The overall visual outcome of inflammatory choroidal neovascularization (CNV) was relatively good, and the recurrence of inflammatory CNV showed higher rates over time after anti-VEGF treatment than previously reported.
Baseline best-corrected visual acuity and retinal pigment epithelium atrophy or absence after treatment are significant predictors for the visual outcome.
Intraretinal hyperreflective foci after anti-VEGF treatment suggests the potential risk of recurrence.
To investigate the long-term clinical features and treatment outcomes of patients with inflammatory choroidal neovascularization (CNV) treated with intravitreal anti−vascular endothelial growth factor (anti-VEGF).
Retrospective, interventional, consecutive case series.
Sixty-five eyes of 65 patients with inflammatory CNV treated with anti-VEGF injections and followed up at least 12 months were included. Retrospective chart review was conducted at a single tertiary referral center.
Study participants were followed up for 60.6 ± 42.8 (range, 16-160) months. Mean age was 33.4 ± 10.8 years, and mean refractive error was −3.94 ± 1.35 D in spherical equivalent. Final best-corrected visual acuity (BCVA) was 0.21 ± 0.20 logMAR after treatment. Patients were treated with bevacizumab (76.9%), ranibizumab (4.6%), aflibercept (3.1%), and drug combinations (15.4%). Systemic corticosteroid or immunosuppressant use was not correlated with visual outcome, required number of anti-VEGF injections, and recurrence. Commonly occurring optical coherence tomography (OCT) features included ellipsoid zone disruption, choroidal hypertransmission, retinal pigment epithelium atrophy or absence (RPEA), intraretinal hyperreflective foci (HRF), choroidal vessel engorgement, focal choroidal excavation, and irregular vascular loops (on OCT angiography). RPEA after treatment (β = 0.238, P = .036) and BCVA (β = 0.267, P = .029) showed significant correlation with final BCVA. A total of 28 patients (43.1%) experienced recurrence; intraretinal HRF after treatment was the single risk factor for recurrence (odds ratio = 2.712, P = .031).
Inflammatory CNV recurrence showed higher rates over time after anti-VEGF treatment than previously reported, even though the overall visual outcome was good. Baseline BCVA and RPEA after treatment are significant predictors for visual outcome. Intraretinal HRF after anti-VEGF treatment suggests the potential risk of recurrence.
I nflammatory choroidal neovascularization (CNV) is a rare but vision-threatening complication of uveitis that results from the disruption of the Bruch’s membrane−retinal pigment epithelium (RPE) complex. , Inflammatory CNV is prevalent in 1.9% of patients with uveitis and in 2.0% to 4.8% of those with posterior uveitis. Inflammatory CNV may occur in several forms of uveitis, including those with infectious and noninfectious etiologies. Conditions that may be complicated by inflammatory CNV include multifocal choroiditis (MFC)/punctate inner choroidopathy (PIC), multiple evanescent white dot syndrome (MEWDS), serpiginous choroiditis, sarcoidosis, Vogt−Koyanagi−Harada (VKH) disease, and toxoplasma retinochoroiditis. ,
Previously, inflammatory CNV was treated using photodynamic therapy (PDT) and focal laser photocoagulation. Since 2007, treatment with intravitreal anti−vascular endothelial growth factor (anti-VEGF) injections has constituted the mainstay in the management of inflammatory CNV, and has shown favorable visual outcomes. Considering the preceding inflammatory condition, treatments with systemic or local corticosteroids or systemic immunosuppressive agents have also been used. , , ,
Inflammatory CNV is usually classic (type 2) CNV, which grows into the subretinal space between the neurosensory retina and RPE. , Multimodal imaging with fluorescein angiography (FA), indocyanine green angiography (ICGA), optical coherence tomography (OCT), and OCT angiography (OCT-A) has enabled the detection of inflammatory CNV. Characteristic OCT features in inflammatory CNV have previously been reported. , , However, differential diagnosis with idiopathic CNV, myopic CNV, and neovascular age-related macular degeneration (AMD) remains challenging. In addition, to date, how the OCT findings influence the visual prognosis and recurrence remains unclear.
