Development and international validation of custom-engineered and code-free deep-learning models for detection of plus disease in retinopathy of prematurity: a retrospective study.

BACKGROUND: Retinopathy of prematurity (ROP), a leading cause of childhood blindness, is diagnosed through interval screening by paediatric ophthalmologists. However, improved survival of premature neonates coupled with a scarcity of available experts has raised concerns about the sustainability of this approach. We aimed to develop bespoke and code-free deep learning-based classifiers for plus disease, a hallmark of ROP, in an ethnically diverse population in London, UK, and externally validate them in ethnically, geographically, and socioeconomically diverse populations in four countries and three continents. Code-free deep learning is not reliant on the availability of expertly trained data scientists, thus being of particular potential benefit for low resource health-care settings. METHODS: This retrospective cohort study used retinal images from 1370 neonates admitted to a neonatal unit at Homerton University Hospital NHS Foundation Trust, London, UK, between 2008 and 2018. Images were acquired using a Retcam Version 2 device (Natus Medical, Pleasanton, CA, USA) on all babies who were either born at less than 32 weeks gestational age or had a birthweight of less than 1501 g. Each images was graded by two junior ophthalmologists with disagreements adjudicated by a senior paediatric ophthalmologist. Bespoke and code-free deep learning models (CFDL) were developed for the discrimination of healthy, pre-plus disease, and plus disease. Performance was assessed internally on 200 images with the majority vote of three senior paediatric ophthalmologists as the reference standard. External validation was on 338 retinal images from four separate datasets from the USA, Brazil, and Egypt with images derived from Retcam and the 3nethra neo device (Forus Health, Bengaluru, India). FINDINGS: Of the 7414 retinal images in the original dataset, 6141 images were used in the final development dataset. For the discrimination of healthy versus pre-plus or plus disease, the bespoke model had an area under the curve (AUC) of 0·986 (95% CI 0·973-0·996) and the CFDL model had an AUC of 0·989 (0·979-0·997) on the internal test set. Both models generalised well to external validation test sets acquired using the Retcam for discriminating healthy from pre-plus or plus disease (bespoke range was 0·975-1·000 and CFDL range was 0·969-0·995). The CFDL model was inferior to the bespoke model on discriminating pre-plus disease from healthy or plus disease in the USA dataset (CFDL 0·808 [95% CI 0·671-0·909, bespoke 0·942 [0·892-0·982]], p=0·0070). Performance also reduced when tested on the 3nethra neo imaging device (CFDL 0·865 [0·742-0·965] and bespoke 0·891 [0·783-0·977]). INTERPRETATION: Both bespoke and CFDL models conferred similar performance to senior paediatric ophthalmologists for discriminating healthy retinal images from ones with features of pre-plus or plus disease; however, CFDL models might generalise less well when considering minority classes. Care should be taken when testing on data acquired using alternative imaging devices from that used for the development dataset. Our study justifies further validation of plus disease classifiers in ROP screening and supports a potential role for code-free approaches to help prevent blindness in vulnerable neonates. FUNDING: National Institute for Health Research Biomedical Research Centre based at Moorfields Eye Hospital NHS Foundation Trust and the University College London Institute of Ophthalmology. TRANSLATIONS: For the Portuguese and Arabic translations of the abstract see Supplementary Materials section.

The incidence and prevalence of recurrent corneal erosion syndrome in London, UK.

BACKGROUND: Recurrent corneal erosion syndrome (RCES) is caused by repeated episodes of corneal epithelial breakdown due to improper adherence of the corneal epithelium to the underlying basement membrane. The most common aetiologies are corneal dystrophy or previous superficial ocular trauma. The incidence and prevalence of the condition is currently unknown. This study aimed to determine the incidence and prevalence of RCES within the London population over a 5-year period in order to better inform clinicians and evaluate how this condition affects ophthalmic service provision. METHODS: A retrospective cohort study over a 5-year period reviewed 487,690 emergency room patient attendances at Moorfields Eye Hospital (MEH) London between 1 January 2015 and 31 December 2019. MEH caters for a local population comprising of around ten regional clinical commissioning groups (CCGs). The data for this study were collected using OpenEyesTM electronic medical records including demographics and comorbidities. The CCGs encompass 41% (3,689,000) of London's total 8,980,000 inhabitants. Using these data the crude incidence and prevalence rates of disease were estimated with results reported per 100,000 population. RESULTS: Out of 330,684 patients, 3623 patients were given a new diagnosis of RCES by the emergency ophthalmology services, and from these, 1056 patients attended outpatient follow-up. The crude annual incidence of RCES was estimated at 25.4 per 100,000, with a crude prevalence rate of 0.96%. There was no statistical difference in annual incidence across the 5-year period. CONCLUSIONS: The period prevalence of 0.96% shows that RCES is not uncommon. There was also a stable annual incidence over the 5-year period, showing no changing trend over the study period. However, identifying the true incidence and period prevalence is a challenging task, as minor cases may heal prior to examination by an ophthalmologist. It is highly likely that RCES is underdiagnosed and therefore underreported.

