosis, staging, and management of all glaucomas, and is irreplaceable in detecting nonglaucomatous causes of ONH cupping and visual field defects similar to those in glaucoma.
To analyze longitudinal changes in peripapillary retinal nerve fiber layer (pRNFL) thicknesses over time in the fellow eyes of patients with unilateral retinal vein occlusion (RVO). A total of 47 patients with unilateral RVO and 47 healthy controls were enrolled. The mean and sectoral pRNFL thicknesses were measured using spectral domain-optical coherence tomography at 1 year intervals, and followed for 3 years. Linear mixed models were performed to calculate and compare the reduction rates of pRNFL thicknesses over time. The mean pRNFL thickness decreased significantly during the 3-year follow-up, with a significant decrease over time in both groups. The reduction rate in mean pRNFL thicknesses was −0.41 μm/year in the control group and −0.68 μm/year in the fellow eyes of RVO group, and the decrease was significantly higher in the fellow eyes of RVO group than in the control group (p < 0.001). Using a multivariate linear mixed model, age (estimate: −0.41, p = 0.011) and hypertension (HTN) (estimate: −6.51, p = 0.014) were significantly associated with the reduction in mean pRNFL thicknesses in fellow eyes of RVO group. The fellow eyes of RVO patients showed a greater reduction in pRNFL thickness over time than normal controls. Age and HTN should be considered as factors to decrease the pRNFL thickness over time in fellow eyes of RVO group.
Purpose of review
To trace the influence of disc hemorrhage studies on our understanding of glaucoma.
Sources
Major articles published during the last 50 years since the rediscovery of disc hemorrhage were identified. A total of 196 articles were selected from 435 articles retrieved using the keywords glaucoma and disc hemorrhage as of August 9 2018 from PubMed.
Recent findings
The main characteristics of disc hemorrhage, including its morphology, recurrence rate, duration, increased incidence in glaucoma, and role in the progression of normal tension glaucoma was well understood by the year 2000. Since then, studies have focused on more sophisticated and accurate methods of elucidating both structural and functional progression, with special attention to the role of the lamina cribrosa. Nevertheless, both the mechanism of disc hemorrhage development and its fuller relationship with glaucoma remain unclear.
Summary
Disc hemorrhage research requires careful study of glaucomatous optic neuropathy. This has been facilitated by recent advances in optical coherence tomography (OCT) angiography and other OCT technologies. Furthermore, animal studies of disc hemorrhage promise new insights into glaucomatous optic neuropathy.
Purpose
To report the prevalence of glaucoma in the Northern Finland Birth Cohort (NFBC) Eye Study.
Methods
Subjects of the population‐based Northern Finland 1966 Birth Cohort (NFBC), aged 45–49 years at the time of the field examination, were randomized to eye screening (50%) and control (50%) groups. The eye examination protocol included best corrected visual acuity (BCVA), measurements of intraocular pressure and central corneal thickness, Humphrey 24‐2 perimetry, stereoscopic optic nerve head (ONH) and retinal nerve fibre layer (RNFL) photography and imaging with optical coherence tomography (OCT), scanning laser polarimetry (GDx) and scanning laser ophthalmoscopy (HRT). The diagnosis of glaucoma was made by two independent general ophthalmologists and three independent glaucoma experts based on the evaluation of the ONH and RNFL photographs and the visual fields.
Results
Totally, 10 321 subjects of the NFBC main study were alive in Finland in 2011, and they were randomized to the NFBC Eye Study group (n = 5155) and the control group (n = 5166). Of the randomized subjects, 3039 of 5155 (59%) responded and had sufficient data for the study. Glaucoma was suspected in 172 subjects (5.7%) at the first phase of the evaluation protocol. The interobserver agreement between two screening ophthalmologists was moderately good (kappa value 0.54 [95% confidence intervals (CI) 0.46–0.61]). Finally, definite glaucoma was found in 33 subjects (1.1% [95% CI 0.8–1.5]).
Conclusion
The study provides up‐to‐date information on the prevalence of glaucoma in a middle‐aged Caucasian population in Finland. The baseline data reported here allows the evaluation of the cost‐effectiveness of screening later on.
Purpose:
The aim of this study was to use a large group of observers to test prior research findings that suggest optic disc size, clinical evaluation of the neuroretinal rim (ISNT rule), and practitioner characteristics influence the accuracy of differentiating normal from glaucomatous optic nerves.
Methods:
Participant observers were optometrists, optometry students, and vision scientists/researchers attending the 2013 American Academy of Optometry Annual Meeting. Each observer viewed and judged six sets of stereoscopic photographs of normal and clinically confirmed glaucomatous optic nerves of different sizes presented in random order. Observers were queried on whether each nerve was glaucomatous or normal, whether the nerve followed the ISNT rule, and whether further evaluation with advanced imaging techniques was indicated.
Results:
Of the 261 observers who participated, 59% were practicing optometrists, 7% were vision scientists, and 34% were residents or students. Of practicing optometrists and vision scientists, half (49%) had more than 15 years of experience, whereas 11% had less than 2 years of experience. Diagnostic accuracy differed based on optic nerve size: average-sized nerves were correctly identified by 90% of subjects, whereas small nerves and large nerves were correctly identified by 42% and 62%, respectively. Notably, only 9% of subjects correctly identified the small glaucomatous nerve, and only 34% correctly identified the large normal nerve. No practitioner characteristics were associated with diagnostic accuracy.
Conclusions:
Accurate identification of glaucomatous optic neuropathy was significantly influenced by optic disc size. This was particularly evident for the large normal nerve and the small glaucomatous nerve. The ISNT rule provided value for differentiating normal from glaucomatous nerves, but its subjective interpretation resulted in considerable intergrader variability. These findings agree with other studies utilizing smaller numbers of observers but larger numbers of optic nerve presentations that disc size and the ISNT rule have value for enhancing accuracy of optic nerve assessment.
Purpose:
To evaluate the characteristics of patients with previously undiagnosed primary open-angle glaucoma (POAG) in Korea.
