Abstract
Purpose
This study aims to describe spectral-domain optical coherence tomography (OCT) characteristics and assess surgical outcomes for bare retinal nerve fiber layer (RNFL) and internal limiting membrane (ILM) tear associated with the epiretinal membrane (ERM).
Methods
This retrospective study comprised 158 consecutive patients with idiopathic ERM with (group A) or without (group B) bare RNFL (19 and 139 patients, respectively) who underwent vitrectomy between March 2016 and May 2019. Surgical videos and pre- and postoperative OCT images were reviewed to characterize bare RNFL and ILM tear and analyze postoperative visual outcomes.
Results
Typically, the torn ILM is seen as a mono- or multi-layered thin scrolled membrane in OCT, correlated frequently with areas of RNFL schisis. On the near-infrared fundus images, a hypo-reflective band corresponding to rolled-up edge of torn ILM and ERM was identified. The bare RNFL area was located adjacent to and peripheral to this band and showed a less-wrinkled retina preoperatively. Mean logMAR BCVA and central foveal thickness (CFT) progressively improved in both groups. However, group A had greater mean CFT change than group B (P = 0.001).
Conclusion
Typical appearances of bare RNFL and ILM tear associated with ERM were identified. Thorough evaluation of preoperative OCT is important for vitreoretinal surgeons to recognize bare RNFL and ILM tear to minimize surgical trauma.
Similar content being viewed by others
Explore related subjects
Discover the latest articles, news and stories from top researchers in related subjects.Avoid common mistakes on your manuscript.
Introduction
Epiretinal membrane (ERM) is a relatively common retinal disorder (prevalence, 2.2–28.9%), characterized by fibrocellular proliferation on the inner retinal surface, above the internal limiting membrane (ILM) [1,2,3]. ERM involving the macular or peri-macular regions can cause visual impairment, metamorphopsia, micropsia, and occasionally monocular diplopia. The mainstay of treatment of symptomatic ERM is surgical removal of the ERM. Current modern vitreoretinal techniques involve small-gauge vitrectomy and intraocular dyes to facilitate visualization and peeling of the posterior hyaloid, ERM, and ILM [4].
In general, the ILM covers the surface of the retinal nerve fiber layer (RNFL). However, in our surgical cases, we observed several unusual forms of ERM. In these cases, after successful vitreous cortex removal, some part of the retina not covered by the ERM was not stained by indocyanine green (ICG) dye, suggesting the absence of the ILM on the RNFL surface. We called this area “bare RNFL.” We found one case series in our literature review, consisting of 23 ERM cases characterized by the presence of a localized ILM tear, among 268 patients followed up for > 6 months [5]. However, this case series was limited by relative low-resolution optical coherence tomography (OCT); moreover, preoperative OCT and intraoperative ICG staining were only performed in a few cases (14 and 13 cases, respectively). Furthermore, there were no data to associate the bare RNFL with postoperative visual outcomes and morphological restorations.
The purpose of the present study was to describe preoperative and postoperative OCT characteristics of the bare RNFL and ILM tear associated with the ERM and to assess and compare the outcomes of pars plana vitrectomy (PPV) surgery in ERM patients with and without the bare RNFL.
Methods
Patients
This study includes a retrospective consecutive chart review of patients with an ERM who underwent PPV performed by a single surgeon (J.G.K.) at Asan Medical Center, Seoul, Korea, between March 2016 and May 2019. The ERM was diagnosed based on a history of blurred or distorted vision, metamorphopsia and/or micropsia, dilated fundoscopic examination, and OCT. Data collected from patient records included age, sex, axial length, pre- and postoperative best-corrected visual acuity (BCVA), and follow-up period.
Patients with a secondary ERM (retinal detachment, diabetic retinopathy, retinal vascular occlusions, uveitis, trauma, etc.), age-related macular degeneration, glaucoma, or media opacity that could significantly interfere with OCT image acquisition, incomplete chart records, and a follow-up period of < 6 months were excluded. The study was approved by the Institutional Review Board of the Asan Medical Center (IRB No. 2019-1066) and adhered to the tenets of the Declaration of Helsinki. Due to the retrospective study design and use of de-identified patient data, the need for obtaining written informed consent was waived.
