Idiopathic multifocal choroiditis (MFC) is an inflammatory disease of unknown etiology with a predilection for otherwise healthy, middle-aged, myopic females.1–4 There is considerable variation in the clinical manifestations of MFC; however, most cases are typified by multiple, punched-out chorioretinal lesions ranging from 50 μm to 350 μm involving the posterior pole and the retinal periphery.4–8 Other manifestations of MFC include vitreous or anterior chamber inflammation, peripheral curvilinear scars, and secondary choroidal neovascularization (NV).4–7 Multimodal imaging (MMI) has improved our understanding of the pathophysiologic mechanisms underlying MFC.
Spaide et al. were the first to identify an association between visual field loss and the OCT finding of outer retinal disruption in 10 eyes of seven patients with MFC.9 Limited follow-up for their cohort prevented them from commenting on the reversibility or persistence of this finding.9 Others have also described outer retinal and retinal pigment epithelium (RPE) involvement in eyes with MFC.8,10,11 Several reports describe classic chorioretinal lesions surrounded by diffuse areas of attenuation or loss of photoreceptors.10,12 In these cases, treatment response varied, with some patients demonstrating a self-limited course, whereas others showed progressive, irreversible photoreceptor damage.10,12
In an initial review of 59 eyes with MFC, all were noted to have outer retinal and/or chorioretinal atrophy. Only nine eyes had areas of hyperautofluorescence associated with ellipsoid zone (EZ) disruption identified in areas that lacked corresponding choroidal or RPE changes. Of these nine cases, the findings persisted in five during an extended follow-up, whereas four experienced late resolution of hyperautofluorescence associated with visual recovery. Based on our initial observations, a multimodal imaging analysis was conducted aimed at identifying biomarkers for potential restoration of visual function in similar cases that might help inform future treatment strategies for MFC.
Patients and Methods
This study was approved by the Western Institutional Review Board (Olympia, WA). It complied with the Health Insurance Portability and Accountability Act of 1996 and followed the tenets of the Declaration of Helsinki.
This was a retrospective review of patients with MFC examined by two physicians (LAY and KBF). Eligible patients were identified by a search through the billing records between January 2012 and December 2016. Follow-up was then obtained through June 2018. MMI and medical records were used to confirm the diagnosis of MFC. Patients with a diagnosis of multiple evanescent white dot syndrome (MEWDS), acute zonal occult outer retinopathy (AZOOR), and other white dot syndromes were excluded based on MMI review. All patients included in this study had undergone a comprehensive systemic workup including tests for sarcoidosis, tuberculosis, syphilis and, in some, histoplasmosis.
All eligible patients had color photography, fundus autofluorescence (FAF), fluorescein angiography (FA), and spectral-domain optical coherence tomography (OCT) available for review. Color photographs, FAF, FA, and indocyanine green angiography (ICGA) images were obtained using the Optos 200Tx (Optos, Dunfermline, UK), and/or Topcon TRC 50IX fundus camera (Topcon Imagenet, Tokyo, Japan), and/or Heidelberg Spectralis HRA+OCT (Heidelberg Engineering, Heidelberg, Germany). The peripheral retina was evaluated using Optos ultra-widefield imaging or by montaging Topcon images. Data and alternative images from some of the patients included in this study have been previously published, though all images in these figures are unique.8,13
Macular OCT was obtained on the Heidelberg Spectralis HRA+OCT (Heidelberg Engineering, Heidelberg, Germany; volume scan between 20° × 15° and 30° × 25° in dimensions). Eye tracking and image registration functions were enabled for all image acquisitions. Enhanced depth imaging OCT scans were used to evaluate choroidal features. Some eyes were imaged with swept-source OCT (Atlantis or Triton OCT; Topcon, Tokyo, Japan). This OCT instrument is not U.S. Food and Drug Administration- approved for sale in the United States.