Methods:
This study examined data from 391 subjects obtained from the 2008-2009 Korean National Health and Nutrition Examination Survey (KNHANES). The KNHANES is a population-based, cross-sectional epidemiological survey. Participants aged 19 years or older completed standardized interviews and dilated ocular examinations, including measurement of intraocular pressure, visual fields with frequency doubling perimetry, and fundus photography. Data from the 361 patients with previously undiagnosed POAG were analyzed and compared with data from the 30 patients with previously diagnosed glaucoma.
Results:
A total of 92.3% of POAG cases were undiagnosed before this study. Adjusted for age and sex, the strongest risk factor for undiagnosed glaucoma was longer elapsed time since last eye doctor visit. Glaucoma patients who had not visited an eye specialist in the last 3 years were 22 times (95% confidence interval, CI, 4.49-105.64, p < 0.001) more likely to have undiagnosed disease compared with patients who had visited an eye specialist in the last month. Another significant factor for previously undiagnosed glaucoma was smaller cup-to-disc ratio (odds ratio, OR, 0.60/0.1 units, 95% CI 0.43-0.85/0.1 units, p = 0.004). The higher vertical cup-to-disc ratio of a subject's two eyes was significantly different between those with previously undiagnosed (0.69) and diagnosed (0.78) POAG (p = 0.001).
Conclusions:
The undiagnosed POAG group had a longer interval from last eye doctor visit and smaller vertical cup-to-disc ratio compared to the diagnosed group.
Optic nerve examination is the mainstay for glaucoma diagnosis and follow-up. Specific features of the optic disc and retinal nerve fiber layer (RNFL) need to be assessed in the context of the disease. Their description is qualitative and subjective and constitutes a challenge for the clinician. This review focuses glaucoma definitions and describes clinical examination protocols and their weaknesses. We also describe the imaging technologies that have been developed in an effort to achieve quantitative measurements which can be used to aid initial glaucoma diagnosis as well as follow-up, provided that their results are critically evaluated in combination with the clinical examination findings. We finally review imaging technologies’ evaluation in clinical as well as in population-based settings.
Objective:
To describe and evaluate the performance of an automated CAD system for detection of glaucoma from color fundus photographs.
Design and setting:
Color fundus photographs of 2252 eyes from 1126 subjects were collected from 2 centers: Aravind Eye Hospital, Madurai and Coimbatore, India. The images of 1926 eyes (963 subjects) were used to train an automated image analysis-based system, which was developed to provide a decision on a given fundus image. A total of 163 subjects were clinically examined by 2 ophthalmologists independently and their diagnostic decisions were recorded. The consensus decision was defined to be the clinical reference (gold standard). Fundus images of eyes with disagreement in diagnosis were excluded from the study. The fundus images of the remaining 314 eyes (157 subjects) were presented to 4 graders and their diagnostic decisions on the same were collected. The performance of the system was evaluated on the 314 images, using the reference standard. The sensitivity and specificity of the system and 4 independent graders were determined against the clinical reference standard.
Results:
The system achieved an area under receiver operating characteristic curve of 0.792 with a sensitivity of 0.716 and specificity of 0.717 at a selected threshold for the detection of glaucoma. The agreement with the clinical reference standard as determined by Cohen κ is 0.45 for the proposed system. This is comparable to that of the image-based decisions of 4 ophthalmologists.
Conclusions and relevance:
An automated system was presented for glaucoma detection from color fundus photographs. The overall evaluation results indicated that the presented system was comparable in performance to glaucoma classification by a manual grader solely based on fundus image examination.
Purpose:
We examined the efficacy of an extended-release drug delivery system, nanosponge (NS) encapsulated compounds, administered intravitreally to lower intraocular pressure (IOP) in mice.
Methods:
Bilateral ocular hypertension was induced in mice by injecting microbeads into the anterior chamber. Hypertensive mice received NS loaded with ocular hypotensive drugs via intravitreal injection and IOP was monitored. Retinal deposition and retinal ganglion cell (RGC) uptake of Neuro-DiO were examined following intravitreal injection of Neuro-DiO-NS using confocal microscopy.
Results:
Brimonidine-loaded NS lowered IOP 12% to 30% for up to 6 days (P < 0.02), whereas travoprost-NS lowered IOP 19% to 29% for up to 4 days (P < 0.02) compared to saline injection. Three bimatoprost NS were tested: a 400-nm NS and two 700-nm NS with amorphous (A-NS) or amorphous/crystalline (AC-NS) crosslinkers. A single injection of 400 nm NS lowered IOP 24% to 33% for up to 17 days compared to saline, while A-NS and AC-NS lowered IOP 22% to 32% and 18% to 26%, respectively, for up to 32 days (P < 0.046). Over time retinal deposition of Neuro-DiO increased from 19% to 71%; Neuro-DiO released from NS was internalized by RGCs.
Conclusions:
A single injection of NS can effectively deliver ocular hypotensive drugs in a linear and continuous manner for up to 32 days. Also, NS may be effective at targeting RGCs, the neurons that degenerate in glaucoma.
Translational relevance:
Patient compliance is a major issue in glaucoma. The use of NS to deliver a controlled, sustained release of therapeutics could drastically reduce the number of patients that progress to vision loss in this disease.
Hereditary optic neuropathies are a prominent cause of blindness in both children and adults. The disorders in this group share many overlapping clinical characteristics, including morphological changes that occur at the optic nerve head. Accurate and prompt clinical diagnosis, supplemented with imaging when indicated, is essential for optimum management of the relevant optic neuropathy and appropriate counseling of the patient on its natural history. Patient history, visual field assessment, optic disc findings and imaging are the cornerstones of a correct diagnosis. This Review highlights the characteristic optic nerve head features that are common to the various hereditary optic neuropathies, and describes the features that enable the conditions to be differentiated.
The vertical cup/disc ratio (CDR) has long been used in the assessment of the glaucoma suspect, though the wide range of CDR values in the normal population limits its use. Cup size is related physiologically to disc size and pathologically to glaucomatous damage. Disc size can be measured at the slit lamp as the vertical disc diameter (DD). The ability of the CDR, in relation to DD, to identify glaucomatous optic discs was investigated.