The ERM patients were divided into two groups as those with (group A) and those without (group B) bare RNFL and ILM tear. Bare RNFL and ILM tear were detected intraoperatively by the surgeon. Lesions suspected to be a torn ILM or bare RNFL were stained twice with ICG dye during surgery (Fig. 1, see Video, Online resource 1). Subsequently, ILM peeling was initiated from the adjacent stained ILM and directed toward the suspected lesion, confirming the absence or presence of the ILM. The results were re-evaluated by a masked observer (J.H.Y) who carefully studied the operation videos to confirm the surgeon’s intraoperative clinical impressions.
Ocular examination
All patients underwent comprehensive ophthalmologic examinations before and 1, 3, and 6 months after surgery. Examinations included slit-lamp biomicroscopy, refraction, dilated fundoscopy, BCVA measurement, and intraocular pressure measurement. The BCVA was measured using a standard Snellen chart, and the Snellen value was converted to the logarithm of the minimum angle of resolution (logMAR) for statistical analyses. Spectral-domain (SD)-OCT with corresponding near-infrared fundus image was also acquired using Spectralis (Heidelberg Engineering Inc., Heidelberg, Germany). The color fundus photographs were reviewed to correlate with the OCT images. The distribution of the bare RNFL area was analyzed. We divided the posterior pole into five subfields, namely the superior, nasal, inferior, temporal, and central parafovea (within 3 mm) according to the Early Treatment Diabetic Retinopathy Study grid. All images were reviewed independently by one investigator (J.H.Y.) who was blinded to all clinical information.
Patients were stratified into four preoperative stages of the ERM based on the OCT B-scan images as previously described: (1) mild-ERM with negligible morphologic or anatomic disruption, (2) ERM associated with more progressive retinal distortion (absence of the foveal pit), (3) presence of an ERM with continuous ectopic inner foveal layers anomalously crossing the central foveal area (well-defined retinal layer), and (4) ERM complicated by significant retinal thickening and remarkable anatomic disruption of the macula (disrupted retinal layers) [6].
Surgical technique
All surgeries were performed by a single surgeon (J.G.K.). In both groups, hybrid 23–25-gauge PPV and ILM peeling were performed. Cataract surgeries were conducted using phacoemulsification techniques with a clear corneal approach. The vitrectomy was performed using the Alcon Constellation® platform (Alcon Laboratories, Fort Worth, Texas, USA). After performing a core vitrectomy, posterior vitreous detachment (PVD) was induced using a vitreous cutter or extensible intraocular pick. In some cases, to ensure complete separation of the vitreous cortex from the retina, 20% diluted triamcinolone acetonide was injected into vitreous cavity. Using the 23-gauge soft-tip aspiration, remnants of the posterior vitreous membrane in the macular area were removed. After staining, the posterior pole with 0.25% ICG, the ILM was grasped with an ILM forceps and peeled off in a circular manner for approximately two disc diameters around the macula. To confirm if there are any residual ILM, 0.25% ICG was injected into the vitreous again and was immediately washed out.
Statistical analysis
All statistical analyses were performed using SPSS version 20.0 (SPSS Inc., Chicago, Illinois, USA). Means and standard deviations were calculated for continuous variables. Intergroup comparisons were evaluated using the Mann-Whitney U test, Fisher’s exact test, or Chi-square test as appropriate. P < 0.05 was considered statistically significant.
Results
Among the 239 eyes that underwent PPV, a total of 158 eyes with ERM were included in the final analysis. Out of the 158 eyes, 19 (12.03%; mean age, 65.26 ± 7.47 years; age range, 54–81 years) were included in group A and 139 (87.97%; mean age, 65.09 ± 8.84 years; age range, 52–86 years) in group B. The mean axial length was 23.68 ± 1.29 mm in group A and 23.72 ± 0.99 in group B. No significant differences were found between the groups with regard to sex, age, refraction, axial length, or proportion of pseudophakic eyes. Group A had a significantly higher grade of ERM than group B (P = 0.001). Patients’ baseline characteristics are summarized in Table 1.
