Identifying the severity of diabetic retinopathy by visual function measures using both traditional statistical methods and interpretable machine learning: a cross-sectional study.

AIMS/HYPOTHESIS: To determine the extent to which diabetic retinopathy severity stage may be classified using machine learning (ML) and commonly used clinical measures of visual function together with age and sex. METHODS: We measured the visual function of 1901 eyes from 1032 participants in the Northern Ireland Sensory Ageing Study, deriving 12 variables from nine visual function tests. Missing values were imputed using chained equations. Participants were divided into four groups using clinical measures and grading of ophthalmic images: no diabetes mellitus (no DM), diabetes but no diabetic retinopathy (DM no DR), diabetic retinopathy without diabetic macular oedema (DR no DMO) and diabetic retinopathy with DMO (DR with DMO). Ensemble ML models were fitted to classify group membership for three tasks, distinguishing (A) the DM no DR group from the no DM group; (B) the DR no DMO group from the DM no DR group; and (C) the DR with DMO group from the DR no DMO group. More conventional multiple logistic regression models were also fitted for comparison. An interpretable ML technique was used to rank the contribution of visual function variables to predictions and to disentangle associations between diabetic eye disease and visual function from artefacts of the data collection process. RESULTS: The performance of the ensemble ML models was good across all three classification tasks, with accuracies of 0.92, 1.00 and 0.84, respectively, for tasks A-C, substantially exceeding the accuracies for logistic regression (0.84, 0.61 and 0.80, respectively). Reading index was highly ranked for tasks A and B, whereas near visual acuity and Moorfields chart acuity were important for task C. Microperimetry variables ranked highly for all three tasks, but this was partly due to a data artefact (a large proportion of missing values). CONCLUSIONS/INTERPRETATION: Ensemble ML models predicted status of diabetic eye disease with high accuracy using just age, sex and measures of visual function. Interpretable ML methods enabled us to identify profiles of visual function associated with different stages of diabetic eye disease, and to disentangle associations from artefacts of the data collection process. Together, these two techniques have great potential for developing prediction models using untidy real-world clinical data.

Retinal Ischaemia in Diabetic Retinopathy: Understanding and Overcoming a Therapeutic Challenge.

BACKGROUND: Retinal ischaemia is present to a greater or lesser extent in all eyes with diabetic retinopathy (DR). Nonetheless, our understanding of its pathogenic mechanisms, risk factors, as well as other characteristics of retinal ischaemia in DR is very limited. To date, there is no treatment to revascularise ischaemic retina. METHODS: Review of the literature highlighting the current knowledge on the topic of retinal ischaemia in DR, important observations made, and underlying gaps for which research is needed. RESULTS: A very scarce number of clinical studies, mostly cross-sectional, have evaluated specifically retinal ischaemia in DR. Interindividual variability on its natural course and consequences, including the development of its major complications, namely diabetic macular ischaemia and proliferative diabetic retinopathy, have not been investigated. The in situ, surrounding, and distance effect of retinal ischaemia on retinal function and structure and its change over time remains also to be elucidated. Treatments to prevent the development of retinal ischaemia and, importantly, to achieve retinal reperfusion once capillary drop out has ensued, are very much needed and remain to be developed. CONCLUSION: Research into retinal ischaemia in diabetes should be a priority to save sight.

Prognostic factors for the development and progression of proliferative diabetic retinopathy in people with diabetic retinopathy.

