Quantitative evaluation of the wide-field fundus photographs in eyes with severe stage 3 and stage 4A retinopathy of prematurity

Abstract

Background

The aim of the study was to investigate the quantitative differences between severe stage 3 and stage 4A retinopathy of prematurity (ROP) by evaluating the pre-treatment fundus photographs.

Methods

Thirty-three eyes with clinical diagnosed as severe stage 3 were classified as severe stage 3 group. Twenty-two eyes with retinal detachment without foveal involvement were classified as stage 4A group. Quantitative measurements were performed on pre-treatment 130 degree fundus photographs.

Results

In the severe Stage 3 group, dense fibrous membranes, vertical tractional bands, and dragging were detected in 18 eyes (55%), in 5 eyes (15%), and in none of the eyes, respectively. In the stage 4A group, dense fibrous membranes, vertical tractional bands, and dragging were detected in 21 eyes (96%), in 22(100%) eyes, and in 17 eyes (77%), respectively. Dragging and vertical tractional bands were higher in the Stage 4A group than in the severe stage 3 group (p = 0.000). Disc-to-fovea distance, the width of the fibrous membranes, the total area of the fibrous membranes, total retinoschisis, and detachment areas were significantly higher in the stage 4A group than in the severe stage 3 group (respectively, p = 0.000,p = 0.006, p = 0.024,p = 0.000).

Conclusions

Fibrous membranes and tractional bands can be detected in severe stage 3 ROP, but the width and the total area of the fibrous membranes and total retinoschisis-detachment area were found to be higher in stage 4A eyes. The dragging of the posterior pole can be an important diagnostic indicator for the diagnosis of stage 4A. Therefore this finding may be a simple auxiliary finding for diagnosis and prognosis of stage 4A ROP.

Background

Retinopathy of prematurity (ROP) is a disease classified into five stages [1]. In the current classification, stage 3 ROP is defined as an extraretinal neovascular proliferation that extends from the ridge into the vitreous [1]. In the presence of exudative or tractional detachment, grading was labeled as Stage 4 ROP [1]. Indirect ophthalmoscopic findings such as loss of choroidal vasculature or of granular pigment epithelium, a groundglass appearance of the detached retina were reported for the diagnoses of stage 4 ROP in the International Classification Retinopathy of Prematurity (ICROP) 2021 report [1].

In current clinical practice, ROP screening and diagnoses were performed mainly with binocular indirect ophthalmoscope (BIO) and contact wide-field fundus imaging devices (WFID). Differential diagnosis can be challenging between severe stage 3 and limited stage 4A ROP by using BIO and WFID [2]. Although handheld optical coherence tomography (OCT) devices provide critical data for the differential diagnoses of severe stage 3, retinoschisis, and retinal detachment, these devices were not widespread enough even in high-income countries [3].

Primary aim of the study was to investigate the quantitative differences between severe stage 3 and stage 4A ROP by evaluating the pre-treatment wide-field fundus photographs (WFPs).

In addition, treatment choice is challenging in the eyes with Stage 4A ROP. Although vitrectomy is a generally accepted and preferred treatment method [4, 5], laser [6, 7] and anti-VEGF [6, 8, 9] treatments have been reported to be applicable in limited cases in recent years. In our clinical practice, we performed laser and/or anti-VEGF(vascular endothelial growth factor) treatment in eyes with limited detachment areas. In the progression of detachment or in eyes with larger areas of detachment, we referred patients to vitrectomy surgery. The secondary aim of the study was to compare the extent of the detachment between the eyes with and without vitrectomy by evaluating pre-treatment WFPs in stage 4A eyes.

Methods

This research was conducted by analyzing the data obtained from the University of Health Sciences, Kanuni Sultan Süleyman Training and Research Hospital, Ophthalmology Department ROP Center. The study was a single-centered retrospective and observational comparative case series. The study was conducted in concordance with the Declaration of Helsinki. The protocol of the study was reviewed and approved by the University of Health Sciences, Kanuni Sultan Süleyman Training and Research Hospital Ethics Committee, approval number is KAEK/2021.04.135. Due to the retrospective study design, informed consent for the study was waived by the ethics committee.

