Convolutional neural networks have recently been applied to ophthalmic diseases; however, the rationale for the outputs generated by these systems is inscrutable to clinicians. A visualization tool ...is needed that would enable clinicians to understand important exposure variables in real time.
To systematically visualize the convolutional neural networks of 2 validated deep learning models for the detection of referable diabetic retinopathy (DR) and glaucomatous optic neuropathy (GON).
The GON and referable DR algorithms were previously developed and validated (holdout method) using 48 116 and 66 790 retinal photographs, respectively, derived from a third-party database (LabelMe) of deidentified photographs from various clinical settings in China. In the present cross-sectional study, a random sample of 100 true-positive photographs and all false-positive cases from each of the GON and DR validation data sets were selected. All data were collected from March to June 2017. The original color fundus images were processed using an adaptive kernel visualization technique. The images were preprocessed by applying a sliding window with a size of 28 × 28 pixels and a stride of 3 pixels to crop images into smaller subimages to produce a feature map. Threshold scales were adjusted to optimal levels for each model to generate heat maps highlighting localized landmarks on the input image. A single optometrist allocated each image to predefined categories based on the generated heat map.
Visualization regions of the fundus.
In the GON data set, 90 of 100 true-positive cases (90%; 95% CI, 82%-95%) and 15 of 22 false-positive cases (68%; 95% CI, 45%-86%) displayed heat map visualization within regions of the optic nerve head only. Lesions typically seen in cases of referable DR (exudate, hemorrhage, or vessel abnormality) were identified as the most important prognostic regions in 96 of 100 true-positive DR cases (96%; 95% CI, 90%-99%). In 39 of 46 false-positive DR cases (85%; 95% CI, 71%-94%), the heat map displayed visualization of nontraditional fundus regions with or without retinal venules.
These findings suggest that this visualization method can highlight traditional regions in disease diagnosis, substantiating the validity of the deep learning models investigated. This visualization technique may promote the clinical adoption of these models.
The purpose of this study is to evaluate the feasibility and patient acceptability of a novel artificial intelligence (AI)-based diabetic retinopathy (DR) screening model within endocrinology ...outpatient settings. Adults with diabetes were recruited from two urban endocrinology outpatient clinics and single-field, non-mydriatic fundus photographs were taken and graded for referable DR ( ≥ pre-proliferative DR). Each participant underwent; (1) automated screening model; where a deep learning algorithm (DLA) provided real-time reporting of results; and (2) manual model where retinal images were transferred to a retinal grading centre and manual grading outcomes were distributed to the patient within 2 weeks of assessment. Participants completed a questionnaire on the day of examination and 1-month following assessment to determine overall satisfaction and the preferred model of care. In total, 96 participants were screened for DR and the mean assessment time for automated screening was 6.9 minutes. Ninety-six percent of participants reported that they were either satisfied or very satisfied with the automated screening model and 78% reported that they preferred the automated model over manual. The sensitivity and specificity of the DLA for correct referral was 92.3% and 93.7%, respectively. AI-based DR screening in endocrinology outpatient settings appears to be feasible and well accepted by patients.
Artificial intelligence technology has advanced rapidly in recent years and has the potential to improve healthcare outcomes. However, technology uptake will be largely driven by clinicians, and ...there is a paucity of data regarding the attitude that clinicians have to this new technology. In June-August 2019 we conducted an online survey of fellows and trainees of three specialty colleges (ophthalmology, radiology/radiation oncology, dermatology) in Australia and New Zealand on artificial intelligence. There were 632 complete responses (n = 305, 230, and 97, respectively), equating to a response rate of 20.4%, 5.1%, and 13.2% for the above colleges, respectively. The majority (n = 449, 71.0%) believed artificial intelligence would improve their field of medicine, and that medical workforce needs would be impacted by the technology within the next decade (n = 542, 85.8%). Improved disease screening and streamlining of monotonous tasks were identified as key benefits of artificial intelligence. The divestment of healthcare to technology companies and medical liability implications were the greatest concerns. Education was identified as a priority to prepare clinicians for the implementation of artificial intelligence in healthcare. This survey highlights parallels between the perceptions of different clinician groups in Australia and New Zealand about artificial intelligence in medicine. Artificial intelligence was recognized as valuable technology that will have wide-ranging impacts on healthcare.
Purpose
To explore the association between age‐related cataract and 10‐year mortality in an adult population in urban China.
