Importance of optical coherence tomography raster scans in early detection of active fellow-eye neovascularization in unilateral neovascular age-related macular degeneration

Abstract

Purpose

To investigate the incidence of and risk factors for failure of detection of active fellow-eye neovascularization on optical coherence tomography(OCT) crosshair scans in patients with unilateral neovascular age-related macular degeneration(AMD).

Methods

In this retrospective study, patients who experienced the development of active neovascularization in the fellow eye during the follow-up period were included(n = 75). Cases in which the neovascularization in the fellow eye could be identified solely through crosshair scans were defined as the crosshair scan detection group(n = 63). Cases in which the aforementioned findings could not be identified through crosshair scans but could be identified through raster scans were defined as the raster scan detection group(n = 12). The factors were compared between the two groups. Risk factors related to undetected neovascularization on crosshair scans were additionally identified.

Results

Active fellow-eye neovascularization, was not detected on OCT crosshair scans in 12 cases(16.0%) but was identified on raster scans in all cases. There was a significant difference in the proportion of neovascularization types between the crosshair scan detection group and the raster scan detection group(P = 0.023). Among the 35 fellow-eye neovascularization cases in patients with type 3 macular neovascularization(MNV), 10(28.6%) were not detected on crosshair scans. Multivariate analysis revealed a significantly higher risk for undetectable fellow-eye neovascularization on crosshair scans in patients with type 3 MNV than in those with typical neovascular AMD(P = 0.037,β = 9.600).

Conclusions

Our findings suggest the need for routine OCT raster scans during fellow-eye examinations in patients with unilateral neovascular AMD, particularly when the first-affected eye is diagnosed with type 3 MNV.

Introduction

Neovascular age-related macular degeneration (AMD) is a sight-threatening disorder that can affect both eyes [1]. Therefore, when one eye is diagnosed, it is not uncommon for the fellow eye to be affected over time, and the incidence of fellow-eye neovascularization is approximately 16.2–20.6% [2, 3]. The incidence is relatively lower in polypoidal choroidal vasculopathy (PCV) [2] and relatively higher in type 3 macular neovascularization (MNV) [4, 5].

Although the introduction of anti-vascular endothelial growth factor has markedly improved treatment outcomes in neovascular AMD, [6] delayed treatment can lead to irreversible damage, thus impeding visual recovery [7]. The visual acuity of the better-seeing eye is closely associated with the patient’s quality of life [8]. In patients with neovascular AMD receiving treatment in both eyes, there is a tendency for the eye diagnosed later to exhibit better visual acuity than the eye diagnosed earlier [9]. Therefore, prompt detection of fellow-eye involvement is crucial not only for preserving visual acuity in the fellow eye [10] but also for preventing deterioration of the patient’s quality of life.

The diagnosis of neovascular AMD involves not only clinical examination but also various multimodal imaging techniques [11, 12]. Among them, optical coherence tomography (OCT) is particularly important [13, 14]. OCT allows for the magnified visualization of retinal cross-sections, enabling the detection of even small amounts of fluid accumulation caused by early active neovascularization. OCT utilizes two main methods: crosshair scans that are horizontally and vertically centered on the fovea, and raster scans that employ multiple horizontal scans to cover the entire macular area.

Crosshair scans offer the advantages of rapid image acquisition and convenient verification of results in clinical practice. However, this approach has the limitation of examining only a partial area of the macula. In contrast, raster scans enable a more detailed assessment of abnormalities across the entire range of the macula but require more time and effort for image acquisition and interpretation. Particularly, when significant damage is present in the macular region, patient fixation during OCT examination can be challenging, leading to a prolonged examination time. Currently, no gold-standard OCT scanning protocol has been established for the evaluation of AMD in clinical practice. In addition, the precise differences in the abilities of crosshair and raster scans for detecting pathologic lesions associated with neovascular AMD have not been fully elucidated.