Morphologic characteristics of ERM-associated bare RNFL and ILM tear
Table 2 summarizes clinical and OCT characteristics of patients with bare RNFL and ILM tear. Typically, the torn ILM is seen in the OCT images as a thin, mono- or multi-layered scrolled membrane. This finding was present in 18 of 19 eyes (94.74%). Additionally, RNFL schisis, which was previously described [7, 8], was present in 17 of 19 eyes (89.47%). Interestingly, RNFL schisis was observed in limited area adjacent to the torn and scrolled ILM. Figure 2 illustrates examples of bare RNFL and ILM tear with regard to the extent and location of RNFL schisis. Occasionally, an intraretinal cyst was observed at the site of schisis. Intraretinal cysts were predominantly located in the ganglion cell layer and the inner nuclear layer (INL).
We found hyporeflective band on near-infrared fundus images corresponding to the edge of the ERM and torn ILM as seen on the OCT image (Fig. 3). A hyporeflective band was characterized by a convex contour toward the center of the fovea and was present in 14 of 19 eyes (73.68%). The bare RNFL was adjacent and peripheral to this hyporeflective band. Only three eyes (15.79%) presented with parafoveal involvement. In the eyes where the bare RNFL did not involve parafovea, the most frequently involved sites were superior (36.84%), inferior (21.05%), nasal (15.79%), and temporal (10.53%) quadrants (Table 3). Postoperatively, torn and scrolled ILM was removed completely on SD-OCT and the corresponding hyporeflective band was faded out. Although the surgery resolved the RNFL schisis almost completely, some irregular areas of dimpling in the RNFL were evident after 1 week, which resolved gradually within a month (Fig. 4).
The preoperative near-infrared fundus image also showed discontinuity of retinal wrinkling at the junction of the bare RNFL area. In addition, the bare RNFL area appeared to be normal retinal thickness preoperatively on the macular thickness map generated using the built-in calculation system of SD-OCT.
Visual outcomes and morphological restoration after surgery
The mean preoperative logMAR value (BCVA) was slightly worse for group A than for group B (0.47 ± 0.31 and 0.32 ± 0.17, respectively; P = 0.083) but significantly improved in both the groups to 0.44 (20/55) ± 0.26 and 0.34 (20/44) ± 0.25, 0.22 (20/33) ± 0.17 and 0.21 (20/32) ± 0.19, and 0.21 (20/32) ± 0.16 and 0.15 (20/28) ± 0.17 at 1, 3, and 6 months after surgery, respectively (Table 4). No significant difference in mean BCVA change was found between the two groups.
The mean central foveal thickness (CFT) progressively improved in both groups, and no significant difference was found between them at 6 months after surgery (P = 0.140, Table 4). However, the mean change in CFT was greater in group A (134.26 ± 97.20 μm) than in group B (61.81 ± 67.81 μm; P = 0.001).
Discussion
Park et al. reported a lower recurrence rate of the ERM in the eyes that had undergone ILM peeling, and since then, several authors have advocated ILM peeling [9,10,11,12], making this maneuver popular in ERM surgeries. During these surgeries, many surgeons injected dyes to stain the ILM for its removal; occasionally, the injection was repeated to ensure adequate staining. There are several causes of negative staining of the ILM, including the ERM and remnant vitreous cortex that can be removed surgically without causing iatrogenic trauma. However, excessive attempts to peel the ILM in the bare RNFL area may damage the underlying retina while grasping or scraping the RNFL, resulting in swelling of the arcuate nerve fiber layer, dissociated optic nerve fiber layer defect, secondary paracentral macular hole, and microcysts in the INL [6, 13,14,15].
In this study, with detailed visualization of the ERM, ILM, and the retinal surface using high-resolution SD-OCT, it was possible to describe the preoperative and postoperative OCT characteristics of the bare RNFL and ILM tear associated with the ERM. Typical appearance of the bare RNFL included torn and scrolled ILM, adjacent RNFL schisis, and intraretinal cyst in the RNFL schisis area. A hyporeflective band corresponding to the rolled-up edge of the torn ILM and ERM was found in the near-infrared fundus images. Postoperatively, these appearances completely resolved on SD-OCT.
Bovey et al. demonstrated the typical appearance of the torn ILM; the fibrous preretinal band consisted of the rolled-up ILM, consistent with our finding in the near-infrared fundus image [5]. However, because of the low-resolution OCT, this study could not characterize any other appearances, such as the inner retinal surface and intraretinal cystic changes. Consequently, the pathomechanism of the bare RNFL and ILM tear associated with ERM remained speculative. They proposed the hypothesis that partial PVD exerts local traction on the ILM, resulting in the detachment of the ILM at this site. However, in our study, an evidence of partial or complete PVD was observed in 13 eyes (68.42%), whereas in 6 eyes (31.58%), the posterior vitreous remained adherent to the macula. Moreover, the fragmented ILM was not observed on the surface of the detached posterior hyaloid membrane.
