HARVARD DEPARTMENT OF OPHTHALMOLOGY
Category : Clinical Research Candidate : Luis Martinez-Velazquez Poster #: C11
Predicting Visual Outcomes in Central Retinal Vein Occlusion ( CRVO ) with Vascular Metrics and Non-Perfusion Area ( NPA ) Measured by Wide-Field Swept-Source OCT-Angiography ( WF SS-OCTA )
Luis A . Martinez-Velazquez1,2 , Grace Baldwin1,2 , Itika Garg1,2 , Ying Cui1,3 , Raviv Katz1 , Rebecca Zeng1,2 , Margaret Duich1 , Diane N . Sayah1 , Jade Moon1,2 , Fillipos Vingopolous1,2 , Hannah E . Wescott1,2 , Thomas Koch1,2 , Nimesh A . Patel2 , David M . Wu2 , John
Purpose : To quantify vascular metrics following central retinal vein occlusion ( CRVO ) using wide-field swept source OCTA ( WF-SS-OCTA ) and evaluate for its associations with visual acuity ( VA ). To determine the possible clinical application of WF-SS-OCTA in monitoring patients that suffered from a CRVO for the development of worsening ischemia .
Methods : A cross-sectional observational study was conducted from March 2019 to December 2021 . WF SS- OCTA ( PLEX ® Elite 9000 , Carl Zeiss Meditec Inc .) 3x3 , 6x6 , and 12x12-mm fovea-centered angiograms were obtained . The ARI Network was used to quantify vessel density ( VD ), vessel skeletonized density ( VSD ), and foveal avascular zone ( FAZ ) area , circularity , and perimeter . Non-perfusion area ( NPA ) was calculated using ImageJ . Mixed effects multiple linear regression analyses and receiver operating characteristics ( ROC ) curve analyses were performed .
Results : The cohort included 156 eyes , with 76 CRVO eyes , including 12 ischemic CRVO , and 80 fellow eyes . All the OCTA vascular metrics , except FAZ area , were found to have significant associations with VA on regression analyses controlling for age , sex , hypertension , hyperlipidemia , and diabetes mellitus . Using stepwise bidirectional elimination , a regression model accounting for age ( β = 0.01 , p = 0.008 ), BMI ( β = -0.01 , p = 0.030 ), FAZ perimeter ( β = 0.26 , p = 0.001 ), FAZ area ( β = -1.20 , p = 0.003 ), and NPA ( β = 0.01 , p < 0.001 ) on 12x12-mm images was best for predicting VA < 20 / 50 . On ROC analysis , a multivariate regression model using NPA to predict VA > 20 / 50 performed best ( area under the curve = 0.91 ).
Conclusions : SS-OCTA can be used to robustly detect microvascular changes in eyes following CRVO , and these changes are strongly associated with the resulting visual acuity . ROC analysis found that out of the OCT-A metrics , NPA was more strongly associated with predicting visual acuity worse than 20 / 50 , supporting that NPA is a strong correlate for ischemia . Our data also supports that vascular metrics from 12x12 angiograms can be used to statistically separate ischemic versus non-ischemic RVOs , supporting the role for SS-OCTA in monitoring and prognosticating patients following CRVO .