Multimodel Imaging Evidence of Traction Component in Lamellar Macular Hole with Epiretinal Proliferation

Lamellar macular hole (LMH) is a frequently seen macular structural changes related to vitreomacular interface abnormalities. In 2006, Witkin et al. [1] first reported the presence of a “thick” epiretinal membrane (ERM) in eyes with LMH on optical coherence tomography (OCT). Parolini et al. [2] described this same unusual appearance as a “dense nontractional” ERM and showed distinct differences from the conventional “tractional” ERM. Pang and associates [3] later named this structure as lamellar hole-associated epiretinal proliferation (LHEP) and described it to be the homogenous medium reflective thick membrane on the epiretinal surface and at LMH edge on spectral-domain OCT (SD-OCT). LHEP was contiguous with the middle retinal layers and originated from the associated retina defect. The clinical implication of LHEP has been studied extensively in recent years. Most studies found the presence of LHEP was associated with deeper and wider retinal defects in LMH, along with more ellipsoid zone (EZ) defects [4‒6]. Furthermore, LHEP could also be found in patients with full-thickness macular hole (FTMH) [3, 4, 7].

The presence of LHEP has been used as one diagnostic criterion to differentiate “degenerative” LMH from “tractional” LMH, a classification proposed by Govetto and associates [8] in 2016. They postulated that 2 distinct entities existed in LMH, which had different OCT morphologies including EZ defects, inner/outer hole diameter ratio, shaped of retinal defects, and presence of LHEP or foveal bump. They emphasized that “degenerative” LMH had no visible signs of epiretinal traction. Although this classification was later adopted by some studies [9‒11], there was still conflicting evidence regarding the presence of tractional components in LMH with LHEP. Some studies showed a high prevalence of coexisting ERM and LHEP [8, 12‒14]. In addition, progression of “tractional” LMH with ERM to “degenerative” LMH with LHEP has been reported [15]. Recently, an OCT-based new classification for LMH has been proposed [16]. In this new classification, LMH was used to describe those lesions previously defined as “degenerative” LMH and ERM foveoschisis was used to describe lesions previously referred to as “tractional” LMH. These aforementioned observations and the evolving classification for LMH suggest that the correlation between epiretinal traction and LMH with LHEP has not been settled.

In the current study, we evaluated the presence of epiretinal traction in patients with LMH by using sequential OCT images, fundus autofluorescence (FAF), and fundus photography, as well as intraoperative findings. By comparing these clinical observations and surgical changes in LMH with and without LHEP, the association of LHEP and epiretinal traction may be better appreciated and understood. In addition, our findings might help comprehending the significance of different morphological features of LMH and further improve the classification of LMHs.

This retrospective case series reviewed patients diagnosed with LMH at National Taiwan University Hospital from August 2005 to July 2019. Patients were divided into two groups according to the presence or absence of LHEP: the LHEP(+) and LHEP(−) groups. The diagnosis of LMH was made according to the criteria proposed by International Vitreomacular Traction Study (IVTS) group [17]. The presence of LHEP, the specific form of epiretinal proliferation defined by Pang and colleagues [3], was carefully identified through reviewing each OCT image. Eyes were further classified into 3 subgroups according to the new OCT-based classification system [16]: LMH, ERM foveoschisis, and mixed subgroups. The study adhered to the tenets of the Declaration of Helsinki and was approved by the Institutional Review Board of the National Taiwan University Hospital.

We collected the demographic data for each patient, including age, gender, spherical equivalent, axial length, and follow-up periods through chart reviews. All patients underwent examinations including best corrected visual acuity (BCVA), slit-lamp examinations, and indirect ophthalmoscopy during the initial visits and follow-up, with regular fundus photography, FAF, and OCT examinations. SD-OCT (RTVue^® Model-RT100 version 3.5; Optovue, Inc., Fremont, CA, USA) with cross-scanning centered on the fovea was done. Eyes with high myopia (i.e., the refractive error ≤−8.00 diopters and/or an axial length ≥26.5 mm), proliferative diabetic retinopathy, retinal vascular occlusion, uveitis, retinitis pigmentosa, and history of receiving vitrectomy surgery were excluded.