Recently, Hussnain et al. reviewed their surgical cases and described similar findings of ILM tear, which they labeled as “ILM dehiscence” [7]. From a series of 41 consecutive eyes with ERM, they found RNFL schisis in over half of the eyes, and over three quarters of these eyes with RNFL schisis had a torn ILM. They also described a protrusion of the RNFL schisis through the ILM dehiscence on a horizontal B-scan and labeled it as the “spaghetti sign.” However, as shown in Fig. 2, vertical OCT B-scan taken at the same location as the horizontal OCT B-scan showing spaghetti sign demonstrated typical appearance of the torn and scrolled ILM and ERM. Therefore, we believe that the spaghetti sign on the SD-OCT indicated not only the protrusion of the RNFL schisis but also the torn and scrolled ILM and ERM. In addition, we believe that the hyporeflective band on the near-infrared fundus image in the present study was identical to what they referred to as hyperreflective areas seen on en-face imaging at the vitreoretinal interface.
It has been postulated that ERM contracture results in tangential traction on the ILM and underlying inner and outer retinal layers [5, 14, 16]. We believe that many findings from the present study supported the hypothesis that the ERM contracture contributed to the occurrence of torn ILM and bare RNFL. The ERM commonly results in centripetal forces and contraction of the posterior pole. As the ERM contracts, the maximum amplitude of the retina and tangential traction force become larger. The tangential traction on the ILM partially separates the RNFL, resulting in fibrillary surface changes (RNFL schisis). Under the tractional force of the ERM, the ILM adhering to the ERM finally detaches from the RNFL. The torn ILM, which is no longer supported by the underlying tissues, folds onto itself. Accordingly, this folding is observed on the side of the macula still covered by the ERM and results in a bare RNFL area located adjacent to and peripheral to the torn and scrolled ILM. The bare RNFL and torn ILM are believed to occur more frequently in grade 3 or 4 ERM (Tables 1 and 4), and the maximal retinal thickness was greater in group A than in group B, indicating that the ERM contraction and tangential traction were important factors in bare RNFL occurrence. Furthermore, in some cases, the excessive tangential traction was sufficient to deform the normal-layered retinal architecture, resulting in intraretinal cystic changes. Interestingly, once the ILM was torn and tangential traction was relieved, the retinal tissue was restored, resulting in the resolution of retinal wrinkling. This phenomenon was well observed in the bare RNFL areas, and RNFL schisis was not observed in the peripheral areas of the bare RNFL. Furthermore, after removing the ERM and ILM, typical appearances of the bare RNFL and torn ILM were resolved in most eyes.
Our findings raise the question whether the strong tangential traction produced by the higher grade ERMs is essential for the ILM tear. Although a grade of ERM was significantly higher in group A, two cases of grade 2 ERM also showed a bare RNFL and torn ILM in this study. A possible explanation is that the tangential tractional force caused by the ERM was asymmetric at the macula because the traction force may vary depending on the location, extent, and thickness of the ERM. In addition, multiple foci of retinal traction in the multidirectional pattern of the ERM could cause the ILM tear.
In this study, most bare RNFL areas were observed within the major vessel arcades and on the outside of the parafovea. Overall, the bare RNFL was most frequently observed in the superior and inferior quadrants (12/19 eyes, 63.16%). Interestingly, in these eyes, peripheral boundaries of the bare RNFL were located around the superior and inferior vascular arcades (Fig. 1). The ILM is known to be very thin over the major retinal vessels [17]. Therefore, we hypothesize that the traction force over major vessel arcades may cause the disruption and tearing of the ILM at this site, resulting in the distribution of the bare RNFL.