This study aimed to investigate the incidence of and risk factors for failure of detection of active fellow-eye neovascularization on OCT crosshair scans in patients with unilateral neovascular AMD. We specifically focused on evaluating potential variations in the detection rate according to the neovascular AMD subtype in the initially affected eye.

Methods

This study enrolled treatment-naïve patients diagnosed with neovascular AMD between January 2019 and December 2020 who were initially treated with three loading injections of ranibizumab (0.5 mg/0.05 mL; Lucentis^®; Genentech Inc., San Francisco, CA, USA) or aflibercept (2.0 mg/0.05 mL; Eylea^®; Regeneron, Tarrytown, NY, USA). Patients who developed fellow-eye neovascularization during the follow-up period were included.

The exclusion criteria were as follows: (1) follow-up period < 12 months after the initial diagnosis, (2) simultaneous diagnosis of neovascular AMD in both eyes, (3) lack of indocyanine green angiography (ICGA) results for the first-involved eye or inability to classify the AMD subtype due to extensive hemorrhage, (4) fellow-eyes with at least one of the following: retinal vascular occlusion; proliferative diabetic retinopathy; or conditions that can lead to secondary choroidal neovascularization such as angioid streak of lacquer crack, refractive error ≥-6.0 diopters, high myopia with an axial length ≥ 26.5 mm, findings of myopic macular degeneration, best-corrected visual acuity (BCVA)≤ 0.1 due to retinal or optic nerve abnormalities, and history of vitreoretinal surgery. The presence of subretinal/intraretinal hemorrhage or retinal thickening during the follow-up period was not considered an exclusion criterion.

Follow-up of the fellow-eye

The treatment and follow-up schedules were consistent with those in our previous study [10]. Follow-up examinations of the fellow eye were conducted at intervals of 2–6 months. Fellow-eye examination routinely included clinical examination, fundus photography, and OCT. In cases where neovascularization was suspected, fluorescein angiography was additionally performed. The decision to perform ICGA and OCT angiography was at the discretion of the attending physician.

All OCT scans were performed using a spectral-domain OCT device (Spectralis™; Heidelberg Engineering GmbH, Heidelberg, Germany). Horizontal and vertical crosshair OCT scans (Fig. 1A, B) aligned on the center of the fovea were performed in conjunction with raster scans covering a 30°x25° field (Fig. 1C). To improve visualization, 80–100 scans were averaged for the crosshair scans. For the raster scan, approximately 10 scans were averaged for a single OCT image and 31 scanning lines were used to cover the region.

Fig. 1

A representative figure showing horizontal (A) and vertical (B) crosshair scans and a raster scans (C) used to detect fellow-eye neovascularization in this study. White arrows in each panel indicate optical coherence tomography scanning lines

Outcome measures

Neovascular AMD lesions in the first-involved eye were classified into three subtypes: PCV, polypoidal lesions with and without branching vascular networks on ICGA; [15] type 3 MNV, presence of characteristic hyper-reflective outer retinal layer lesions associated with intraretinal edema with or without subretinal fluid or sub-retinal pigment epithelial fluid on OCT, accompanied with focal hyperfluorescence with late leakage on angiography at the neovascularization site [16]; and typical neovascular AMD without features of PCV or type 3 MNV. All image analyses were performed by a single experienced examiner (J.H.K.).

Fellow-eye neovascularization was defined as the occurrence of active neovascularization, indicated by the presence of intraretinal/subretinal fluid, pigment epithelium detachment (PED), or hemorrhages in association with the neovascularization. Non-exudative MNV [17, 18] was not considered fellow-eye neovascularization in this study.

OCT images taken at the time of incident neovascularization were reviewed in patients with fellow-eye neovascularization. Cases in which the presence of intraretinal/subretinal fluid, PED, or hemorrhages related to neovascularization could be identified solely through crosshair scans were defined as the crosshair scan detection group. In contrast, cases in which the aforementioned findings could not be identified on crosshair scans but could be detected on raster scans were defined as the raster scan detection group. The following factors were compared between the two groups: age, sex, type of neovascularization in the initially involved eye (typical neovascular AMD, PCV, or type 3 MNV), duration between the diagnosis of the first-involved eye and fellow-eye neovascularization, and interval of fellow-eye examination immediately before the diagnosis of neovascularization. Multivariate analysis was performed to identify the risk factors associated with undetected neovascularization on crosshair scans.

