Paper MER
Paper MER
Paper MER
ANDREA GOVETTO, ROBERT A. LALANE, III, DAVID SARRAF, MARTA S. FIGUEROA, AND
JEAN PIERRE HUBSCHMAN
E
PURPOSE: To describe the presence of continuous PIRETINAL MEMBRANE (ERM) FORMATION IS A COM-
ectopic inner foveal layers associated with epiretinal mon retinal condition characterized by fibrocellular
membranes (ERMs) and to present a new optical coher- proliferation at the vitreoretinal interface, above the
ence tomography (OCT) staging system of ERMs. internal limiting membrane, with a prevalence that ranges
DESIGN: Retrospective multicenter observational case between 2.2% and 28.9%, which increases with aging.1–3
series. ERMs are typically idiopathic, but can also be associated
METHODS: Clinical charts and spectral-domain OCT with retinal vascular or inflammatory diseases, trauma,
images of 194 eyes of 172 consecutive patients diagnosed tumors, intraocular surgery, or retinal detachment.4 The
with ERMs were reviewed and analyzed. precise pathophysiology of this clinical entity is not
RESULTS: The presence of continuous ectopic inner completely defined, although the proliferation of hyalo-
foveal layers was identified in 63 out of 194 eyes cytes in the setting of anomalous posterior vitreous detach-
(32.5%) and this morphology was significantly associated ment and vitreoschisis has been suggested as a possible
with lower visual acuity. ERMs were divided into 4 mechanism in the early development of idiopathic ERMs.4
stages. Stage 1 (43 out of 194 eyes, 22.1%) ERMs were Various classification schemes have been proposed,5–7
mild and thin and a foveal depression was present. Stage including the original categorization by Gass, the most
2 (88 out of 194 eyes, 45.4%) ERMs were associated widely used.7 Owing to dramatic improvements in image
with widening of the outer nuclear layer and loss of the resolution and data acquisition speed, spectral-domain
foveal depression. Stage 3 (51 out of 194 eyes, 26.3%) optical coherence tomography (SDOCT) has driven a
ERMs were associated with continuous ectopic inner transformative change in the study of ERMs, providing
foveal layers crossing the entire foveal area. In stages 1, rapid cross-sectional imaging of the retina with near-
2, and 3 all retinal layers were clearly defined on OCT. histologic detail. OCT has become the gold standard in
Stage 4 ERMs (12 out of 194 eyes, 6.2%) were thick the evaluation of ERMs, but despite its universal applica-
and associated with continuous ectopic inner foveal tion, a globally accepted OCT-based classification system
layers. In addition, retinal layers were disrupted. Visual is still lacking.8
acuity progressively declined from stage 1 through stage In recent years OCT studies of the foveal microstructure
4 (P < .001). associated with ERMs have flourished in order to identify
CONCLUSIONS: The presence of continuous ectopic in- the anatomic changes that may cause loss in visual acu-
ner foveal layers in ERMs is a newly described OCT ity,9–17 including disruption of the inner segment
finding associated with significant vision loss and is an ellipsoid zone and photoreceptor outer segments.9–12
essential element of a novel OCT-based grading scheme More recently attention has shifted to the study of the
of ERMs that may influence visual prognosis. (Am J inner retinal anatomy. In this regard, various reports have
Ophthalmol 2017;175:99–113. Ó 2016 Elsevier Inc. shown a significant association of the thickness of the
All rights reserved.) inner retinal layers with vision loss in eyes with ERM.13–17
This study performed an in-depth SDOCT analysis of
the development and evolution of ERMs and identified
novel morphologic features that impacted visual acuity
Supplemental Material available at AJO.com.
prognosis and were critical in the development of a new
Accepted for publication Dec 7, 2016. OCT-based staging system.
From the Retina Division (A.G., R.A.L., J.P.H.) and Retinal Disorders
and Ophthalmic Genetics Division (D.S.), Stein Eye Institute, University
of California Los Angeles, Los Angeles, California; Greater Los Angeles
VA Healthcare Center, Los Angeles, California (D.S.); and Retina
Division, Ophthalmology Department, Ramon y Cajal University METHODS
Hospital, Madrid, Spain (M.S.F.).