Healing of Chemical Injury-Related Persistent Corneal Epithelial Defects With Topical Insulin.

PURPOSE: The purpose of this study was to evaluate the effectiveness of insulin eye drops for treating refractory persistent epithelial defects (PEDs). METHODS: A prospective, single-center, case series was performed from March 2020 to September 2021. All patients were prescribed insulin eye drops for refractory PEDs that failed on maximum standard medical treatment (including serum eye drops). The drops were used 4 times/day. Patients were followed up at 2 weekly intervals with full slitlamp examination and serial anterior segment photography. The primary end point was resolution of the epithelial defect. RESULTS: Eleven eyes of 10 patients were treated with insulin eye drops. The mean age of the cohort was 45.4 + 25 years with a mean follow-up of 195.7 + 114.3 days after re-epithelization. The most common causative condition was chemical injury (n = 5, 60%). Mean baseline PED defect was 41.3 + 55.2 mm 2 . Nine of 11 eyes (82%) fully re-epithelized within a mean time of 62.3 + 34.6 days (range 14-112). In 2 patients who did not achieve re-epithelization, one had a reduction in size from 12.25 mm 2 to 4.5 mm 2 and the other had no response. No recurrence in defect was observed in the group that had fully re-epithelized. CONCLUSIONS: This study showed that the use of topical insulin eye drops led to a successful resolution of PED in 9 of 11 cases. We demonstrate the use of insulin for closure of PEDs in chemical eye injury. Larger controlled studies are required to further evaluate this novel therapy.

Finger-Prick Autologous Blood (FAB) Eye Drops for Dry Eye Disease: Single Masked Multi-Centre Randomised Controlled Trial.

Purpose: To investigate the quantitative and qualitative efficacy of finger-prick autologous blood (FAB) eye drops versus conventional medical therapy for the treatment of severe dry eye disease (DED). Methods: Two centre, single masked, randomised controlled trial. Sixty patients in total were recruited with thirty patients (sixty eyes) treated with FAB eye drops four times per day in addition to their conventional DED treatment, and thirty patients (fifty-eight eyes) served as control subjects on conventional treatment alone. Ocular surface disease index (OSDI), Schirmer's test, fluorescein ocular staining grade (OCSG) Oxford schema and fluorescein tear film break-up time (TBUT), were performed at baseline, at 4 and 8 weeks. Results: OSDI scores significantly decreased in the FAB arm by greater than -17.68 (-37.67 to -2.96, p=0.02) compared to the control arm. There were greater improvements in OCSG and TBUT in the FAB arm but these were non-significant (p>0.05). Conclusion: This feasibility study demonstrates adding FAB eye drops to conventional medical therapy for DED improves mean OSDI symptom score compared to conventional medical therapy alone. It may have particular use in settings where serum is unobtainable. An adequately powered and well-designed randomised trial is needed to further evaluate the long-term clinical benefit of FAB.

The management of neurotrophic keratitis.

PURPOSE OF REVIEW: The management of neurotrophic keratitis (NK) has evolved in the last decade. The present article reviews updated management guidelines of this entity, as well as future innovations in the field. RECENT FINDINGS: The advent of confocal microscopy has allowed for the first time to image corneal nerves. In addition, multiple novel topical treatments such as nerve growth factor have improved the prognosis of this disease, with many other in the pipeline. Finally, corneal nerve restoration is now possible with corneal neurotization procedures. SUMMARY: Many novel treatments based on agents that stimulate nerve regrowth are now available to treat NK. Improvement in neurotization procedures could also address advanced stages of this disease with surgery.

Capturing the Occult Central Retinal Artery Occlusion Using Optical Coherence Tomography.

