Early Outcomes of Intravitreal Aflibercept 8 mg in Eyes Previously Treated with Aflibercept 2 mg for Neovascular Age-Related Macular Degeneration with AI-Based Biomarker Quantification

Abstract

Introduction

To evaluate the early outcomes of aflibercept 8 mg (afl8) treatment in patients with neovascular age-related macular degeneration (nAMD) previously treated with aflibercept 2 mg (afl2).

Methods

A retrospective, observational, monocentric Swiss study. Patients with nAMD who had received ≥ 3 consecutive afl2 injections and switched to afl8 because of persistent or recurrent fluids, or to extend treatment intervals, were included in the study. All patients started a loading phase of 3-monthly afl8 injections, followed by a treat-and-extend regimen. Outcome measures included changes in best-corrected visual acuity (BCVA), maximal pigment epithelial detachment (PED) height, central subfield thickness (CST), optical coherence tomography (OCT) biomarkers quantified using artificial intelligence, and treatment intervals until month 6.

Results

Fifty-two eyes of 44 patients who concluded the loading phase were included in this analysis. Mean age was 80.2 ± 8.5 years; 73% of patients were females. At month 6, BCVA was unchanged, PED and CST height experienced a significant decrease by – 18.5 ± 12.9 μm (p = 0.0005) and – 14.0 ± 3.5 μm (p = 0.0042), respectively. Volumes of intraretinal fluid (IRF), subretinal fluid (SRF), and PED decreased (IRF: 4.4 ± 15.6–3.8 nl; SRF: 15.7 ± 35.7–10.3 ± 34.0 nl; PED: 268.2 ± 423.2–252.2 ± 484.1 nl). Mean treatment intervals increased by 1.7 ± 0.5 weeks from the last assigned interval and by 0.6 ± 0.2 weeks from previous maximal fluid-free intervals after switching (p = 0.0005 and p = 0.18, respectively). One mild vitritis was observed and resolved with vitrectomy and topical drops without decreased visual acuity.

Conclusion

Our real-world study supports the potential short-term benefits of afl8 in improving anatomical and durability outcomes in patients with recurrent nAMD. These findings also highlight the added value of data-driven evaluations. Long-term studies are needed to confirm the effectiveness and durability of afl8 in this population.

Supplementary Information

The online version contains supplementary material available at 10.1007/s40123-026-01320-5.

Key Summary Points

Why carry out this study?

Neovascular age-related macular degeneration (nAMD) is a leading cause of irreversible vision loss in older populations, with many patients experiencing persistent or recurring fluids despite frequent injections

The CANDELA and PULSAR clinical trials demonstrated the non-inferiority of aflibercept 8 mg compared with aflibercept 2 mg. Switching to a higher molecular dose could improve anatomical stability, therefore requiring fewer injections

Limited real-world evidence exists regarding the outcomes of switching therapy from aflibercept 2 mg to its higher dose option

What was learned from the study?

Switching to the higher dose of aflibercept resulted in tight fluid control and treatment interval extension, particularly in poor responders

Aflibercept 8 mg is well tolerated in patients previously treated with aflibercept 2 mg

AI-based biomarker quantification could provide a deeper understanding of treatment outcomes

Introduction

Neovascular age-related macular degeneration (nAMD) is a macular pathology leading to a progressive deterioration of central visual acuity [1]. Although nAMD accounts for a small percentage of total AMD cases, it represents a leading cause of blindness in developed nations, especially among individuals aged > 60 years [2]. Since the advent of treatments targeting vascular endothelial growth factor (VEGF), the incidence of legal blindness and visual impairment has declined [3].

Anti-VEGF and angiopoietin-2 (Ang-2) inhibition therapies are the standard for nAMD management. However, some patients show inadequate response or develop resistance over time, requiring monthly injections [4, 5]. The narrow treatment intervals impose a substantial burden on patients, caregivers, and clinicians, often limiting their practicality in real-world settings [6]. The suboptimal response to treatment is well documented in the literature, and it is estimated that 20% of eyes will never achieve complete retinal dryness despite continuous therapy [7].