Inquiries to Andrea Govetto, Retina Division, Stein Eye Institute,
University of California Los Angeles, 100 Stein Plaza, Los Angeles, CA A RETROSPECTIVE, OBSERVATIONAL, AND CONSECUTIVE
90095-7002; e-mail: a.govetto@gmail.com chart review of patients diagnosed with ERM and seen by
be clearly identified owing to significant alterations in the layers anomalously crossing the central foveal area. The
normal retinal anatomy. The presence of a foveal depres- foveal depression was absent, and widening of the outer nu-
sion was only present in 43 out of 194 eyes (22.1%). clear layer was often less pronounced if compared with
These specific morphologic features were essential in stage 2 ERMs. In stage 3 ERMs all retinal layers were
developing a 4-stage classification scheme of ERMs, with clearly identified on OCT (Figure 3, Center-bottom;
clinical and prognostic relevance, as illustrated in Figure 3. Supplemental Video, available at AJO.com).
Stage 1 (43 out of 194 eyes, 22.1%) was defined as the Finally, stage 4 (12 out of 194 eyes, 6.2%) was defined as
presence of a mild ERM with negligible morphologic or an ERM complicated by significant retinal thickening and
anatomic disruption. All retinal layers were clearly identi- remarkable anatomic disruption of the macula. Continuous
fied with easily distinguishable boundaries (Figure 3, Top). ectopic inner foveal layers extending from the INL and IPL
In all stage 1 eyes, the foveal depression was identified, and crossing the entire foveal area were again identified.
although it was shallower than the unaffected fellow eye Retinal layers were noted to be significantly distorted,
in patients with unilateral ERM. disorganized, and not clearly identified with OCT
Stage 2 (88 out of 194 eyes, 45.4%) was defined as the (Figure 3, Bottom).
presence of ERMs associated with more progressive retinal In the ERM staging process 2 masked graders indepen-
distortion. While the foveal depression was lost and a char- dently evaluated all the OCT scans of the included
acteristic stretching of the outer nuclear layer was present, ERMs, and the interobserver agreement was 89.12%
all retinal layers were defined and clearly identified with with a Cohen kappa coefficient of 0.82 (excellent agree-
OCT (Figure 3, Top-center). ment).
Stage 3 (51 out of 194 eyes, 26.3%) was defined as the BCVA correlation was performed and was greater in
presence of an ERM with continuous ectopic inner foveal stage 1 (0.025 6 0.06 logMAR, 20/21 Snellen equivalent)
Age (y) 69.1 6 7.5 70.8 6 9.4 70.2 6 8.7 68.5 6 8.7 .664a
Central foveal thickness (mm) 321.6 6 60.1 412.4 6 49.8 497.37 6 64 639.25 6 93.5 <.001a
Outer nuclear layer thickness (mm) 169.89 6 66 267.47 6 58 230.77 6 61 - <.001a
Ectopic inner foveal layer thickness (mm) - - 161.55 6 63 - -
Cystoid macular edema 0/43 (0%) 2/88 (2.3%) 7/51 (13.7%) 5/12 (41.7%) <.001b
Ellipsoid disruption 1/43 (2.3%) 10/88 (11.4%) 11/51 (21.6%) 9/12 (75%) <.001b
Cotton ball sign 3/43 (7%) 21/88 (23.9%) 11/51 (21.6%) 0/12 (0%) .034b
BCVA, logMAR (Snellen) 0.02 6 0.6 (20/21) 0.14 6 0.13 (20/27) 0.33 6 0.17 (20/43) 0.61 6 0.26 (20/81) <.001c
and 2 (0.14 6 0.13 logMAR, 20/27 Snellen equivalent) regression model (P ¼ .001). This significance was further
and significantly better than eyes in stage 3 (0.33 6 0.17 confirmed with a multivariate linear regression model
logMAR, 20/42 Snellen equivalent) and 4 (0.61 6 0.26 controlling for other possible causes of reduced BCVA,
logMAR, 20/81 Snellen equivalent), with a P value such as tractional CME, central foveal thickness, and
<.001 (Table 2). disruption of the ellipsoid zone, suggesting that the sole
The presence of ectopic inner foveal layers was signifi- presence of ectopic inner foveal layers may be an indepen-
cantly associated with lower BCVA in a univariate linear dent risk factor for lower visual acuity (P ¼ .001).