Charts of the patients who underwent treatment for ROP between May 2018 and Dec 2021 in the referral ROP center were retrospectively analyzed. We retrospectively evaluated the WFPs of the eyes whose clinical diagnoses were severe stage 3 ROP and Stage 4A ROP. During clinical practice Stage 3 subgrouping based on according to the illustrations published in the ICROP 1984 [10]. Exclusion criteria are as follows: (1) Eyes with labeled as mild/moderate stage 3 ROP during clinical practice. (2) Eyes with labeled as stage 4B ROP during clinical practice. (3) The eyes whose pre-treatment WFPs were not available. All eyes who met the study criteria between May 2018 and Dec 2021 were included in the study.

According to the clinical diagnosis that was settled by performing an examination with BIO and by evaluating WFPs during clinical practice, eyes were evaluated in two subgroups. Eyes with dense fibrous membranes near the ridge, elevated retinal vessels/retinoschisis, and vertical tractional bands extending into the vitreous and without retinal detachment were classified as severe stage 3 ROP. Eyes with limited tractional and exudative retinal detachment without foveal involvement were grouped as stage 4A ROP. Narrowing of the temporal vascular arcades and/or displacement of the macular center was defined as dragging.

All WFPs were captured with the 130° wide-angle PanoCam PRO (Visunex, Fremon CA, USA) device. Laser and intravitreal bevacizumab injection treatments were performed by Author1 (A1) and A2. Vitreoretinal surgeries performed by A3. Evaluation and the quantitative measurements of the WFPs were performed by A1. Quantitative measurements were made using Image J (National Institutes of Health, Bethesda, MD, USA) software.

Horizontal disc diameter (DD), optic disc-to-fovea (DF) distance, disc area (DA), width and total area of the fibrous membranes, the total retinoschisis and detachment areas were quantitatively measured in units of pixels. With the 130-degree camera attachment of the PanoCam PRO, each WFP image corresponds to 4,160 pixels horizontally and 3,120 pixels vertically.

Disc diameter, DF distance, DA, width of the fibrous membranes, total area of the fibrous membranes, and total retinoschisis and detachment areas were compared between groups in terms of pixels. The ratio of these aforementioned quantitative parameters to DD and DA is calculated. The measurement technique was presented in Figs. 1, 2 and 3. Measurement technique of the some of these parameters were described in our previous studies [11,12,13].

Fig. 1

Distance and area measurements were measured in units of pixels with the Image J software. a, b The image is magnified to perform more precise measurements with the Image J software. Measured pixels are not affected by the magnification made for the measurement. c, d 130-degree fundus photograph corresponds to 4160*3210 pixels. a Disc diameter (DD) measured along the line extending from the optic disc center to the fovea. The horizontal DD is the measurement between the points d₁ and d₂. b The disk area (DA) measurement method is shown. The optic disc border was drawn manually by the author. The Image J program calculates the disc area according to the drawn area in terms of pixels. c Optic disc-to-fovea (DF) distance measurement in an eye without macular dragging is shown. The measurement between the points d_c and f is the DF distance. d Note the increase in DF distance in an eye with macular dragging

Fig. 2

The measurement technique of the retinoschisis and retinal detachment is shown. The margins of the area were drawn by the author. Image J software calculates the area in units of pixels a Elevated retinal vessels and retinoschisis are present in the temporal periphery. The area circled with yellow line presents the retinoschisis. Image J software calculated the area in pixels. b Retinal detachment is present in the inferior temporal quadrant. The measurement technique of the area of the detachment is shown

Fig. 3

In this figure, insufficient pupillary dilation, intravitreal hemorrhage, and vitreous clouding secondary to severe plus disease adversely affected the image quality. a, b, c, d The ridges of the fibrous bands were drawn by the authors, and each area was calculated with Image J software. For 360-degree extending fibrous bands, multiple wide-field fundus photographs were evaluated. Fibrous bands in all quadrants were carefully evaluated in different photographs, and their margins were manually drawn. We calculated the total area of fibrous bands in that eye by summing the results. Care was taken to ensure no skipped areas and that the same area was not measured twice. a The region where the fibrous band is widest is indicated by the white dashed line in the lower nasal quadrant

For the subgroup analyses of eyes with Stage 4A ROP, the eyes that were treated with vitrectomy during follow-up were grouped as vitrectomy-required eyes; the eyes that were treated with laser, intravitreal bevacizumab (IVB), or combined IVB and laser were grouped as vitrectomy-nonrequired eyes. The ratios of the total area of the fibrous membranes/DA and detachment area/DA were compared between the eyes with and without vitrectomy.