Fundus photography and FAF were analyzed for the presence of ERM, vascular stretching or displacement, surface striation, or wrinkling. Sequential images were carefully examined for any changes of vascular position within the arcade. All preretinal changes, including the presence of ERM, attached posterior hyaloid, and vitreomacular traction, were recorded, along with the presence of ellipsoid zone defect and other LMH related features (as listed below) in the OCT images. The diameter of the inner and the outer opening of LMH, the thickness of the thinnest base, and the maximum height of the retina at central foveal area (Fig. 1) were manually measured by the built-in software caliper.

The presence of ERM and/or attached posterior hyaloid traction, in addition to those directly visible lesions in OCT images or fundus pictures, were also based on: (1). evidence of macular vascular traction, (2). the presence of surface wrinkling/folds in en face OCT image, and (3). the identification of non-indocyanine green (ICG) stained area during surgery. The presence of macular vascular traction was determined by FAF and/or fundus photography images (Fig. 2) or by comparing the vascular pattern and positional changes from sequential fundus images or the pre- and postoperative images in those patients who underwent vitrectomy.

To precisely identify the change of the vascular pattern, we first converted the raw FAF or color fundus images into 8-bit gray-scale images and paired images centered on the same macular area were chosen for comparison. The 2 paired images were later assigned with different colors, red and gray, and merged into one image by using optic disc and major vessels as references for alignment. Difference in the macular vascular patterns, identified with the split of different colored vessels, was recorded as macular vascular traction. The image processing was performed using ImageJ version 1.52a (National Institutes of Health, Bethesda, MD, USA).

Indications of operation for LMHs included BCVA worse than 20/40 and evidence of epiretinal traction or progression of LMH. All surgeries were done with 23-gauge or 25-gauge system (Constellation^® Vision System; Alcon Laboratories Inc., Fort Worth, TX, USA). The surgical techniques for LMH have been described elsewhere [4, 18]. In brief, posterior vitreous detachment was induced if not already present, and complete posterior hyaloid and ERM removal were performed. The internal limiting membrane was peeled after stained with diluted ICG dye (25 mg in 15 mL of 5% glucose water). The staining pattern of ICG was recorded. In patients with LHEP, the tissue was trimmed and either left untouched or embedding into the LMH [19]. Finally, air-fluid exchange was performed and perfluoropropane (C3F8) was used for tamponade with prone positioning recommended.

The baseline demographic data and image findings were compared between the LHEP(+) and LHEP(−) groups, as well as between the 3 subgroups. BCVA was tested with the Snellen acuity chart, and the results were converted into a logMAR scale for calculation and comparison. We used the independent t test to compare continuous variables between 2 groups and the analysis of variance (ANOVA) for the comparison between 3 groups. The Pearson χ² test or Fisher’s exact test was used for categorical variables. All statistical analyses were performed using IBM SPSS Statistics version 20 (SPSS Inc., Chicago, IL, USA). A p value of < 0.05 was considered as statistically significant.

In our study, 100 patients and 108 eyes with LMH were enrolled, in which 53 eyes were LHEP(+) and 55 eyes were LHEP(−) in OCT images. The mean follow-up time for the patients was 53.5 ± 34.5 months. The mean ages were similar in the LHEP(+) and LHEP(−) groups (p = 0.86). There was no significant difference between the two groups in mean axial length, spherical equivalent, and initial and final BCVA. The comparison of demographic data and clinical findings between the two groups is summarized in Table 1.

The presence of yellowish pigment at fovea (33% and 31%, p = 1.00, respectively) and the evidence of macular vascular traction (including vascular pattern and position changes; 92% and 84%, p = 0.36, respectively) were similar between LHEP(+) and LHEP(−) eyes. The prevalence of ERM and/or attached posterior hyaloid, evidenced by multimodal images, was 100% in both groups (p = 1.00).