88 normal, 53 early glaucoma, and 59 ocular hypertensive subjects underwent stereoscopic optic disc photography and clinical biometry. Photographs were analysed in a masked fashion by computer assisted planimetry. The relation between vertical cup diameter and DD was explored by linear regression, and expressed in terms of CDR. The upper limit of normal was defined by the 95% prediction intervals of this regression (method 1) and by the upper 97.5 percentile for CDR (method 2). The sensitivity and specificity of CDR to identify an optic disc as glaucomatous was tested with these disc size dependent and disc size independent cut offs in small, medium, and large discs.
The CDR was related to DD by the equation CDR = (-1.31 + (1.194 x DD))/DD. The sensitivity in small, medium, and large discs was 80%, 60%, and 38% respectively for method 1 and 33%, 67%, and 63% respectively for method 2. Specificity was 98.9% (method 1) and 97.7% (method 2).
The CDR, relative to disc size, is useful clinically, especially to assist in identifying small glaucomatous discs.
An analysis of data from the Advanced Glaucoma Intervention Study (AGIS) has found eyes reported to have partial optic disc rim notching (not to the edge) at baseline to have less risk of subsequent visual field loss than eyes with no notching. Because this is counterintuitive and because classification of notching had not been defined in the AGIS protocol, we have assessed AGIS ophthalmologists interobserver and intraobserver agreement on notching.
Fourteen glaucoma subspecialists classified notching in 26 pairs of stereoscopic disc photographs of eyes with mild to severe glaucomatous optic neuropathy. They classified images as showing either no notching, notching not to the edge, or notching to the edge. Several hours later, 10 of them classified the same images a second time.
In an analysis of interobserver agreement, of 26 stereoscopic images, a plurality of ophthalmologists classified notching as absent in 9 (35%), as present but not to the edge in 7 (27%), and as present and not to the edge in 10 (38%). All 14 ophthalmologists (100%) agreed on the classification of 7 (27%) of the images, and 13 of the 14 ophthalmologists (93%) agreed on the classification of 4 additional images (15%). Of these 11 images with at least 93% agreement, notching was reported as absent in 3 (27%) and to the edge in 8 (73%). In the remaining 15 images, there was substantial disagreement about whether notching was present and, if so, whether it was to the edge. In an analysis of intraobserver agreement, none of the 10 ophthalmologists who completed the viewing a second time classified all eyes exactly the same as the first time, though 5 ophthalmologists made 4 or fewer reclassifications. Overall, 80% of the original classifications were reproduced on second reading. Of the initial classifications that were not reproduced, slightly more than half were first classified as having notching not to the edge.
Without definitions or examples of optic disc rim notching, the glaucoma subspecialists had relatively high intraobserver agreement but were likely to disagree with each other in characterizing the degree of disc rim notching. We recommend development of a standard photographic classification of disc rim notching. The classification should be tested for inter- and intra-observer agreement.
To compare monoscopic and stereoscopic assessment of the optic disc using novel software for the digital stereoscopic analysis of optic disc stereopairs.
Software was developed for the stereoscopic display of digital optic disc images using an interlaced display method. Neuroretinal rim width was determined at 10 degree intervals around the optic disc using a custom (stereoscopic) cursor whose depth was adjusted to that of Elschnig's rim. Measurements were taken, first viewing the disc monoscopically and at a separate sitting, stereoscopically.
Measurements were made in 35 eyes from 35 patients (1260 estimates for each observer) using three observers. The mean cup to disc ratio (CDR) ranged from 0.57 to 0.66 (SD 0.13-0.14) for monoscopic viewing compared with 0.64 to 0.69 (SD 0.12-0.14) for stereoscopic viewing. Stereoscopic assessments gave higher CDRs in temporal, superior, nasal, and inferior aspects of the optic disc (p<0.001, Mann-Whitney U test). Agreement between observers in estimating CDR was high for monoscopic assessment (intraclass correlation coefficient 0.74 (CI 0.72 to 0.76) increasing to 0.80 (0.78 to 0.82) for stereoscopic assessment.
Digital stereoscopic optic disc assessment provides lower estimates of neuroretinal rim width and higher levels of interobserver agreement compared with monoscopic assessments.
To determine the repeatability and agreement between a digital camera (monoscopic and stereoscopic images) and the Heidelberg Retina Tomograph (HRT; Heidelberg Engineering, Heidelberg, Germany) in determining cup-to-disc ratio. A secondary purpose was to determine the monoscopic and stereoscopic thresholds that maximize specificity and sensitivity when compared with the HRT.
Community living participants aged between 70 and 79 years had their optic discs imaged with a digital nonmydriatic retinal camera (NMRC) and the HRT. Intraclass correlation coefficients (ICCs) and 95% tolerance limits of change were used to determine repeatability characteristics of the instruments. The agreement between the HRT- and NMRC-derived area cup-to-disc ratios was assessed using weighted kappa statistics and receiver operator characteristic (ROC) curves.
The agreements between the monoscopic and stereoscopic images and HRT were assessed for 1238 and 1173 eyes, respectively. The reliability measures for both NMRC techniques and HRT were almost perfect (ICC = 0.84-0.99) with narrow tolerance limits of change (9.2%-18.4%) and very small systemic biases (P < 0.05). The agreement between the HRT and both NMRC techniques was substantial, with a weighted kappa = 0.83. The HRT gave a marginally larger area cup-to-disc ratio than the monoscopic and stereoscopic images by 0.008 and 0.006, respectively (P < 0.001). The areas under the ROC curves for both NMRC techniques were 0.98, indicating excellent discriminating characteristics (P < 0.001). An area cup-to-disc ratio cutoff of > or =0.5 for the monoscopic and stereoscopic NMRC was highly specific (94.1% and 91.6%) and sensitive (87.5% and 97.2%, respectively) in determining an HRT-derived area cup-to-disc ratio >0.6.
The monoscopic and stereoscopic digital images showed excellent repeatability and demonstrated substantial agreement with the HRT. The results indicate that the digital NMRC could be a reliable and useful instrument for assessing area cup-to-disc ratio and screening for glaucoma-suspect eyes in the community.
