Imaging the Choroid and Choroidal Neovascularization in Eyes with Inflammation

The choroidal vasculature is the primary source of both oxygen and blood-borne nutrients to the outer retina, including the retinal pigment epithelium (RPE), photoreceptors, and otherwise avascular fovea.^1,2 This high flow system also provides primary ocular access for the circulating, or humoral, immune system. It should be no surprise, therefore, that the choroid is a common site of inflammation, or that recent improvements in choroidal imaging have resulted in rapid advances in our understanding of the pathogenesis of ocular inflammation and its complications. Traditional tools for choroidal imaging have included B-scan ultrasonography, computerized axial tomography (CAT), magnetic resonance imaging (MRI), and indocyanine green angiography (ICGA). More recently, technological improvements in structural optical coherence tomography (OCT) have revealed important morphologic and functional characteristics of the choroid in physiological and pathological conditions, including those leading to secondary choroidal neovascularization (CNV). These advances include significant progress in OCT algorithms and devices that impact image acquisition, processing, display and analysis, such as Enhanced Depth Imaging-OCT (EDI-OCT), en face-OCT, Swept-Source-OCT (SS-OCT), and OCT-Angiography (OCT-A).^3–8 Three original articles in this issue of Ocular Immunology & Inflammation (OII) employ multimodal imaging to evaluate morphologic findings of the choroid or CNV in patients with, or at increased risk for, ocular inflammation.^9–11

Duru et al.⁹ used EDI-OCT to measure subfoveal and perifoveal choroidal thickness in 117 patients with rheumatoid arthritis (RA) and 46 age-matched, healthy control subjects seen at a referral center in Ankara, Turkey. A total of 74 (63.2%) of the patients with RA had active disease (DAS-28 >2.60). None of the patients with RA had evidence of past or present ocular inflammation. Subfoveal choroidal thickness was approximately 70 μm (~25%), thinner in eyes with RA as compared with controls. A difference of comparable relative magnitude was also measured 1500 μm temporal and nasal to the fovea. None of these differences were correlated with RA disease activity, however. While the authors made no mention of controlling for the use of systemic corticosteroids, which are known to produce vasoconstriction and systemic hypertension in a sizable proportion of patients,¹² the mean arterial blood pressure was comparable in the two groups.

The same group from Ankara used an identical study design to evaluate subfoveal and perifoveal choroidal thickness in 58 patients with systemic lupus erythematosus (SLE) and compared these findings to measurements taken in an identical number of healthy control subjects.¹⁰ Differences in choroidal thickness measurements, both under and adjacent to the fovea, were of similar absolute and relative magnitudes as those seen in patients with RA. The authors hypothesized that decreased choroidal thickness in RA and SLE patient populations may have resulted from chronic choroidal vascular inflammation, perhaps producing narrowing or occlusion of the vessels. Neither population had visual symptoms or overt signs of either uveitis or retinal vasculitis, however. A number of studies have employed EDI-OCT to measure choroidal thickness in eyes with uveitis, with most showing increased thickness in the setting of active disease.^1,3,6,7

Wu et al.¹¹ retrospectively examined the demographics, causes of uveitis, and lesion characteristics of 150 eyes (166 lesions) of 125 patients with uveitis complicated by the development of CNV. Patients were seen over an 11-year period at a referral center in Huizhou, Guangdong, China. All 166 lesions were described as having a “classic” morphology based on fluorescein angiography (FA) and, of the 31 eyes imaged with OCT, the lesions were subretinal, or Type 2, in all cases. One in five patients had bilateral CNV and about one in ten eyes had two lesions. The vast majority of lesions were subfoveal (39.8%), juxtafoveal (20.5%), or extrafoveal (36.8%), with only a minority occurring near the arcade vessels (1.8%) or disc (1.2%). The most common causes of uveitis associated with the development of CNV were punctate inner choroidopathy (PIC; 50.4%) and multifocal choroiditis (MFC; 22.4%), followed by Vogt–Koyanagi–Harada (VKH) disease (8.0%), and other uveitic conditions (19.2%; 12.8% of whom had inflammation of unspecified type).

Many now consider PIC and MFC in the absence of panuveitis – as opposed to MFC with panuveitis (MFCPU) – to be the same, or related, conditions.^13,14 This plausible possibility accepted, the observation by Wu et al.¹¹ that collected MFC/PIC lesions contributed disproportionately to their total uveitic CNV population (72.8%) would seem to be supported by the proportion of MFC/PIC cases reported in several uveitic CNV cohorts, including the 2009 multinational study by Mansour et al.¹⁵ (43.8%); the 2011 Pan-American study by Arevalo et al.¹⁶ (45.5%); and the 2013 Systemic Immunosuppressive Therapy for Eye disease (SITE) study by Baxter et al.¹⁷ (37.0%). Moreover, patients with MFC/PIC are at particularly high risk for developing CNV, especially when the characteristic chorioretinitis lesions involve Zone 1.¹⁸ In a large, retrospective study of 65 eyes of 41 patients with MFC/PIC seen in a large retina referral practice in New York, for example, Fung et al.¹⁹ observed CNV at presentation in 61.3% of eyes, with 32.6% developing new or recurrent CNV during follow-up (range: 0–343 months; mean 92.6 months; median 84.0 months).

Using structural OCT and angiography to distinguish purely inflammatory MFC/PIC lesions from those associated with CNV can be challenging. Recently, Baumal et al. showed that OCT-A may be helpful in making this determination.²⁰ Moreover, while the study by Wu et al.¹¹ identified only Type 2 lesions in eyes with uveitic CNV using structural OCT and FA in a subgroup of their study patients, Spaide et al.¹⁴ and Amer et al.²¹ used multimodal imaging to show that MFC/PIC-associated CNV lesions may be Type 1 (sub-RPE), Type 2, or mixed—often with continued extension into the overlying outer retina, producing distinctive, finger-like, hyper-reflective columns on OCT – a finding that Hoang et al.²² have termed the “pitch-fork sign.”

Together, these three studies highlight both the utility of multimodal imaging and the expanding spectrum of choroidal findings in patients with ocular inflammation. Choroidal neovascularization, while a relatively infrequent complication of most types of uveitis, can compromise vision quickly and so should be sought in eyes with suggestive symptoms or signs. This is especially true for eyes with MFC/PIC, and particularly when the posterior pole is involved. New imaging tools, most notably enhancements in structural OCT and the advent of OCT-A, provide great promise for investigating the causes and complications of choroidal inflammation. Several solicited reviews dedicated to the application of multimodal imaging to uveitis will appear in OII over the coming months.

Declaration of Interest

DF is an employee at Genentech, Inc. and receives stock/stock options from Roche. SM is a consultant to Novartis and Bayer. KBF is a consultant to Optovue, Heidelberg Engineering, and Optos. The other authors have no financial conflicts.

Funding

Supported in part by The Pacific Vision Foundation (ETC) and The San Francisco Retina Foundation (ETC).