Use of an artificial intelligence-generated vascular severity score improved plus disease diagnosis in retinopathy of prematurity.

OBJECTIVE: To evaluate whether providing clinicians with an artificial intelligence-based vascular severity score (AI-VSS) improves consistency in diagnosis of plus disease in retinopathy of prematurity (ROP). DESIGN: This is a multi-reader diagnostic accuracy imaging study. PARTICIPANTS: Eleven ROP experts (4 pediatric ophthalmologists, 7 retina specialists), 9 of which had been in practice for 10 or more years. METHODS: Retcam (Natus Medical Incorporated) fundus images were obtained from premature infants during routine ROP screening as part of the Imaging and Informatics in ROP study between January 2012 and July 2020. From all available exams, a subset of 150 eye exams from 110 infants were selected for grading. An AI-VSS was assigned to each set of images using the i-ROP DL system. The clinicians were asked to diagnose plus disease for each exam and assign an estimated VSS (range 1-9) at baseline, and then again one month later with AI-VSS assistance. A reference standard diagnosis (RSD) was assigned to each eye exam from the i-ROP study based on 3 masked expert labels and the ophthalmoscopic diagnosis. MAIN OUTCOME MEASURE: Mean linearly weighted kappa for plus disease diagnosis compared to the RSD. Area under the receiver operating characteristic and precision-recall curves (AUROC, AUPR) for 1-9 labels compared to RSD for plus disease. RESULTS: Expert agreement improved significantly from substantial (κ: 0.69 [0.59, 0.75]) to near perfect (κ: 0.81 [0.71, 0.86]) when AI-VSS was integrated. Additionally, there was a significant improvement in plus disease discrimination as measured by mean [95% confidence interval] AUROC (0.94 [0.92, 0.96] to 0.98 [0.96, 0.99], difference: 0.04 [0.01, 0.06]) and AUPR (0.86 [0.81, 0.90] to 0.95 [0.91, 0.97], difference: 0.09 [0.03, 0.14]). CONCLUSIONS: Providing ROP clinicians with an AI-based measurement of vascular severity in ROP was associated with both improved plus disease diagnosis and improved continuous severity labeling, as compared to a reference standard diagnosis for plus disease. If implemented in practice, AI-VSS could reduce inter-observer variability and standardize treatment for infants with ROP.

Ocular and developmental outcomes of a dosing study of bevacizumab for retinopathy of prematurity.

PURPOSE: To report 2-year ocular and developmental outcomes for infants receiving low doses of intravitreal bevacizumab for type 1 retinopathy of prematurity (ROP). METHODS: A total of 120 premature infants (mean birthweight, 687 g; mean gestational age, 24.8 weeks) with type 1 ROP were enrolled in a multicenter, phase 1 dose de-escalation study. One eye per infant received 0.25 mg, 0.125 mg, 0.063 mg, 0.031 mg, 0.016 mg, 0.008 mg, 0.004 mg, or 0.002 mg of intravitreal bevacizumab; fellow eyes when treated received one dosage level higher. At 2 years, 70 of 120 children (58%) underwent ocular examinations; 51 (43%) were assessed using the Bayley Scale of Infant and Toddler Development. RESULTS: Correlation coefficients for the association of total dosage of bevacizumab with Bayley subscales were -0.20 for cognitive (95% CI, -0.45 to 0.08), -0.15 for motor (95% CI, -0.41 to 0.14), and -0.19 for language (95% CI, -0.44 to 0.10). Fourteen children (21%) had myopia greater than -5.00 D in one or both eyes, 7 (10%) had optic nerve atrophy and/or cupping, 20 (29%) had strabismus, 8 (11%) had manifest nystagmus, and 9 (13%) had amblyopia. CONCLUSIONS: In this study cohort, there was no statistically significant correlation between dosage of bevacizumab and Bayley scores at 2 years. However, the sample size was small and the retention rate relatively low, limiting our conclusions. Rates of high myopia and ocular abnormalities do not differ from those reported after larger bevacizumab doses.