Methods
A total of 1405 participants aged 50 years or older were examined ...at baseline in the Guangzhou Liwan Eye Study. All participants were invited to attend a 10‐year follow‐up visit. Cataract cases were defined as either having visible lens opacity confirmed with direct ophthalmoscope under pupil dilation or previous history of cataract surgery. Visual impairment (VI) was defined as a visual acuity of 20/40 or worse in the better‐seeing eye with habitual correction if worn. Body mass index (BMI) was based on anthropometric data. A brief questionnaire regarding family income, educational attainment and medical history of systemic disease was administered. Mortality rates were compared using the log‐rank test and Cox proportional hazards regression models.
Results
Among 1405 participants examined at baseline, 957 participants (68.1%) had visible lens opacity or history of cataract surgery. After 10 years, 320 (22.8%) participants died. The 10‐year mortality rate was significantly higher in participants with cataract than in those without (30.1% versus 7.14%, log‐rank p < 0.05). After adjusting for age, gender, family income, educational attainment, BMI, history of diabetes and hypertension and presence of VI, presence of cataract predicted a nearly threefold increase in the risk of mortality (HR, 2.99; 95% CI, 1.89–4.71).
Conclusions
Our findings that age‐related cataract is a predictor for poorer survival compared to those without may imply that cataract is a biomarker of ageing and frailty.
Abstract
This study investigated the diagnostic performance, feasibility, and end-user experiences of an artificial intelligence (AI)-assisted diabetic retinopathy (DR) screening model in real-world ...Australian healthcare settings. The study consisted of two components: (1) DR screening of patients using an AI-assisted system and (2) in-depth interviews with health professionals involved in implementing screening. Participants with type 1 or type 2 diabetes mellitus attending two endocrinology outpatient and three Aboriginal Medical Services clinics between March 2018 and May 2019 were invited to a prospective observational study. A single 45-degree (macula centred), non-stereoscopic, colour retinal image was taken of each eye from participants and were instantly screened for referable DR using a custom offline automated AI system. A total of 236 participants, including 174 from endocrinology and 62 from Aboriginal Medical Services clinics, provided informed consent and 203 (86.0%) were included in the analysis. A total of 33 consenting participants (14%) were excluded from the primary analysis due to ungradable or missing images from small pupils (n = 21, 63.6%), cataract (n = 7, 21.2%), poor fixation (n = 2, 6.1%), technical issues (n = 2, 6.1%), and corneal scarring (n = 1, 3%). The area under the curve, sensitivity, and specificity of the AI system for referable DR were 0.92, 96.9% and 87.7%, respectively. There were 51 disagreements between the reference standard and index test diagnoses, including 29 which were manually graded as ungradable, 21 false positives, and one false negative. A total of 28 participants (11.9%) were referred for follow-up based on new ocular findings, among whom, 15 (53.6%) were able to be contacted and 9 (60%) adhered to referral. Of 207 participants who completed a satisfaction questionnaire, 93.7% specified they were either satisfied or extremely satisfied, and 93.2% specified they would be likely or extremely likely to use this service again. Clinical staff involved in screening most frequently noted that the AI system was easy to use, and the real-time diagnostic report was useful. Our study indicates that AI-assisted DR screening model is accurate and well-accepted by patients and clinicians in endocrinology and indigenous healthcare settings. Future deployments of AI-assisted screening models would require consideration of downstream referral pathways.
