Choroidal Changes after Intravitreal Aflibercept, Faricimab, Brolucizumab, and High-Dose Aflibercept in Treatment-Naïve Neovascular Age-Related Macular Degeneration Authors: Sam Razavi, Matteo Menna, Serena Fragiotta, Sarah Van Herreweghe, Andrea Lazzerini, Marco Battista, Federico Beretta, Riccardo Sacconi, Giuseppe Querques Published: 2026 ABSTRACT PurposeTo compare choroidal structural changes after a loading phase of different anti-VEGF agents in treatment-naïve neovascular age-related macular degeneration (nAMD).DesignRetrospective, monocentric, observational cohort study.SubjectsFifty-six eyes of 56 treatment-naïve patients with nAMD.Methods, Intervention, or TestingPatients received 3 consecutive monthly intravitreal injections of aflibercept 2 mg, aflibercept 8 mg, brolucizumab, or faricimab. Multimodal imaging was performed at baseline and at 3 months using swept-source OCT. Central choroidal thickness (CCT) and choroidal vascularity index (CVI) were quantified. Changes from baseline were analyzed within and between treatment groups, with adjustment for age.Main Outcome MeasuresChange in CCT and CVI after the loading phase.ResultsAll treatment groups demonstrated a significant reduction in CCT at 3 months (all P ≤ 0.006). The greatest mean CCT reduction was observed with brolucizumab, followed by aflibercept 8 mg, faricimab, and aflibercept 2 mg. Age-adjusted analysis showed a significant effect of treatment on CCT change, with a significant treatment-by-age interaction (F(3, 48) = 6.24; P = 0.001; η2 = 0.197). Choroidal vascularity index decreased significantly only in the aflibercept 8 mg group (mean change: –4.9; P = 0.01), driven by a preferential reduction in the luminal component (P = 0.002). Brolucizumab predominantly affected the stromal component (P = 0.006) without significant change in CVI.ConclusionsAnti-VEGF therapy induces significant choroidal thinning in treatment-naïve nAMD during the loading phase. However, different agents demonstrate distinct choroidal compartmental effects. These findings suggest drug-specific patterns of choroidal remodeling, potentially influenced by molecular and pharmacologic properties, warranting further prospective evaluation to determine their clinical implications.Financial Disclosure(s)Proprietary or commercial disclosure may be found in the Footnotes and Disclosures at the end of this article. Age-related macular degeneration (AMD) is a leading cause of visual impairment and severe vision loss. The neovascular form (neovascular AMD [nAMD]) is characterized by the development of macular neovascularization (MNV), resulting in sudden central visual acuity loss, and, in many cases, leading to permanent visual disability and legal blindness. This deterioration profoundly affects the quality of life and functional independence.1 In pathological angiogenesis, the overexpression of VEGF-A is a key event that stimulates endothelial cell proliferation, migration, and new vessel formation.2, 3, 4 Anti-VEGF therapy has transformed the management of nAMD, demonstrating both efficacy and safety in enhancing visual and anatomical outcomes in affected patients.5,6 These drugs act by binding VEGF ligands and preventing their interaction with VEGF receptors, thereby blocking downstream signaling pathways that drive angiogenesis. Inhibiting VEGF leads to regression of MNV and is associated with reduced vascular permeability, leakage, and retinal edema.7, 8, 9 Because VEGF-A binds to its receptors (VEGFR1 and VEGFR2) on choroidal endothelial cells, supporting vessel survival and function, it is reasonable to anticipate a potential impact on choroidal thickness (CT) during anti-VEGF treatment.10, 11, 12 Several clinical studies have reported a significant reduction in CT after anti-VEGF treatment, particularly with the use of aflibercept.13,14 A previous study by 2 of the authors (S.R. and G.Q.) demonstrated significant modifications in CT after intravitreal ranibizumab treatment,15 further supporting the need to understand the impact of different anti-VEGF agents on choroidal vasculature. Treatment