Statistically significant differences in central foveal in stages 1 and 2 and highest in stages 3 and 4, with statis-
thickness were encountered between the 4 subgroups. tically significant differences between the 4 groups
Retinal thickness was lowest in eyes with stage 1 ERMs (P < 2.001, Table 2).
and was greatest in eyes with stage 4 ERMs (P < .001, OCTA was performed in 9 eyes with stage 1, 18 eyes with
Table 2). stage 2, and 14 eyes with stage 3 ERMs. In stage 1 ERMs
The prevalence of tractional CME and ellipsoid disrup- variations in the FAZ size were minimal, while in stage 2
tion increased according to the ERM stage and was lowest and stage 3 ERMs the FAZ was significantly reduced if
compared with the unaffected fellow eye (Figure 4, Top, progressed directly to stage 3. No cases of ERM regression
Center, Bottom). Notably, this finding was more pro- were noted.
nounced in stage 3 ERMs, in which the FAZ was nearly The progression to more advanced stages was associated
entirely absent in certain cases (Figure 4, Bottom). with a substantial and statistically significant drop in
Macular-foveal capillaries were not identified in any of BCVA in all cases, which was particularly evident in those
these cases. OCTA was also performed in 2 stage 4 membranes that progressed from type 2 to type 3
ERMs, but the scans were excluded from the analysis owing (P ¼ .012), as shown in Figure 7 and Table 3.
to poor imaging quality. Interestingly, during the follow-up period the thickness
The majority of eyes with ERMs in which surgery was of the outer nuclear layer increased significantly in
deferred (108 out of 131, 82.5%) failed to show signs of eyes with ERMs that evolved from type 1 to type 2
morphologic progression to later stages during the mean (P ¼ .008), while in eyes with type 2 ERMs that
follow-up of 22.2 6 18.3 months (Table 3). Incidence progressed to type 3, significant thinning of the same layer
of tractional CME and ellipsoid disruption did not was identified (P ¼ .011), presumably owing to the devel-
change significantly in the 4 groups (P ¼ .5 and opment of ectopic inner foveal layers (Table 4). The
P ¼ .125, respectively), but in stage 2 ERMs the pres- thickness of the ectopic inner foveal layers was not
ence of the cotton ball sign was significantly higher at assessed in stage 4 ERMs, as the measurements were not
the end of the follow-up period (P ¼ .02). Functionally, considered reliable.
even in the absence of anatomic progression a slight but
statistically significant decrease in visual acuity was
registered in ERMs stage 2 (P < .001) and 3
(P ¼ .018) at the end of the follow-up period, as DISCUSSION
demonstrated in Table 3.
Anatomic progression to more advanced stages was THE TRACTIONAL STRESS INDUCED BY ERMS CAN INVOLVE
described in 23 out of 131 ERMs (17.5%). Specifically, 9 all retinal layers, causing a broad clinical spectrum of macular
out of 32 stage 1 ERMs (28.1%) progressed to stage 2 complications that ranges from increased retinal thickness
(Figure 5, Top), 10 out of 77 stage 2 ERMs (13%) with or without macular schisis to the formation of lamellar
progressed to stage 3 (Figure 5, Center), and 2 out of 17 and full-thickness macular defects.21–25 Causes of visual
stage 3 ERMs (11.1%) progressed to stage 4 during the reduction include macular distortion and/or edema,26 but
follow-up (Figure 5, Bottom). In 1 case, it was possible to an accurate prediction of vision loss may be elusive.
record the gradual evolution of an ERM through stages 1, Factors such as ellipsoid zone disruption9–12 and
2, and 3 (Figure 6). Finally, in 1 case a stage 1 ERM alterations in the photoreceptor outer segment length27
have also been correlated with lower preoperative and post- irregularity index as a valid tool for postoperative visual
operative visual acuity. prognosis in eyes with ERMs.17
More recently, the role of the inner retinal layers in vi- The present study confirms that the inner retinal layers of
sual acuity loss has been studied more closely, as these the macula may be especially sensitive to tractional stress
more proximal layers may be the primary affected site of and illustrates that ERM formation may significantly alter
ERM-associated mechanical stress. Ganglion cell–inner the inner foveal microanatomy. We propose that the chronic
plexiform layer complex thickening was significantly asso- anteroposterior and centripetal traction caused by the ERM
ciated with visual acuity reduction in various reports,15,16 may induce the displacement and reorganization of the inner
while some authors have proposed the inner retinal retinal layers, creating a continuous floor of inner retinal
noted in the inner nuclear layer. The ellipsoid zone is preserved. In this case the visual acuity remained stable. (Center) Case 2: pro-
gression from stage 2 to stage 3 epiretinal membrane. After 3 years of follow-up, ectopic inner foveal layers continue across the entire
foveal area. The outer nuclear layer appears thinner, and is no longer in contact with the internal limiting membrane. The ellipsoid
zone is preserved. The visual acuity decreased from 20/20 to 20/40 Snellen equivalent. (Bottom) Case 3: progression from stage 3 to
stage 4 epiretinal membrane. After 3 years of follow-up, the retinal layers are disrupted, with ill-defined borders. The epiretinal mem-
brane is thicker, with clear signs of traction affecting the underlying retina. There are alterations of the external limiting membrane
and the ellipsoid zone. Visual acuity remained stable.