Statistical analysis used the SPSS (SPSS Inc, PASW Statistics for Windows, Version, 18.0, Chicago, USA). Normality analysis of the data was performed using the Shapiro–Wilk test. Normally distributed data were compared with parametric tests (independent t-test), and other parameters were analyzed by nonparametric tests (Mann Whitney-U). Categorical variables were compared with the chi-square. A p-value of < 0.05 was considered significant. In the text and tables, normally distributed data and data compared with parametric tests were presented as mean ± std, and non-normally distributed data were presented as median (25th to 75th percentile values).

Results

Thirty-six infants fulfilled the study criteria, and WFPs of 55 eyes of 36 infants underwent further quantitative analyses for the study. Thirty-three eyes were classified as severe Stage 3 ROP, and twenty-two eyes were classified as stage 4A ROP. The main demographic characteristics are presented in Table 1.

Table 1 Comparison of demographic parameters between groups

Comparison of fundoscopic findings that were noted during clinical practice is presented in Table 2. Dense fibrous membranes, vertical tractional bands, and dragging were higher in the Stage 4A group than in the severe stage 3 group (p = 0.000). Comparison of quantitative parameters that were measured on WFPs between severe Stage 3 ROP and Stage 4A ROP is presented in Table 3. Disc-to-fovea distance, width of the fibrous membranes, total area of the fibrous membranes, total retinoschisis and detachment areas were significantly higher in the Stage 4A group than in the severe stage 3 group (respectively, p = 0.000,p = 0.006, p = 0.024,p = 0.00).

Table 2 Comparison of ophthalmoscopic findings between severe stage 3 ROP and stage 4A ROP

Table 3 Comparison of quantitative parameters measured on wide-field fundus photographs

Out of the 33 eyes that were diagnosed as severe Stage 3 ROP, as primary treatment, 25 eyes underwent intravitreal bevacizumab monotherapy, four eyes underwent laser, and four eyes underwent IVB with laser combination. Out of the 25 eyes that underwent IVB monotheraphy, five eyes underwent additional laser treatment during follow-up, four eyes underwent re-injection, and one eye underwent vitrectomy (For the severe subhyaloid hemorrhage that prevents posterior pole visualization).

Twenty-two eyes were diagnosed as stage 4A ROP. Five eyes (one laser, three IVB, and one vitrectomy for vitreous hemorrhage) underwent treatment before the diagnosis of Stage 4A. The remaining 17 eyes were presented with stage 4A ROP without any prior treatment. The mean age at the 4A ROP diagnosis was 41.6 ± 3.4 weeks postmenstrual. After the diagnosis of stage 4A ROP, two eyes underwent IVB, seven eyes underwent laser, nine eyes underwent laser and IVB combination, one eye underwent vitrectomy (without prior IVB or laser), one eye underwent IVB following subsequent vitrectomy (This eye was treated with laser for stage 3 ROP in another center and referred for the progression to Stage 4A ROP. Details were presented as patient 35 in additional files (e Fig. 1 and e Table 1)), two eyes underwent IVB and laser following subsequent vitrectomy. During vitrectomy surgeries, peripheral avascular retina was ablated with laser. Totally, in 20 eyes, peripheral avascular retina was ablated with laser, and in eight eyes laser was performed to the posterior to the detachment to prevent posterior spread of the detachment. Macula was attached at the final examination in all eyes with a mean follow-up of 47.9 ± 49.8 weeks. Dragging of the optic disc and the posterior pole was present in 16 eyes during the final examination.

During follow-up in none of the eyes, progressive tractional detachment was observed. In one eye, exudative total retinal detachment was observed after vitrectomy, which was treated with scleral buckle and re-vitrectomy with heavy silicon oil injection (Details were presented as patient 35 in e Fig. 1 and e Table 1). Peripheral small lenticular opacity, which did not obscure the optic axis, developed in the eye that underwent vitrectomy due to subhyaloid hemorrhage in severe Stage 3 ROP group.