of nAMD has evolved significantly with the introduction of newer anti-VEGF agents.16 Aflibercept 2 mg (Eylea, Bayer), high-dose aflibercept 8 mg (Eylea HD, Bayer), brolucizumab (Beovu, Novartis), and faricimab (Vabysmo, Roche) are currently the primary treatment options.7,17,18 While various studies have compared the efficacy of these drugs, the existing literature on this topic remains fragmentary, with few studies providing direct comparisons among different anti-VEGF molecules. The aim of this study was to compare and analyze choroidal changes after intravitreal injections of aflibercept 2 mg, faricimab, brolucizumab, and aflibercept 8 mg in treatment-naïve patients affected by nAMD. Methods This monocentric, retrospective, observational, and longitudinal study included patients with treatment-naïve nAMD, defined as no prior anti-VEGF therapy, who received a loading dose of anti-VEGF treatment, including aflibercept 2 mg (Eylea, Bayer), aflibercept 8 mg (Eylea HD, Bayer), brolucizumab (Beovu, Novartis), and faricimab (Vabysmo, Roche). Patients were treated at Centre Ophtalmologie Transparence Tours between January 2024 and March 2025. The study was approved by the Institutional Review Board of the institution and adhered to the tenets of the Declaration of Helsinki. Written informed consent was obtained from all participants prior to inclusion in the study. Patients aged ≥50 years were eligible if they had treatment-naïve MNV secondary to AMD with active exudation (subretinal fluid, intraretinal fluid, or both) regardless of MNV type, warranting anti-VEGF therapy. Inclusion required complete clinical data and multimodal imaging comprising, at minimum, structural OCT, fluorescein angiography, and/or indocyanine green angiography to confirm the presence of MNV, or alternatively, OCT angiography demonstrating the presence of a neovascular network. Patients were excluded if they had an MNV secondary to other causes (i.e., myopia, inflammatory, pachychoroid disease spectrum), poor dioptric media interfering with imaging evaluation, macular atrophy or complete outer retina and retinal pigment epithelium atrophy according to the Classification of Atrophy Meeting criteria,19 history of intraocular surgery within the previous 6 months, and any other vitreoretinal and inflammatory disorders. Clinical and Multimodal Imaging Diagnosis Patients were evaluated by a senior retinal specialist (S.R.), who established the diagnosis of MNV according to the Consensus on Neovascular Age-Related Macular Degeneration Nomenclature.20 The diagnosis was based on clinical evidence of a neovascular disease in the context of AMD, confirmed on structural OCT, OCT angiography, and fluorescein angiography/indocyanine green angiography as needed. Color fundus photograph, OCT, and OCT angiography were performed using swept-source Deep Range Imaging OCT Triton (Topcon Corporation). Analysis of 3-dimensional macula and 12-line radial scan pattern with follow-up function were performed. OCT B-scans were obtained using 16/16 signal averaging, and only images with a quality index ≥93/100 were included to ensure optimal image quality.21 The built-in software automatically compensates for motion artifacts, but the images were also checked for artifacts by a single operator (S.V.H.). Central macular thickness was automatically calculated by the built-in software. All OCT acquisitions and measurements were conducted at a single center using the same device (Topcon Triton OCT) and by a single experienced operator to ensure consistency. Central choroidal thickness (CCT) was measured using the device’s built-in software, which automatically segments the choroid from the outer border of Bruch’s membrane to the choroid–scleral junction. A thickness map is generated, with CCT defined as the mean CT within the central 1-mm circle of the ETDRS grid. The methodology used to quantify CT in the present study is consistent with that previously described by 2 of the authors (S.R. and G.Q.).15 The choroidal vascularity index (CVI) was defined as the ratio of the luminal to the total