Previous studies did not evaluate the relationship between the bare RNFL/ILM tear, visual outcomes, and morphological restoration. Interestingly, although our results showed a significantly difference in terms of preoperative ERM grade, preoperative CFT, and preoperative maximum retinal thickness in the macular region between the two groups, there was no significant difference in the preoperative BCVA, which is consistent with Hirano et al.’s findings [18]. At baseline, group A tended to have a higher grade of ERM with a more disrupted retinal architecture, and this difference contribute to the greater improvement in the CFT in group A. However, despite of the greater anatomical improvement in group A, BCVA improvements were not significantly different between the two groups. We believe that the bare RNFL and torn ILM did not seem to influence the visual recovery or cause any deterioration in the preoperative visual acuity.
This study had several limitations. First, data were retrospectively collected by reviewing the medical records. Further prospective study using ultra-wide field OCT will provide a better understanding of the pathophysiologic mechanism and characteristics of bare RNFL and ILM tear. Second, a relatively small number of eyes had ILM tear and our findings should be confirmed in a larger cohort. Third, because most patients were seen at a tertiary care referral and visited our clinic with symptomatic ERM, we could not evaluate the prevalence of ILM tear in the general population. Lastly, we did not analyze the specimens histopathologically nor did we evaluate the functional visual outcomes, such as visual field, microperimetry, or severity of metamorphopsia.
In conclusion, we demonstrated typical appearances of the bare RNFL and ILM tear associated with the ERM. Contracture of the ERM may induce an ILM tear and bare RNFL, which can be seen clinically during PPV as a patch of negative ICG staining. Rolled-up edges of the torn ILM and ERM can also be observed as a hyporeflective band with a discontinuity in retinal wrinkling in the near-infrared fundus image and as a mono- or multi-layered thin scrolled membrane in the OCT images. Attempting to peel some tissues in the bare RNFL area may not be safe as it may cause unnecessary trauma and worsen the visual outcomes. Therefore, it is important that vitreoretinal surgeons meticulously analyze preoperative fundus and OCT images to recognize bare RNFL and torn ILM and minimize unnecessary surgical maneuvers in such areas. We expect that a further study with a larger patients’ cohort, could address the exact effect of bare RNFL on postoperative outcomes and improve the understanding of the underlying pathophysiologic mechanism of bare RNFL and ILM tear.
Availability of data and material
The data used to support the findings of this study are available from the corresponding author upon request.
References
Ng CH, Cheung N, Wang JJ, Islam AF, Kawasaki R, Meuer SM, Cotch MF, Klein BE, Klein R, Wong TY (2011) Prevalence and risk factors for epiretinal membranes in a multi-ethnic United States population. Ophthalmology 118:694–699. https://doi.org/10.1016/j.ophtha.2010.08.009
Cheung N, Tan SP, Lee SY, Cheung GCM, Tan G, Kumar N, Cheng CY, Wong TY (2017) Prevalence and risk factors for epiretinal membrane: the Singapore epidemiology of eye disease study. Br J Ophthalmol 101:371–376. https://doi.org/10.1136/bjophthalmol-2016-308563
You Q, Xu L, Jonas JB (2008) Prevalence and associations of epiretinal membranes in adult Chinese: the Beijing eye study. Eye (London, England) 22:874–879. https://doi.org/10.1038/sj.eye.6702786
Hernandez F, Alpizar-Alvarez N, Wu L (2014) Chromovitrectomy: an update. J Ophthalmic Vis Res 9:251–259
Bovey EH, Uffer S (2008) Tearing and folding of the retinal internal limiting membrane associated with macular epiretinal membrane. Retina (Philadelphia, PA) 28:433–440. https://doi.org/10.1097/IAE.0b013e318150d6cf
Govetto A, Lalane RA 3rd, Sarraf D, Figueroa MS, Hubschman JP (2017) Insights into epiretinal membranes: presence of ectopic inner foveal layers and a new optical coherence tomography staging scheme. Am J Ophthalmol 175:99–113. https://doi.org/10.1016/j.ajo.2016.12.006
Hussnain SA, Sharma T, Hood DC, Chang S (2019) Schisis of the retinal nerve fiber layer in epiretinal membranes. Am J Ophthalmol. https://doi.org/10.1016/j.ajo.2019.06.003