The recent expansion of available agents for the treatment of nAMD has made it possible to switch molecules in challenging cases of inadequate therapeutic response [8, 9]. Time-to-market differences of molecules, and especially their different doses and mechanisms of action, have permitted their denomination into first- and second-generation agents. The first anti-VEGF generation, namely bevacizumab, aflibercept 2 mg (afl2), and ranibizumab, has demonstrated efficient treatment outcomes [10]. Molecular switching in case of potential resistance mechanisms of poor responders has been accurately described among first-generation agents. Specifically, switching to afl2 resulted in improved anatomical outcomes and extended treatment intervals [11]. Resistance mechanisms could be linked to the persistent stimulation of proangiogenic factors and the complement system, or the activation of new pathogenic pathways [6].

However, the frequent injections required in poor responders represented a considerable strain, indicating the requirement for therapeutic advances. Second-generation agents, namely faricimab, brolucizumab, and aflibercept 8 mg (afl8), were then introduced, increasing the possibility of reaching extended durability outcomes by focusing on higher molecular dosing [12] or dual-pathway activity [13]. Recent studies highlighted the benefits of switching suboptimal responders to more recent agents [13]. Patient outcomes significantly improved with the adjusted treatment mechanisms and dynamics [12]. In this context, both the CANDELA and PULSAR clinical trials demonstrated not only the non-inferiority results of afl8 treatment outcomes compared with afl2 but also the potential of therapeutic benefit [14, 15]. The analysis of a higher dose molecule is of particular interest in treatment-resistant patients to determine the potential of treatment interval extension.

Assessing treatment efficacy is becoming increasingly important to characterize molecular effects and to understand their suitability for individual cohorts. Determining the most effective therapeutic management for patients with nAMD remains a complex and unresolved clinical challenge. The use of artificial intelligence (AI) can enhance clinical analysis of treatment molecules, therefore providing an in-depth understanding of their outcomes [16]. AI has emerged as a transformative tool in optical coherence tomography (OCT) imaging analysis, offering objective, fully automated capabilities to detect and quantify compartmental retinal fluids and their location within key macular zones, translating into optimal treatment outcome understanding and prediction [17, 18].

This study aims to evaluate the efficacy of afl8 in treatment-resistant patients with nAMD and provide meaningful insights from a real-world setting. In the management of nAMD, where treatment efficacy and treatment intervals are closely linked to retinal fluid status [19], we utilized advanced AI-based analysis of OCT biomarkers to comprehensively evaluate the early anatomical response to afl8 intravitreal injections in the first 6 months of treatment.

Methods

Study Description

This observational, retrospective, single-arm study was conducted at the Retina Unit of the Swiss Visio Montchoisi Ophthalmology Center in Lausanne, Switzerland. The study design was approved by the local ethics committee (Commission cantonale d'éthique de la recherche sur l'être humain CER-VD—part of Swissethics) for data collection and analysis (CERVD: 2024–01491) and was conducted in accordance with the principles of the Helsinki Declaration and its amendments. Informed consent was collected prior to patient inclusion in the study.

Inclusion Criteria

Patients with an active macular neovascularization (MNV) secondary to AMD and who received ≥ 3 consecutive afl2 intravitreal injections and who were switched to afl8 because of persistent or recurring fluids, or to extend treatment intervals, were enrolled in the study. Both eyes were included if eligible.

Study Process and Treatment Regimen

A full ophthalmological examination was performed at each visit, including ETDRS best corrected visual acuity (BCVA) and spectral domain optical coherence tomography (SD-OCT, Heidelberg Inc®). MNV types were collected from electronic medical records whenever a prior evaluation was available before initiating any anti-VEGF treatment.

After switching, patients received 3 monthly afl8 injections (loading phase), in accordance with the Swiss label, aiming for complete dryness prior to interval extension. This was followed by a Treat & Extend regimen (TER) with a treatment interval adjustment of 2 weeks to a maximum treatment interval of 16 weeks primarily based on disease activity, notably the presence or absence of intraretinal fluid (IRF) and subretinal fluid (SRF), as well as BCVA assessments. Sub-retinal pigment epithelium (RPE) fluid and pigment epithelium detachment (PED) were also considered if they showed signs of increased disease activity; however, their stable appearance did not preclude treatment interval extension. Monthly injections were continued in case of fluid persistence after loading.

Study Outcome

Clinical functional and anatomical measurements were collected monthly until the end of the loading phase and at month 6: BCVA in ETDRS (functional outcomes), CST, PED height, and presence of IRF, SRF, and sub-RPE fluid (anatomical outcomes). Additionally, using a validated and CE-marked AI-based quantification tool (RetinAI Discovery CORE®; version 5.4), biomarkers such as outer nuclear layer, Henle fiber layer (ONL + HFL), myeloid zone (MZ), photoreceptors (EZ + OPR + IZ), retinal pigment epithelium (RPE), choriocapillaris and choroidal stroma (CC + CS), IRF, SRF, and PED were quantified (anatomical outcomes). Notably, all OCT cube images were transferred to a secure cloud-based system and then processed for the automated quantification. Data on the performance and repeatability of OCT biomarker quantification have been previously published [20, 21].