Glaucomatous optic neuropathy is classified by morphologic changes in the intrapapillary and parapapillary region of the optic nerve head and the retinal nerve fibre layer. These changes can be evaluated using descriptive optic nerve head variables which are the size and shape of the optic disc; size, shape and pallor of the neuroretinal rim; size of the optic cup in relation to the area of the disc; configuration and depth of the optic cup; cup-to-disc diameter ratio and cup-to-disc area ratio; position of the exit of the central retinal vessel trunk on the lamina cribrosa surface; presence and location of splinter-shaped haemorrhages; occurrence, size, configuration and location of parapapillary chorioretinal atrophy; diffuse and/or focal decrease of the diameter of the retinal arterioles; and visibility of the retinal nerve fibre layer. Assessment of these variables is useful for the early detection of glaucomatous optic nerve damage, to follow-up patients with glaucoma, to differentiate various types of the chronic open-angle glaucomas, and to get hints for the pathogenesis of glaucomatous optic nerve fibre loss.
Development of a standardized internet-based system to self-assess skills in optic disc examination for glaucoma risk assessment.
Prospective internet-based observational study.
Total of 197 participants (glaucoma subspecialists, general ophthalmologists and trainees) from 22 countries.
Forty-two optic disc images demonstrating a range of features were selected from 2500 monoscopic disc photographs of normal and glaucomatous eyes. Images were presented to clinicians via website (http://www.gone-project.com). Participants were asked to assess nine topographic features and make a subjective assessment of glaucoma likelihood.
Inter-observer agreement using kappa (κ) or weighted kappa (κ(w) ).
There was substantial level of inter-observer agreement between glaucoma subspecialists for assessment of glaucoma likelihood (κ(w) = 0.63). Inter-observer agreement was high for haemorrhage (κ= 0.83) and substantial for disc size, disc shape, cup:disc ratio, peripapillary atrophy and cup shape (κ(w) = 0.59-0.68). Subspecialists had stronger inter-observer agreement for glaucoma likelihood and for most disc characteristics than did trainees: the greatest difference being the assessment for retinal nerve fibre layer loss. Analysis of individual disc answers from ophthalmology trainees showed that discs leading to lower agreement of glaucoma likelihood tend to produce lower agreement for the assessment of cup:disc ratio, cup shape, cup depth and retinal nerve fibre layer. Discs with features of moderate to deep cup or cup:disc ratio between 0.6 and 0.8 also lead to lower agreement in glaucoma likelihood.
This internet-based system is a readily accessible and standardized tool, for clinicians globally, that permits self-assessment and benchmarking of skills in optic disc examination.
Beta-Zone parapapillary atrophy (PPA) occurs more commonly in eyes with glaucoma. Rates of glaucomatous visual field (VF) progression in eyes with and without beta-zone PPA at the time of baseline assessment were compared.
Retrospective, comparative study.
Two hundred forty-five patients from the New York Glaucoma Progression Study.
Subjects with glaucomatous optic neuropathy and repeatable VF loss were assessed for eligibility. Eyes with a Heidelberg Retina Tomograph II (HRT) examination, at least 5 visual field tests after the HRT in either eye, optic disc photographs, and <6 diopters of myopia were enrolled. beta-Zone PPA was defined as a region of chorioretinal atrophy with visible sclera and choroidal vessels adjacent to the optic disc. Global rates of VF progression were determined by automated pointwise linear regression analysis. Univariate analysis included age, gender, ethnicity, central corneal thickness (CCT), refractive error, baseline mean deviation, baseline intraocular pressure (IOP), mean IOP, IOP fluctuation, disc area, rim area, rim area-to-disc area ratio, beta-zone PPA area, beta-zone PPA area-to-disc area ratio, and presence or absence of beta-zone PPA.
The relationship between beta-zone PPA and the rate and risk of glaucoma progression.
Two hundred forty-five eyes of 245 patients (mean age, 69.6+/-12.3 years) were enrolled. The mean follow-up was 4.9+/-1.4 years and the mean number of VFs after HRT was 9.3+/-2.7. beta-Zone PPA was present in 146 eyes (65%). Eyes with beta-zone PPA progressed more rapidly (-0.84+/-0.8 dB/year) than eyes without it (-0.51+/-0.6 dB/year; P<0.01). Multivariate regression showed significant influence of mean IOP (hazard ratio [HR], 1.11; P<0.01), IOP fluctuation (HR, 1.17; P = 0.02), and presence of beta-zone PPA (HR, 2.59; P<0.01) on VF progression. Moderate (0.5-1.5 dB/year; P = 0.01) and fast (>1.5 dB/year; P = 0.08) global rates of progression occurred more commonly in eyes with beta-zone PPA than in eyes without it. Thinner CCT (<525 microm) had a weak but significant correlation with presence of beta-zone PPA (kappa = 0.13).
Eyes with beta-zone PPA are at increased risk for glaucoma progression and warrant close clinical surveillance.
To determine the diagnostic accuracy of judging optic disc photographs for glaucoma by ophthalmologists.
Evaluation of diagnostic test and technology.
A total of 243 of 875 invited ophthalmologists in 11 European countries.
We determined how well each participant classified 40 healthy eyes and 48 glaucomatous eyes with varying severity of the disease on stereoscopic slides. Duplicate slides were provided for determining intraobserver agreement. All eyes were also imaged with the GDx with variable corneal compensation (GDx-VCC) (Carl Zeiss Meditec AG, Jena, Germany) and the Heidelberg Retina Tomograph (HRT) I (Heidelberg Engineering GmbH, Heidelberg, Germany). Diagnostic accuracies of clinicians were compared with those of the best machine classifiers.
Accuracy of classification, expressed as sensitivity, specificity, and overall accuracy. Intraobserver agreement (kappa).
The overall diagnostic accuracy of ophthalmologists was 80.5% (standard deviation [SD], 6.8; range, 61.4%-94.3%). The machine classifiers outperformed most observers in diagnostic accuracy; the GDx-VCC nerve fiber indicator and the HRT's best classifier correctly classified 93.2% and 89.8% of eyes, respectively. The intraobserver agreement (kappa) varied between -0.13 and 1.0 and was on average good (0.7).
In general, ophthalmologists classify optic disc photographs moderately well for detecting glaucoma. There is, however, large variability in diagnostic accuracy among and agreement within clinicians. Common imaging devices outperform most clinicians in classifying optic discs.
Proprietary or commercial disclosure may be found after the references.