choroidal area. Choroidal vascularity index was calculated on the subfoveal horizontal scan using postprocessing using the open-source software Fiji ImageJ2 (version 2.3.0/1.53f, https://imagej.net/). The region of interest was traced with the polygon tool between the outer boundary of the retinal pigment epithelium and the choroid–scleral junction. The analysis followed the method described by Sacconi et al22,23 applying a Niblack’s auto local threshold for image binarization. Dark pixels identified the luminal choroidal area, while all pixels within the region of interest represented the total choroidal area. All imaging analyses were performed by a single expert retinal specialist (S.R.). To ensure consistency and reduce measurement bias, a second independent retinal specialist (R.S.) reviewed a subset of images and all cases in which the primary grader expressed uncertainty. Discrepancies were resolved by consensus. All imaging acquisitions and analyses were performed at baseline, defined as time of diagnosis, and 1 month after completion of the loading phase. Given the real-world retrospective setting, treatment allocation was not randomized. The choice of anti-VEGF agent was primarily determined by drug availability in France during the study period, following a temporal sequence (aflibercept 2 mg, brolucizumab, faricimab, and aflibercept 8 mg), and by the treating physician’s discretion. Statistical Analysis Quantitative data were summarized as mean ± standard deviation. Normality was evaluated using the Shapiro-Wilk test, with a P value <0.05 indicating a deviation from normality. Changes in quantitative variables before and after the loading phase were analyzed using paired t-tests or Wilcoxon signed-rank tests when normality assumptions were not met. To compare CCT and CVI changes among treatment groups, ΔCCT values (month 3–baseline) were calculated, and an analysis of variance was performed on these delta values. When significant, post hoc pairwise comparisons were corrected for multiple testing. Homogeneity of variances was evaluated using Levene test. A 2-sided P value threshold of 0.05 was considered statistically significant. All statistical analyses were conducted using RStudio software (version 2022.07.1; RStudio, PBC, URL: http://www.rstudio.com/). Results Fifty-six eyes of 56 patients with treatment-naïve nAMD fulfilled the inclusion and exclusion criteria and were included in the study. Among these, 23 eyes received aflibercept 2 mg (mean age = 85 ± 5.0), 9 eyes received brolucizumab 6 mg (mean age = 78.7 ± 9.0), 10 eyes received aflibercept 8 mg (mean age = 78.4 ± 7.0), and 14 eyes received faricimab 6 mg (mean age = 80 ± 9.2). A descriptive table of baseline patient characteristics and MNV subtypes is presented in Table 1. Although the overall analysis of variance showed a significant age difference among groups, post hoc Tukey tests did not detect significant pairwise differences after adjustment for multiple comparisons. No relevant ocular or systemic adverse events were observed after the loading phase in any treatment group.Table 1Baseline Characteristics of the Population EnrolledAfliberceptBrolucizumabAflibercept 8 mgFaricimabP ValueTotal eyes2391014-Age (yrs)85 ± 5.078.7 ± 9.078.4 ± 7.080 ± 9.20.03Gender (male/female)9/144/55/54/100.74Subfovéal CT (μm)203 ± 75.2187.2 ± 77.5188.5 ± 60.0187.1 ± 58.90.87Choroidal vascularity index (CVI)31.2 ± 12.627.9 ± 14.233.7 ± 6.529.5 ± 13.20.74MNV subtypes, n (%) Type 1 MNV12 (52.2%)5 (55.6%)6 (60%)7 (50%)1∗ Type 2 MNV7 (30.4%)2 (22.2%)3 (30%)5 (35.7%)0.91∗ Type 3 MNV4 (17.4%)2 (22.2%)1 (10%)2 (22.2%)0.82∗CT = choroidal thickness; MNV = macular neovascularization.Data are expressed as mean ± standard deviation. P values are calculated using analysis of variance test with Levene test for homogeneity of variances.