Primary Outcomes

The primary outcome was defined as the mean change in treatment intervals before and after 6 months of switching to afl8. Specifically, for a better assessment of pre-switch treatment durability, two treatment intervals were measured: the maximal fluid-free interval, defined as the longest continuous period (in weeks) during the pre-switch treatment phase in which OCT showed absence of disease activity; the last assigned treatment interval, defined as the time in weeks between the last injection of afl2 and the first injection with afl8.

Secondary Outcome

The secondary outcome was the qualitative and quantitative assessment of functional and anatomical outcomes after 6 months of treatment, as well as safety endpoints at each visit. Additionally, a durability sub-analysis was performed in a subgroup of patients classified as poor responders, defined as those whose maximum fluid-free interval was ≤ 6 weeks during the treatment period preceding the switch, indicating persistent or frequently recurring disease activity despite treatment. A similar methodological approach for data collection and outcome analysis was described in our recent study focusing on treatment-naïve patients [22].

Statistics

Sample Size

The sample size was determined assuming a normally distributed variable both at baseline and the end of the study. We considered an expected standard deviation of 15 patients and an average difference of approximately 7.5 patients to produce an effect size of 0.5. With an effect size of ρ = 0.5, a significance level of the test at α = 5%, and a test power of 80%, we obtained a minimum sample size of 28 patients.

Analysis

Descriptive statistics were performed on all study variables. Chi-square test was applied to evaluate the change in qualitative variables over time (IRF, SRF, and sub-RPE status). Normality of data distribution was evaluated using the Kolmogorov-Smirnov test. The statistical difference between continuous variables was assessed using the Wilcoxon signed-rank test. GraphPad Prism software (GraphPad Software, San Diego, CA) was used for statistical calculations. All data are presented as mean ± standard deviation (SD). A P-value ≤ 0.05 was considered statistically significant.

Results

Baseline Demographics

Fifty-two eyes of 44 patients were included in the study. The cohort was predominantly represented by females (73%), with a mean age of 80.2 ± 8.5 years. The mean number of injections before switching molecules amounted to 32.6 ± 18.7 injections.

Eyes (n = 32) were qualitatively classified according to the type of MNVs. Twenty-one eyes were type 1 MNV (65.7%), five eyes were type 2 MNV (15.6%), five eyes were type 3 MNV (15.6%), and one eye was characterized by a mixed 1 and 2 type of MNV (3.1%). The MNV type of the remaining eyes is unknown. PEDs were present in 46 (88.5%) eyes and were classified as fibrous only (84.9%), predominantly fibrous (6.5%), predominantly serous (4.3%), and serous only (4.3%); 58.7% of PEDs were located subfoveally. At baseline, 37.2% and 45.1% of patients presented with IRF and SRF, respectively. Baseline characteristics are summarized in Table 1.

Table 1.

Baseline characteristics

Baseline characteristics	n = 44 patients, 52 eyes
Mean age, years	80.2 ± 8.5
Sex, females (%)	38 (73%)
BCVA (ETDRS letters)	72.8 ± 14.7
MNV characteristics (n = 32)
Type 1 (occult)	21 (65.7%)
Type 2 (classic)	5 (15.6%)
Type 1 and 2 (mixed)	1 (3.1%)
Type 3 (retinal angiomatous proliferation)	5 (15.6%)
Fluid type
Intraretinal fluid	37.2%
Subretinal fluid	45.1%
PED (n = 46) type
Fibrous	39 (84.9%)
Predominantly fibrous	3 (6.5%)
Predominantly serous	2 (4.3%)
Serous	2 (4.3%)
PED location
Subfoveal	27 (58.7%)
Non-subfoveal	19 (41.3%)

BCVA best-corrected visual acuity, MNV macular neovascularization, PED pigment epithelial detachment

Functional Outcomes

At baseline, 4 eyes had a BCVA ≤ 50 ETDRS letters, while 47 eyes had a BCVA ≥ 51 ETDRS letters. More than half of the cohort had a BCVA of > 76 letters. One eye did not have BCVA recorded at baseline.