Imaging data from each visit were evaluated. Alterations to the choroid, RPE, and retina were qualitatively studied by correlating MMI data. In select cases, SD-OCT was correlated with registered confocal scanning laser ophthalmoscopy, FAF, and FA. The OCT B-scans were analyzed from hyperautofluorescent lesions seen on FAF, referred to as the lesional zone, and to the junction between the hyperautofluorescent lesions and the surround area, known as the paralesional zone (Figure 1). The integrity of each hyperreflective outer retinal band (RPE, EZ, external limiting membrane [ELM], outer nuclear layer [ONL], and the outer plexiform layer [OPL]) was analyzed, with loss of RPE defined as an attenuation of the RPE/Bruch's membrane complex associated with choroidal hypertrasmission.14 The EZ and the ELM were both graded as intact or disrupted and the shape of ELM descent at the lesion border was graded as flat, curved, reflected, or scrolled.15 The thickness of the ONL and OPL were measured using the caliper tool within the Heidelberg software. The slope of the OPL was recorded as absent or present.15
Near infrared fundus autofluorescence of the right eye of a patient with persistent hyperautofluorescence (A) with the lesion of interest demarcated in green and the corresponding optical coherence tomography (OCT) image (B). This demonstrated the loss of outer retinal layers, including the ellipsoid zone (EZ), external limiting membrane (ELM), and outer nuclear layer (ONL) in the lesional zone (C; arrowhead to arrowhead) and a flat ELM descent and outer plexiform layer (OPL) slope at the junctional zone (C; arrows). In D, we observe the border of a hyperautofluorescent zone in a patient with transient hyperautofluorescence. The OCT (E) demonstrates a disruption of the EZ in the lesional area (F; arrowhead to arrowhead) without ELM descent or OPL slope at the junction between the lesional and paralesional areas (E; green vertical line).
Fifty-nine eyes from 37 patients (29 women) with a mean age of 38.5 years (median age: 34.0 years; range: 17 years to 76 years) at disease presentation met eligibility criteria. Thirty-three patients self-identified as white, two as Asian, one as African-American, and one as Hispanic. Demographic and clinical data from the entire cohort are summarized in Tables 1 and 2. Areas of outer retinal and/or chorioretinal atrophy with or without associated regions of persistent hyperautofluorescence were present in all eyes, with nine eyes (15.3%) of six patients having regions of hyperautofluorescence associated with EZ disruption but without choroidal or RPE loss visible on MMI. For these nine eyes, the hyperautofluorescence and EZ changes were reviewed at each visit and were found to either be transient (Group 1) with resolution occurring within 1 to 4 months (mean: 73.8 days; range: 28 days to 125 days) or to persist (Group 2) for as long as 8 years. Group 1 included four eyes (6.8%) of 3 patients (mean follow-up: 1.3 years; Figures 2 and 3). Group 2 included five eyes (8.5%) of three patients (mean follow-up: 4.6 years; Figure 4). For the purpose of this analysis, the 50 eyes of the remaining 31 patients are referred to as having “typical MFC.”
Summary of Demographic, Clinical, and Treatment Features for the Entire Cohort: Eyes With Transient Hyperautofluorescence Associated With EZ Loss (Group 1) and Eyes With Persistent Hyperautofluorescence and EZ Loss (Group 2)
Summary of Age, Visual Acuity Findings, and Duration of Disease for the Entire Cohort:Eyes With Transient Hyperautofluorescence Associated With EZ Loss (Group 1) and Eyes With Persistent Hyperautofluorescence and EZ Loss (Group 2)
Fundus autofluorescence (FAF) (A) of the right eye of a patient with transient hyperautofluorescence at first presentation (FAF as seen in Figure 1D) demonstrating widespread hyperautofluorescence with hypoautofluorescent lesions near the disc (B; arrow) that correspond with the lesions seen on the color fundus photo (B). The optical coherence tomography (OCT) through the hyperautofluorescent area (C) demonstrates disruptions of the ellipsoid zone (EZ) and the external limiting membrane (ELM) with the junction between the lesional and paralesional regions demarcated with a vertical line. Figures D-F are from 1 month later. One month after the initiation of oral steroids, there is partial resolution of the hyperautofluorescence (D), no change on color fundus imaging (E) and restoration of the EZ and ELM on the OCT (F) when compared with same cut from 1 month prior (C).
Color photographs (A, C) and fundus autofluorescence (FAF) (B,D) of a patient with transient hyperautofluorescence, which resolved by 9 weeks after initial presentation. There is very little evidence of change in the color photos, but the FAF demonstrates marked resolution of the abnormal hyperautofluorescence seen at presentation.
Fundus autofluorescence (FAF) (A) of the left eye of a patient with persistent hyperautofluorescence from 2013, with a multizonal distribution of the hyperautofluorescent lesions. The color fundus image (B) shows retinal pigment epithelium (RPE) changes inferior to the disc. Optical coherence tomography (OCT) through the patchy lesion (C) demonstrates a loss of ellipsoid zone and outer retinal layers (arrow). There is a descent of the outer plexiform layer (OPL) at the junctional zone (vertical line). Figures D-F are from 3 years later, in 2016. There is a subtle reduction in the hyperautofluorescence in the nasal aspect of the patchy lesion (D) which otherwise retains a consistent size and distribution. The color image demonstrates some RPE migration nasal to the disc (E) but is otherwise unchanged. The OCT remained stable (F).