∗Fisher exact test (Freeman-Halton extension for 2 × 4 tables). Table 1Baseline Characteristics of the Population EnrolledAfliberceptBrolucizumabAflibercept 8 mgFaricimabP ValueTotal eyes2391014-Age (yrs)85 ± 5.078.7 ± 9.078.4 ± 7.080 ± 9.20.03Gender (male/female)9/144/55/54/100.74Subfovéal CT (μm)203 ± 75.2187.2 ± 77.5188.5 ± 60.0187.1 ± 58.90.87Choroidal vascularity index (CVI)31.2 ± 12.627.9 ± 14.233.7 ± 6.529.5 ± 13.20.74MNV subtypes, n (%) Type 1 MNV12 (52.2%)5 (55.6%)6 (60%)7 (50%)1∗ Type 2 MNV7 (30.4%)2 (22.2%)3 (30%)5 (35.7%)0.91∗ Type 3 MNV4 (17.4%)2 (22.2%)1 (10%)2 (22.2%)0.82∗CT = choroidal thickness; MNV = macular neovascularization.Data are expressed as mean ± standard deviation. P values are calculated using analysis of variance test with Levene test for homogeneity of variances.∗Fisher exact test (Freeman-Halton extension for 2 × 4 tables). Baseline Characteristics of the Population Enrolled CT = choroidal thickness; MNV = macular neovascularization. Data are expressed as mean ± standard deviation. P values are calculated using analysis of variance test with Levene test for homogeneity of variances. Fisher exact test (Freeman-Halton extension for 2 × 4 tables). Differences in best-corrected visual acuity and central macular thickness before and after the loading phase are summarized in Table 2.Table 2Changes in BCVA and CMT Pretreatment and PosttreatmentAflibercept, N = 23Brolucizumab, N = 9Aflibercept 8 mg, N = 10Faricimab, N = 14BCVA pre0.48 ± 0.40.42 ± 0.10.45 ± 060.5 ± 0.5BCVA post0.34 ± 0.30.35 ± 0.10.13 ± 0.10.3 ± 0.3P∗0.020.110.0090.01CMT pre283.9 ± 54.5332.8 ± 115.9270.7 ± 43.5331.9 ± 105CMT post232.1 ± 30.7212.4 ± 40.7230.6 ± 29.5283.6 ± 53P∗<0.0010.0050.040.01BCVA = best-corrected visual acuity; CMT = central macular thickness.P∗: paired t-test to assess difference between baseline (pre) and after the loading phase (post). Best-corrected visual acuity is expressed in logarithm of the minimum angle of resolution units. Table 2Changes in BCVA and CMT Pretreatment and PosttreatmentAflibercept, N = 23Brolucizumab, N = 9Aflibercept 8 mg, N = 10Faricimab, N = 14BCVA pre0.48 ± 0.40.42 ± 0.10.45 ± 060.5 ± 0.5BCVA post0.34 ± 0.30.35 ± 0.10.13 ± 0.10.3 ± 0.3P∗0.020.110.0090.01CMT pre283.9 ± 54.5332.8 ± 115.9270.7 ± 43.5331.9 ± 105CMT post232.1 ± 30.7212.4 ± 40.7230.6 ± 29.5283.6 ± 53P∗<0.0010.0050.040.01BCVA = best-corrected visual acuity; CMT = central macular thickness.P∗: paired t-test to assess difference between baseline (pre) and after the loading phase (post). Best-corrected visual acuity is expressed in logarithm of the minimum angle of resolution units. Changes in BCVA and CMT Pretreatment and Posttreatment BCVA = best-corrected visual acuity; CMT = central macular thickness. P∗: paired t-test to assess difference between baseline (pre) and after the loading phase (post). Best-corrected visual acuity is expressed in logarithm of the minimum angle of resolution units. All treatment groups demonstrated a significant reduction in CCT at 3 months. In the aflibercept 2 mg group, mean baseline CCT was 203 ± 75.2 μm, decreasing to 184.2 ± 76.3 μm at month 3 (P < 0.001). In the brolucizumab group, CCT decreased from 187.2 ± 77.5 μm to 155.8 ± 68.2 μm (P < 0.001). In the aflibercept 8 mg group, CCT decreased from 188.5 ± 60.0 μm to 161.1 ± 60.8 μm (Wilcoxon: P = 0.006). In the faricimab group, baseline CCT was 187.1 ± 58.9 μm, decreasing to 165.7 ± 62.4 μm at month 3 (P < 0.001). The greatest mean reduction was observed with brolucizumab (–31.4 ± 13.4 μm), followed by aflibercept 8 mg (–27.4 ± 21.9 μm), faricimab (–21.3 ± 9.9 μm), and aflibercept 2 mg (–18.7 ± 12.1 μm) (Fig 1). An analysis of covariance test was performed to evaluate differences in ΔCT among treatment groups while controlling for age. The analysis revealed a significant main effect of treatment group, F(3, 48) = 6.465, P < 0.001, η2 = 0.204, indicating that the magnitude of CT change differed among treatment groups after adjusting for age. Age itself was also a significant covariate, F(1, 48) = 8.85, P = 0.005, η2 = 0.093. Importantly, there was a significant group ∗ age interaction, F(3, 48) = 6.24, P = 0.001, η2 = 0.197, indicating that the effect of treatment on ΔCT varied across patient ages. Baseline CT was not significantly associated with ΔCT (P = 0.139), and no significant interaction with treatment