Kim JS, Chhablani J, Chan CK, Cheng L, Kozak I, Hartmann K, Freeman WR (2012) Retinal adherence and fibrillary surface changes correlate with surgical difficulty of epiretinal membrane removal. Am J Ophthalmol 153:692–697.e691–692. https://doi.org/10.1016/j.ajo.2011.08.042
Park DW, Dugel PU, Garda J, Sipperley JO, Thach A, Sneed SR, Blaisdell J (2003) Macular pucker removal with and without internal limiting membrane peeling: pilot study. Ophthalmology 110:62–64. https://doi.org/10.1016/s0161-6420(02)01440-9
Shimada H, Nakashizuka H, Hattori T, Mori R, Mizutani Y, Yuzawa M (2009) Double staining with brilliant blue G and double peeling for epiretinal membranes. Ophthalmology 116:1370–1376. https://doi.org/10.1016/j.ophtha.2009.01.024
Kwok A, Lai TY, Yuen KS (2005) Epiretinal membrane surgery with or without internal limiting membrane peeling. Clin Exp Ophthalmol 33:379–385. https://doi.org/10.1111/j.1442-9071.2005.01015.x
Schechet SA, DeVience E, Thompson JT (2017) The effect of internal limiting membrane peeling on idiopathic epiretinal membrane surgery, with a review of the literature. Retina (Philadelphia, PA) 37:873–880. https://doi.org/10.1097/iae.0000000000001263
Watanabe A, Arimoto S, Nishi O (2009) Correlation between metamorphopsia and epiretinal membrane optical coherence tomography findings. Ophthalmology 116:1788–1793. https://doi.org/10.1016/j.ophtha.2009.04.046
Chang S, Gregory-Roberts EM, Park S, Laud K, Smith SD, Hoang QV (2013) Double peeling during vitrectomy for macular pucker: the Charles L. Schepens Lecture. JAMA Ophthalmol 131:525–530. https://doi.org/10.1001/jamaophthalmol.2013.2176
Modi A, Giridhar A, Gopalakrishnan M (2017) Spectral domain optical coherence tomography-based microstructural analysis of retinal architecture post internal limiting membrane peeling for surgery of idiopathic macular hole repair. Retina (Philadelphia, PA) 37:291–298. https://doi.org/10.1097/iae.0000000000001129
Jung JJ, Hoang QV, Ridley-Lane ML, Sebrow DB, Dhrami-Gavazi E, Chang S (2016) Long-term retrospective analysis of visual acuity and optical coherence topographic changes after single versus double peeling during vitrectomy for macular epiretinal membranes. Retina (Philadelphia, PA) 36:2101–2109. https://doi.org/10.1097/iae.0000000000001055
Kishi S, Numaga T, Yoneya S, Yamazaki S (1986) Epivascular glia and paravascular holes in normal human retina. Graefes Arch Clin Exp Ophthalmol 224:124–130. https://doi.org/10.1007/bf02141484
Hirano M, Morizane Y, Kanzaki Y, Kimura S, Hosokawa M, Shiode Y, Doi S, Toshima S, Takahashi K, Hosogi M, Fujiwara A, Takasu I, Shiraga F (2019) En face image-based analysis of retinal traction caused by epiretinal membrane and its relationship with visual functions. Retina (Philadelphia, PA). https://doi.org/10.1097/iae.0000000000002569
Funding
This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.
Author information
Authors and Affiliations
Corresponding author
Ethics declarations
Conflict of interest
The authors declare that they have no conflict of interest.
Ethical approval
The study was approved by the Institutional Review Board of the Asan Medical Center (IRB No. 2019-1066) and adhered to the tenets of the Declaration of Helsinki. Due to the retrospective study design and use of de-identified patient data, the need for obtaining written informed consent was waived.
Code availability
Not applicable.
Additional information
Publisher’s note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Electronic supplementary material
Representative example of bare retinal nerve fiber layer (RNFL) and internal limiting membrane (ILM) tear associated with the epiretinal membrane. After staining the posterior pole with 0.25% indocyanine green, unstained area of the ILM, suggesting a bare RNFL, was observed.
ESM 1
(WMV 32614 kb)
Rights and permissions
About this article
Cite this article
Yeo, J.H., Kim, JG. Bare retinal nerve fiber layer and internal limiting membrane tear associated with epiretinal membrane: OCT findings and visual outcomes. Graefes Arch Clin Exp Ophthalmol 258, 1379–1387 (2020). https://doi.org/10.1007/s00417-020-04684-x
Received:
Revised:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s00417-020-04684-x