When compared to the typical MFC cohort, patients with transient hyperautofluorescence due to EZ disruption (Group 1) were numerically more likely to have received systemic treatment and less local treatment than those lacking this finding. All three patients in Group 1 and 1 of 3 patients in Group 2 received systemic treatment, compared with only nine of 31 in the typical MFC cohort. As a corollary, the typical MFC cohort received more local treatment, including intravitreal anti-vascular endothelial growth factor (VEGF) injections (18 of 50 eyes; 36%), photodynamic therapy (six of 50 eyes; 12%), focal laser (four of 50 eyes; 8%) and sub-Tenon's triamcinolone (nine of 50 eyes; 18%). In the nine eyes in Groups 1 and 2, none were treated with anti-VEGF therapy. One of the patients in Group 2 had a presumed macular CNV that was treated with a vitrectomy and submacular surgery when she was 13 years old. Since then, she has received no additional therapy in that eye.
There was a trend toward improved vision in Group 1 when comparing vision at first versus last visit, as three out of four eyes (75%) experienced an improvement in vision, and the fourth remained stable at 20/20 (logMAR 0.0). Despite essentially the same mean, median and range of initial visions in Group 1 and 2, none of the eyes in Group 2 had an improvement in vision at the final visit. Three of five eyes (60%) in Group 2 had a reduction in vision, and the other two eyes (40%) remained stable. Both of these groups differed from the typical MFC patients, who had essentially stable vision with a mean of 20/45 (logMAR 0.35) at first presentation compared with 20/41 (logMAR 0.31) at last visit. This was consistent with the vision of the group in total (Table 1).
There were no recurrences in the nine eyes of the patients with hyperautofluorescence associated with EZ disruption, and no alterations in the choroid or RPE (Groups 1 and 2) during the period recorded. This differs from the recurrence rate in the 50 eyes in the typical MFC cohort, in which 12 (24%) showed evidence of evolution or recurrence during the period in which data were available.
The three patients with the transient hyperautofluorescence associated with EZ disruption (Group 1) also had a disruption of the ELM in the lesional zone but without ELM descent at the junction. The transient lesions had a mean ONL thickness of 59 μm (median: 61 μm) and a mean OPL thickness of 18 μm (median: 18 μm). This was notably different when compared with the three patients who had persistent hyperautofluorescence associated with ELM disruption (Group 2), two of whom had no visible ONL or OPL. Only one eye in Group 2 had a visible ONL or OPL, but the thickness of both layers was significantly reduced (10 μm and 15 μm, respectively) as compared with those in the transient lesions. At the junction of the persistent hyperautofluorescent zone, a flat ELM descent and an OPL slope was noted in all cases.
Idiopathic MFC is a chronic, progressive inflammatory disease that involves the outer retina and choroid and may present with both hypo- and hyperautofluorescent lesions.1,8,16,17 Hyperautofluorescence is commonly seen as a marker of disease activity, appearing either as a new lesion, at the margins of previously quiescent lesions, or prior to the formation of macular NV.2,12,18–22 Over time, or following treatment with immunosuppressive agents, these hyperautofluorescent lesions typically fade in intensity.18–20,22,23 Hypoautofluorescent lesions frequently correlate to areas of RPE atrophy and outer retinal atrophy.2,12,18–21
The presence of hyperautofluorescence as a proxy for disease activity is not exclusive to MFC; however, with multimodal imaging, especially FAF, it becomes possible to discriminate MFC from other entities that may present with a similar clinical picture, including AZOOR, MEWDS, and AMPPE.20,23–26 Our patients may have features similar to these other entities, especially AZOOR, but in the latter, the zone of hyperautofluorescence is thought to be secondary to accumulation of fluorescent material in the subretinal space.27–33 In contrast, the areas of hyperautofluorescence in our cohort correlated with sites of selective disruption of the ellipsoid zone and the interdigitation zone without associated disturbances in the RPE or choroid.32 The clinical stability of this subset of MFC patients also helps to differentiate it from AZOOR, as in the latter though there may be a trizonal FAF, there is usually progression of the disease over time.32
Four of the five eyes with the persistent hyperautofluorescence and EZ disruption described in this study had hyperautofluorescent rings (Figures 1 and 3) similar to those seen in a number of other diseases including retinitis pigmentosa, enhanced S-cone syndrome, and autoimmune retinopathy.34–37 The etiology of these findings range from thinning and disorganization of the outer retina and likely increased deposition of lipofuscin in the RPE cells in S-cone syndrome to increased photoreceptor phagocytosis and resultant lipofuscin accumulation within the RPE in others.34,36 Hyperautofluorescent rings are not unique to MFC, but the presence of stable hyperautofluorescent rings without associated RPE changes or progression seen in three of our patients are a novel finding in MFC.
