The bifurcation and bump signs: optical coherence tomography (OCT) findings overlying neurosensory retinal detachment


 Purpose To expand the description of a common OCT finding of outer plexiform layer (OPL) splitting overlying neurosensory retinal detachment from various causes, along with a thickened outer plexiform bump at the transition from attached to detached retina.Methods etrospective review of our practice’s image teaching collection and database of retina journal articles looking for representative examples of OPL splitting overlying macular detachment. Each of our patient’s scans were then analyzed to see if the location or angle of detachment influenced these findings.Results We analyzed 12 eyes (12 patients) with splitting of the OPL within paracentral detached retina (the “bifurcation sign”) from various causes. A localized thickening of the OPL (the “bump sign”) was present in 6 of these eyes. Although explicitly described in only one publication, the bifurcation sign could be found in numerous prior publications within our journal database. The bump sign, also explicitly described in one prior publication, was also seen in most of these published cases. Surprisingly, these findings appeared unrelated to either the angle of detachment or nasal versus temporal location. We then synthesized our findings with the current literature to help elucidate what this can teach us about current OCT nomenclature for the outer plexiform and outer nuclear layers.Conclusion OPL changes are commonly found with retinal detachment. We suggest a modified classification of the OCT OPL and outer nuclear layers which is only applicable when the normal OCT anatomy is altered either by the OCT technician or overlying retinal detachment.


Introduction
In 2009, Ahlers et al showed that 89% of 18 eyes with acute central serous chorioretinopathy (CSC) had diffuse hyperre ectivity in the outer nuclear layer (ONL) and outer plexiform layer (OPL) when imaged with high de nition optical coherence tomography (OCT). [1] Ahlers et al were unsure whether these were real morphologic changes or secondary to altered re ectivity from changes in the orientation of the neurosensory retina. Otani et al [2] and Lujan at al [3] independently and simultaneously more formally suggested that the angle of the OPL with the incident OCT laser beam alters the OPL smoothness and thickness in the normal macula. The normal nasal OPL appears thicker with a roughened outer margin and temporally has a thinner and smoother outer margin. This re ectivity is reversed nasally and temporally depending on how the angle of the incident OCT beam is oriented by horizontally displacing the OCT beam's entry position toward the nasal or temporal edges of the pupil. Otani et al felt that this phenomenon was due to the OPL being a heterogenous structure: the outer two-thirds consisting of the Henle's layer component (photoreceptor axons interleaved with Muller cell processes) and the inner one-third consisting of the synaptic region. When the incident beam strikes Henle's layer more obliquely, only the thin synaptic layer is visualized but when the incident beam strikes Henle's layer more perpendicularly, the visualized OPL becomes thicker. Lujan et al also noticed that signi cant changes in Henle's layer re ectivity appeared in two patients with macular pathology from age-related macular degeneration and CSC.
Ouyang et al later expanded on these ndings, identifying six different phenotypes of Henle's layer variable thickness and re ectivity: bright, columnar, dentate, delimited, indistinct and dark. [4] The delimited variant was shown in their gure Id, where Henle's layer was iso-hypore ective with the ONL with a hyperre ective outer boundary.
We recently noticed this delimited OPL appearance, which we call the "bifurcation sign", in many of the OCTs of our patients with CSC ( Figure I). We also noticed that many of our eyes also contained a thickening of the OPL at the in ection from attached to detached retina, which we call the "bump sign".
We therefore searched through our practice's retinal image and journal article portable document format (PDF) teaching collection to look for similar cases to try to better characterize these ndings. We then synthesized Disorders that routinely alter the neurosensory retinal anatomy, such as optic nerve pit maculopathy, Vogt-Koyanagi-Harada syndrome, and diabetic traction detachment, were excluded since this would have confounded our evaluation of the OPL. No attempt was made to include every such example in our collection.
From this review of approximately 1700 images, we selected 12 representative eyes (12 patients) with macular detachment from various causes, including CSC (cases 1, 2, 4, 5, 7, 8, 9, 10 and 12), choroidal nevus (case 11), uveal malignant melanoma (case 3) and loculated uid following retinal reattachment surgery (case 6). For cases 1, 2, 3, 4, 5, 6, 8, 10 and 11 we reviewed our practice's retinal image teaching collection. Cases 7, 9 and 12 were donated by Dr. Richard Spaide with similar selection criteria. Most reviewed and selected images were from eyes with CSC since this was by far the most common entity in our collection that met our selection criteria. All patients gave written permission to use their images for educational purposes.
Each of the 12 patient B-scans was analyzed for the following: angle of detached retina with respect to the retinal pigment epithelium (RPE) at the transition from attached to detached retina (retinal plane at the level of the OPL, RPE plane at the RPE apex), nasal versus temporal location of the OPL splitting, whether the OPL splitting was located on the steeper or shallower slope of the detachment, and the presence of an OPL bump. If the area of transition from attached to detached retina was outside the eld of view of the OCT, this was estimated by extrapolating the planes of visible RPE and neurosensory retina until they intersected at the presumed transition point.
As part of the Retina Rocks collection, our practice also maintains a database of over 4,500 key retina-related journal article PDFs led by disease state. We browsed through publications dealing with exudative or rhegmatogenous retinal detachment looking for OCT images showing OPL splitting within the detachment.