group was observed (P = 0.274). Clinical examples are reported in Figures 2 and 3.Figure 1Subfoveal choroidal thickness (CT) changes according to anti-VEGF molecules and molar dose. Top panel illustrates the molecular configuration of aflibercept 2 mg, faricimab, aflibercept 8 mg, and brolucizumab, arranged from left to right in order of increasing molar dose (red triangle). The bar graph depicts the absolute mean change in subfoveal CT from baseline to month 3, expressed in micrometers (μm), with error bars representing standard deviation. The table below reports the corresponding mean ΔCT values (baseline–month 3) with 95% confidence intervals (CIs). Increasing choroidal thinning is observed with higher molar dosing across agents.Figure 2Representative case of a patient treated with aflibercept 8 mg. (A) Baseline examination showing a color fundus photograph with superimposed choroidal thickness map. The green line indicates the location of the corresponding B-scan. (B) Three months after treatment, a marked reduction in choroidal thickness is observed, with the chorioscleral junction indicated by white triangles.Figure 3Representative case of a patient treated with brolucizumab. (A) Baseline examination showing a color fundus photograph with superimposed choroidal thickness map. The green line indicates the location of the corresponding B-scan. (B) Three months after treatment, a marked reduction in the choroidal profile (white arrows) as also confirmed in the choroidal thickness map. Subfoveal choroidal thickness (CT) changes according to anti-VEGF molecules and molar dose. Top panel illustrates the molecular configuration of aflibercept 2 mg, faricimab, aflibercept 8 mg, and brolucizumab, arranged from left to right in order of increasing molar dose (red triangle). The bar graph depicts the absolute mean change in subfoveal CT from baseline to month 3, expressed in micrometers (μm), with error bars representing standard deviation. The table below reports the corresponding mean ΔCT values (baseline–month 3) with 95% confidence intervals (CIs). Increasing choroidal thinning is observed with higher molar dosing across agents. Representative case of a patient treated with aflibercept 8 mg. (A) Baseline examination showing a color fundus photograph with superimposed choroidal thickness map. The green line indicates the location of the corresponding B-scan. (B) Three months after treatment, a marked reduction in choroidal thickness is observed, with the chorioscleral junction indicated by white triangles. Representative case of a patient treated with brolucizumab. (A) Baseline examination showing a color fundus photograph with superimposed choroidal thickness map. The green line indicates the location of the corresponding B-scan. (B) Three months after treatment, a marked reduction in the choroidal profile (white arrows) as also confirmed in the choroidal thickness map. The analysis of CVI demonstrated a significant reduction from 33.7 ± 6.5% to 28.7 ± 8.3% (P = 0.01) in the aflibercept 8 mg group. In the aflibercept 2 mg, the CVI change from 31.2 ± 12.6% and 30 ± 13.8% did not reach the statistical difference (P = 0.26), likewise for brolucizumab (27.9 ± 14.2% vs. 27.9 ± 16%, P = 0.98) and faricimab (29.5 ± 13.2% vs. 29 ± 12.5%, P = 0.78). No significant differences in ΔCVI were found among treatment groups in the unadjusted analysis (1-way analysis of variance: F(3, 52) = 1.50, P = 0.20) or after adjusting for age (analysis of covariance: F(3, 48) = 0.44, P = 0.73, η2 = 0.03). A subanalysis of the CVI parameters analyzed is reported in Table 3. This analysis showed differences among treatment groups in terms of the choroidal compartments affected. In particular, aflibercept 8 mg had a greater effect on the luminal area (P = 0.002), while brolucizumab demonstrated a significant effect on stromal area (P = 0.006).Table 3Subanalysis for Choroidal Vascularity Index Parameters, Including Total, Luminal, and Stromal AreasAflibercept, N = 23Brolucizumab, N = 9Aflibercept 8 mg, N = 10Faricimab, N = 14P∗ ValueTotal area (pixels) Baseline6044 ± 17255908 ± 15025745 ± 14335622 ± 15050.87 Posttreatment5902 ± 20985129 ± 14305197 ± 12325306 ± 15680.56 P†0.380.010.0040.01 Delta CVI–1.20.05–4.9–0.890.24 95% CI–4.3, 0.93–6.9, 1.6–7.6, –2.7–3.1, 1.9Luminal area (pixels) Baseline2107 ± 12471815 ± 12071986 ± 8211783 ± 11480.83 Posttreatment2055 ± 15191617 ± 11531565 ± 8051676 ± 11450.67 P†0.680.270.0020.23Stromal area (pixels) Baseline3937 ± 6294092 ± 4693759 ± 7723838 ± 5740.66 Posttreatment3847 ± 7023511 ± 4333632 ± 5343630 ± 5700.47 P†0.140.0060.320.07CI = confidence interval; CVI = choroidal vascularity index.Data are expressed as mean ± standard deviation. P∗ values are calculated using analysis of variance test with Levene test for homogeneity of variances or P† as a paired t-test between baseline and posttreatment.Bold values indicate statistically significant within-group changes from baseline to posttreatment, with P < 0.05. Table 3Subanalysis for Choroidal Vascularity Index Parameters, Including Total, Luminal, and Stromal AreasAflibercept, N = 23Brolucizumab, N = 9Aflibercept 8 mg, N = 10Faricimab, N = 14P∗ ValueTotal area (pixels) Baseline6044 ± 17255908 ± 15025745 ± 14335622 ± 15050.87 Posttreatment5902 ± 20985129 ± 14305197 ± 12325306 ± 15680.56 P†0.380.010.0040.01 Delta CVI–1.20.05–4.9–0.890.24 95% CI–4.3, 0.93–6.9, 1.6–7.6, –2.7–3.1, 1.9Luminal area (pixels) Baseline2107 ± 12471815 ± 12071986 ± 8211783 ± 11480.83 Posttreatment2055 ± 15191617 ± 11531565 ± 8051676 ± 11450.67 P†0.680.270.0020.23Stromal area (pixels) Baseline3937 ± 6294092 ± 4693759 ± 7723838 ± 5740.66 Posttreatment3847 ± 7023511 ± 4333632 ± 5343630 ± 5700.47 P†0.140.0060.320.07CI = confidence interval; CVI = choroidal vascularity index.Data are expressed as mean ± standard deviation. P∗ values are calculated using analysis of variance test with Levene test for homogeneity of variances or P† as a paired t-test between baseline and posttreatment.Bold values indicate statistically significant within-group changes from baseline to posttreatment, with P < 0.05. Subanalysis for Choroidal Vascularity Index Parameters, Including Total, Luminal, and Stromal Areas CI = confidence interval; CVI = choroidal vascularity index. Data are expressed as mean ± standard deviation. P∗ values are calculated using analysis of variance test with Levene test for homogeneity of variances or P† as a paired t-test between baseline and posttreatment. Bold values indicate statistically significant within-group changes from baseline to posttreatment, with P < 0.05. Discussion The present study compared the effect of different anti-VEGF agents in modulating the choroid. It has already been demonstrated that anti-VEGF therapy reduces CT in typical treatment-naïve nAMD during the loading dose.24 This effect has been demonstrated to be superior in aflibercept when compared to ranibizumab, but also a tendency of a greater effect observable with brolucizumab has been hypothesized.24,25 However, whether the choroidal thinning observed during anti-VEGF therapy is a direct pharmacologic effect or part of the natural course of AMD remains debated.26 In this context, evidence of an inverse correlation between the number of injections and CT supports the hypothesis that cumulative anti-VEGF exposure may contribute to enhanced choroidal thinning.27 Nevertheless, the heterogeneity of available studies and the diversity of anti-VEGF agents examined result in a limited understanding of this phenomenon and its pathogenic and prognostic implications. Our study contributes to clarifying this issue by confirming a reduction in CCT after the loading phase for all examined anti-VEGF agents, after accounting for age. Notably, the treatment factor explained approximately 20% of the variance in ΔCT after adjusting for age (η2 = 0.204), indicating a clinically relevant contribution of drug type to choroidal remodeling. Age also emerged as an independent predictor of ΔCT, and the significant treatment group × age interaction underscores that the magnitude of CT change is age-dependent and that the relative impact of each drug varies across the age