After 3 monthly injections, visual acuity remained stable from a value of 72.8 ± 14.7 ETDRS letters to 71.8 ± 15.0 ETDRS letters (p = 0.1594). Moreover, it did not vary significantly at month 6 (72.9 ± 16.2; p = 0.5248).

Anatomical Outcomes

OCT biomarkers were quantified both manually and using AI. After the loading phase, CST decreased from 244.2 ± 67.0  to 222.0 ± 50.4 μm (p < 0.0001), and PED height decreased from 174.2 ± 111.6 to 162.1 ± 121.7 μm (p < 0.0001) (Fig. 1A, B).

Fig. 1.

Anatomical outcomes. Changes in means ± SD of CST (A) and PED height (B) in µm. CST central subfield thickness, PED pigment epithelial detachment

Qualitative reductions in IRF, SRF, and sub-RPE were observed in 43.2%, 66.7%, and 59.3%, respectively, of patients after the loading phase, with a > 40% decrease for all fluid types after the first injection (Fig. 2).

Fig. 2.

Qualitative analysis of fluid dynamics. Proportion of patients with fluids during the loading phase. IRF intraretinal fluid, SRF subretinal fluid, RPE retinal pigment epithelium

Clinically significant OCT biomarkers quantified by AI include IRF, SRF, PED, and HRF. After the loading phase, IRF decreased from 4.4 ± 15.6 nl to 1.2 ± 5.8 nl (p = 0.0822), SRF decreased from 15.7 ± 35.7 to 2.6 ± 8.1 nl (p = 0.0750), and PED significantly decreased from 268.2 ± 423.2 to 242.5 ± 461.4 nl (p < 0.0013) (Fig. 3). HRF significantly decreased from 0.4 ± 0.5 to 0.3 ± 0.2 nl (p = 0.0047) (Table 2).

Fig. 3.

Quantitative analysis of fluid dynamics. IRF intraretinal fluid, SRF subretinal fluid, PED pigment epithelial detachment

Table 2.

OCT biomarkers: early retinal anatomical remodeling

	AI-quantified biomarkers
	Baseline		Month 3		Month 6
Parameters	Mean	SD	Mean	SD	Mean	SD	P-value 0–3	P-value 0–6
CC-CS	188.0	65.7	173.4	61.9	185.7	86.5	< 0.0001	0.0015
EZ-OPR-IZ	25.2	6.3	24.5	6.4	25.8	8.2	0.0001	0.0001
HRF	0.4	0.5	0.3	0.2	0.3	0.3	0.0047	0.0241
IRF	4.4	15.6	1.2	5.8	3.8	14.1	0.0822	0.6327
MZ	20.9	4.3	20.6	4.5	21.5	5.5	0.0243	0.0313
ONL-HFL	57.1	7.1	57.3	7.1	58.7	12.0	0.2049	0.1016
PED	268.2	423.2	242.5	461.4	252.2	484.1	0.0013	0.1189
RPE	20.9	3.2	20.5	3.4	21.4	3.9	0.0031	0.0486
RT	294.5	28.8	291.1	30.0	299.6	52.8	< 0.0001	0.1569
SRF	15.7	35.7	2.6	8.1	10.3	34.0	0.075	0.1788

Means ± SD of all AI-quantified biomarkers at baseline, month 3, and month 6

CC-CS choriocapillaris + choroidal stroma, EZ-OPR-IZ photoreceptors, HRF hyperreflective foci, IRF intraretinal fluid, MZ myeloid zone, ONL-HFL outer nuclear layer + Henle fiber layer, PED pigment epithelial detachment, RPE retinal pigment epithelium, RT retinal thickness, SRF subretinal fluid

Statistically significant values (P ≤ 0.05) are shown in bold

Other AI-quantified biomarkers were CC-CS, EZ-OPR-IZ, ONL-HFL, MZ, RPE, and RT. CC-CS significantly decreased from 188.0 ± 65.7 to 173.4 ± 61.9 μm (p < 0.0001), EZ-OPR-IZ significantly decreased from 25.2 ± 6.3 to 24.5 ± 6.4 μm (p = 0.0001), ONL-HFL decreased from 57.1 ± 7.1 to 57.3 ± 7.1 μm (p = 0.2049), MZ significantly decreased from 20.9 ± 4.3 to 20.6 ± 4.5 μm (p = 0.0243), RPE significantly decreased from 20.9 ± 3.2 to 20.5 ± 3.4 μm (p = 0.0031), and RT significantly decreased from 294.5 ± 28.8 to 291.1 ± 30.0 μm (p < 0.0001) (Table 2).

