Correlation Between Clinical Signs and Optical Coherence Tomography With Enhanced Depth Imaging Findings in Patients With Birdshot Chorioretinopathy

Importance  Birdshot chorioretinopathy (BCR) is a bilateral posterior uveitis that typically requires aggressive therapy to prevent loss of vision. Clinical signs of disease activity may be subtle and visual acuity is often preserved despite significant loss of visual function. Optical coherence tomography with enhanced depth imaging (OCT-EDI), a new technology that allows visualization of structures posterior to the retinal pigment epithelium, may be a useful tool to monitor disease activity in these patients.

Objective  To determine the correlation between symptoms and signs of disease activity in BCR and specific findings on OCT-EDI.

Design, Setting, and Participants  Retrospective medical record review of 14 patients treated for BCR in the uveitis clinic at Northwestern University. All patients underwent OCT-EDI (58 scans). Clinical symptoms of photopsias/vibrating vision and signs of macular edema, vitreous haze, and retinal vasculitis were graded; a second grading scale was developed for the evaluation of OCT-EDI. Individual scans of each eye of each patient at each point were graded in a masked fashion.

Exposure  Optical coherence tomography with EDI in BCR.

Main Outcomes and Measures  Spearman rank correlation of clinical measures to OCT-EDI measures.

Results  The most frequent score in each clinical category was 0 (inactive). In those BCR patients with symptoms (21 eye examinations), the subjective complaint of photopsias/vibrating vision was associated with the objective finding of suprachoroidal fluid on OCT-EDI (P = .003), and the frequency and severity of photopsias correlated with the thickness of the fluid band (Pearson product moment correlation, 0.39). Two of the clinical markers of disease activity measured in this study (vasculitis and vitreous haze) also showed a significant Spearman rank correlation with the presence and amount of suprachoroidal fluid on OCT-EDI (vasculitis, 0.45 [P < .001]; vitreous haze, 0.59 [P < .001]).

Conclusions and Relevance  The presence of suprachoroidal fluid on OCT-EDI appears to correlate with the subjective complaints of photopsias in patients with BCR and other more easily assessed clinical features such as vasculitis and vitreous haze. Optical coherence tomography with EDI may be a useful tool for objective monitoring of BCR.

Birdshot chorioretinopathy (BCR) is a bilateral posterior uveitis characterized by creamy yellow choroidal lesions, mild vitritis, and vasculitis. Common complaints include floaters, vibrating vision, photopsias, decreased contrast sensitivity, and decreased night vision.¹ Diagnosis is based on this clinical presentation and supported by testing for the HLA-A29 allele, results of which are positive in more than 95% of patients.² In patients with subtle clinical findings, the use of indocyanine green and fundus autofluorescence can highlight lesions that are not apparent clinically.^3,4 Although a small subset of patients may not require treatment, most patients require aggressive therapy to reduce the risk of permanent and severe visual disability.^5-9

Visual acuity is not an adequate marker of disease activity in BCR, because it is often preserved despite significant loss of visual function.¹ For this reason, other tests of visual function are typically performed at least yearly in these patients. Serial electroretinography (ERG) is used to detect changes in overall retinal function,¹⁰ although fluctuations are common and may complicate interpretation of the results. In addition, ERG is not always readily available, is a lengthy procedure, and is often difficult for the patient. Automated or kinetic perimetry is often performed to detect central and peripheral field loss. Visual field defects are common in patients with BCR, even those with good visual acuity.^5,11 Reduction in ERG or visual field testing results reflect changes in retinal function during the intervening months or years since the last test but do not provide real-time measures of disease activity. Assessment of current disease activity relies on results of the clinical examination, with adjunctive imaging tools such as optical coherence tomography (OCT) to quantify macular thickness and fluorescein angiography to detect retinal vasculitis and macular leakage. Until recently, however, choroidal disease activity has not been accessible to imaging. Ocular coherence tomography with enhanced depth imaging (OCT-EDI) has enabled improved visualization of structures posterior to the retinal pigment epithelium (RPE).¹² This imaging approach has been used recently to characterize posterior segment findings in uveitis, including structural changes in the choroid in a cross-sectional study of patients with active and quiescent BCR.¹³ The purpose of the present study is to determine whether chorioretinal changes on OCT-EDI correlate with clinical measures of disease activity in patients with BCR.