There are no large-scale cohort studies or randomized clinical trials that have reported on OCT characteristics and treatment outcomes, because of the low prevalence of inflammatory CNV. Furthermore, most studies on inflammatory CNV have been conducted over relatively shorter time periods (from 4 weeks to 24 months). Here, we aim to evaluate the long-term clinical characteristics and treatment outcomes of inflammatory CNV, with special focus on OCT features, in a relatively large case series. In addition, we aim to study the treatment responses to various intravitreal anti-VEGF agents and the recurrence while on pro re nata (PRN) treatment.
PATIENTS AND METHODS
Our study was a retrospective, interventional, consecutive case series of patients with active inflammatory CNV who were treated between January 2006 and December 2020 at the Seoul St. Mary’s Hospital. Patients with proven diagnoses of inflammatory CNV who were treated with intravitreal anti-VEGF injections and received follow-up for at least 12 months after initial presentation were included. Exclusion criteria included (1) any previous treatment for CNV including intravitreal injections, PDT, and laser photocoagulation; (2) vitreoretinal surgery in the preceding 12 months; (3) presence of other concomitant ocular diseases (i.e., retinal vein occlusion, glaucoma, epiretinal membrane, diabetic retinopathy); (4) symptoms likely related to neovascular AMD rather than to inflammatory CNV; and (5) presence of significant media opacity that deteriorated the quality of multimodal imaging.
The study design followed the principles of the Declaration of Helsinki, and all protocols were approved by the Institutional Review Board of the Catholic University of Korea. The requirement for written informed consent was waived because of the retrospective nature of this study.
PATIENT EVALUATION AND TREATMENT
Patient characteristics, including age, sex, follow-up duration, medical history, and history of previous ocular treatments, were retrieved from medical records. All patients underwent comprehensive baseline and follow-up ophthalmic examinations, including best-corrected visual acuity (BCVA), slit-lamp examination, pneumatic tonometry, and fundus examination. Follow-up examinations were performed for the study participants in the clinic, at <3-month intervals, even in asymptomatic cases. In cases of active uveitis, active CNV or any self-reported symptoms, the inspection interval was shortened as deemed appropriate. The BCVA was converted to the logarithm of the minimal angle of resolution (logMAR) units before analysis. Eyes were scanned with either a spectral domain OCT (Spectralis HRA OCT; Heidelberg Engineering) in patients who first presented before 2017, or with swept-source OCT (DRI-OCT; Topcon Corp) in those who first presented after 2017. We used the same OCT device throughout the follow-up period for each individual patient. In addition, we performed FA and ICGA (HRA-2; Heidelberg Engineering). Owing to the retrospective nature of this study, OCT-A (DRI-OCT; Topcon Corp), images could be obtained only when available for patients seen after 2017.
Inflammatory CNV was identified using multimodal imaging methods, including FA, ICGA, OCT, and OCT-A. The diagnosis was based on the presence of active CNV in eyes with either noninfectious or infectious uveitis. Active CNV was defined by the presence of any of the following findings: dye leakage on FA/ICGA, intraretinal/subretinal/subRPE fluid on OCT, and retinal hemorrhage on fundus examination. Quiescent CNV detected by OCT-A without hemorrhage, fluid on OCT, or dye leakage on FA/ICGA were considered inactive CNV. An intravitreal anti-VEGF injection was administered in cases with activity on OCT, including the presence of subretinal, intraretinal, or subRPE fluid, occurring either individually or in combination. Patients were treated with the PRN regimen, and the selection and interval of anti-VEGF drugs were decided upon by the physician. Anti-VEGF treatments were discontinued after complete anatomical resolution of activity on OCT. Recurrence was defined as the presence of fluid on the structural OCT, newly seen at least 3 months after the complete resolution of the previous fluid.