At month 6, CST and PED height significantly decreased to 230.2 ± 63.5 μm and 155.7 ± 124.5 μm (p = 0.0042 and p = 0.0005, respectively) (Fig. 1). Most AI-quantified biomarkers significantly decreased (Table 2). Moreover, the proportion of patients with IRF, SRF, and sub-RPE decreased by 40.5%, 62.2%, and 66.7%, respectively (‘Supplementary Material’).

Treatment Intervals and Number of Injections

At the end of the loading phase, 46% of eyes (n = 24) were eligible for starting a Treat and Extend regimen, while the remaining eyes continued monthly injections.

Before switching to afl8, the maximal fluid-free interval reached was 6.2 ± 2.2 weeks, and the mean last treatment interval was 5.4 ± 1.9 weeks.

After 6 months of treatment, 298 injections were made. A mean last-assigned interval duration of 7.1 ± 2.3 weeks was reached with a mean number of 5.7 ± 0.8 injections, establishing an average increase of 1.7 ± 0.5 weeks (p = 0.0005) from the last interval before switch, and of 0.6 ± 0.2 weeks (p = 0.1780) from the previous maximal fluid-free interval.

In a subgroup analysis of patients classified as poor responders (n = 33), treatment intervals significantly increased from 5.2 ± 1.0 weeks to 7.0 ± 2.3 weeks ( p = 0.0003) following 6 months of afl8 therapy (Fig. 4).

Fig. 4.

Durability outcomes. Maximal fluid-free interval before and after switching to aflibercept 8 mg

Discontinuation Rate and Safety Profile

Forty-eight eyes concluded the 6-month observational time point. Two patients (3 eyes) were lost to follow-up after concluding the loading phase. One eye of one patient was switched back to the previous molecule because of a lack of anatomical improvement.

Intraocular inflammation was recorded in one patient who developed a vitritis and benefited from posterior vitrectomy, intraocular antibiotic injection, and topical treatment.

Discussion

Maintaining long-term disease control with fewer injections remains a major challenge in nAMD management, particularly for patients who have already undergone multiple anti-VEGF treatments. In this context, the ability to extend treatment intervals without sacrificing efficacy is of considerable value, as it may reduce the cumulative burden of frequent intravitreal injections and clinical visits [6].

nAMD management has been revolutionized over the years [5], positively impacting patient prognosis [9] and attenuation of patient burden [6]. However, some patients exhibit a poor response or develop treatment resistance over time, necessitating frequent injections to manage the disease [4, 5]. Anti-VEGF therapy may be ineffective either from the outset or after an initially favorable response [6]. A suboptimal initial effect can arise from various clinical factors, such as incorrect diagnosis or underlying genetic predisposition [6]. Additionally, pharmacological factors—including drug tolerance, tachyphylaxis, and structural changes in the neovascular complex—may contribute to the reduced efficacy [6]. Resistance may also be driven by alternative angiogenic mechanisms or activation of other pathogenic pathways, such as compensatory growth factors, persistent complement system activation, and ongoing inflammation [6]. These mechanisms can lead to the development of persistent or recurrent exudation, even after an initial therapeutic response. The formulation of a higher dose molecule could overcome resistance mechanisms in frequently treated patients.

The CANDELA and PULSAR clinical trials demonstrated that a higher dose of aflibercept has the potential for therapeutic benefits [14, 15]. Specifically, they demonstrated that afl8, administered at 12- or 16-week intervals following three initial monthly doses, delivers outcomes and safety profiles non-inferior and clinically equivalent to the standard 2 mg dose given every 8 weeks. However, the promising evidence of these studies lacks generalizability. The sole focus on naive patients’ results makes the evidence not applicable to pre-treated patients. Thus, gathering real-world data on afl8 is especially important for understanding its effectiveness and safety in poor responders who have previously received treatment with afl2.

Real-world analyses on cohorts switched to afl8 have already been published in the literature, however, with a general focus on short-term functional and anatomical outcomes and limited insights on durability [23–26].

Our study focuses on switched patients previously treated exclusively with afl2. Unsurprisingly, no functional improvements were observed in our cohort, confirming once again the state of chronic macular damage that characterizes switchers [27, 28]. However, we report improved anatomical outcomes, including reduction of retinal subfield thickness, PED height, and fluid volumes. The effects on our cohort align with the findings of Borchert et al., who also showed anatomical restructuring without visual gains [27].