BACKGROUND: Diabetic retinopathy (DR) is characterised by neurovascular degeneration as a result of chronic hyperglycaemia. Proliferative diabetic retinopathy (PDR) is the most serious complication of DR and can lead to total (central and peripheral) visual loss. PDR is characterised by the presence of abnormal new blood vessels, so-called "new vessels," at the optic disc (NVD) or elsewhere in the retina (NVE). PDR can progress to high-risk characteristics (HRC) PDR (HRC-PDR), which is defined by the presence of NVD more than one-fourth to one-third disc area in size plus vitreous haemorrhage or pre-retinal haemorrhage, or vitreous haemorrhage or pre-retinal haemorrhage obscuring more than one disc area. In severe cases, fibrovascular membranes grow over the retinal surface and tractional retinal detachment with sight loss can occur, despite treatment. Although most, if not all, individuals with diabetes will develop DR if they live long enough, only some progress to the sight-threatening PDR stage.  OBJECTIVES: To determine risk factors for the development of PDR and HRC-PDR in people with diabetes and DR. SEARCH METHODS: We searched the Cochrane Central Register of Controlled Trials (CENTRAL; which contains the Cochrane Eyes and Vision Trials Register; 2022, Issue 5), Ovid MEDLINE, and Ovid Embase. The date of the search was 27 May 2022. Additionally, the search was supplemented by screening reference lists of eligible articles. There were no restrictions to language or year of publication.  SELECTION CRITERIA: We included prospective or retrospective cohort studies and case-control longitudinal studies evaluating prognostic factors for the development and progression of PDR, in people who have not had previous treatment for DR. The target population consisted of adults (≥18 years of age) of any gender, sexual orientation, ethnicity, socioeconomic status, and geographical location, with non-proliferative diabetic retinopathy (NPDR) or PDR with less than HRC-PDR, diagnosed as per standard clinical practice. Two review authors independently screened titles and abstracts, and full-text articles, to determine eligibility; discrepancies were resolved through discussion. We considered prognostic factors measured at baseline and any other time points during the study and in any clinical setting. Outcomes were evaluated at three and eight years (± two years) or lifelong.  DATA COLLECTION AND ANALYSIS: Two review authors independently extracted data from included studies using a data extraction form that we developed and piloted prior to the data collection stage. We resolved any discrepancies through discussion. We used the Quality in Prognosis Studies (QUIPS) tool to assess risk of bias. We conducted meta-analyses in clinically relevant groups using a random-effects approach. We reported hazard ratios (HR), odds ratios (OR), and risk ratios (RR) separately for each available prognostic factor and outcome, stratified by different time points. Where possible, we meta-analysed adjusted prognostic factors. We evaluated the certainty of the evidence with an adapted version of the GRADE framework.   MAIN RESULTS: We screened 6391 records. From these, we identified 59 studies (87 articles) as eligible for inclusion. Thirty-five were prospective cohort studies, 22 were retrospective studies, 18 of which were cohort and six were based on data from electronic registers, and two were retrospective case-control studies. Twenty-three studies evaluated participants with type 1 diabetes (T1D), 19 with type 2 diabetes (T2D), and 17 included mixed populations (T1D and T2D). Studies on T1D included between 39 and 3250 participants at baseline, followed up for one to 45 years. Studies on T2D included between 100 and 71,817 participants at baseline, followed up for one to 20 years. The studies on mixed populations of T1D and T2D ranged from 76 to 32,553 participants at baseline, followed up for four to 25 years.  We found evidence indicating that higher glycated haemoglobin (haemoglobin A1c (HbA1c)) levels (adjusted OR ranged from 1.11 (95% confidence interval (CI) 0.93 to 1.32) to 2.10 (95% CI 1.64 to 2.69) and more advanced stages of retinopathy (adjusted OR ranged from 1.38 (95% CI 1.29 to 1.48) to 12.40 (95% CI 5.31 to 28.98) are independent risk factors for the development of PDR in people with T1D and T2D. We rated the evidence for these factors as of moderate certainty because of moderate to high risk of bias in the studies.  There was also some evidence suggesting several markers for renal disease (for example, nephropathy (adjusted OR ranged from 1.58 (95% CI not reported) to 2.68 (2.09 to 3.42), and creatinine (adjusted meta-analysis HR 1.61 (95% CI 0.77 to 3.36)), and, in people with T1D, age at diagnosis of diabetes (< 12 years of age) (standardised regression estimate 1.62, 95% CI 1.06 to 2.48), increased triglyceride levels (adjusted RR 1.55, 95% CI 1.06 to 1.95), and larger retinal venular diameters (RR 4.28, 95% CI 1.50 to 12.19) may increase the risk of progression to PDR. The certainty of evidence for these factors, however, was low to very low, due to risk of bias in the included studies, inconsistency (lack of studies preventing the grading of consistency or variable outcomes), and imprecision (wide CIs). There was no substantial and consistent evidence to support duration of diabetes, systolic or diastolic blood pressure, total cholesterol, low- (LDL) and high- (HDL) density lipoproteins, gender, ethnicity, body mass index (BMI), socioeconomic status, or tobacco and alcohol consumption as being associated with incidence of PDR. There was insufficient evidence to evaluate prognostic factors associated with progression of PDR to HRC-PDR.  AUTHORS' CONCLUSIONS: Increased HbA1c is likely to be associated with progression to PDR; therefore, maintaining adequate glucose control throughout life, irrespective of stage of DR severity, may help to prevent progression to PDR and risk of its sight-threatening complications. Renal impairment in people with T1D or T2D, as well as younger age at diagnosis of diabetes mellitus (DM), increased triglyceride levels, and increased retinal venular diameters in people with T1D may also be associated with increased risk of progression to PDR. Given that more advanced DR severity is associated with higher risk of progression to PDR, the earlier the disease is identified, and the above systemic risk factors are controlled, the greater the chance of reducing the risk of PDR and saving sight.