spectrum. The age–CT relationship has been confirmed by multiple large studies,28, 29, 30 with a more pronounced age-related thinning observed after the age of 50.31,32 In our series, a consistent trend toward a greater thinning in the brolucizumab subgroup was noted. This differential effect may be partly explained by molecular characteristics of the drugs themselves. Several newer anti-VEGF agents have been introduced in clinical practice, and all 4 agents evaluated in our study (aflibercept 2 mg, aflibercept 8 mg, faricimab, and brolucizumab) bind to VEGF-A. These drugs differ markedly in molecular size, with brolucizumab having a molecular weight of approximately 26 kDa, faricimab about 110 kDa, and both aflibercept formulations roughly 115 kDa, which may influence their intravitreal pharmacokinetics and tissue penetration.33 Consequently, these differences in molecular weight translate into substantial differences in molar dose. Brolucizumab (26 kDa) delivers approximately 0.31 μmol per injection; faricimab (110–149 kDa) provides around 0.05 to 0.07 μmol per injection; standard aflibercept (115 kDa) delivers approximately 0.035 μmol, whereas aflibercept HD, with the same molecular weight, provides roughly 0.14 μmol per injection.34 This combination of lower molecular weight and higher molar dose may enhance pharmacodynamic activity at the choroidal level, potentially accounting for the more pronounced CCT reduction observed in the brolucizumab subgroup, followed by aflibercept 8 mg, faricimab, and aflibercept 2 mg. Therefore, these pharmacologic differences provide a plausible mechanistic framework linking molecular size, molar dose, and tissue penetration to the extent of choroidal thinning observed after anti-VEGF therapy. Beyond pharmacokinetic differences, it should also be considered that faricimab is a bispecific antibody targeting both VEGF-A and angiopoietin-2, and thus its clinical efficacy is not solely dependent on VEGF inhibition or molar dose. Angiopoietin-2 blockade is known to modulate vascular stability, inflammation, and endothelial permeability through VEGF-independent pathways, potentially contributing to disease control without necessarily inducing marked choroidal thinning.7,17,35 Therefore, differences in CCT reduction should not be interpreted as reflecting a linear hierarchy among anti-VEGF agents but rather as the result of distinct different mechanisms and tissue-specific effects. In our study, the analysis of CVI added further insights. Although all anti-VEGF agents produced a significant reduction in CCT, changes in CVI varied across molecules. A decrease in CVI reflects a disproportionate reduction in the luminal (vascular) component relative to the total choroidal area, which was observed in eyes receiving aflibercept 8 mg. In this subgroup, both CVI and luminal area significantly differed before and after treatment. This may reflect a preferential effect of aflibercept on the vascular lumen, possibly driven by its high VEGF-binding capacity and stronger antipermeability action.7,36 Because VEGFA is a strong regulator of vascular permeability and vascular leakage,8,9 its inhibition may cause a more pronounced vascular lumen reduction. However, this effect may be only transitory considering that long-term results on aflibercept 2 mg have demonstrated to affect both luminal and stromal components.37 Although producing the greatest overall choroidal thinning, brolucizumab did not significantly alter CVI. However, further analysis of the CVI parameters revealed a significant effect of brolucizumab on the stromal component. This pattern may explain the lack of a significant difference in CVI, as the formula reflects the luminal changes that were irrelevant in this subgroup. This further corroborates previous findings hypothesizing a possible influence of brolucizumab on the stromal component.25 As postulated by the authors, brolucizumab may better penetrate into the choroidal stroma thanks to its smaller size, leading to a most effective fluid resorption at this anatomical level. The reduction of fluid and