As part of additional analyses, the BCVA and central retinal thickness (CRT) of the fellow eye at the time of neovascularization identification were compared between the crosshair and raster scan detection groups. Additionally, BCVA and CRT values were compared between the three groups according to the type of neovascularization in the initially involved eye. For the initially involved eye, the proportion of cases in which the presence of intraretinal/subretinal fluid, PED, or hemorrhages related to neovascularization could be identified solely through crosshair scans at the time of the initial diagnosis of neovascularization was evaluated. In addittion, this proportion was compared among the different types of MNV.

Statistical analysis

Data are presented as mean ± standard deviation or number (%), where applicable. BCVA values were converted to the logarithm of the minimum angle of resolution (logMAR) value for analysis. Lesion reactivation was compared between the groups using the Mann–Whitney U test, chi-square test, or Fisher’s exact test. Multivariate analysis was performed using binary logistic regression analysis. Statistical analyses were performed using a commercially available software package (Statistical Package for the Social Sciences [SPSS] for Windows, version 21.0; IBM, Armonk, NY, USA). Statistical significance was set at P < 0.05.

Results

A total of 75 patients were included, and the mean follow-up period was 36.6 ± 8.5-month. Figure 2 shows the timing of fellow eye neovascularization in the included patients. Among these, the first-involved eye was diagnosed with typical neovascular AMD, PCV, and type 3 MNV in 25, 15, and 35 patients, respectively. The mean interval between the initial diagnosis and fellow-eye neovascularization was 19.2 ± 10.1 months.

Fig. 2

Kaplan–Meier graph showing the timing of fellow-eye neovascularization in typical neovascular age-related macular degeneration (n = 25, solid line), polypoidal choroidal vasculopathy (n = 15, dashed line), and type 3 macular neovascularization (n = 35, dotted line)

All 75 fellow-eye neovascularizations were detected on raster scans, whereas 12 (16.0%) were not detected on crosshair scans. Figures 3 and 4 show representative cases where fellow-eye neovascularization was detected only on raster scans. The type of neovascularization (P = 0.023) and interval of fellow-eye examination immediately before neovascularization (P = 0.046) was significantly different between the crosshair scan detection group (n = 63) and the raster scan detection group (n = 12) (Table 1). Among the 35 fellow-eye neovascularization cases in patients with type 3 MNV, 10 (28.6%) were not detected on crosshair scans. In the typical neovascular AMD and PCV groups, the proportions were 4.0% (1 of 25 cases) and 6.7% (1 of 15 cases), respectively. There was no difference in age (P = 0.789), sex (P = 0.490), or interval between the diagnosis of the first-involved eye and fellow-eye neovascularization (P = 0.351) between the two groups.

Fig. 3

Representative case of fellow-eye neovascularization in type 3 macular neovascularization (MNV). A patient was diagnosed with type 3 MNV in the left eye. Thirteen months after the initial diagnosis, there is no sign of fellow-eye neovascularization on horizontal (A) and vertical (B) crosshair optical coherence tomography (OCT) scan images. However, raster scan images show intraretinal hyperreflective foci (C, short arrow) accompanied by adjacent retinal edema (D, short arrow) inferior to the fovea. Type 3 MNV is observed on indocyanine-green angiography (E, arrow), and OCT angiography (F, arrow). Long horizontal arrows on panels C and D indicate the OCT scanning lines

Fig. 4

Representative case of fellow-eye neovascularization in type 3 macular neovascularization (MNV). A patient was diagnosed with type 3 MNV in the left eye. Twelve months after the initial diagnosis, there is no sign of fellow-eye neovascularization on horizontal (A) and vertical (B) crosshair optical coherence tomography (OCT) images. However, raster scan images show an intraretinal hyperreflective lesion (C, short arrow) accompanied by adjacent retinal edema and disruption of the retinal pigment epithelium infero-temporal to the fovea. Type 3 MNV is observed on indocyanine-green angiography (D, arrow), and OCT angiography (E, arrow). The long horizontal arrow on panel C indicates the OCT scanning line