ANTECEDENTES: La retinopatía diabética (RD) se caracteriza por la degeneración neurovascular como consecuencia de la hiperglucemia crónica. La retinopatía diabética proliferativa (RDP) es la complicación más grave de la RD y puede provocar una pérdida total (central y periférica) de la visión. La RDP se caracteriza por la presencia de vasos sanguíneos de neoformación anormales, neovascularización, en la papila óptica (NVP) o en cualquier otra parte de la retina (NVE). La RDP puede evolucionar a una RDP con características de alto riesgo (RDP­CAR), que se define por la presencia de NVP de más de un cuarto a un tercio del área discal más hemorragia vítrea o prerretiniana, o hemorragia vítrea o prerretiniana que oscurece más de un área papilar. En los casos graves, crecen membranas fibrovasculares sobre la superficie retiniana y se puede producir un desprendimiento de retina por tracción con pérdida de la visión, a pesar del tratamiento. Aunque la mayoría de las personas con diabetes, si no todas, desarrollarán RD si viven lo suficiente, solo algunas llegan a la fase de RDP, que pone en peligro la vista. OBJETIVOS: Determinar los factores de riesgo de aparición de la RDP y RDP­CAR en personas con diabetes y RD. MÉTODOS DE BÚSQUEDA: Se hicieron búsquedas en el Registro Cochrane central de ensayos controlados (Cochrane Central Register of Controlled Trials, CENTRAL; que contiene el Registro de ensayos del Grupo Cochrane de Salud ocular y de la visión [Cochrane Eyes and Vision]; 2022, número 5), Ovid MEDLINE y Ovid Embase. La fecha de búsqueda fue el 27 de mayo de 2022. Además, la búsqueda se complementó con el cribado de las listas de referencias de los artículos elegibles. No hubo restricciones en cuanto al idioma ni al año de publicación. CRITERIOS DE SELECCIÓN: Se incluyeron estudios de cohortes prospectivos o retrospectivos y estudios longitudinales de casos y controles que evaluaran los factores pronósticos para la aparición y la progresión de la RDP, en personas que no habían recibido tratamiento previo para la RD. La población de interés estaba formada por adultos (≥18 años de edad) de cualquier sexo, orientación sexual, etnia, nivel socioeconómico y ubicación geográfica, con retinopatía diabética no proliferativa (RDNP) o RDP sin llegar a RDP­CAR, diagnosticada según la práctica clínica habitual. Dos autores de la revisión examinaron de forma independiente los títulos y resúmenes, así como los artículos completos, para determinar la elegibilidad; las discrepancias se resolvieron mediante debate. Se tuvieron en cuenta los factores pronósticos medidos al inicio del estudio y en cualquier otro punto temporal durante el estudio y en cualquier contexto clínico. Los desenlaces se evaluaron a los tres y ocho años (± dos años) o de por vida. OBTENCIÓN Y ANÁLISIS DE LOS DATOS: Dos autores de la revisión extrajeron de forma independiente los datos de los estudios incluidos mediante un formulario de extracción de datos que se desarrolló y evaluó antes de la etapa de obtención de datos. Las discrepancias se resolvieron mediante debate. Para evaluar el riesgo de sesgo se utilizó la herramienta Quality in Prognosis Studies (QUIPS). Se realizaron metanálisis en grupos clínicamente relevantes utilizando un enfoque de efectos aleatorios. Se proporcionaron los cociente de riesgos instantáneos (CRI), los odds ratios (OR) y las razones de riesgos (RR) por separado para cada factor pronóstico y desenlace disponibles, estratificados por diferentes puntos temporales. Cuando fue posible, se realizó un metanálisis de los factores pronósticos ajustados. La certeza de la evidencia se evaluó con una versión adaptada del método GRADE. RESULTADOS PRINCIPALES: Se han examinado 6391 registros. A partir de estos se identificaron 59 