In 6 eyes of the stage 4A group (2 eyes treated with anti-VEGF monotherapy, one eye treated with laser, three eyes treated with laser and anti-VEGF) and in one eye of the stage 3 group (treated with laser), follow-up time was lower than three months.

Comparison of quantitative measurements between the vitrectomy-required and non-required eyes is presented in Table 4.

Table 4 Comparison of quantitative parameters of eyes with Stage 4A ROP between vitrectomy required and non-requried eyes

Main clinical data is summarized in the additional file 2 (e Table 1). The clinical course of the included eyes were presented in detail with WFPs in e Fig. 1.

Discussion

Dense fibrous membranes, elevated retina vessels/retinoschisis, and vertical tractional bands can be detected in severe stage 3 ROP, but the width and the total area of the fibrous membranes and total retinoschisis-detachment area were found to be higher in stage 4A eyes. The dragging of the posterior pole can be an important diagnostic indicator for the diagnosis of stage 4A. Measurement of the ratio of the DF distance to disc diameter may be a quantitative finding that indicates the presence and severity of the dragging of the macula. In addition, at the time of diagnoses, postmenstural age (PMA) and DF distance were higher in Stage 4A group. For these findings, in the presence of dragging in an infant with advanced PMA, the presence of Stage 4A ROP should be evaluated cautiously.

In the ICROP 1984 and ICROP 2005 Stage 3 was subdivided into three grades such as mild, moderate, and severe stage 3 ROP [10, 14]. In the third revision, it is stated that stage 3 ROP represents varying degrees of extraretinal neovascular tissue [1]. In a previous study performed with a handheld-OCT device, tractional retinoschisis term was identified for the three eyes that had progressive ROP despite laser photocoagulation [15]. The authors reported that handheld-OCT devices may provide valuable insight into the interaction of the retina, vitreous, and the ridge in patients with progressive stage 3 ROP. In the eyes with stage 3 ROP, the extent of the extraretinal neovascular tissue (by clock hours), thickness, the severity and the height of the vertical component, whether it is accompanied by retinoschisis and its distance from the fovea may be prognostic factors. Therefore, during clinical practice, we did not evaluate all eyes with stage 3 ROP in the same category.

In 2006, Patel reported that, in one eye diagnosed as stage 4A ROP with indirect ophthalmosopy, OCT revealed macular detachment and microcystic changes in the inner retina [3]. Patel advocated that the visibility of choroid vessels at the macula is essential for differentiating the detached and attached retina. In premature babies with less pigmented retinal pigment epithelium, the choroidal vessels are distinguishable through the transparent attached retina. The detachment of the retina in the macular region may prevent visualization of the choroidal vessels due to the intraretinal and subretinal fluid. In a recent report [1], 5 eyes were diagnosed as stage 4A based on clinical examination; OCT imaging revealed retinoschisis and no detachment in the area of perceived retinal elevation. They reported that these eyes progressed to stage 4A after primary laser treatment and regression of the retinoschisis in all eyes within the two months following IVB treatment. That report advocates the possible overdiagnosis of stage 4A based on indirect ophthalmoscopy compared to the OCT-based diagnoses. In addition, it supports the positive treatment response to the IVB treatment on retinoschisis simulating retinal detachment.

Differentiating the stage 4A ROP from severe stage 3 ROP can be challenging without OCT. Due to the lack of widespread use of commercially available handheld-OCT devices, the majority of clinicians have to make this distinction by using BIO and WFID in clinical practice. Therefore, we consider the present study important in terms of presenting the photographic features of severe stage 3 and stage 4A cases. However, since we did not differentiate between detachment and retinoschisis using OCT, we recommend that our results should be evaluated with caution. For all that, pre-treatment and post-treatment photographs of the included patients were presented as additional files in order for each reader could better assess the severity of the disease of the patients presented in the study.