Given that durability is a critical testing point for switch patients, some studies have reported the potential of afl8 to extend treatment intervals [24, 29]. Notably, Bala et al. analyzed a large cohort of 209 previously treated eyes switched to afl8 and reported an interval increase in the early treatment phases [29]. Similarly, our population experienced an increased treatment interval after switch. Specifically, patients described as poor responders had a mean treatment interval extension of almost 2 weeks. These outcomes could probably be attributable to the tight fluid control achieved with afl8. Indeed, we showed effective drying of compartmental fluids after the first injection and stabilization of disease activity.

Recent technological advances have led multiple research groups to embrace AI for OCT image analysis, specifically for fluid volume quantification [17, 30, 31]. In our study, we used an advanced AI-based segmentation to quantify additional OCT biomarkers for a comprehensive assessment of anatomical outcomes. Beyond fluid quantification, AI enabled a detailed characterization of retinal layers, PED morphology, and microstructural alterations that provide a more comprehensive understanding of anatomical responses in nAMD. Previous studies demonstrated the relevance of such quantitative biomarkers for morpho-functional outcomes [32].

Although injections are necessary to stabilize the disease, their association with ocular adverse events has been extensively described in the literature [33]. Notably, the most reported adverse events include infectious endophthalmitis (0.09%–2.9%), rhegmatogenous retinal detachment (up to 0.67%), and subconjunctival hemorrhage (approximately 10%) [34]. Clinical trials have already demonstrated the similarity between afl2 and afl8 regarding the safety profile [14, 15]. In our study, the incidence of adverse events was 0.33% and remained minimal throughout the observation period.

This study has several limitations that should be considered. As a single-center study, the findings may not be fully generalizable to other clinical settings. The retrospective design and the lack of a control cohort limit causal inference. Moreover, the sample size justification represents an intrinsic constraint, and the findings should therefore be interpreted as exploratory. The follow-up period was limited to 6 months, restricting insights into long-term outcomes, particularly regarding treatment durability. Future multicentric, randomized studies with longer follow-up are needed to confirm our results.

Conclusion

In our real-world study, switching to afl8 was associated with early anatomical improvements and longer treatment intervals in patients with recurrent nAMD, including those with limited durability before the switch. This study also underlines how data-driven analysis contributes to more accurate therapeutic evaluations, moving beyond conventional observational metrics. Afl8 could be a promising alternative when switching therapy to poor responders; however, the study is restricted to a short-term follow-up, and a comprehensive long-term analysis is needed to confirm both the effectiveness and durability of afl8 in this population.

Supplementary Information

Below is the link to the electronic supplementary material.

Author Contributions

Conception and design: Nicolò Bartolomeo, Konstantinos Kitsos-Kalyvianakis, Aude Ambresin. Data collection: Nicolò Bartolomeo, Konstantinos Kitsos Kalyvianakis, Baptiste Crozat, Yannic Pannatier Schuetz, Anna Chiara Nascimbeni, Daniela Gallo Castro, Aude Ambresin. Analysis and interpretation: Nicolò Bartolomeo, Konstantinos Kitsos-Kalyvianakis, Mamadou Pathé Barry, Aude Ambresin. All authors participated in drafting and critically revising the manuscript and approved the final version for submission. The authors had full control over the content and final decision on all aspects of this publication. All authors attest that they meet the current ICMJE criteria for authorship.

Funding

The study was partially financially supported by Bayer (Schweiz) AG (CH). Most of the study, including the Rapid Service Fee, was funded by the Swiss Visio Retina Research Center.

Data Availability

The datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request.

Declarations

Conflict of Interest

Nicolò Bartolomeo, Konstantinos Kitsos-Kalyvianakis, Yannic Pannatier Schuetz, Anna Chiara Nascimbeni, Daniela Gallo Castro, Baptiste Crozat, Mamadou Pathé Barry, and Aude Ambresin have nothing to disclose.

Ethical Approval

The study design was approved by the local ethics committee (Commission cantonale d'éthique de la recherche sur l'être humain CER-VD -part of Swissethics) for data collection and analysis (CERVD: 2024–01491) and was conducted in accordance with the principles of the Helsinki Declaration and its amendments. Informed consent was collected prior to patient inclusion in the study.