Central macular thickness (CMT) was measured using either circular map analysis provided by the spectral domain OCT software or the macular volume scan of the swept-source OCT device. Location of the inflammatory CNV was determined according to the Early Treatment Diabetic Retinopathy Study grid as either subfoveal (within concentric circle with 0.5-mm radius at central fovea), parafoveal (within a 0.5- to 1.5-mm radius from the fovea), perifoveal (within a 1.5- to 3-mm radius from the fovea), extrafoveal (outside 3 mm radius concentric circle), or peripapillary (situated around the optic disc). The presence or absence of the following features was evaluated on OCT by a retinal specialist (Y.H.P.): presence of focal choroidal excavation (FCE), choroidal hypertransmission, intraretinal hyperreflective foci (HRF), perilesional RPE atrophy or absence (RPEA), and disruption of the ellipsoid zone (EZ; border between the inner and outer segments) bands.
The FCE was defined as an area of choroid concavity detected on OCT. , Choroidal hypertransmission was defined as vertical bands of increased reflectivity extending to the RPE and choroid. The RPEA was defined as either a complete absence of RPE, a degradation of the RPE layer reflectivity and thickness, or an RPE contour break. , The HRF was defined as a small, discrete, well-circumscribed lesion with reflectivity equal to or more than the RPE layer. ,
Categorical data were expressed as absolute numbers, and continuous data were expressed as mean ± standard deviation (95% confidence interval). The Kruskal−Wallis test was performed to compare clinical characteristics according to etiological diagnoses. The Wilcoxon matched-pairs signed rank test and the McNemar test were used to compare the patients’ clinical parameters before and after undergoing anti-VEGF treatment. Binary logistic regression analysis was performed to identify factors associated with inflammatory CNV recurrence. Univariate and multivariate linear regression analyses were performed to evaluate predictors of final BCVA in eyes with inflammatory CNV following treatment. The cumulative probabilities of recurrence after CNV resolution were assessed using Kaplan−Meier analysis. A log-rank test was used to compare the survival rates between groups using different anti-VEGF agents.
All statistical analyses were performed using SPSS for Windows (version 24.0; IBM Corp). Kaplan−Meier survival curves were created using GraphPad Prism version 8.4.2 for Windows . Results with P values of <.05 were considered statistically significant.
DEMOGRAPHIC AND CLINICAL CHARACTERISTICS
During the study period, a total of 3,615 patients with uveitis visited the clinic. Among these, 65 eyes from 65 patients met the inclusion criteria and were included in the analysis. The estimated percentage of inflammatory CNV, among the eyes with uveitis, was 1.80%. The mean age of the patients was 33.4 ± 10.8 years, and 49 patients (75.4%) were female. Study participants were followed up for 60.6 ± 42.8 months (range, 12-160 months). Baseline demographics and disease characteristics of the study population are provided in Table 1 . The mean logMAR BCVA was 0.49 ± 0.31 (Snellen equivalent: 20/63) at baseline, and the mean refractive error in spherical equivalents was −3.94 ± 1.35 diopter. Of the studied eyes of the participants, only 7 eyes (10.8%) showed intraocular inflammation at the time of diagnosis of inflammatory CNV. The inflammatory CNV were mostly subfoveal (46.2%) in location, followed by parafoveal (32.3%), perifoveal (12.3%), peripapillary (6.2%), and extrafoveal (3.1%) locations. The most common etiology of uveitis leading to inflammatory CNV was PIC/MFC (38 eyes, 58.5%). Other etiologic diagnoses included acute retinal pigment epitheliitis (ARPE) (7.7%), toxoplasmosis (6.2%), MEWDS (4.6%), sarcoidosis (4.6%), serpiginous choroiditis (1.5%), and idiopathic uveitis (16.9%). There were no significant differences in the total number of administered anti-VEGF injections ( P = .414), recurrence of inflammatory CNV ( P = .062), changes in BCVA ( P = .129) and changes in CMT ( P = .462) between the diagnostic groups ( Table 2) .