ObjectivesTo investigate the association between glaucoma and 10-year mortality rate in an adult population in China.DesignPopulation-based cohort study.SettingThe Liwan Eye Study, ...China.Participants1405 baseline participants aged 50 years and older were invited to attend a 10-year follow-up examination.Primary and secondary outcome measuresThe International Society of Geographic and Epidemiologic Ophthalmology criteria was used to define glaucoma. Detailed information of mortality was confirmed using the Chinese Centre for Disease Control and Prevention. Presenting visual impairment (PVI) was defined as a presenting visual acuity of less than 20/40 in the better-seeing eye. The 10-year mortality rates were compared using the log-rank test. Cox proportional hazards regression models were used to investigate the association between glaucoma and mortality.ResultsA total of 1372 (97.7%) participants with available gonioscopic data were included in the analysis. Of these, 136 (9.9%), 33 (2.4%) and 21 (1.5%) participants had primary angle closure (PAC) suspect (PACS), PAC and PAC glaucoma (PACG), and 29 (2.1%) had primary open angle glaucoma (POAG). After 10 years, 306 (22.3%) participants were deceased. The 10-year mortality was significantly associated with PACG (HR, 2.15, 95% CI 1.14 to 4.04, p=0.018) but not associated with PAC (HR, 1.27, 95% CI 0.67 to 2.39, p=0.463), PACS (HR, 1.32, 95% CI 0.95 to 1.83, p=0.099) and POAG (HR, 0.74, 95% CI 0.36 to 1.49, p=0.395) when age and gender were adjusted for. This association was no longer statistically significant (HR, 1.60, 95% CI 0.70 to 3.61, p=0.263) when covariables, such as income, education, body mass index, PVI, history of diabetes and hypertension, were adjusted for. Larger vertical cup-to-disc ratio (VCDR >0.30) was only a significant risk factor in multivariable analysis (HR, 1.60, 95% CI 1.11 to 2.33, p=0.011).ConclusionsPACG was significantly associated with higher long-term mortality, but this association was likely to be confounded by other systemic risk factors. VCDR >0.3 was the only independent predictor, implying that it may be a marker of ageing and frailty.
We investigate the impact of parental myopia on spherical equivalent (SE) progression and axial length (AL) elongation.
Children and their parents were invited for annual examinations from 2006 ...(baseline). Cycloplegic autorefraction and AL were measured at each visit. Parental refractive status was determined using refraction data from their baseline visit. Children were classified into five groups: no myopic parents (non-non), only one moderately myopic parent (non-moderate), only one highly myopic parent (non-high), two moderately myopic parents (moderate-moderate), and one moderately myopic or more severe and one highly myopic parent (moderate-high/high-high). The relationship between progression of SE and AL with parental refractive status was estimated by linear mixed-effects models. Data from 2006 to 2017 were analyzed in the current study.
A total of 1831 children were enrolled (mean age, 11 ± 2.7 years; mean standard error, -0.49 ± 2.16 diopters D at baseline. Myopia progressed faster for children with parental myopia (non-non group as reference, all P < 0.05), while AL elongation mirrored the change in SE (all P < 0.001 except for non-mod group P = 0.12). As for the age-specific change in SE and AL, children in the mod-high/high-high group presented with the fastest progression. Children with highly myopic parents were at higher risks of being highly myopic during adulthood (odds ratio = 13.98 and 25.71 for non-high and mod-high/high-high groups; both P < 0.001).
SE progresses and AL elongates at a faster rate at an earlier age in children with parental myopia. Children with highly-myopic parents have higher risks of being highly myopic during adulthood.
To investigate the prevalence and associations of visual impairment (VI) and major eye diseases with chronic kidney disease (CKD) in the United States.
Cross-sectional study.
We investigated the ...prevalence and associations of VI and major eye diseases with CKD among 5,518 participants aged 40 years or older in the 2005-2008 National Health and Nutrition Examination Survey. An estimated glomerular filtration rate of lower than 60 mL/min/1.73 m2 was defined as CKD. Corrected visual acuity of worse than 20/40 in the better-seeing eye was defined as VI. Major eye diseases, including any ocular disease, any objectively determined ocular disease, cataract surgery, any retinopathy, diabetic retinopathy (DR), age-related macular degeneration (AMD), and glaucoma were evaluated from questionnaire or retinal photographs using standardized grading protocols.
The prevalence of VI and major eye diseases were approximately 2- to 7-fold higher in participants with CKD than in those without (all P < .05). After controlling for multiple confounders, the presence of CKD was associated with VI (odds ratio OR: 2.01, 95% confidence interval CI: 1.14-3.54), any ocular disease (OR: 1.65, 95% CI: 1.22-2.22), any objectively determined ocular disease (OR: 1.52, 95% CI: 1.06-2.19), any retinopathy (OR: 1.70, 95% CI: 1.18-2.45), and DR (OR: 2.34, 95% CI: 1.23-4.42). There was no association of CKD with cataract surgery, AMD, or glaucoma. A significant association between CKD and any ocular disease was observed among nondiabetic participants. The presence of CKD was closely related to VI and any retinopathy among diabetic participants.
This nationally representative sample of the US population demonstrated high prevalence and strong associations of VI and major eye diseases with CKD, highlighting the importance of ocular screening among CKD patients and potential common pathogenesis underlying these conditions.