Multimodal imaging is invaluable not only for differentiating MFC from other diseases, but also for helping to discriminate one form of MFC from another. The classic features of MFC include the presence of multiple chorioretinal lesions, which can be located in the macula or the periphery, peripheral curvilinear streaks, peripapillary changes (either atrophy or hyperpigmentation), choroidal neovascularization, epiretinal membranes, cystoid macular edema, and both anterior and posterior uveitis.1,5,38–40 The proposed etiology of the patterns of autofluorescence in MFC, especially hyperautofluorescence, is varied.10 For example, Freund et al. suggested that hyperautofluorescence in areas of outer retinal disruption may be the result of a window defect.21 They proposed that reduced photopigment density in areas of affected outer segments reduces absorption of excitation light and fluorescent wavelengths originating from RPE lipfuscin.21 Mantovani et al. reported that EZ disruption corresponded with hyperautofluorescent lesions.22 Spaide et al. found that inactive lesions showed either absent or normal autofluorescence; active NV was marked by a ring of hyperautofluorescence, and late NV had a varied phenotype, ranging from hypo- to hyperautofluorescence.11 Thus, there exists a range of explanations for hyperautofluorescence patterns in MFC, and it is likely that each of these anatomic alterations is related to different clinical courses and visual outcomes. These papers underscore the clinical value of multimodal imaging in the recognition and management of MFC.
In this retrospective cohort analysis of 59 eyes with MFC, we found that nine eyes (15.3%) demonstrated a pattern of hyperautofluorescence on FAF imaging associated with EZ disruption without observable alterations at the level of the choroid or RPE. Of these nine eyes, five showed persistent outer retinal loss and hyperautofluorescence with thinning of the ONL, whereas four demonstrated resolution of the hyperautofluorescence with restoration of the EZ and a preserved ONL thickness. The persistent lesions were notable in that they contrast with the prevailing understanding that hyperautofluorescent lesions are a mark of disease activity and will thus resolve with disease quiescence (Figure 1).12,18–21 Other studies have described outer retinal atrophy without choroidal involvement in MFC, but the details of these have never been parsed out, nor the prevalence reported.8,10
A strength of this study is the wide range of imaging modalities used to evaluate the morphologic characteristics of outer retinal changes, which we believe allows for a more nuanced description of MFC and provides novel insights regarding the pathophysiology of the disease. By registering confocal scanning laser ophthalmoscope-derived FAF images with OCT we were also able to precisely examine the correlation between FAF changes and alterations to outer retinal bands on OCT. In addition, although a small cohort limits statistical significance, there was a trend indicating that patients with transient hyperautofluorescence have stable or improved vision after resolution of the out retinal loss, whereas those with persistent hyperautofluorescence are likely to lose vision and, at best, will preserve the vision they have. A close study of a patient's OCT can aid in differentiating these two groups. Because only patients with preservation of the OPL and ONL appear to have potential for visual recovery, our personal experience is that immunosuppression in these cases has the potential to help restore outer retinal function and possibly prevent progression to a more complete form of irreversible outer retinal atrophy. Although we lack sufficient data to make treatment recommendations based on anecdotal cases, we believe our colleagues should consider these two distinctly different patterns when evaluating and treating similar patients.
We acknowledge several limitations of this study, including its retrospective design and small sample size. Although there were some differences in treatment paradigms between the groups, the small number of patients limits the conclusions one may draw from this. Although we did not observe any changes in the FAF patterns after a mean of 4.6 years of follow-up in these patients, it is possible that, with longer review, changes in FAF patterns might become apparent. We also cannot entirely exclude that RPE changes below the resolving power of the imaging modalities used in this report contributed to the hyperautofluorescence changes seen in these eyes. Furthermore, biochemical changes within RPE cells, such as upregulation of fluorophores in the absence of structural changes, may have accounted for the hyperautofluorescence. Postmortem histologic evaluation of MFC cases will help address these limitations and improve our understanding of this disease.