Low- and Very Low-Dose Bevacizumab for Retinopathy of Prematurity: Reactivations, Additional Treatments, and 12-Month Outcomes.

PURPOSE: Low-dose and very low-dose intravitreal bevacizumab (IVB) have been reported to be successful in short-term treatment of type 1 retinopathy of prematurity (ROP), down to an initial dose of 0.004 mg. We now report 12-month outcomes for these infants. DESIGN: Masked, multicenter, dose de-escalation study. PARTICIPANTS: One hundred twenty prematurely born infants with type 1 ROP. METHODS: A cohort of 120 infants with type 1 ROP in at least 1 eye from 2 sequential dose de-escalation studies of low-dose IVB (0.25 mg, 0.125 mg, 0.063 mg, and 0.031 mg) or very low-dose IVB (0.016 mg, 0.008 mg, 0.004 mg, and 0.002 mg) to the study eye; the fellow eye (if also type 1) received 1 dose level higher of IVB. After primary success or failure at 4 weeks, clinical management was at investigator discretion, including all additional treatment. MAIN OUTCOME MEASURES: Reactivation of severe ROP by 6 months corrected age, additional treatments, retinal and other ocular structural outcomes, and refractive error at 12 months corrected age. RESULTS: Sixty-two of 113 study eyes (55%) and 55 of 98 fellow eyes (56%) received additional treatment. Of the study eyes, 31 (27%) received additional ROP treatment, and 31 (27%) received prophylactic laser therapy for persistent avascular retina. No trend toward a higher risk of additional ROP treatment related to initial IVB doses was found. However, time to reactivation among study eyes was shorter in eyes that received very low-dose IVB (mean, 76.4 days) than in those that received low-dose IVB (mean, 85.7 days). At 12 months, poor retinal outcomes and anterior segment abnormalities both were uncommon (3% and 5%, respectively), optic atrophy was noted in 10%, median refraction was mildly myopic (-0.31 diopter), and strabismus was present in 29% of infants. CONCLUSIONS: Retinal structural outcomes were very good after low- and very low-dose IVB as initial treatment for type 1 ROP, although many eyes received additional treatment. The rate of reactivation of severe ROP was not associated with dose; however, a post hoc data-driven analysis suggested that reactivation was sooner with very low doses.

Timely recognition of retinal hemorrhage in pediatric abusive head trauma evaluation.

BACKGROUND: Distinct patterns of retinal hemorrhages (RHs) are suggestive of abusive head trauma in the context of unexplained intracranial injury. Current recommendations encourage an eye examination within 48 hours of admission due to the rapid resolution of RH. The purpose of this study was to identify clinical factors associated with a delay in funduscopic examination outside the recommended 48 hours. METHODS: Retrospective chart review was completed on all inpatient consultations by the Child Protection Team with evidence of intracranial injury on computed tomography or magnetic resonance imaging over 3 years at a large children's hospital. Extracted data included demographic characteristics, history of intubation, pediatric intensive care unit (PICU) admission, extraventricular drain placement, seizures, use of vasopressor support, and presence of other injuries. Descriptive statistics were used to describe the patient population, clinical characteristics, and outcomes. Multivariate logistic regression was used to identify factors associated with delayed eye examinations. RESULTS: A total of 203 patients met inclusion criteria. Of those, 39 (19.2%) had a delay in initial funduscopic examination. Multivariate analyses revealed that PICU admission, surgical intervention, and seizure activity were significant predictors of delayed examination after controlling for multiple clinical factors. Neurosurgical consultation was shown to be protective against a delayed examination. CONCLUSIONS: Rapid resolution of RH may occur in child abuse. Prompt ophthalmology examinations and neurosurgery consultation when child abuse is suspected help avoid a delay in diagnosis.

Overminus Lens Therapy for Children 3 to 10 Years of Age With Intermittent Exotropia: A Randomized Clinical Trial.