Table 1 Comparison between the crosshair scan detection and cases raster scan detection groups

Multivariate analysis revealed a significantly higher risk of undetectable fellow-eye neovascularization on crosshair scans in patients with type 3 MNV than in those with typical neovascular AMD (P = 0.037, β = 9.600)(Table 2). Other factors, including age (P = 0.995), sex (P = 0.757), interval between the diagnosis of the first-involved eye and fellow-eye neovascularization (P = 0.292), and interval of fellow-eye examination immediately before neovascularization (P = 0.180) were not significantly associated with the risk of undetectable fellow-eye neovascularization.

Table 2 Analysis of risk factors associated with the undetected fellow-eye neovascularization on crosshair scans

The mean fellow-eye logMAR BCVA and CRT when neovascularization was noted were 0.33 ± 0.38 (Snellen equivalent, 20/42) and 377.3 ± 144.4 μm, respectively. BCVA was significantly better (mean logMAR BCVA, 0.17 ± 0.13 [20/29] vs. 0.36 ± 0.41 [20/45], P = 0.023) and CRT was significantly lower (mean, 274.4 ± 52.2 μm vs. 394.7 ± 148.4 μm, P = 0.001) in the raster scan detection group than in the crosshair scan detection group. The mean fellow-eye BCVA when neovascularization was detected was 0.31 ± 0.21(20/40), 0.44 ± 0.74 (20/55), and 0.29 ± 0.23(20/38) in the typical neovascular AMD, PCV, and type 3 MNV groups, respectively. The mean fellow-eye CRT was 395.8 ± 147.1 μm, 393.0 ± 208.3 μm, and 353.4 ± 106.1 μm in the typical neovascular AMD, PCV, and type 3 MNV groups, respectively. There was no significant difference in BCVA (P = 0.846) and CRT (P = 0.721) between the three groups.

In the initially involved eyes, pathological findings were detected on cross-hair scans in 73 eyes (97.3%) when neovascularization was first diagnosed. There was no difference in the detection rate based on OCT cross-hair scans among the different neovascularization subtypes (typical neovascular AMD vs. PCV vs. type 3 MNV = 100% vs. 100% vs. 93.9%, respectively, P = 0.685).

Discussion

In this study, the majority of active fellow-eye neovascularization that developed during the follow-up of unilateral neovascular AMD was detected on crosshair scans. However, approximately one-sixth of neovascularization cases could not be detected solely through crosshair scanning, and raster scanning was required for accurate diagnosis. Two factors, the type of neovascularization in the initially affected eye and the interval between fellow eye examinations, were significantly associated with undetected fellow-eye neovascularization on crosshair scanning.

In particular, the risk of failure to detect fellow-eye neovascularization on crosshair scan was 9.6 times higher in type 3 MNV, than in typical neovascular AMD. We believe that the primary reason for this finding is the characteristic lesion distribution in type 3 MNV. It is well known that type 3 MNV lesions typically develop with foveal sparing [19, 20]. The absence of a deep capillary plexus in the foveal region, which is postulated to be the origin of type 3 MNV, may contribute to the distinctive distribution pattern of the lesions [19].

Type 3 MNV is frequently accompanied by prominent macular edema and PED [21, 22]. However, during its early stages, the extent of the lesion is typically confined to a small area, presenting as small hyperreflective foci with mild edema in the surrounding region [16, 18]. For prompt diagnosis of type 3 MNV, OCT should be performed directly at the site of the lesion. This is particularly crucial when a lesion develops relatively distant from the fovea. Type 3 MNV is a progressive disease [22] with the potential for a relentless course, making the prevention of undertreatment a key aspect in its management [23, 24]. Given the potential for worse prognosis with disease progression, [25] early diagnosis is crucial for improving the treatment outcomes and prognosis. Considering the markedly high risk of fellow-eye neovascularization in type 3 MNV, it is imperative to perform raster scans and meticulously examine all acquired images during fellow-eye screening, even if it entails additional time and effort.