Aims: To report spectral-domain optical coherence tomography (OCT) findings in cases of impending or occult central retinal artery occlusion (CRAO) in which a diagnosis other than CRAO was made on initial presentation.Methods: Retrospective, observational case series of patients diagnosed with CRAO for whom on initial presentation fundal examination and OCT findings were deemed unremarkable and/or a diagnosis other than CRAO was made. OCT images from the initial presentation were then reviewed for evidence of inner retinal ischaemia.Results: In total, 214 cases of CRAO were identified. Eleven patients (5.14%) had been given an alternative initial diagnosis at their first presentation in casualty and were included. The age range was 20-84 years and 81% (9/11) were male. On review of initial OCT imaging performed in casualty, all cases had evidence of inner retinal ischaemia.Conclusions: CRAO is an ophthalmic emergency which leads to vision loss which is often irreversible. Examination of the fundus may be normal early in the course of the disease and therefore a timely diagnosis may be missed. This case series reports the OCT findings of inner retinal ischaemia in patients with occult or impending CRAO which may aid in the early diagnosis and referral to stroke services.

Management and outcomes of the small pupil in cataract surgery: iris hooks, Malyugin ring or phenylephrine?

PURPOSE: To investigate outcomes for small versus large pupils in cataract surgery using different pupil expansion techniques. METHODS: Retrospective case-series reviewing 20,175 patients' cataract surgery electronic medical records at Moorfields Eye Clinic in Bedford Hospital NHS Trust from January 2010 to April 2020. Outcomes such as visual acuity (VA), intraocular pressure, intraoperative, post-operative complications were recorded and small pupil expansion device outcome. RESULTS: One thousand, four hundred twenty-six patients were identified as having small pupil (SP). Of these, 1110 patients (77.8%) had interventions to expand the pupil including 447 (31.3%) with intracameral phenylephrine (IC PE) alone, 194 (13.6%) with iris hooks and 469 (32.9%) with a Malyugin ring. The large pupil (LP) group had a statistically significant greater gain in VA than the SP group (p < 0.05). SPs had a significantly higher rate of intraocular complications including posterior capsular rupture (PCR) with vitreous loss (OR 2.75, p < 0.001). There was also a significantly higher rate of post-operative complications such as corneal oedema (OR 2.64, p < 0.001) and anterior uveitis (OR 2.11, p < 0.001) in the SP group. However, VA improvement and complications between the different pupil expansion groups showed no significant differences (p > 0.05) except for a greater rate of iris tears in the Malyugin group (p < 0.05). CONCLUSION: To date, this is the largest reported case-series comparing Malyugin rings and iris hooks with other pupil expansion techniques. The various techniques to expand pupil size appear to be safe and equally effective in improving VA with a similar rate of complications except for a greater rate of iris tears with Malyugin ring.

Finger-Prick Autologous Blood in the Treatment of Persistent Corneal Epithelial Defects.

PURPOSE: Autologous hemoderivative eye drops have a role in the management of persistent epithelial defects (PEDs), but their use may be limited by cost and availability. Finger-prick autologous blood (FAB) treatment uses whole capillary blood, obtained from a sterilized fingertip, as an alternative form of hemoderivative eye drop therapy. To date, 1 report has described the safe and effective use of FAB for dry eye and PEDs. We report the results of 10 eyes (10 patients) treated with FAB for PEDs. METHODS: Ten patients with PEDs in 1 eye for a mean of 259 ± 201 days due to diabetic neurotrophic keratopathy (n = 3), herpetic keratitis (n = 3), postpenetrating keratoplasty (n = 1), keratoconjunctivitis sicca (n = 1), postradiotherapy (n = 1), and neuropathic ulcer (n = 1) were treated with FAB 4 times a day for 28 days in addition to conventional therapies. All patients had been unsuccessfully treated with conventional therapy before commencing on FAB. None of the patients had received any surgical treatment for PED. RESULTS: At day 28, the PED had healed in 60% (n = 6) of the eyes. In 1 eye, the PED reduced in size by half. Thirty percent (n = 3) of patients had incomplete follow-up data at the end of the study. CONCLUSIONS: FAB in combination with conventional treatment may be successfully used in the management of refractory PEDs. No adverse effects arising from FAB treatment were observed.

Computer analysis of individual cataract surgery segments in the operating room.