During the last three decades, scleral buckle and lens sparing vitrectomy (LSV) was reported as the main treatment modality for stage 4A ROP [16,17,18,19]. Pre-surgical laser and IVB applications were reported to be with higher success of the surgery for the stage 4A ROP [5, 20, 21]. In a recent study that evaluates the parameters affecting the postoperative success of surgery for stage 4A ROP, it is reported that the presence of preoperative laser, anti-VEGF or combined treatment was found to be a higher anatomical and functional success [5]. Postoperative vitreous hemorrhage was significantly lower in eyes that had anti-VEGF treatment [5]. In two recent studies, it has been reported that preoperative laser treatment reduces the risk of iatrogenic retinal tear and lensectomy [5, 21]. Xu et al. [20] reported that less surgery time, lens removal ratio, higher anatomical success, better visual outcomes in eyes that underwent IVB prior to the surgery.

Sukgen and Kocluk [6] reported that stage 4A ROP less than six clock hours may regress without surgical treatment. Out of the 18 eyes with Stage 4 ROP treated with laser, eight eyes underwent LSV. They reported that, while all eyes that underwent LSV had ≥ 6 clock hours detachment, in the eyes that did not require LSV, 90% of the eyes had < 6 clock hours detachment [6]. In addition, in four eyes treated with intravitreal ranibizumab and whose detachment area was lower than six clock hours, they reported that LSV did not require. Narnaware and Bawankule reported 93% success with laser treatment on 14 eyes [7]. In thirteen eyes with successful treatment, the detachment was reported to be less than 4 clock hours. In the unsuccessful eye, the detachment was reported to be 6 clock hours.

Hyaloidal tractions that cause progression from stage 3 to stage 5 may originate from the posterior hyaloid near the fibrous membranes. Different tractional vectors may cause the progression of stage 3 ROP to advanced stages arising from various sites such as the ridge and the lens, the ridge and the optic disc, the ridge and the vitreous base, and circumferential traction along the ridge. According to our clinical experience before the study, we hypothesized that more posterior, thicker, and extended as clock hours fibrous membrane has more potential to progress further stages. Our study showed that the width of the fibrous bands, total area of the fibrous membranes, and total retinoschisis + detachment area were higher in the Stage 4A group when compared to severe stage 3.

Our study may not have a sufficient sample size to comprehensively compare Stage 4A cases that required vitrectomy and those that did not. In addition, the requirement for vitrectomy may be affected by our preferences and conditions during the clinical practice of the included eyes. We consider that higher ratio of width of the band/ DD, total area of ​​the fibrous membranes/DA, total retinoschisis + detachment area/ DA, detachment area/DA and lower ratio of distance from detachment-attached retina border to the fovea / DD in may be associated with requirement of vitrectomy. For all that, in the present study only the higher area of the detachment was found statistically higher in the vitrectomy required eyes. Further studies with larger sample sizes may provide more robust data on this subject.

The finding that higher optic disc to fovea distance in the vitrectomy non-required eyes may be a controversial finding. This finding may indicate that we had a lower tendency to recommend vitrectomy with a sequel finding such as dragging. In addition, it may be thought that some of the eyes with macular dragging may not deteriorate without vitrectomy, but our study is not sufficient to make a statement on this subject. The clinical courses of the eyes with dragging and whether the dragging is increased or decreased were not compared in our study. We consider that performing pre-treatment and post-treatment comparisons of the WFPs in a larger series of stage 4A ROP that underwent vitrectomy provides further information on this subject.

In a recent study that evaluated OCT findings of the eyes that were clinically diagnosed as stage 4A ROP, significant part of them did not have OCT evidence of retinal detachment [3]. The authors offered the term of stage 4A-schisis by OCT to describe this subgroup [3]. In our study, we consider that the subgroup of the vitrectomy-required Stage 4A eyes had more severe disease than those vitrectomy-nonrequired stage 4A eyes. The fact that the majority of the vitrectomy-nonrequired stage 4A eyes responded well to IVB, laser or both of them supports this assumption. We consider that numerous factors may have been influential on the course of stage 4A ROP. These are the presence of the OCT-proven detachment or retinoschisis, how far away the detachment or schisis area is from the fovea, and the extent of the detached or schisized retina. Although we evaluated the majority of them in the present study, we may have overdiagnosed detachment for the eyes with real retinoschisis due to the lack of OCT imaging. Variability and inconsistencies in the diagnoses of Stage 4A ROP may affect the treatment preferences of the clinicians and the reported treatment success of the studies.