Morphometrics, a branch of morphology, represents the study of size and shape components of biological form and their variation in the population. Assessment of optic disc morphology is essential in the diagnosis and management of many ophthalmic disorders. Much work has been performed to characterize size-related parameters of the optic disc; however, limited information is available on shape variation in the general population. In contrast to optic disc or cup sizes, which are conceptually meaningful variables with a defined unit of measurement, there are few metric constructs by which to quantify, visualize and interpret variation in optic disc or cup shape. This has significance in ophthalmic diseases with a genetic basis as recent evidence has suggested that optic disc shape may be heritable. Conventional optic disc shape measures of 'ovality' and 'form-factor' reduce a complex structure to a single number and eliminate information of potential diagnostic relevance from further analyses. The recent advent of 'geometric morphometrics', a branch of statistics that incorporates tools from geometry, biometrics and computer graphics in the quantitative analysis of biological forms, has enabled spatial relationships in shape data to be retained during analysis. The analytical methods employed in geometric morphometrics can be separated into two distinct groups: landmark-based (e.g. Procrustes analysis, thin-plate splines) and boundary outline techniques (e.g. Fourier analysis). In this review, we summarize current approaches to the study of optic disc morphology, discuss the underlying theory of geometric morphometrics within the context of analytical techniques and then explore the contemporary relevance of the subject matter to several biological fields. Finally we illustrate the potential application of geometric morphometrics to the specific problem of optic disc shape and glaucoma assessment.
A significant amount of variability exists between observers in designating cup/diac (C/D) ratios. Further, different methods are used for evaluating the size of the cup. On method uses a combination of contour and color, the other specifically measures contour and pallor separately. This study confirms that these methods yeild different numerical results. Interestingly, the "cupping/pallor" observers show no significant difference among one another in C/D determination which may support the notion that cupping observations are more consistent than "standard" C/D ratios. Though both groups feel the cup is larger on stereo compared to monocular viewing, the "cupping/pallor" group demonstrated much greater differences in this regard. However, the evaluation of pallor by the latter group corresponded closely to the C/D ratios of the "standard" group both for nonstereo and stereo. An individual observer is reasonably consistent on repeat evaluation, but at times, inconsistency is quite substantial. These inconsistencies lead to the conclusion that C/D ratios are an inexact method of recording the status of a disc. Except for considerable changes over time, this numerical method is probably not reliable in checking for small disc changes. However, the disc alone can provide clues as to whether it is physiologic or pathologic. Nevertheless, even with expert observers, significant variability exists in interpretation of an optic disc in this regard. Individuals who evaluate the disc cup by cupping and pallor tend to call discs more pathologic than other observers. The clinical implications of this study suggest that certain observers are more accurate than others and certain discs are more easily evaluated than others. Yet, no one method seems foolproof and no specific criteria can as yet distinguish a normal from an abnormal disc. Stereo color transparencies are evaluated more accurately than are nonstereo color prints. Vessel detail and pallor patterns are probably the most useful disc characteristics to observe. In this regard, a C/D ratio does nothing to indicate whether a disc is normal or not. A statement as to the observer's opinion should be made for each disc evaluation in addition to recording the appearance of the disc.
The authors studied intraobserver and interobserver agreement, under monoscopic and stereoscopic conditions, in estimating vertical cup-to-disc ratios and in assessing whether a disc had glaucomatous damage. Six glaucoma experts evaluated 75 optic disc photographs under both viewing conditions. The experts also re-evaluated 25 photographs. Intraobserver agreement in estimating vertical cup-to disc ratios was high (median weighted kappa, 0.79). Interobserver agreement in estimating vertical cup-to-disc ratios was moderate (stereoscopic median weighted kappa, 0.67); individual experts differed by as much as 0.2 disc diameters (DD) monoscopically and 0.16 DD stereoscopically. The observers estimated larger vertical cup-to-disc ratios when evaluating the same discs under stereoscopic conditions than under monoscopic conditions. Intraobserver agreement in assessing glaucomatous disc damage was substantial (median kappa, 0.76). Interobserver agreement in assessing glaucomatous damage was moderate (stereoscopic median kappa, 0.50). This study confirms the ability of experts to reliably evaluate the optic disc within themselves and emphasizes the need for developing standardized methods for interobserver evaluation of the optic disc in glaucoma.
Glaucomatous optic nerve damage is typically associated with intrapapillary changes, such as neuroretinal rim loss. In this study, parapapillary chorioretinal atrophy was evaluated in 691 normal eyes, 1081 glaucomatous eyes, and 31 eyes with ocular hypertension. It was significantly larger and occurred more often in the glaucomatous eyes (parapapillary atrophy area, 1.07 +/- 0.83 mm2) (mean +/- SD) than in the normal eyes (0.55 +/- 0.64 mm2) or in the eyes with ocular hypertension (0.55 +/- 0.37 mm2). These differences were significant also for eyes with moderate glaucomatous damage (0.86 +/- 0.62 mm2). Parapapillary chorioretinal atrophy was associated with shallow glaucomatous cupping, diffuse nerve fiber loss, a marked tessellated fundus, and only moderately elevated intraocular pressure. It increased with a decreasing neuroretinal rim area. It showed a spatial correlation to neuroretinal rim loss inside the optic disc. In unilateral glaucoma, it was larger in the affected eye than in the unaffected eye. Parapapillary chorioretinal atrophy is associated with glaucoma.
Glaucomatous optic nerve damage has generally been associated with high cup-to-disc ratios. Fifteen eyes of nine patients with increased intraocular pressure and glaucomatous visual field loss but low cup-to-disc ratios are reported. The optic disc area was significantly (P less than 0.01) smaller than in 429 normal subjects and 556 glaucoma patients with high cup-to-disc ratios. Parapapillary chorioretinal atrophy was significantly larger and retinal nerve fiber bundles were significantly less visible than in the normal group. The latter two parameters were not significantly different in the glaucoma groups with low and high cup-to-disc ratios when the groups were matched for mean perimetric loss. The authors conclude that in eyes with small optic discs, glaucomatous optic nerve damage may be indicated more sensitively by parapapillary changes than by cup-to-disc ratios. Glaucomatous eyes with small optic nerve heads can have misleadingly low cup-to-disc ratios.