|Number of eyes||65|
|Follow-up period, mo||60.6 ± 42.8|
|Age, y||33.4 ± 10.8|
|Sex, n (%)|
|Baseline best-corrected visual acuity (logMAR)||0.49 ± 0.31|
|Refractive error (spherical equivalent), diopter||−3.94 ± 1.35|
|Location of CNV|
|Etiology of uveitis|
|Serpiginous choroiditis||1 (1.5%)|
|Anti-VEGF agents received, n (%)|
|Bevacizumab monotherapy||50 (76.9%)|
|Ranibizumab monotherapy||3 (4.6%)|
|Aflibercept monotherapy||2 (3.1%)|
|Switch a||10 (15.4%)|
|Total number of anti-VEGF injections||3.6 ± 2.5|
|Time interval between anti-VEGF injections, mo||2.0 ± 0.9|
|Concomitant immunosuppressive treatment, n (%)||6 (9.2%)|
|Concomitant systemic steroid treatment, n (%)||28 (43.1%)|
|Recurrence of inflammatory CNV, n (%)||28 (43.1%)|
|Time interval to recurrence, mo||35.8 ± 30.3|
a Any combination of the abovementioned anti-VEGF agents. Data are expressed as mean ± SD (95% confidence interval).
|Diagnosis||Total||Total Number of Anti-VEGF Injections||Recurrence of Inflammatory CNV||Mean Change in BCVA||Mean Change in CMT|
|MFC/PIC||38||3.8 ± 2.9||16 (42.1%)||–0.24 ± 0.42||–45.3 ± 23.2|
|ARPE||5||2.4 ± 0.5||2 (40%)||–0.20 ± 0.50||–74.3 ±53.4|
|Toxoplasmosis||4||2.3 ± 1.5||0||–0.35 ± 0.13||–74.3 ± 60.5|
|MEWDS||3||5.3 ± 2.5||3 (100%)||–0.60 ± 0.43||–66.7 ± 54.6|
|Sarcoidosis||3||3.67 ± 0.58||3 (100%)||–0.70 ± 0.17||–101.3 ± 30.0|
|Idiopathic||11||3.27 ± 2.32||4 (36.4%)||–0.13 ± 0.41||–53.9 ± 40.7|
|P value a||—||0.414||0.062||0.129||0.462|
MORPHOLOGIC CHARACTERISTICS OF INFLAMMATORY CHOROIDAL NEOVASCULARIZATION
Table 3 shows the clinical and morphologic parameters before and after undergoing anti-VEGF treatment. Of the 65 eyes, 45 (69.2%) were followed up using spectral domain OCT (Spectralis HRA OCT), and 20 (30.8%) with swept-source OCT (DRI-OCT). The mean logMAR BCVA improved from a baseline value of 0.49 ± 0.31 (Snellen equivalent: 20/63) to 0.21 ± 0.20 (Snellen equivalent: 20/32) after a mean of 3.6 injections ( P = .001). Mean CMT decreased from a baseline value of 295.1 ± 60.7 μm to 244.7 ± 37.8 μm after a mean of 3.6 injections ( P < .001).
|OCT Parameters||Before Anti-VEGF||After Anti-VEGF||P Value|
|Clinical parameters, mean ± SD|
|Best-corrected visual acuity, logMAR||0.49 ± 0.31||0.21 ± 0.20||.001 a|
|Central macular thickness, μm||295.1 ± 60.7||244.7 ± 37.8||<.001 a|
|OCT parameters, no.|
|Focal choroidal excavation||8||13||.130 b|
|Choroidal hypertransmission||38||41||.989 b|
|Choroidal vessel engorgement||10||3||.013 b|
|RPE atrophy or absence||23||19||.546 b|
|EZ disruption||40||21||.002 b|
|Intraretinal hyperreflective foci||18||9||.024 b|
|OCT-A pattern of CNV, no.|
|Irregular vascular loops||9||—||—|
|Branching networks of small vessels||6||—||—|