estudios (87 artículos) elegibles para inclusión. Treinta y cinco fueron estudios de cohortes prospectivos, 22 fueron estudios retrospectivos, 18 de los cuales fueron de cohortes y 6 se basaron en datos de registros electrónicos, y 2 fueron estudios retrospectivos de casos y controles. Veintitrés estudios evaluaron a participantes con diabetes tipo 1 (DT1), 19 con diabetes tipo 2 (DT2) y 17 incluyeron poblaciones mixtas (DT1 y DT2). Los estudios sobre la DT1 incluyeron entre 39 y 3250 participantes al inicio del estudio, con un seguimiento de 1 a 45 años. Los estudios sobre la DT2 incluyeron entre 100 y 71 817 participantes al inicio del estudio, con un seguimiento de 1 a 20 años. Los estudios sobre poblaciones mixtas de DT1 y DT2 variaron entre 76 y 32 553 participantes al inicio del estudio, con un seguimiento de 4 a 25 años. Se encontró evidencia que indicó que los niveles más altos de hemoglobina glucosilada (hemoglobina A1c [HbA1c]) (OR ajustado que varió de 1,11 [intervalo de confianza (IC) del 95%: 0,93 a 1,32] a 2,10 [IC del 95%: 1,64 a 2,69]) y los estadios más avanzados de retinopatía (OR ajustado que varió entre 1,38 [IC del 95%: 1,29 a 1,48] y 12,40 [IC del 95%: 5,31 a 28,98]) son factores de riesgo independientes para el desarrollo de RDP en personas con DT1 y DT2. La evidencia para estos factores se consideró de certeza moderada debido al riesgo moderado a alto de sesgo en los estudios. También hubo alguna evidencia que indicó varios marcadores de enfermedad renal (por ejemplo, nefropatía [OR ajustado que varió entre 1,58 (IC del 95% no proporcionado) y 2,68 (2,09 a 3,42)] y creatinina [metanálisis ajustado CRI 1,61 (IC del 95%: 0,77 a 3.36)]), y, en las personas con DT1, la edad en el momento del diagnóstico de la diabetes (< 12 años) (estimación de la regresión estandarizada 1,62; IC del 95%: 1,06 a 2,48), el aumento de los niveles de triglicéridos (RR ajustado 1,55; IC del 95%: 1,06 a 1,95) y los diámetros venulares retinianos mayores (RR 4,28; IC del 95%: 1,50 a 12,19) podrían aumentar el riesgo de progresión a RDP. Sin embargo, la certeza de la evidencia para estos factores fue de baja a muy baja, debido al riesgo de sesgo en los estudios incluidos, la inconsistencia (falta de estudios que impide la calificación de consistencia o desenlaces variables) y la imprecisión (IC amplios). No hubo evidencia importante ni consistente que apoyara que la duración de la diabetes, la presión arterial sistólica o diastólica, el colesterol total, las lipoproteínas de baja (LDL) y alta (HDL) densidad, el sexo, el origen étnico, el índice de masa corporal (IMC), el nivel socioeconómico o el consumo de tabaco y alcohol estuvieran asociados con la incidencia de RDP. No hubo evidencia suficiente para evaluar los factores pronósticos asociados con la progresión de la RDP a RDP­CAR. CONCLUSIONES DE LOS AUTORES: Es probable que el aumento de la HbA1c se asocie con la progresión a la RDP; por lo tanto, mantener un control adecuado de la glucosa durante toda la vida, independientemente del estadio de gravedad de la RD, podría ayudar a prevenir la progresión a la RDP y el riesgo de sus complicaciones que ponen en peligro la vista. La insuficiencia renal en personas con DT1 o DT2, así como una menor edad en el momento del diagnóstico de la diabetes mellitus (DM), el aumento de los niveles de triglicéridos y el aumento de los diámetros venulares retinianos en personas con DT1 también se podrían asociar con un mayor riesgo de progresión a RDP. Dado que la gravedad más avanzada de la RD se asocia con un mayor riesgo de progresión a RDP, cuanto antes se identifique la enfermedad y se controlen los factores de riesgo sistémicos mencionados, mayores serán las posibilidades de reducir el riesgo de RDP y conservar la vista.