Using planimetric analysis of stereoscopic optic disk photographs, we studied 21 optic nerve heads of 11 subjects who shared the common feature of optic cups that were larger than the mean + 2 S.D. within the normal population. A comparison of these findings to those of 571 normal optic disks and 706 optic nerve heads in eyes with chronic primary open-angle glaucoma showed the following morphologic characteristics: (1) abnormally large optic disk area (mean +/- S.D., 4.49 +/- 0.56 mm2), (2) large cup/disk ratios with the horizontal ratio (0.78 +/- 0.03) significantly (P less than .001) larger that the vertical (0.71 +/- 0.03), (3) increased incidence of cilioretinal arteries, (4) normal neuroretinal rim area (2.06 +/- 0.35 mm2), (5) normal neuroretinal rim configuration, inferiorly (0.43 +/- 0.08 mm) broader (P less than .001, Wilcoxon test) than superiorly (0.33 +/- 0.06 mm), smallest (P less than .0001) temporally (0.20 +/- 0.04 mm), (6) normal form of zone alpha (irregular hypopigmentation and hyperpigmentation) of the parapapillary chorioretinal atrophy with its widest extension in the temporal horizontal area, (7) no zone beta (visible large choroidal vessels and sclera), (8) normal caliber of the parapapillary retinal vessels, and (9) normal parapillary retinal nerve fiber layer. These characteristics are helpful in the differentiation of primary and secondary large cups.
Reliable measures of optic nerve status are important parameters for epidemiologic and clinical studies of glaucoma. Stereo fundus photography has been used to reduce interobserver variation, but little data have been reported quantifying the level of agreement in these situations. This article examines chance-corrected intraobserver and interobserver agreement for horizontal and vertical cup:disc ratios and width of the narrowest remaining neuroretinal rim on stereo, color, fundus photographs. Intraobserver agreement was substantial to almost perfect for both readers (kappas between 0.82 and 0.86 for horizontal and vertical cup:disc ratios and 0.71 for the neuroretinal rim). Interobserver agreement was also substantial (kappas of 0.71 for horizontal cup:disc ratio, 0.74 for vertical cup:disc ratio, and 0.58 for the neuroretinal rim). Both intraobserver and interobserver agreement were best for glaucoma patients followed by ocular hypertensives and controls. Dichotimizing cup:disc ratios into suspicious or not suspicious using a range of cutoff points (greater than or equal to 0.5, greater than or equal to 0.6, and greater than or equal to 0.7 disc diameters [DD]) had little effect on agreement levels. Estimation of the status of the optic nerve is complex, requiring judgment about the shape and structure of the cup. Agreement is optimized by using stereo photographs, using the same observer and, when evaluating progression, comparing photographs from different points in time at a single sitting.
Four hundred and fifty-seven unselected normal human optic nerve heads of 319 subjects (163 men, 156 women, mean age 42.7 +/- 19.6 years) were evaluated by magnification-corrected morphometry of optic disc photographs. Mean optic disc surface measured 2.69 +/- 0.70 mm2 (0.80-5.54 mm2), mean diameter horizontally 1.76 +/- 0.31 mm (0.91-2.61 mm), and vertically 1.92 +/- 0.29 mm (0.96-2.91 mm). The form was slightly vertically oval. Optic cup area averaged 0.72 +/- 0.70 (0.00-3.41 mm2), mean horizontal cup diameter 0.83 +/- 0.58 mm (0.00-2.08 mm) and mean vertical diameter 0.77 +/- 0.55 mm (0.00-2.13 mm). The cup was significantly (P less than 0.0001) larger in discs with steep "punched-out" cups (1.37 +/- 0.62 mm2) compared to discs having cups with temporal flat slopes (0.59 +/- 0.39 mm2). Neuroretinal rim area ranged from 0.80 to 4.66 mm2 (mean 1.97 +/- 0.50 mm2), and was significantly correlated (P less than 0.00001) to the optic disc area. It was broadest in the inferior optic disc region (P less than 0.001), followed by the superior, nasal and temporal (P less than 0.001) regions. Horizontal cup/disc ratio (mean 0.39 +/- 0.28, minimum 0.00, maximum 0.87) was larger in 426 (93.2%) optic discs than the vertical one (mean 0.34 +/- 0.25, minimum 0.00, maximum 0.85). Concerning optic disc area, side differences of 0.10 mm2 or less were detected in 28%, of 0.2 mm2 or less in 46% and of 0.50 mm2 or less in 80% (cumulative frequencies). Concerning neuroretinal rim area, side differences of 0.10 mm2 or less were found in 31%, of 0.20 mm2 or less in 52% and of 0.50 mm2 or less in 84%.(ABSTRACT TRUNCATED AT 250 WORDS)
To determine the agreement among optometrists, ophthalmologists, and ophthalmology residents in assessing glaucomatous optic nerve damage. The authors also determined the sensitivity of each group of observers for identifying glaucomatous optic nerve damage.
Six optometrists, six general ophthalmologists, and six third-year ophthalmology residents evaluated 75 stereoscopic optic disc photographs. Observers estimated the vertical cup:disc ratio (C:D) and assessed the presence of glaucomatous damage. Agreement among and within observers was estimated by the kappa statistic (KW, k). The sensitivity and specificity for the identification of glaucomatous optic nerve damage were determined for each group of participants.
Intraobserver agreement (KW 0.69-0.79) was greater than interobserver agreement (KW 0.56-0.68) in assessing the C:D ratio and glaucomatous optic nerve damage for optometrists, ophthalmologists, and residents. Interobserver agreement for ophthalmologists (KW 0.68) was substantial and significantly higher than for optometrists (KW 0.56) and residents (KW 0.56) when estimating the C:D ratio. Ophthalmologists and residents had higher sensitivity (78%) in identifying glaucomatous optic nerve damage than did optometrists (56%). The specificity for all three groups was relatively poor (range, 47%-60%).
The moderate interobserver agreement across all three groups of observers suggests the need to develop standardized criteria for assessing glaucomatous optic disc damage. Ophthalmologists in this study have a higher interobserver agreement in estimating the C:D ratio and are more sensitive than optometrists in assessing glaucomatous optic nerve damage.