Case Reports
Figures for all cases were labelled to illustrate the various OPL and ONL details of interest, including a single OPL in attached retina (red arrow), a thickened OPL bump at the in ection from attached to detached retina (yellow arrow), a split hyperre ective OPL surrounding a hypore ective OPL region (white arrows), and a single OPL band overlying detachment (orange arrow).
Case I ( Figure II The serous detachment is over 6mm in length and therefore occupies the entire OCT scan. The OPL is a single layer on the far nasal and entire temporal edges of the serous detachment but splits into 2 layers towards the nasal fovea. The normal OPL layer is seen as a single hyperre ective line in the attached temporal retina. There is a thickened bump in the OPL at the in ection from attached to detached retina. The OPL then bifurcates along the shallower ascending portion of the detachment and fuses into a single layer along the steeper descending portion of the detachment. The normal OPL layer is seen as a single hyperre ective line in the attached temporal retina. The OPL is bifurcated more centrally and fused more distally along both the nasal and temporal portions of the detachment.

Results
The Table summarizes the major OCT ndings for each case.
We selected 12 representative eyes (12 patients) from our practice's retinal image teaching collection that exhibited a hyperre ective bifurcation of the OPL sandwiching a band of variable re ectivity associated with macular detachment due to CSC, choroidal nevus, uveal malignant melanoma, and loculated uid following successful retinal reattachment surgery. Within the area of detachment where the OPL was not bifurcated, most of these eyes had a thickened, diffusely hyperre ective OPL approximating the same height as the bifurcation.
Splitting of the OPL layer occurred with an angle of detachment between 10 (case 8) and 57 (case 2) degrees. But we also found that the OPL was a single layer with an angle of detachment between 12 (case 4) and 38 (case 6) degrees.
In half of our cases (6 eyes, 6 patients), the OPL had a thickened hyperre ective bump overlying the in ection from attached to detached retina. The OPL then more or less maintained this increased thickness compared to that seen in attached retina. When visualized, this bump appeared at the junction of the steeper detached edge (cases 1, 2, 5, 6) as well as on both sides of shallowly detached retina (cases 10 and 11).
We then browsed through our practice's PDF retina journal database looking for references dealing with retinal detachment and subretinal uid. We found numerous instances of the OCT bump and splitting, although the bump was explicitly described only once 1 and the splitting explicitly described only twice. [1,4] Bifurcation of the OPL was found with CSC, [5,6,3,7] CSC with type 1 macular neovascularization, [8] drusen with overlying subretinal uid, [9] unilateral acute idiopathic maculopathy, [10] rhegmatogenous retinal detachment, [11] loculated uid following retinal detachment repair, [12,13] and subretinal uid associated with choroidal nevus, [4] dome macula, [14] retinal vein occlusion, [15] Kearns-Sayre Syndrome, [16]  Most of our eyes exhibited the well-known phenomenon of a thickened, diffusely hyperre ective OPL, what Ouyang et all termed the bright OPL phenotype. We also noticed a common, but rarely previously discussed, OPL bifurcation overlying retinal detachment. We found that the OPL split along the steeper, shallower and within both portions of equally shallowly detached retina. Splitting was noted with an angle of detachment as low as 10 degrees (case 8) and as high as 57 degrees (case 2). Similarly, OPL splitting was absent with an angle as low as 12 degrees and as high as 38 degrees (case 6). In addition, splitting was found nasally, temporally and on both sides of the detachment. In other eyes with detachment there was no OPL splitting. We expected that the split OPL should be seen within a given angle of detachment, since others have suggested that the OPL is best visualized when the incident OCT beam is most perpendicular to it. [3,4] However, we saw this nding regardless of location (nasal vs temporal) or angle of detachment. We are not sure why this was the case.
Our cases con rmed prior observations that the height of the hyperre ective paracentral OPL in attached retina was thinner and the outer nuclear layer thicker than the region between the area of apparent OPL splitting or hyperre ectivity within detached retina. Curcio et al provided retinal and choroidal layer thicknesses for the central and parafoveal macula.