the hydrostatic pressure within the stroma have been associated with the greater ability of brolucizumab to resolve pigment epithelial detachment. The choroidal stroma is composed by extravascular tissue containing collagen, elastic fibers, fibroblast, non-neovascular smooth cells, and melanocytes.38 In the choroidal stroma, a significant VEGF immunoreactivity has been observed.39 Mesenchymal-like cell phenotype has been also described in the choroidal stroma, where they may secrete antiangiogenic growth factor pigment epithelium-derived factor/Serpin F1 inhibiting the proliferation of choroidal vascular cells.40 The choroidal effects of aflibercept observed in our study may have relevant clinical implications within the pachychoroid disease spectrum. Previous work by Koizumi et al13 demonstrated a significant reduction in CT after aflibercept treatment in eyes with nAMD, reporting a diffuse effect on CT regardless of the subtype of MNV.13 Our finding of a reduction in the luminal component, as assessed by CVI, further emphasized a direct effect of the drug on the vascular compartment of the choroid. This may therefore support the potential role of aflibercept, particularly the high-dose formulation, in conditions characterized by choroidal congestion and dilation, such as central serous chorioretinopathy and peripapillary pachychoroid syndrome, for which emerging evidence has been reported.41 In parallel, Govetto et al42 reported that subfoveal CT is reduced in nAMD compared with non-nAMD, a finding that may be partly attributable to the cumulative effects of anti-VEGF therapy. Taken together, these observations highlight a potential advantage of aflibercept in modulating choroidal remodeling in eyes with pachychoroid features. This study has several limitations that should be acknowledged. First, its retrospective and monocentric design may introduce selection bias and limit the generalizability of the findings. The sample size, particularly within the brolucizumab and aflibercept 8 mg subgroups, was relatively small. Additionally, follow-up was restricted to the loading phase (3 months), which does not allow assessment of long-term choroidal remodeling or sustained CVI changes. Moreover, other potential confounders, such as axial length, systemic factors, or diurnal variation, were not controlled for and may have influenced choroidal metrics. Despite these limitations, the present study further corroborates the role of anti-VEGF agents in causing choroidal thinning in nAMD regardless of agent subtype. In this retrospective analysis, all evaluated anti-VEGF agents produced significant reductions in CCT after the loading phase, with brolucizumab demonstrating the greatest overall effect. However, changes in CVI varied among agents, giving a further key for interpretation of the anti-VEGF impact on the choroidal vasculature. In particular, aflibercept 8 mg, with its higher molar dose and higher binding affinity for VEGF, produces a measurable constriction on the luminal component due to a preferential action on choroidal vascular endothelial cells. Brolucizumab, with its very small molecular size, may achieve deeper penetration into the choroidal stroma. At this level, its effect can promote fluid resolution, which is likely reflected by a contraction of the stromal component leaving the CVI unchanged. The combination of choroidal thinning and divergent CVI responses suggests that individual anti-VEGF molecules may influence choroidal vascular and stromal compartments differently, likely reflecting differences in molecular size, molar dose, and tissue penetration. Finally, the distinct effects of the different anti-VEGF agents on the luminal and stromal choroidal components may also influence the pattern and efficiency of fluid resorption after treatment. This compartment-specific response could therefore be clinically useful when tailoring therapy according to the predominant distribution of fluid. Further prospective studies are needed to clarify the long-term clinical implications of these distinct choroidal remodeling patterns.