Imaging Features Associated with Persistent Avascular Retina in Retinopathy of Prematurity

Abstract

The use of anti-VEGF therapy as the primary treatment for retinopathy of prematurity (ROP) has generated attention to the potential for long-term avascularity of the peripheral retina after initial treatment. The most recent International Classification of Retinopathy of Prematurity, Third Edition, formally termed this finding “persistent avascular retina” (PAR) and recognized that it can occur after spontaneous regression or anti-VEGF treatment.1Chiang M.F. Quinn G.E. Fielder A.R. et al.International Classification of Retinopathy of Prematurity, Third Edition.Ophthalmology. 2021; 128: e51-e68Abstract Full Text Full Text PDF PubMed Scopus (249) Google Scholar A potential benefit of anti-VEGF primary treatment compared with laser is increased vascularized area, potentially enlarging the functional visual field long term. However, this treatment also has the potential to reactivate and cause atypical ROP-related retinal detachments in infancy, as well as increasing the long-term risk of detachment.2Hu J. Blair M.P. Shapiro M.J. et al.Reactivation of retinopathy of prematurity after bevacizumab injection.Arch Ophthalmol. 2012; 130: 1000-1006Crossref PubMed Scopus (238) Google Scholar A better understanding of which babies are at highest risk of PAR could help risk stratify who would benefit most from more frequent examinations and possibly laser. Recent work has demonstrated that a higher vascular severity score (VSS), derived from the artificial intelligence–based Imaging and Informatics in ROP deep learning algorithm, is associated with more posterior disease, higher disease stage and extent, and higher risk of treatment failure with laser.3Gupta K. Campbell J.P. Taylor S. et al.A quantitative severity scale for retinopathy of prematurity using deep learning to monitor disease regression after treatment.JAMA Ophthalmol. 2019; 137: 1029-1036Crossref PubMed Scopus (50) Google Scholar In this report, we evaluate whether imaging-based quantitative metrics such as the VSS or the measured posterior extent of retinal vascularization at the time of treatment with intravitreal bevacizumab (IVB) may be associated with disease reactivation or failure to fully vascularize (PAR). We performed a retrospective consecutive case series of 66 eyes from 33 infants with type 1 ROP. This study was performed in accordance with the Health Insurance Portability and Accountability Act of 1996 and respected the tenets of the Declaration of Helsinki. Institutional Review Board Ethics Committee approval was obtained from the University of California, San Diego, and Rady Children’s Hospital. Informed consent was not required for this retrospective study. Demographic details are summarized in Table S1 (available at www.aaojournal.org). Retinal fundus images were acquired via a RetCam (Natus) or the Panocam (Visunex Medical Systems). Treatment and re-treatment decisions were determined via single-provider clinical examination (E.N.); re-treatment was recommended if there was reactivation of neovascularization or worsening plus disease or if vascular growth failed to reach zone III by 50 weeks postmenstrual age (PMA). The extent of retinal vascularization and rate of vascular growth after IVB (0.625 mg) were measured (H.R-H., E.N.). Briefly, the mean of 5 measurements—taken from the center of the optic disc, through the macula, to the anterior edge of the ridge—was calculated from retinal fundus images using ImageJ (https://imagej.nih.gov) (Fig S1, available at www.aaojournal.org). Measurements were converted from pixels to millimeters using a standardized pixel conversion calculator (https://pixelcalculator.com/en). Eye-level VSS at the time of IVB and laser (if applicable) treatment was calculated using the Imaging and Informatics in ROP DL system.4Campbell J.P. Singh P. Redd T.K. et al.Applications of artificial intelligence for retinopathy of prematurity screening.Pediatrics. 2021; 147e2020016618Crossref Google Scholar,5Chen T.A. Shields R.A. Bodnar Z.H. et al.A spectrum of regression following intravitreal bevacizumab in retinopathy of prematurity.Am J Ophthalmol. 