occurs and, in such context, the elongation and packing of and we were able to confirm the presence of continuous
cones and other adaptation phenomena may negate the ectopic inner foveal layers in all scans through the central
need for a clear optical zone in the fovea. fovea. Moreover, the eye tracking system was applied in all
The present study allows the analysis of this issue from those eyes with progression to stage 3 ERMs and illustrated,
a different perspective, as in our case the presence of in all cases, the development of the ectopic inner foveal
ectopic inner foveal layers occurred in patients with a layers across the fovea in a precisely aligned B-scan at base-
presumably physiological pit prior to ERM formation, line and follow-up.
in which the foveal cones developed without any over- In conclusion, this report gives new insights into the
lying retinal tissue. In stage 3 and 4 ERMs, the ectopic pathophysiology of ERMs and the visual significance of
inner foveal layer interposed between the afferent light the optic pit by describing a previously unreported clinical
and the photoreceptors, and may obstruct or degrade entity referred to as ectopic inner foveal layers. The pres-
the visual image projected on the cone photoreceptors. ence of this acquired condition, the pathophysiology of
Moreover, the chronic displacement of the retinal layers which is still poorly defined, was associated with significant
may cause photoreceptor damage and deformation, vision loss and is the centerpiece of a new OCT-based
contributing to the visual alterations and distortion in grading scheme of ERMs that may influence the surgical de-
eyes with ERM. cision in these patients.
The analysis of surgical results in those eyes that un- Larger, prospective studies with longer follow-up are
derwent pars plana vitrectomy and ERM peel was not necessary to support our hypothesis and reduce bias, while
the objective of this descriptive study, and will be further investigations are required for a better understand-
assessed in a future study. However, preliminary evidence ing of the pathologic mechanisms leading to the forma-
suggests that the ectopic inner foveal layers may persist tion of the ectopic inner foveal layers. Improvements in
even after surgical ERM removal, as illustrated in OCT resolution and OCTA technology will be useful
Figure 9. tools for this purpose. Furthermore, the analysis of surgical
Limitations of this study include its retrospective nature, results will be critical to explore possible postoperative
the limited follow-up, and the lack of high-density macular functional and anatomic differences between the 4 stages
raster scanning and OCTA for all cases. It is possible that of ERMs.
the normal foveal anatomy may have been missed by the We hope that our results and proposed staging scheme
standard 20 3 15-degree macular raster, biasing the ERM will encourage future study so as to achieve a broader
staging process. High-density tracking was, however, consensus, and ultimately improve the management of
performed in stage 3 ERMs in a significant subset of eyes ERMs.
FUNDING/SUPPORT: THIS WORK WAS SUPPORTED BY AN UNRESTRICTED INSTITUTIONAL GRANT FROM RESEARCH TO PRE-
vent Blindness (RPB), New York, New York, USA. Financial Disclosures: David Sarraf: consultant for Bayer (Leverkusen, Germany), Genentech
(San Francisco, CA, USA), and Optovue (Freemont, CA, USA). Research grants from Allergan (Parsippany-Troy Hills, NJ, USA), Heidelberg (Heidel-
berg, Germany), Optovue (Freemont, CA, USA), and Regeneron (Tarrytown, NY, USA). Speaker bureau for Optovue and Novartis. Marta S. Figueroa:
consultant for Alcon (Fort Worth, TX, USA), Novartis (Basel, Switzerland), Allergan (Parsippany-Troy Hills, NJ, USA), Bayer (Leverkusen, Germany).
Jean Pierre Hubschman: consultant for Alcon (Fort Worth, TX, USA), Pixium-Visium (Paris, France), Allergan (Parsippany-Troy Hills, NJ, USA), and
Avalanche Biotechnologies (Menlo Park, CA, USA). The following authors have no financial disclosures: Andrea Govetto and Robert A. Lalane III.