IMPORTANCE: This is the first large-scale randomized clinical trial evaluating the effectiveness and safety of overminus spectacle therapy for treatment of intermittent exotropia (IXT). OBJECTIVE: To evaluate the effectiveness of overminus spectacles to improve distance IXT control. DESIGN, SETTING, AND PARTICIPANTS: This randomized clinical trial conducted at 56 clinical sites between January 2017 and January 2019 associated with the Pediatric Eye Disease Investigator Group enrolled 386 children aged 3 to 10 years with IXT, a mean distance control score of 2 or worse, and a refractive error between 1.00 and -6.00 diopters (D). Data analysis was performed from February to December 2020. INTERVENTIONS: Participants were randomly assigned to overminus spectacle therapy (-2.50 D for 12 months, then -1.25 D for 3 months, followed by nonoverminus spectacles for 3 months) or to nonoverminus spectacle use. MAIN OUTCOMES AND MEASURES: Primary and secondary outcomes were the mean distance IXT control scores of participants examined after 12 months of treatment (primary outcome) and at 18 months (3 months after treatment ended) assessed by an examiner masked to treatment group. Change in refractive error from baseline to 12 months was compared between groups. Analyses were performed using the intention-to-treat population. RESULTS: The mean (SD) age of 196 participants randomized to overminus therapy and 190 participants randomized to nonoverminus treatment was 6.3 (2.1) years, and 226 (59%) were female. Mean distance control at 12 months was better in participants treated with overminus spectacles than with nonoverminus spectacles (1.8 vs 2.8 points; adjusted difference, -0.8; 95% CI, -1.0 to -0.5; P < .001). At 18 months, there was little or no difference in mean distance control between overminus and nonoverminus groups (2.4 vs 2.7 points; adjusted difference, -0.2; 95% CI, -0.5 to 0.04; P = .09). Myopic shift from baseline to 12 months was greater in the overminus than the nonoverminus group (-0.42 D vs -0.04 D; adjusted difference, -0.37 D; 95% CI, -0.49 to -0.26 D; P < .001), with 33 of 189 children (17%) in the overminus group vs 2 of 169 (1%) in the nonoverminus group having a shift higher than 1.00 D. CONCLUSIONS AND RELEVANCE: Children 3 to 10 years of age had improved distance exotropia control when assessed wearing overminus spectacles after 12 months of overminus treatment; however, this treatment was associated with increased myopic shift. The beneficial effect of overminus lens therapy on distance exotropia control was not maintained after treatment was tapered off for 3 months and children were examined 3 months later. TRIAL REGISTRATION: ClinicalTrials.gov Identifier: NCT02807350.

Short-term Outcomes After Very Low-Dose Intravitreous Bevacizumab for Retinopathy of Prematurity.

Importance: Intravitreous bevacizumab (0.25 mg to 0.625 mg) is commonly used to treat type 1 retinopathy of prematurity (ROP), but there are concerns about systemic toxicity, particularly the risk of neurodevelopmental delay. A much lower dose may be effective for ROP while reducing systemic risk. Previously, after testing doses of 0.25 mg to 0.031 mg, doses as low as 0.031 mg were found to be effective in small cohorts of infants. Objective: To find the lowest dose of intravitreous bevacizumab effective for severe ROP. Design, Setting, and Participants: Between April 2017 and May 2019, 59 premature infants with type 1 ROP in 1 or both eyes were enrolled in a masked, multicenter, dose de-escalation study. In cohorts of 10 to 14 infants, 1 eye per infant received 0.016 mg, 0.008 mg, 0.004 mg, or 0.002 mg of intravitreous bevacizumab. Diluted bevacizumab was prepared by individual research pharmacies and delivered using 300-µL syringes with 5/16-inch, 30-guage fixed needles. Analysis began July 2019. Interventions: Bevacizumab intravitreous injections at 0.016 mg, 0.008 mg, 0.004 mg, or 0.002 mg. Main Outcomes and Measures: Success was defined as improvement by 4 days postinjection and no recurrence of type 1 ROP or severe neovascularization requiring additional treatment within 4 weeks. Results: Fifty-five of 59 enrolled infants had 4-week outcomes completed; the mean (SD) birth weight was 664 (258) g, and the mean (SD) gestational age was 24.8 (1.6) weeks. A successful 4-week outcome was achieved for 13 of 13 eyes (100%) receiving 0.016 mg, 9 of 9 eyes (100%) receiving 0.008 mg, 9 of 10 eyes (90%) receiving 0.004 mg, but only 17 of 23 eyes (74%) receiving 0.002 mg. Conclusions and Relevance: These data suggest that 0.004 mg may be the lowest dose of bevacizumab effective for ROP. Further investigation is warranted to confirm effectiveness of very low-dose intravitreous bevacizumab and its effect on plasma vascular endothelial growth factor levels and peripheral retinal vascularization.