In OCT images, more LHEP(+) eyes had EZ defect (53%) than LHEP(−) eyes (18%, p < 0.001). The LHEP(+) eyes had similar mean diameter of the inner opening of the hole compared to LHEP(−) eyes (p = 0.57) but a larger mean diameter of the outer opening of the LMH (p = 0.001), therefore a smaller ratio of the inner/outer diameter compared (p = 0.026). The mean thickness at the thinnest base of the hole was thinner in LHEP(+) eyes than in LHEP(−) eyes (p < 0.001); while the average thickness of the highest retina at fovea area was similar in both groups (p = 0.77). The imaging findings of the patients are summarized in Table 1.

In our series, 30 LHEP(+) and 19 LHEP(−) eyes underwent surgical interventions. The change in logMAR BCVA after the surgery was −0.23 ± 0.37 (11.5 ETDRS letters) in the LHEP(+) group and −0.28 ± 0.46 (14 letters) in the LHEP(−) group (p = 0.67). Decreased vascular traction in the macular area was noted in 88% of the LHEP(+) eyes and 100% of LHEP(−) eyes (p = 0.27). The maximum retinal thickness at fovea decreased in both groups after the surgery (p = 0.33). More eyes in the LHEP(−) group (4/7, 57%) had EZ recovered after the surgery compared with the LHEP(+) group (3/18, 17%; p = 0.031). The findings of the patients who received surgical interventions are summarized in Table 2.

There were 32 (30%) eyes classified as LMH subgroup, 67 (62%) eyes as ERM foveoschisis subgroup, and 9 (8%) eyes as mixed subgroup. Both baseline and final BCVA were the worst in the mixed subgroup, followed by the LMH subgroup, and the best in the ERM foveoschisis subgroup (p = 0.001 and 0.094, at baseline and final visit). Macular vascular tractions were evidenced similarly in 3 subgroups (80% in LMH, 92% in ERM foveoschisis, 88% in mixed; p = 0.26). All of the eyes had ERM and/or attached posterior hyaloid (p = 1.00). The presence of LHEP was more frequently found in the LMH compared with ERM foveoschisis or the mixed subgroup (p = 0.002). EZ defect was found in 20 (63%) eyes in the LMH, 13 (19%) eyes in the ERM foveoschisis, and 5 (56%) eyes in the mixed subgroup (p < 0.001).

In OCT images, the LMH subgroup had largest inner and outer opening and thinnest maximum retinal thickness among the 3 subgroups. In contrast, the ERM foveoschisis subgroup had smallest inner and outer opening, and greatest base and maximum retinal thickness. Among the 49 eyes that underwent vitrectomy, 17 were classified as LMH, 27 as ERM foveoschisis, and 5 as mixed subgroup. There was no significant difference between the three subgroups regarding the change of BCVA after surgery (p = 0.21). The clinical findings of the three subgroups are summarized in Table 3.

In this study, we investigated the association of LHEP with epiretinal traction. The results showed that among all patients who had LMH with LHEP, 92% exhibited fundus-photographically evident macular vascular traction; the coexistence of LHEP and ERM or attached posterior hyaloid in OCT images was 100%. In addition, the macular vascular traction, as evidenced by the shift of vascular distribution pattern during follow-up or after surgery (Fig. 2), was noted in 88% of eyes with LHEP; the average of the highest retinal thickness also decreased 89.8 μm after surgery. If any of the above findings was taken as a sign of the presence of epiretinal traction, our results showed that 100% of eyes with LMH and LHEP in our study had some degree of epiretinal traction.

The coexistence of ERM and LHEP has been noted since the earliest report by Witkin and colleagues [1], which they presented cases showing “thickened ERM” and traditional ERM at the same time. This simultaneous expression of ERM and LHEP was further documented in almost every study regarding LHEP [2‒4, 15], with the prevalence ranging from 14% to 100% [8, 12‒14, 20]. The presence of coexisting ERM and tractional components could be missed if B-scan OCT images were not oriented to the existing parafoveal ERM [9, 21], and therefore the prevalence underestimated. By using multimodal imaging and orienting the OCT scanning area toward the ERM, Hirano and associates [9] demonstrated a higher prevalence of ERM. In addition to typical ERM, posterior hyaloid could also generate tractional forces on the retinal surface and was found to coexist with LHEP in LMH in different conditions [1, 22‒24]. In the current study, with the combination of OCT and other image modalities, we found a coexisting ERM and/or posterior hyaloid with LHEP in every case (Fig. 3). Based on these findings, we postulate that tractional components are involved, although might in different degrees, in the development of most, if not all, of the LMH.