PURPOSE: Objective feedback is important for the continuous development of surgical skills. Motion tracking, which has previously been validated across an entire cataract procedure, can be a useful adjunct. We aimed to measure quantitative differences between junior and senior surgeons' performance in three distinct segments. We further explored whether automated analysis of trainee surgical videos through PhacoTracking could be aligned with metrics from the EyeSi virtual reality simulator, allowing focused improvement of these areas in a controlled environment. METHODS: Prospective cohort analysis, comparing junior vs. senior surgeons' real-life performance in distinct segments of cataract surgery: continuous curvilinear capsulorhexis (CCC), phacoemulsification, and irrigation and aspiration (I&A). EyeSi metrics that could be aligned with motion tracking parameters were identified. Motion tracking parameters (instrument path length, number of movements and total time) were measured. t-test used between the two cohorts for each component to check for any significance (p < 0.05). RESULTS: A total of 120 segments from videos of 20 junior and 20 senior surgeons were analysed. Significant differences between junior and senior surgeons were found during CCC (path length p = 0.0004; number of movements p < 0.0001; time taken p < 0.0001), phacoemulsification (path length p < 0.0001; number of movements p < 0.0001; time taken p < 0.0001), and irrigation and aspiration (path length p = 0.006; number of movements p = 0.013; time taken p = 0.036). CONCLUSION: Individual segments of cataract surgery analysed using motion tracking appear to discriminate between junior and senior surgeons. Alignment of motion tracking and EyeSi parameters could enable independent, task specific, objective and quantitative feedback for each segment of surgery thus mirroring the widely utilised modular training.

The feasibility of finger prick autologous blood (FAB) as a novel treatment for severe dry eye disease (DED): protocol for a randomised controlled trial.

INTRODUCTION: Patients with severe dry eye disease (DED) often have limited treatment options with standard non-surgical management focused on the use of artificial tears for lubrication and anti-inflammatory drugs. However, artificial tears do not address the extraordinary complexity of human tears. Crudely, human tears with its vast constituents is essentially filtered blood. Blood and several blood-derived products including autologous serum, have been studied as tear substitutes. This study proposes to test the use of whole, fresh, autologous blood obtained from a finger prick for treatment of severe DED. METHODS AND ANALYSIS: The research team at the two participating sites will approach patients with severe DED for this study. Recruitment will take place over 12 months and we expect to recruit 60 patients in total. The primary outcome of this feasibility study is to estimate the proportion of eligible patients approached who consent to and comply with study procedures including treatment regimen and completion of required questionnaires. The secondary outcome measures, although not powered for in this feasibility, include corneal inflammation (assessed by the Oxford corneal staining guide), patient pain and symptoms scores (assessed by the Ocular Surface Disease Index Score), and objective signs of DED as indicated by visual acuity (assessed by Schirmer's test, tear break-up time, lower and/or upper tear meniscus height measurement). Other secondary outcomes include patients' quality of life (assessed using the validated EQ-5D-5L Questionnaire), cost to the National Health Service (NHS) and patient (assessed via use of NHS services and privately purchased over-the-counter treatment related to DED) and safety measure of pressure within the eye (assessed by the Intraocular Pressure (IOP) Score). ETHICS AND DISSEMINATION: This protocol and any subsequent amendments, along with any accompanying material provided to the participant in addition to any advertising material used in this trial have been approved by the East of England - Cambridgeshire and Hertfordshire Research Ethics Committee (REC reference: 17/EE/0508). Written approval from the committee was obtained and subsequently submitted to the respective Trust's Research and Development (R&D) office with final NHS R&D approval obtained. Data obtained from this study will be published in a suitable peer-review journal and will also presented at international ophthalmic conferences including the American Academy of Ophthalmology, the Royal College of Ophthalmology Annual Congress, the Association for Research and Vision and Ophthalmology, and the European Society of Cataract and Refractive Surgery. Information will be provided to patient groups and charities such as the Sjogren's Society and the Royal National Institute of Blind People. This will also be shared with the study participants as well as with relevant patient groups and charities. TRIAL REGISTRATION NUMBER: NCT03395431; Pre-results.

Dexamethasone Intravitreal Implant Therapy for Retinal Vein Occlusion Macular Oedema and Conversion to Ranibizumab in Clinical Practice.

PURPOSE: Evaluation of outcomes in retinal vein occlusions (RVOs) for: (1) multiple repeat dexamethasone (DEX) injections and (2) conversion from DEX to ranibizumab. METHODS: We conducted a retrospective study evaluating outcomes of multiple DEX injections and those requiring conversion to ranibizumab at Moorfields Eye Hospital, Bedford, UK. All patients had undergone a complete ophthalmic work-up. RESULTS: Patients (n = 129) had a mean follow-up of 19.9 months. The mean improvement in central retinal thickness was 312 µm after final DEX (p = <0.0001). Mean peak best corrected visual acuity (BCVA) after final DEX was an improvement of 16 ETDRS letters (p < 0.0001). Forty-nine patients were converted and received a mean of 9.37 ranibizumab injections with a mean improvement in BCVA of 15 ETDRS letters (p < 0.0001) compared with final DEX. CONCLUSIONS: This study supports the use of ranibizumab in eyes previously treated with DEX and provides long-term efficacy and safety data for multiple DEX injection.