Cortical Visual Impairment in Congenital Cytomegalovirus Infection.

PURPOSE: To describe the presentation, evolution, and long-term outcome of cortical visual impairment (CVI) in patients with symptomatic congenital cytomegalovirus (CMV) infection, and to identify risk factors for the development of CVI in patients with symptomatic congenital CMV. METHODS: Retrospective subanalysis of a long-term prospective cohort study with data gathered from 1982 to 2013. RESULTS: Eleven of 77 (14.3%) patients with symptomatic CMV, 0 of 109 with asymptomatic CMV, and 51 control patients had CVI. Overall, patients with symptomatic CMV had worse vision than patients with asymptomatic CMV, who in turn had worse vision than control patients. Microcephaly, intracranial calcification, dilatation of ventricles, encephalomalacia, seizure at birth, optic atrophy, chorioretinitis/retinal scars, strabismus, and neonatal onset of sensorineural hearing loss were risk factors associated with CVI. CONCLUSIONS: CVI may result from symptomatic congenital CMV infection. The relationship of CVI and its risk factors in patients with CMV suggests the potential to predict the development of CVI through predictive modeling in future research. Early screening of CVI in children born with symptomatic congenital CMV can facilitate educational, social, and developmental interventions. [J Pediatr Ophthalmol Strabismus. 2019;56(3):194-202.].

Clinical Pearl: The "Rule of 8".

Preschool-aged children typically perform worse than 20/20 on visual acuity testing in the absence of ocular or visual pathway abnormalities due to non-ocular issues, including concentration, cooperation, and/or confidence. As a result, vision screening for this age group can be a challenge. The "Rule of 8" is an easy-to-remember, highly effective mnemonic that clinicians can use to help differentiate children who are performing visual acuity testing at age-appropriate levels from those who may need further evaluation. [J Pediatric Ophthalmol Strabismus. 2017;54:e75-e76.].

Assessment of Lower Doses of Intravitreous Bevacizumab for Retinopathy of Prematurity: A Phase 1 Dosing Study.

Importance: Intravitreous bevacizumab (0.25 to 0.625 mg) is increasingly used to treat type 1 retinopathy of prematurity (ROP), but there remain concerns about systemic toxicity. A much lower dose may be effective while reducing systemic risk. Objective: To find a dose of intravitreous bevacizumab that was lower than previously used for severe ROP, was effective in this study, and could be tested in future larger studies. Design, Setting, and Participants: Between May 2015 and September 2016, 61 premature infants with type 1 ROP in 1 or both eyes were enrolled in a masked, multicenter, phase 1 dose de-escalation study. One eye of 10 to 14 infants received 0.25 mg of intravitreous bevacizumab. If successful, the dose was reduced for the next group of infants (to 0.125 mg, then 0.063 mg, and finally 0.031 mg). Diluted bevacizumab was delivered using 300 µL syringes with 5/16-inch, 30-gauge fixed needles. Interventions: Bevacizumab injections at 0.25 mg, 0.125 mg, 0.063 mg, and 0.031 mg. Main Outcomes and Measures: Success was defined as improvement in preinjection plus disease or zone I stage 3 ROP by 5 days after injection or sooner, and no recurrence of type 1 ROP or severe neovascularization requiring additional treatment within 4 weeks. Results: Fifty-eight of 61 enrolled infants had 4-week outcomes completed; mean birth weight was 709 g and mean gestational age was 24.9 weeks. Success was achieved in 11 of 11 eyes at 0.25 mg, 14 of 14 eyes at 0.125 mg, 21 of 24 eyes at 0.063 mg, and 9 of 9 eyes at 0.031 mg. Conclusions and Relevance: A dose of bevacizumab as low as 0.031 mg was effective in 9 of 9 eyes in this phase 1 study and warrants further investigation. Identifying a lower effective dose of bevacizumab may reduce the risk for neurodevelopmental disability or detrimental effects on other organs.