We were granted permission by the institutional review board of Northwestern University to perform a retrospective review of the existing medical records of patients with a diagnosis of BCR. All patients with BCR had undergone testing with OCT-EDI during the study period. Diagnosis was based on the clinical picture of bilateral posterior uveitis with creamy yellow choroidal lesions, vitritis, vasculitis, and HLA-A29 positivity.

All patients underwent a complete eye examination that included dilated fundus biomicroscopy performed by one of us (D.A.G.). A grading scale was developed for the clinical evaluation of disease activity. Because many patients were unable to separate their subjective complaints, the symptoms of vibrating vision and photopsias were considered together. Shimmering vision/photopsias were graded 0 if absent, 1 if present only occasionally, 2 if assessed by the patient as significant but not continuous, and 3 if present almost continuously and interfering with the patient’s normal activities. Cystoid macular edema (CME) was defined as the presence of macular thickening with or without cyst formation on the results of clinical examination or OCT-EDI, or as the presence of a petaloid pattern of leakage on fluorescein angiogram. If the OCT scans were reviewed by the clinical grader, choroidal imaging findings were hidden. Cystoid macular edema was graded as 0 if absent, 1 if mild with preservation of foveal contour, 2 if marked with loss of foveal architecture, and 3 if edema extended outside of the macula. This grading scale was used because BCR is often associated with diffuse retinal edema. Analysis of the data based on central macular thickness gave similar results. Vasculitis was graded clinically and based on fluorescein angiography findings, if available. The vasculitis of BCR is typically a mild periphlebitis. Representative photographs of grades 1 and 2 vasculitis are shown in Figure 1. Vitreous haze was graded using a published standardized grading system.¹⁴ One of us (D.A.G.) who was masked to the OCT-EDI results graded each eye of every patient at each visit using these clinical scales.

Each eye underwent a macular cube scan using a commercially available device (Spectralis HRA+OCT; Heidelberg Engineering). Each eye underwent scanning using a 20 × 20 degree cube, including 25 B-scans, with the mean of each calculated 16 times. For each eye, a total of 3 OCT B-scans were graded, 1 through the fovea, 1 in the superior, and 1 in the inferior macula (scans 3, 21, and the foveal scan).

We developed a grading scale for the evaluation of OCT-EDI scans (Figure 2). For each eye at each visit, two of us (A.D.B. and A.A.F.) who were masked to the clinical records and disease activity status of the patients independently graded the scans. The OCT-EDI grading for each scan was compared between graders, and any scan that received a score with an intergrader difference of greater than 1 step was reevaluated by both readers to achieve a consensus. Each eye was graded independently from the other. For each study visit, the following 4 variables were scored for each OCT-EDI scan: presence of suprachoroidal fluid, RPE integrity, ellipsoid zone integrity, and retinal thickening. Suprachoroidal fluid was graded on a scale of 0 to 2 based on the presence of a hyporeflective band between the choroid and sclera. Grade 0 denoted an absence of the band; 1, a thin region whose vertical extent was less than that of the overlying RPE; and 2, a band that was thicker than the overlying RPE in the same scan. Disruption of RPE and ellipsoid zone integrity were graded based on the percentage of disruption in a particular scan. Grade 0 was assigned to scans with an intact layer (RPE or ellipsoid zone); 1, disruption of less than 50% of the lateral extent of the scan; 2, disruption of 51% to 90% of the lateral extent of the scan; and 3, disruption of 91% or greater of the layer. Retinal thickness was graded based on the lateral extent of CME. Grade 0 was assigned in the absence of CME; 1, CME in the fovea or less than 50% of perifoveal region; 2, CME in the fovea and perifoveal regions or foveal CME greater than 50% of lateral extent of the scan; and 3, loss of retinal architecture and an end-stage appearance. This grading system generated 6 possible data points for each variable per study visit per eye (3 OCT-EDI scans per eye, each graded by 2 individuals). The scores for each OCT-EDI variable were summed. The maximum score per eye for suprachoroidal fluid was 12, because the highest score for a particular scan was 2. For the other variables (RPE, retinal thickness, and ellipsoid zone), the maximum score for a particular study eye was 18.

Clinical measures were compared with OCT-EDI measures using a Spearman rank correlation coefficient. For each OCT-EDI measure, a Bonferroni correction was made for the multiple correlations with the 4 clinical variables. Therefore, P ≤ .0125 (.05/4) was considered statistically significant. Statistical analysis was performed using commercially available software (SAS OnlineDoc, version 9.3, 2011; SAS Institute Inc).