The cornea and the optic disc form the anteroposterior opening of the sclera. This study evaluated whether an abnormal shape of the optic disc is associated with an abnormal configuration of the cornea measured as corneal astigmatism.
The study design was a cross-sectional one.
The study included 882 subjects (430 women, 452 men) with a mean age of 45.9 +/- 13.6 years (mean +/- standard deviation; range, 8-87 years) and a mean refractive error of -1.09 +/- 2.76 diopters (range, -21.0 diopters to +7.0 diopters).
Corneal astigmatism was determined by keratometry, and the optic disc was analyzed morphometrically by planimetric evaluation of optic disc photographs.
Corneal astigmatism, ratio of minimal-to-maximal disc diameter, and optic disc form factor were measured.
The amount of corneal astigmatism was significantly (P < 0.001) correlated with an increasingly elongated optic disc shape. Corneal astigmatism was significantly (P < 0.01) higher in eyes with tilted discs. It was significantly (P = 0.006) smaller in eyes with an almost circular disc shape. Amblyopia was significantly (P < 0.05) associated with an elongated optic disc shape and high corneal astigmatism. The axis of corneal astigmatism was correlated with the orientation of the longest disc diameter. The optic disc was significantly (P < 0.05; chi-square test) more often horizontally oval in eyes with a steeper horizontal corneal meridian than in eyes with a steeper vertical corneal meridian. Correspondingly, the disc was significantly (P < 0.05) more often vertically oval in eyes with a steeper vertical corneal meridian than in eyes with a steeper horizontal corneal meridian.
An abnormal optic disc shape is significantly correlated with corneal astigmatism. Especially in young children, if an abnormal optic disc shape is found on routine ophthalmoscopy, refractometry should be performed to rule out corneal astigmatism and to prevent amblyopia. The direction of the longest optic disc diameter can indicate the axis of corneal astigmatism.
Optic nerve diseases, such as the glaucomas, lead to changes in the intrapapillary and parapapillary region of the optic nerve head. These changes can be described by the following variables: size and shape of the optic disk; size, shape, and pallor of the neuroretinal rim; size of the optic cup in relation to the area of the disk; configuration and depth of the optic cup; ratios of cup-to-disk diameter and cup-to-disk area; position of the exit of the central retinal vessel trunk on the lamina cribrosa surface; presence and location of splinter-shaped hemorrhages; occurrence, size, configuration, and location of parapapillary chorioretinal atrophy; diffuse and/or focal decrease of the diameter of the retinal arterioles; and visibility of the retinal nerve fiber layer (RNFL). These variables can be assessed semiquantitively by ophthalmoscopy without applying sophisticated techniques. For the early detection of glaucomatous optic nerve damage in ocular hypertensive eyes before the development of visual field loss, the most important variables are neuroretinal rim shape, optic cup size in relation to optic disk size, diffusely or segmentally decreased visibility of the RNFL, occurrence of localized RNFL defects, and presence of disk hemorrhages.
Demonstrating that optometrists can make valid and reliable assessments of optic disc features is an important prerequisite for establishing schemes for shared care/co-management. Previous studies have estimated observer variability among experts in the assessment of optic disc cupping, but there has been a paucity of information on observer variability amongst optometrists. This paper describes a study to investigate intra- and inter-observer variability for a range of disc features, as graded by both ophthalmologists and optometrists. Five observers (three optometrists and two ophthalmologists) graded 48 stereo-pairs of optic disc photographs from 48 patients on two separate occasions. Each observer graded the following features: vertical and horizontal C/D ratios, narrowest rim width, the presence/absence of a disc haemorrhage, focal pallor of the neuroretinal rim, peri-papillary atrophy, the steepness of the cup-edge and the presence/absence of the cribriform sign. The average intra- and inter-observer standard deviation (SD) of differences are, respectively, 0.11 and 0.19 for the vertical C/D ratios and 0.10 and 0.18 for the horizontal C/D ratios. For the vertical C/D ratio the average weighted kappa (kappa w) is 0.79 within observers and 0.46 between observers. Percentage agreements for the presence/absence of a disc haemorrhage range from 96 to 100% (average kappa = 0.92) within observers and from 90 to 98% (average kappa = 0.77) between observers. For other disc features, average kappa w values range from 0.67 to 0.71 within observers and from 0.23 to 0.46 between observers. Intra- and inter-observer comparisons (within and between different professionals) across all disc features are comparable for the optometrists and ophthalmologists, thus demonstrating that optometrists can make valid assessments of disc features. The implications for shared care are discussed.
To compare stereophotographic and monophotographic optic disc assessments made using a digital optic disc stereo camera.
Stereo digital optic disc photographs of 150 selected patients who had presented to a glaucoma clinic were assessed by two masked observers on separate occasions using (1) the stereophotographs and a stereoviewer, (2) a single image from the same stereopair. Results were analysed for both right and left eyes separately. 95% tolerance limits for change (TC) and intraclass correlation coefficients (ICC) were calculated and a multivariate analysis using a general linear model for repeated measures was performed.
A total of 201 optic disc images of 150 patients (84 females, 108 left eyes) were analysed. Mean age of patients was 64 years. The results for right eyes are as follows (similar results were obtained for left eyes). Intraobserver (stereoscopic compared to monoscopic) measurements of: horizontal cup:disc ratios (CDR), ICC = 0.5995 and 0.7269, TC = 34% and 27%; vertical CDR, ICC = 0.8298 and 0.817, TC = 25% and 27%; area CDR, ICC = 0.7757 and 0.8259, TC = 28% and 25%; circumference CDR, ICC = 0.7618 and 0.8103, TC = 28% and 25%. Interobserver measurements of: horizontal CDR, ICC stereoscopic (SS) = 0.7287; monoscopic (MS) = 0.5030; TC SS = 30%; MS = 32%; vertical CDR, ICC SS = 0.8439; MS = 0.7106; TC SS = 25%; MS = 31%; area CDR, ICC SS = 0.8392; MS = 0.6276; TC SS = 26%; MS = 32%; circumference CDR, ICC SS = 0.8433; MS = 0.6438, TC SS = 26%; MS = 31%. Systematic bias between observers and between methods was within acceptable limits.