Assessment of Plasma Cytokine Profile Changes in Bevacizumab-Treated Retinopathy of Prematurity Infants.

PURPOSE: We compared changes of plasma angiogenesis cytokine profiles in infants who were treated with intravitreal injection of bevacizumab (IVB) for type 1 retinopathy of prematurity (ROP) with age-matched preterm non-ROP infants. METHODS: Thirteen infants with type 1 ROP and 13 age-matched preterm non-ROP infants were included. Blood samples were collected prior to treatment (time 0) and 6 weeks after the treatment (time 42). Plasma levels of nine cytokines from the angiogenesis growth factor panel and seven soluble cytokine receptors were measured using a magnetic multiplex assay. RESULTS: Plasma cytokine profiles changed from time 0 to time 42 in both groups. In bevacizumab-treated ROP infants, the following plasma angiogenesis growth factor and soluble cytokine receptor levels decreased significantly: soluble VEGF-A (sVEGF-A; P = 0.0001), sVEGF-D (P = 0.04), angiopoietin-2 (Ang-2; P = 0.002), sVEGF receptor 1 (R1) and R2 (P = 0.005), soluble IL-6 receptor (sIL-6R; P = 0.002), soluble glycoprotein 130 (spg130; P = 0.0001), and soluble TNF receptor (sTNFR) I and II (P = 0.0001). The following factors and receptors increased significantly: sVEGF-C (P = 0.05), placental growth factor (PlGF; P = 0.02), endothelin-1 (ET-1; P = 0.0001), and FGF-1 (P = 0.02). At time 42, sVEGF-A, sgp130, sIL-6R, sTNFR I, and sTNFR II were lower, and ET-1 level was higher, in bevacizumab-treated ROP infants compared to age-matched non-ROP infants. CONCLUSIONS: The results suggest that bevacizumab treatment resulted in significant angiogenic cytokine profile changes in infants with severe ROP. The long-term clinical impact of these changes should be studied carefully.

Pharmacokinetics of bevacizumab and its effects on serum VEGF and IGF-1 in infants with retinopathy of prematurity.

PURPOSE: To measure serum levels of bevacizumab and to compare serum levels of free vascular endothelial growth factor (VEGF) and insulin-like growth factor-1 (IGF-1) in infants who were treated with either intravitreal injection of bevacizumab (IVB) or laser for type 1 retinopathy of prematurity (ROP). METHODS: Twenty-four infants with type 1 ROP were randomized into three treatment groups: IVB at 0.625 mg per eye per dose, IVB at 0.25 mg per eye per dose, and laser. Blood samples were collected prior to treatment and on posttreatment days 2, 14, 42, and 60. Weekly body weights were documented from birth until 60 days post treatment. Serum levels of bevacizumab, free VEGF, and IGF-1 were measured with enzyme-linked immunosorbent assay (ELISA). RESULTS: Serum bevacizumab was detected 2 days after the injection, peaked at 14 days, and persisted for up to 60 days with half-life of 21 days. Area under the curve (AUC) analysis showed that systemic exposure to bevacizumab was variable among the subjects and was dose dependent. Serum free VEGF levels decreased in all three subgroups 2 days post treatment, with more significant reductions found in both IVB-treated groups, P = 0.0001. Serum IGF-1 levels were lower in both IVB-treated groups. CONCLUSIONS: Clearance of bevacizumab from the bloodstream in premature infants takes at least 2 months. Although serum free VEGF levels decreased following either laser or bevacizumab treatment, the reductions were more significant in the IVB-treated groups. Potential long-term effects of systemic exposure to bevacizumab in infants need to be studied further.