[20] Mean central foveal thickness for the OPL synaptic, OPL pedicles, OPL Henle, and ONL (ONL rods plus ONL cones) were 0, 0, 22, and 34 µm, respectively. Mean thickness 1mm nasal to the macular center for the OPL synaptic, OPL pedicles, OPL Henle, and ONL (ONL rods plus ONL cones) were 6, 7, 29, and 27 µm, respectively. Mean thickness 1mm temporal to the macular center for the OPL synaptic, OPL pedicles, OPL Henle, and ONL (ONL rods plus ONL cones) were 7, 6, 55, and 32 µm, respectively. OPL Henle in the paracentral macula is thus about 70% thicker than the ONL. Current OCT technology does not allow for us to accurately measure the height of these bands into µm resolution, but the relative heights of these regions in our eyes were most consistent with Curcio's measurements within the area of detachment ( Figure XIV). Each of our patients' altered OPL, whether it was a bump, bifurcation, or solid hyperre ective band, all had the same height approximating its true anatomic thickness. Thus, altering the angle of the OPL with the incident OCT laser beam, either by manipulating the beam's location within the patient's pupil or by nature presenting us with a retinal detachment, may allow for better separation of these layers and a truer correlation with anatomic measurements.
We therefore suggest a revised OCT OPL and ONL classi cation scheme applicable only when the normal retinal anatomy is altered either by the OCT technician or overlying retinal detachment ( Figure XV). The innermost hyperre ective band, the OPL Dendritic Zone, includes the synaptic and terminal subzones of horizontal and bipolar cell processes and terminals of the rods and cones. The anatomic thickness of this region is only 13-14 µm,[20] and it is unclear what components contribute to its hyperre ectivity. The thickened hyperre ective region at the in ection from attached to detached retina, as well as the hypore ective band within detachment, is the OPL Henle's Zone. This area may also appear as a solid hyperre ective band. The border between Henle's and the ONL is the outermost hyperre ective band, followed by the hypore ective ONL.
We prefer to call these bands, zones, to emphasize that these are heterogeneous cellular structures.
Our retina image teaching collection, built from patients seen at our practice, is composed of images selected for their quality and how well each exempli es and teaches something about each entity. No attempt has been made over the years to include every example of every patient we have examined. Similarly, when retrospectively looking through this collection for images with OPL splitting, we made no attempt to include every example we found. Thus, we can make no conclusion regarding how often these ndings are present with retinal detachment.
In summary, we expand on prior descriptions of splitting of the OPL within a macular detachment (bifurcation sign) and a thickened bump within the OPL at the in ection from attached to detached retina (bump sign).
These ndings were common in our practice's retinal image teaching collection and were also found unmentioned in many published references dealing with retinal detachment and subretinal uid. As previously described, the appearance of what we currently call the OPL and ONL on OCT changes depending on the angle of the incident OCT laser with the retinal microstructures. However, contrary to prior explanations for this phenomenon, the varied OPL phenotypes inexplicably appear to be independent of the incident angle of the illuminating OCT laser. This altered appearance may give a clearer picture of the true anatomy.

Declarations
Funding / Financial Disclosures: The authors did not receive support from any organization for the submitted work.
Con icts of interest/Competing interests: The authors have no relevant nancial or non-nancial interests to disclose.
Ethics approval / Consent to participate / publication: The University of Pikeville Ethics Committee con rmed that no IRB approval was required. All patients gave written permission to have their images used for educational purposes, including publication.
Availability of data and material: Not applicable        The OPL then bifurcates along the steeper ascending portion of the detachment (white arrows) and fuses into a single layer along the shallower descending portion of the detachment (orange arrow).          Similarly, the ONL in attached retina (yellow bar) appears thinner within the area of detachment. hyperre ective region at the in ection from attached to detached retina, as well as the hypore ective band within detachment, is the OPL Henle's Zone. This area may also appear as a solid hyperre ective band. The border between Henle's and the ONL is the outermost hyperre ective band, followed by the hypore ective ONL.