This study using a digital stereo camera indicates that there may be little benefit of stereoscopic imaging over monoscopic imaging despite demonstrating small but inconsistent differences between both observers and methods.
To determine the interobserver and intraobserver agreement in the recognition of different patterns of glaucomatous optic disk damage and evaluate if these patterns changed over time in patients followed for a number of years.
Patients with early to moderate glaucoma (n = 105) were consecutively enrolled to participate in a prospective observational study. In the first part of the present study, optic disk photographs obtained closest to patient's entry date in the prospective study were classified in a masked fashion by three observers according to the pattern of optic disk damage into one of the following: (1) focal, (2) myopic, (3) senile sclerotic, (4) concentric cup enlargement, (5) normal appearance, or (6) miscellaneous (those disks that did not qualify for any of the other groups). The observers were also asked to assign a confidence score for each classification, ranging from 1 (low confidence) to 5 (highest confidence). The three observers reclassified the photographs after a minimal period of 2 months, in order to assess intraobserver agreement. In a second part of the study, one observer reviewed, in a masked fashion, all the optic disk photographs that had been taken during the routine follow-up of the 105 patients in order to evaluate whether the classification of disk pattern changed over time.
Intraobserver agreement yielded kappa values (95% confidence interval [CI]) from 0.51 (CI, 0.40 to 0.62) to 0.85 (CI, 0.77 to 0.93) depending on the observer. Interobserver agreement kappa values between all three observers was 0.40 (CI, 0.35 to 0.46), but it improved if photographs classified with moderate or higher degrees of confidence were included (0.52 [CI, 0.44 to 0.60]). During a mean follow-up period of 8.2 +/- 4.8 years, 41.6% of the eyes were always classified into the same group and 23.6% of the eyes were classified at least once into two or more of the four patterns of damage (groups 1 to 4).
Intraobserver agreement on the pattern of optic disk damage was generally very good, with some variation among the observers. Interobserver agreement was reasonable and increased proportionally with the confidence in the classification. During long-term follow-up, the pattern of optic disk damage usually did not change. This type of classification can probably be used accurately in clinical practice.
To determine if a computer-based stereoscopic teaching program could improve optic cup/disc ratio (CDR) agreement between student observers and an expert.
Experimental study.
Six student observers (A-F) assessed at least 30 digital stereoscopic optic disc images using a digital stereoscopic analysis program. All observers made 36 CDR measurements (every 10 degrees ) for each disc image that they assessed.
Disc images were divided into 3 sets of 10 images (image sets 1, 2, and 3). Observers A, B, and C determined CDRs for all 3 image sets. Set 1 was examined without access to a teaching program, set 2 using a teaching program that allowed the observer to view an expert assessment of the optic disc, and set 3 after using the teaching program. Observers D, E, and F evaluated image sets 1 and 3 only and did not have access to the teaching program. Ten months later, these same observers (D-F) viewed image set 2 using the teaching program and then reassessed image set 3.
Standard deviation (SD) of CDR differences between observers and an expert, the percentage of CDR observations differing > or =0.20 from those of the expert, and intraclass correlation coefficients (ICCs) between observers.
Before teaching, the average SD of differences between all observers and the expert was 0.10. After teaching, the average SD of differences between all observers and the expert was 0.057. For observers D, E, and F, the average SD of differences for image set 3 without teaching was 0.074. The percentage of all observer measurements that differed > or =0.20 from those of the expert for image set 1 was 16.7%; for set 3 after teaching, 1.7%; and for set 3 for observers D, E, and F without teaching, 10%. Interobserver ICC values for all observers were 0.37 before teaching and 0.76 after. For observers D, E, and F, the ICC value for set 3 without access to teaching was 0.69.
This study suggests that an interactive digital stereoscopic teaching program improves agreement between observers and an expert when assessing CDRs.
Characteristics of misclassified discs in the European Optic Disc Assessment Trial (EODAT) Diagnosis and pathogenesis of glaucomatous optic neuropathy: morphological aspects
- Nj Reus
- Hg Lemij
- Eodate Group
- Wm Budde
Reus NJ, Lemij HG, Group EODATE.
Characteristics of misclassified discs in the
European Optic Disc Assessment Trial (EODAT).
Invest Ophthalmol Vis Sci. 2008;49:3627.
13. Jonas JB, Budde WM. Diagnosis and
pathogenesis of glaucomatous optic neuropathy:
morphological aspects. Prog Retin Eye Res.
2000;19(1):1-40.
Glaucomatous optic neuropathy evaluation project: a standardized Internet system for assessing skills in optic disc examination
- Y X Kong
- M A Coote
- O Neill
Kong YX, Coote MA, O'Neill EC, et al.
Glaucomatous optic neuropathy evaluation project:
a standardized Internet system for assessing skills
in optic disc examination. Clin Experiment
Ophthalmol. 2011;39(4):308-317.
The optic nerve head in hereditary optic neuropathies
- O Neill
- Ec Mackey
- Da Connell
- Pp Hewitt
- Aw Danesh-Meyer
- Hv Crowston
O'Neill EC, Mackey DA, Connell PP, Hewitt AW,
Danesh-Meyer HV, Crowston JG. The optic nerve
head in hereditary optic neuropathies. Nat Rev
Neurol. 2009;5(5):277-287.
Characteristics of misclassified discs in the European Optic Disc Assessment Trial (EODAT)
- N J Reus
- H G Lemij
- Eodate Group
Reus NJ, Lemij HG, Group EODATE.
Characteristics of misclassified discs in the
European Optic Disc Assessment Trial (EODAT).
Invest Ophthalmol Vis Sci. 2008;49:3627.
Ophthalmoscopic evaluation of the optic nerve head
- P G Sanfilippo
- A Cardini
- A W Hewitt
- J G Crowston
- D A Mackey
Sanfilippo PG, Cardini A, Hewitt AW, Crowston
JG, Mackey DA. Optic disc morphology: rethinking
shape. Prog Retin Eye Res. 2009;28(4):227-248.
15. Jonas JB, Budde WM, Panda-Jonas S.
Ophthalmoscopic evaluation of the optic nerve
head. Surv Ophthalmol. 1999;43(4):293-320.