Fourteen patients underwent a complete clinical examination and testing with OCT-EDI on the same day from July 23, 2012, through January 29, 2013, with a total of 58 individual eye examinations. All patients were white and had a positive HLA-A29 allele. Twelve patients were women, and the mean age at presentation was 58 (range, 44-68) years. Each patient included in this study underwent a mean of 2 (range, 1-5) separate examinations on each eye during this period.

Table 1 shows the frequency of clinical scores assigned to eyes in the study. The overall clinical grading ranged from 0 to 3 for flashes and CME, from 0 to 2 for vasculitis, and from 0 to 1 for vitreous haze. The most frequent score in each clinical category was 0 (no clinical activity), ranging from 57% of eye visits for flashes to 79% of eye visits for vitreous haze in treated patients.

The frequency of scores given to each eye based on OCT-EDI findings is shown in Table 2. The OCT-EDI grading ranged from 0 to 12 for suprachoroidal fluid and 0 to 18 for the other variables. Similar to the clinical scores, the most frequent score was 0 in 3 of the 4 categories (suprachoroidal fluid, RPE disruption, and retinal thickening), ranging from 52% of scans for RPE disruption to 71% of scans for retinal thickening. Only 7% of scans received a score of 0 for ellipsoid layer disruption; the most frequently assigned score in this category was 6 (19% of eyes).

The relationship between the clinical measures and OCT-EDI grading is shown in Table 3. The presence and thickness of the suprachoroidal fluid band on OCT-EDI findings correlated with the following 3 of the 4 measured clinical signs: photopsias (P = .003), vasculitis (P < .001), and vitreous haze (P < .001). The presence of CME correlated with RPE disruption (P = .002) and disruption of the ellipsoid layer (P = .008).

Birdshot chorioretinopathy is a progressive posterior uveitis that often results in significant ocular morbidity. Loss of visual acuity is attributed to factors such as CME and retinochoroidal atrophy,¹⁵ although measurement of visual acuity is not sufficient to assess degrees of visual impairment.¹ Other clinical measures that have been proposed in patients with BCR include contrast sensitivity,¹⁶ visual field testing,¹⁷ and measurements of quality of life.¹⁸

To date and to our knowledge, no objective, widely available, rapid, noninvasive tool is available to help clinicians quantify disease activity to guide therapeutic interventions. Clinical testing in BCR can be thought of in terms of tests that aid in the diagnosis and those that assess disease activity. Tests that can aid in the diagnosis include flow cytometry or polymerase chain reaction for the HLA-A29 allele, in addition to indocyanine green angiography and fundus autofluorescence imaging. Both imaging studies can reveal lesions that are not apparent clinically,^3,19 whereas fundus autofluorescence can also highlight RPE atrophy not visualized on fundus examination.⁴ Assessments that can help to monitor disease activity include fluorescein angiography, OCT, ERG, contrast sensitivity, and visual field testing. Fluorescein angiography can reveal vasculitis that is not apparent clinically, and OCT allows monitoring of CME. Serial ERG and visual field testing are used to monitor changes in retinal function. Signs of disease progression on ERG have been characterized.^20-22 Visual field changes are common, and multiple patterns of visual loss have been described.¹¹

The subjective complaints of vibrating vision and central photopsias are common in patients with BCR.¹ Although many ophthalmologists will escalate treatment based on these subjective complaints,¹ some rely on documented deterioration on the ERG or the visual field before declaring treatment failure. However, because ERG and visual field testing are typically obtained annually to biannually, patients may experience significant and irreversible loss of visual function before appropriate treatment modifications are made. These tests are difficult to interpret in view of significant variability between visits, even in patients without active retinal disease.^23,24 The use of OCT-EDI has recently uncovered the presence of suprachoroidal fluid in patients with BCR.¹³ In the present study we further examined the different OCT-EDI variables to determine their correlation with clinical grading. The association of suprachoroidal fluid with symptomatic photopsias and vibrating vision may provide an objective measure of this complaint that physicians can use to gauge disease activity more accurately and quantitatively and adjust therapy. We do not suggest that the suprachoroidal fluid is responsible for the visual symptoms, which are most likely retinal in origin; rather, this represents an association. The presence and degree of suprachoroidal fluid in this study correlated not only with the subjective complaints of central photopsias and shimmering vision, but also with the clinical signs of retinal vasculitis and vitreous haze, which suggests that the presence of suprachoroidal fluid may serve as a generalized biomarker of disease activity in BCR.