2019; 198: 63-69Abstract Full Text Full Text PDF PubMed Scopus (31) Google Scholar To assess risk factors for PAR, the cohort was divided into 2 groups: “PAR” (eyes that received IVB and subsequent laser photocoagulation) and “No PAR” (eyes that fully vascularized and did not require subsequent laser photocoagulation). Shapiro-Wilk normality tests were used to assess whether birth weight (BW), gestational age (GA), and PMA at IVB treatment came from normal distributions. Given the results, t tests or Mann–Whitney tests were used to assess group differences. To account for inter-eye correlations, generalized estimating equations (GEEs) were used to evaluate whether significant differences existed between the median (interquartile range [IQR]) VSS of eyes that received subsequent laser and those that did not, adjusting for BW, GA, and PMA. Likewise, GEE was used to evaluate whether there were differences in zone and the extent of retinal vascularization between eyes in each group. A post hoc analysis further divided eyes that developed PAR into 3 groups based on the number of weeks passed until re-treatment was performed: 0–6 weeks, 6–12 weeks, and 12 or more weeks. Again, GEE was used to determine if the extent of retinal vascularization was associated with time until reactivation. Although BW, GA, and PMA were not statistically different between groups, the median (IQR) VSS in the PAR group was significantly higher (8.2 [2.1]) than in the No PAR group (7.0 [3.6], P < 0.001, Fig 2A). Likewise, the median (IQR) length of vascularized retina for all eyes was 310.4 (90.0) mm but was shorter in the PAR group (285.3 [94.2] mm) compared with the No PAR group (364.6 [126.3] mm, P < 0.001, Fig 2B). This coincided with the finding that zone I ROP was significantly associated with PAR (P < 0.001). A post hoc analysis revealed that the time to reactivation for eyes that would eventually develop PAR was associated with the extent of the retinal vasculature measured at initial IVB treatment (Fig 2C). The median (IQR) length of vascularized retina for eyes requiring re-treatment at 0–6 weeks, 6–12 weeks, and 12 or more weeks was 194.0 (41.9), 269.0 (88.4), and 310.0 (69.2), respectively. All groups were significantly different from one another (P < 0.001). Persistent avascular retina is not uncommon after spontaneous regression of ROP and may be higher in patients with anti-VEGF.5Chen T.A. Shields R.A. Bodnar Z.H. et al.A spectrum of regression following intravitreal bevacizumab in retinopathy of prematurity.Am J Ophthalmol. 2019; 198: 63-69Abstract Full Text Full Text PDF PubMed Scopus (31) Google Scholar Recent multicenter case series have demonstrated late complications such as retinal breaks, retinal detachments, and vitreous hemorrhage in patients with a history of ROP with or without treatment.6Hamad A.E. Moinuddin O. Blair M.P. et al.Late-onset retinal findings and complications in untreated retinopathy of prematurity.Ophthalmol Retina. 2020; 4: 602-612Abstract Full Text Full Text PDF PubMed Google Scholar These risks, plus reports of late reactivation after anti-VEGF, have led some providers to recommend combination therapy with deferred laser for reactivation or PAR, although there are no consensus guidelines for when to treat either.7Kim J. Kim S.J. Chang Y.S. Park W.S. Combined intravitreal bevacizumab injection and zone I sparing laser photocoagulation in patients with zone I retinopathy of prematurity.Retina. 2014; 34: 77-82Crossref PubMed Google Scholar The results of this study suggest that quantitative assessment of imaging features, using either objective assessment of vascular severity or measurement of the posterior extent of retinal vascularization, may help identify babies at highest risk of reactivation or PAR. These results are consistent with prior clinical studies that zone I eyes are at highest risk of treatment failure or reactivation. Notably, in this study, although the clinical grading of zone did not differ for the majority of eyes, objective assessment of the posterior vascular border did, suggesting that objective metrics of disease severity may be more precise than clinical examination and that ophthalmic imaging may be beneficial. Download .pdf (.11 MB) Help with pdf files Fig S1 Download .pdf (.07 MB) Help with pdf files Table S1

Topics & Concepts