It is reported that paradoxical vascular-fibrotic reaction may cause progressive tractional retinal detachment within the first two weeks after intravitreal anti-VEGF injection [22,23,24]; therefore, the vitreoretinal team should be kept on standby as emergency surgery may be required. According to the reported case reports, post anti-VEGF tractional detachments can be evaluated in two categories in relation to initial treatment and by starting time. The former one is the rapid progression of the detachment [22, 23], and the latter one is the insidious reactivation after the full regression of the ophthalmoscopic visible extraretinal neovascular proliferation [25, 26]. Subsequently, Yonekawa et al. [24] reported progressive tractional retinal detachment in eyes with thirty-five eyes treated with anti-VEGF in an international, multicentric case series. They described two new tractional detachment types on eyes treated with anti-VEGFs: (1) Midperipheral detachment with tight circumferential vectors, (2) Very posterior detachment with prepapillary contraction. In their series, 49% of the eyes were detached within four weeks. In the case reports with rapid progression, published photographs suggest that pre-treatment fibrovascular membranes were extended by 12 clock hours [22, 23].

In this present report, although in some cases, dragging of the posterior pole and fibrotic tissue deteriorated after the primary IVB treatment, in none of the eyes, progressive tractional detachment was observed. In the right eye of patient 24, we consider that dragging and detachment occurred due to unresponsiveness to the IVB treatment instead of paradoxical fibrotic reaction. Because after the initial IVB treatment, plus did not regress, and the disease progressed in a sequential manner following stages 2, 3 and 4A. In the right eye of patient 30, progressive fibrotic tissue and detachment were observed two weeks following the IVB treatment. Although this case can be labeled as a paradoxical fibrotic reaction after the IVB treatment, the fibrotic tissue and detachment were limited and regressed after the laser treatment. In addition, in the right eye of patient 31, although the dragging of the temporal vascular arcade and fibrotic tissue progressed after the IVB treatment, the detachment fully regressed two weeks after the IVB treatment. In both eyes of patient 36, fibrotic membranes were extended in 12 clock hours. Both eyes underwent laser and IVB treatment. During follow-up, due to the high risk of progressive tractional detachment right eye underwent LSV one week after the initial treatment. The fellow eye underwent additional laser posterior to the fibrotic tissue. In both eyes, no progressive tractional detachment was observed during the follow-up.

The major limitations of our study were that the final diagnosis was not made with the guidance of OCT device and retrospective design with a small sample. Therefore, the distinction between detachment and retinoschisis was made only by examination that was performed with BIO and the evaluation of the WFPs. We may tend to diagnose detachment over retinoschisis when in doubt. Another limitation is that all quantitative measurements were performed by a single observer. Therefore, interobserver and intraobserver variability of the quantitative measurements was not evaluated.

Further studies that evaluate the quantitative and qualitative parameters of fundus photographs and OCT findings of severe stage 3 and stage 4A may help to determine additional diagnostic signs that may prescience the presence of tractional retinoschisis and retinal detachment. In this way, it can be possible to determine the criteria for stage 3 with tractional retinoschisis, which can be controlled with laser and/or IVB without requiring surgery, and stage 4A, which has a risk of progression and will require surgery. Upcoming studies, with the assistance of artificial intelligence, analyze WFP for automated ROP staging, which may evaluate quantitatively and faster the thickness of the ridge, location, and extent of the ridge, retinoschisis, and detachment.

In our patient group with stage 4A ROP and limited detachment area, in the eyes treated with laser and IVB, macula was attached in final examination but in the majority of them dragging of the posterior pole was observed. In our limited sample size and with several limitations, our study suggests that laser may be considerable in the eyes with limited detachment area. Laser application to the posterior of the detachment border may prevent the progression of the detachment. During clinical practice, in eyes with temporal dragging of the macula, presence of the detachment should be evaluated carefully by BIO. In low-resource settings where the OCT is not available, the dragging of the optic disc and macula may be a cautionary finding for the presence of Stage 4A, and the higher detachment area may indicate the higher requirement of the PPV. Further multicenter, prospective studies with larger sample sizes are needed for the better categorization and standardization of severe stage 3 and stage 4A ROP, and these studies may provide more evidence for the treatment strategies for stage 4A ROP.