In 2016, Govetto and associates [8] separated the LMH into degenerative and tractional types, with the presence of LHEP as the cardinal feature of degenerative LMH. Evidence proposed by the authors to support the degenerative nature of these LMH were the noncontractile property of LHEP, less frequently associated with tractional ERM, and the lack of intraretinal schisis. While their morphological descriptions regarding LMH with and without LHEP were in accordance with most studies, there is, however, evidence suggesting that LMH presenting with LHEP might not be purely degenerative in nature, such as frequent coexistence of tractional ERM and LHEP [8, 12‒14], LMH with initial tractional morphology which progressed to “degenerative” morphology [3, 15], and both anatomical and functional improvement after surgical removal of epiretinal tractions in LMH with LHEP [4, 19, 25, 26]. In a new OCT-based classification, the authors renamed the degenerative and tractional LMHs as LMH and ERM foveoschisis, in which they also considered that retinal traction was rarely evident in LMH (equivalent to degenerative LMH) in contrast to ERM foveoschisis [16]. Nevertheless, the findings from our current study, by using this new classification, further demonstrated the involvement of epiretinal traction in LMH.

One of the major reasons to support the classification by Govetto et al. [8] and Hubschman et al. [16] was the lack of retinal vascular traction or retinal thickening in those classified as degenerative LMH or LMH. Hirano et al. [9] showed no vascular pattern change in en face OCT image in degenerative LMH. However, after carefully comparing the serial fundus photography and the images before and after operation in non-operated and operated eyes, we found most cases of degenerative LMH had vascular pattern changes as a sign of epiretinal traction (Fig. 2). En face OCT image also showed prominent retinal wrinkling, folds, and traction in degenerative LMH (Fig. 3). In addition, we found that retinal thickening, one of the optional criteria to support the diagnosis of ERM foveoschisis [16], also improved after vitrectomy in cases with LHEP, indicating the presence of retinal traction in LMH with LHEP. Furthermore, Parolini et al. reported that the presence of alpha-smooth muscle cells could be found in both tractional and dense membranes, although in different concentration [2]. Bringmann et al. also demonstrated that OCT-evident tractional forces were involved in the development of degenerative LMH [26]. One possible explanation for the differences in structural changes between LMHs with and without LHEP may be the different traction direction and forces. We postulate that in those LMH with degenerative morphology and LHEP, the traction force on the fovea is homogenous, mainly tangential in direction by the tightly adherent ERM. This mode of traction may be more likely to induce deeper and wider central foveal defect while minimize the vascular pattern change and retinal elevation, hence result in the “degenerative” morphology reported in previous study.

The involvement of tractional forces in degenerative LMH or LMH was further supported by longitudinal follow-up of patients with LMH. In our observation, some cases of LMH, starting with ERM and tractional morphologies in OCT images, could gradually develop LHEP and showed “degenerative” morphology, with some of them further progressed to FTMH (Fig. 4, 5). Our observation was in accordance with previous reported from Compera and associated [15], which also demonstrated a case which progressed from LMH with ERM to LMH with LHEP, and finally to FTMH with spontaneous closure. Similar cases had also been documented by Pang et al. [3]. More importantly, these observations were in line with a recent longitudinal study reported by Lee et al. [27] which illustrated variable tractional pathways in the development of idiopathic LMHs and showed that some of the cases evolved into degenerative configuration in the later stage. This sequential change of LMH help elucidate the possible pathogenesis of LMH with and without LHEP, that is, LMH initially developed as a result of epiretinal traction with OCT showing tractional morphologies, and when the retinal defect became deeper and wider (i.e., when Muller cells were involved), LHEP occurred, as a reaction of Muller cell to the retinal tissue loss, with corresponding OCT morphological changes (Fig. 5). A recent review article by Wu and Bradshaw. [28] also supported the hypothesis that tractional force which disrupts the Muller cells is the initial event in both ERM foveoschisis and LMH. Less often, a direct insult from vitreomacular traction may lead to retinal tissue loss and a deep retinal defect, with subsequent development of LHEP, such as occurred in the aborted process of FTMH [29, 30] (Fig. 4). Thus, the presence or absence of LHEP might indicate different stages of LMH and different severity of retinal defects, with those exhibiting LHEP to be in a later process of LMH formation or with more severe retinal defects.