Although most patients in this study had clinically inactive disease (grade 0 in 57%-79% of eye examinations), disruption of the ellipsoid zone was encountered in most of the patients (62% of scans received a grade of ≥5), suggesting that photoreceptor disruption is widespread, even in quiescent eyes. We had expected that disruption of the ellipsoid layer would be the most relevant OCT anatomic biomarker of flashes, but in this small patient sample, we found no direct correlation, perhaps owing to our inability to differentiate recent from long-standing disruption in this layer. A longitudinal study with long-term follow-up of patients may reveal characteristics of photoreceptor changes on OCT that better correlate with flashes.

This retrospective review included a small number of patients with BCR seen during a few months. Most of the eyes received a score of 0 in every clinical category and in 3 of the 4 OCT-EDI categories, because most of the patients were receiving therapy rather than being treatment naive. The small number of eyes with disease activity is a limiting factor in this study and suggests that only a few patients drove the correlation. In addition, the study was not designed to measure a correlation between 2 eyes of a single individual at a single visit. Several of the patients had been treated with local therapy in 1 eye, so in these patients, the OCT findings did not correlate. As well, these grading scales were developed for this study only and have not been validated. The scale has not been validated and is not intended to be used in the clinic setting at this time. A single clinician at a single center performed retrospective clinical grading of all patients in this study, and a prospective validation study with defined clinical criteria is needed. In addition, BCR is a progressive disease; therefore a prospective study that follows up patients for several years through periods of activity and quiescence is needed to better characterize the sensitivity of OCT-EDI to variations in disease activity. Choroidal thickness was not evaluated in this study, because this factor is highly variable and the small sample size would not allow for a useful analysis. Future studies are needed to explore choroidal thickness and its correlation with disease activity in a larger patient population.^25,26

The presence of suprachoroidal fluid on OCT-EDI may correlate with the subjective complaints of central photopsias and vibrating vision in patients with BCR, and with other clinical features such as vasculitis and vitreous haze. Complaints of vibrating vision have been correlated with increased lesion pigmentation,¹ although the lesion grading system requires a fair amount of sophistication and time to perform. To our knowledge, no anatomic or objective corollary to the complaint of vibrating, shimmering vision has been demonstrable previously in real time. Visual field loss and ERG represent damage accrued during intervening months or years, leading to loss of neuronal function. In this study, the presence of suprachoroidal fluid on OCT-EDI correlated with the subjective complaints of central photopsias and vibrating vision. If validated in other studies, this variable may be an objective and quantifiable end point that can be used to guide clinical decision making. Future studies with larger groups of patients, prospective data collection, and prospective clinical grading with cross validation in other populations are needed. Earlier identification of patients at risk for visual loss may allow earlier intervention and avoid end-stage disease with permanent loss of vision.

The presence of suprachoroidal fluid on OCT-EDI appears to correlate with the subjective complaints of photopsias in patients with BCR as well as with other more easily assessed clinical features such as vasculitis and vitreous haze. Optical coherence tomography with EDI may be a useful tool for objectively monitoring patients with BCR.

Submitted for Publication: June 1, 2013; final revision received January 30, 2014; accepted February 9, 2014.

Corresponding Author: Andrea D. Birnbaum, MD, PhD, Department of Ophthalmology, Northwestern University Feinberg School of Medicine, 645 N Michigan Ave, Chicago, IL 60611 (andrea-birnbaum@northwestern.edu).

Published Online: May 15, 2014. doi:10.1001/jamaophthalmol.2014.877.

Author Contributions: Dr Birnbaum had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Study concept and design: Birnbaum, Fawzi, Goldstein.

Acquisition, analysis, or interpretation of data: All authors.

Drafting of the manuscript: All authors.

Critical revision of the manuscript for important intellectual content: Fawzi, Goldstein.

Statistical analysis: Rademaker.

Administrative, technical, or material support: Birnbaum, Fawzi, Goldstein.

Study supervision: Goldstein.

Conflict of Interest Disclosures: None reported.

Funding/Support: This study was supported in part by an unrestricted grant from Research to Prevent Blindness.

Role of the Sponsor: The funding source had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.