PREDICTIVE FACTORS ASSOCIATED WITH ANATOMICAL AND FUNCTIONAL OUTCOMES AFTER PANRETINAL PHOTOCOAGULATION IN PEOPLE WITH PROLIFERATIVE DIABETIC RETINOPATHY.

PURPOSE: To determine effects of baseline characteristics and laser type performed on outcomes in people with proliferative diabetic retinopathy (PDR) undergoing panretinal photocoagulation (PRP). METHODS: Medical records of all consecutive patients with PDR naïve to PRP, identified using an electronic database, evaluated at the Macula Clinic, Belfast Health and Social Care Trust, receiving their first PRP between January 1, 2016, and June 30, 2017, and followed for a minimum of 6 months after stabilization of PDR, were retrospectively reviewed. Outcomes included time to stabilization after PRP, progression of PDR, and mean change in the best-corrected visual acuity from baseline to the last follow-up. Cox regression was used to estimate hazard ratios (HRs) for the effect of baseline characteristics and type of laser on outcomes after treatment. RESULTS: One hundred and fourteen patients (135 eyes) with a mean age of 56.9 (SD: 12.9) years were included, 67% males. People receiving pattern or mixed laser had a statistically significantly delayed stabilization (HR: 0.57, P = 0.020; and HR: 0.44, P ≤ 0.001, respectively) and increased risk of progression (HR: 1.87, P = 0.045; and HR: 1.96, P = 0.028, respectively) when compared with those receiving standard laser. Among other potential predictors in multivariable regression analysis, only vitreous hemorrhage and fibrosis or traction at baseline increased the risk of progression (HR: 1.58, P = 0.047; and HR: 4.29, P ≤ 0.001, respectively). Baseline characteristics and type of laser had no statistically significant effect on vision. CONCLUSION: These findings should be considered when selecting laser treatment, planning surveillance, and counselling patients with PDR undergoing PRP.

Anatomic-Functional Correlates in Lesions of Retinal Vein Occlusion.