The surgical results of LMH with and without LHEP varied among studies. The inferior outcomes in LMH with LHEP reported by some investigators were further used to support the idea that these LMHs were driven by “degenerative” process rather than tractions, and hence the surgical removal of epiretinal tractions could not benefit these patients [2, 5, 6, 12, 31]. However, our previous study had demonstrated significant visual improvement, similar to LMH without LHEP, in LMH with LHEP after surgical removal of visible tractions [4]. Several other studies had demonstrated improved vision in patients with LMH and LHEP after operation when LHEP was preserved during epiretinal traction removal [10, 11, 14, 18, 19, 26, 32]. The visual and anatomical improvements after peeling of ERM/posterior hyaloid and internal limiting membrane in these cases had indicated that tractional component was still involved in the pathogenesis of LMH with LHEP, therefore removal of epiretinal traction could benefit these patients. Furthermore, our study demonstrated that ERM frequently coexist in LMH in the new OCT classification system and these patients could still benefit from surgical removal of epiretinal tractions.

The main limitation of current study was its retrospective nature. Furthermore, not every patient had OCT images before the formation of LMH to support our hypothesis regarding the formation and progression of LMH, thus evidence from studies with images prior to LMH formation and long-term follow-up are required to support our hypothesis. In addition, the exact mechanism that caused LHEP formation, besides the well-known association with deeper and wider retinal defects, remained unclear. However, our study had demonstrated the coexistence of epiretinal traction and concurrent ERM in those LMH with LHEP as well as those without LHEP. All cases in our study had undergone regular high-resolution OCT and fundus photography examinations, and multimodal imaging were used to more accurately access the presence of epiretinal traction. Sequential changes of images from different modality also provided additional information to evaluate the presence of tractional components. These observations provided useful information to help better understand LMH.

We investigated the clinical findings of LMH with and without LHEP and showed evidence of epiretinal traction in both conditions. Our observations indicated that epiretinal traction is one of the typical findings rather than the exception in LMH showing LHEP, in “degenerative” LMH or in LMH under the new OCT morphological criteria. The findings from present study could improve our current understanding of LMH and other vitreomacular interface disorders. Our findings suggest the term “degenerative” should be used with caution while describing LMH and should be limited to those truly presents no tractional components [33, 34]. In addition, patients with lesions that fit the new OCT morphological criteria of LMH do not preclude the presence of epiretinal traction, and surgical intervention could still benefit these patients. Further studies are still required to better understand the exact mechanism of LHEP formation as well as the etiology of LMH.

This retrospective review study was conducted with approval from the National Taiwan University Hospital Institutional Review Board (approval number 202004056RINC) and in accordance with the tenets of the Declaration of Helsinki. Waiver of informed consent was approved by the National Taiwan University Hospital Institutional Review Board due to the retrospective design of this study.

The authors have no conflicts of interest to declare.

This study did not receive any funding.

Y.-T.S.: data acquisition and analysis and drafting the manuscript; C.-M.Y.: study design, drafting and critical revision of the manuscript; T.-T.L.: study design, data analysis, and interpretation, drafting, and revision of the manuscript. All authors agreed to be accountable for all aspects of the work submitted and approved the final manuscript.

All data generated or analyzed during this study are included in this article. Further inquiries can be directed to the corresponding author.