We propose implementing the aforementioned examination protocol even in patients with non-neovascular AMD accompanied by reticular pseudodrusen. Reticular pseudodrusen is frequently observed in eyes diagnosed with type 3 MNV [19, 26]. Moreover, the presence of pseudodrusen is speculated to be associated with the development of type 3 MNV [21]. Therefore, special attention should be paid during regular check-ups to the potential development of type 3 MNV in eyes with non-neovascular AMD accompanied by reticular pseudodrusen, and OCT raster scans should be routinely performed.

In recent years, significant progress has been made in the application of artificial intelligence (AI) in the field of ophthalmology, including the development of AI-based automated interpretation of OCT images for the detection of fluid in neovascular AMD [27]. Furthermore, an AI-enabled tool for treatment decisions in neovascular AMD is under investigation [28]. Based on the findings of this study, we believe that the development of sophisticated techniques for accurate analysis of raster scan images is crucial in establishing an AI model for determining treatment strategies in patients with type 3 MNV.

In this study, fellow-eye neovascularization was not detected on crosshair scans when the interval between fellow-eye examinations was shorter. It is possible that more frequent screenings may have led to the detection of neovascularization at an earlier stage, thus influencing study outcomes. In a previous study with type 3 MNV, a longer fellow-eye examination interval was associated with poor visual acuity and greater visual deterioration of the fellow eye at neovascularization [10]. To date, there is no established consensus for the optimal interval of fellow-eye examinations in cases of unilateral neovascular AMD. Both the present study and a previous study [10] underscore the importance of frequent fellow-eye examinations for the timely detection of neovascularization.

In contrast to the detection of fellow eye neovascularization, the diagnosis of neovascularization in the initially involved eye showed sufficient results with cross-hair scans alone in most patients, regardless of the type of neovascularization. Continuous fellow eye examinations during the treatment of unilateral neovascularization, regardless of the presence of symptoms, may allow for the early diagnosis of fellow eye neovascularization. In contrast, at the initial diagnosis, most patients visited the hospital with symptoms such as visual disturbances, suggesting that the disease was in a somewhat progressive state. This difference may lead to variations in the detection rates when using cross-hair scans alone.

This study focused on detecting neovascularization using OCT. However, in clinical practice, fundus examinations or photography are routinely used to identify retinal abnormalities. Because type 3 MNV are often accompanied by small retinal hemorrhages, [21, 29] a thorough examination of the fundus can help diagnose early type 3 MNV. In addition, advancements in OCT technology can reduce the time required to obtain raster scan images. For example, swept-source OCT consumes less time for image acquisition compared to the spectral domain OCT [30]. This advancement in OCT technology are expected to aid in detecting subtle abnormalities in the macular area through denser scanning lines.

This study had several limitations. First, this retrospective study was based on treatment data obtained in a clinical setting. Therefore, the follow-up protocol was not strictly controlled. Second, owing to the absence of monthly follow-up, the detection of fellow-eye neovascularization may have been delayed in some patients. Third, OCT angiography, which exhibits good sensitivity for detecting neovascularization in its early stages, was not routinely performed for fellow-eye monitoring. Fourth, since our study assessed active neovascularization, which is accompanied by exudation and hemorrhage, the results are not valid for subclinical neovascularization. Fourth, the sample size was relatively small. Lastly, all patients were Korean.

In conclusion, in unilateral neovascular AMD, 16% of active fellow-eye neovascularization was not detected on OCT crosshair scans but was identified using raster scans. The risk of undetected neovascularization on crosshair scans is markedly high when the initially affected eye is diagnosed with type 3 MNV. These findings suggest the need for routine OCT raster scans during fellow-eye examinations in cases of unilateral neovascular AMD, particularly in patients with type 3 MNV.