Purpose: To evaluate anatomic-functional associations at sites of retinal lesions in retinal vein occlusion (RVO). Methods: This pilot, prospective, observational study was conducted at the Northern Ireland Clinical Research Facility (NICRF) of Queen's University and the Belfast Health and Social Care Trust, Northern Ireland, between August 1, 2018, and September 30, 2019. The study included 10 treatment-naïve patients with RVO (10 RVO eyes and 10 fellow eyes). There were 81 points/sites assessed for each eye at baseline; six patients were re-assessed 6 months after anti-vascular endothelial growth factor therapy at the same locations. We investigated associations between retinal sensitivity and presence of structural RVO lesions, including retinal ischemia, hemorrhages, intraretinal fluid (IRF) and subretinal fluid outside the foveal/parafoveal regions. Comparisons were made between RVO eyes and fellow eyes at baseline, and between RVO eyes at baseline and at 6 months after treatment. Regression models were used to investigate anatomic-functional associations. Results: At baseline, strong associations were found between reduced retinal sensitivity and presence of ischemia (estimate = -2.08 dB; P < 0.001), intraretinal fluid (estimate = -7.82 dB; P < 0.001), and subretinal fluid (estimate = -8.66 dB; P < 0.001). Resolution of subretinal fluid but not intraretinal fluid was associated with improved function (estimate = 2.40 dB [P = 0.022]; estimate = 1.16 dB [P = 0.228], respectively). However, reperfusion of ischemic retina, observed in 31 of 486 points (6%) 6 months after anti-vascular endothelial growth factor therapy, was associated with a further decrease in retinal sensitivity (estimate = -2.34 dB; P = 0.035). Conclusions: Retinal sensitivity was decreased at sites of RVO lesions. Decreased function at sites of retinal ischemia did not recover after treatment, even when reperfusion occurred.

Quantitative Parameters from OCT Angiography in Patients with Diabetic Retinopathy and in Those with Only Peripheral Retinopathy Compared with Control Participants.

Purpose: To describe the differences in a range of quantitative OCT angiography (OCTA) metrics across early stages of diabetic retinopathy (DR), providing robust effect estimates as well as sensitivity and specificity. Design: Cross-sectional study with population-based sampling. Participants: Four hundred forty-one eyes from 296 individuals: 328 control eyes (no diabetes mellitus [DM] and no DR), 55 eyes with DM and no DR, and 58 eyes with early nonproliferative DR. Methods: Multimodal retinal imaging included color fundus photography, color Optomap ultra-widefield imaging, and spectral-domain OCT (Spectralis OCT2; Heidelberg Engineering GmbH) with the OCTA module. All images were graded for the presence and severity of DR features. OCTA images were assessed manually for inclusion based on quality. Binary OCTA metrics were assessed after 3-dimensional projection artifact removal including from the nerve fiber layer vascular plexus, superficial vascular plexus (SVC), and deep vascular plexus (DVC) by Early Treatment Diabetic Retinopathy Study (ETDRS) grid, foveal avascular zone (FAZ) area, FAZ minimum and maximum diameter, perimeter length, and circularity. Main Outcome Measures: Diabetes mellitus and DR status and presence or absence of DR in the retinal periphery. Results: The reduction in vessel densities in participants with DM and manifest DR compared with control participants tended to be twice that of those with DM, but no DR, compared with control participants. Some evidence of spatial heterogeneity in vessel reductions was found in those yet to develop DR, whereas those with manifest DR had significant reductions across the ETDRS grid. The FAZ perimeter and circularity were impacted most significantly by DM, and those with DR showed decreased multispectral fractal dimensions compared with control participants. Eyes with peripheral DR had reduced vessel density compared with those with DM and no DR only in the superior outer, temporal inner, and temporal outer regions in the DVC and SVC. The area under the receiver operating characteristic curve ranged between 0.48 and 0.73. Conclusions: Significant differences in OCTA metrics can be found in those with DM before manifest DR using commercially available equipment with minimal image postprocessing. Although diagnostic performance was poor, these metrics may be useful for measuring change over time in clinical trials.