Beyond pelvic pathology: retinal microvascular rarefaction as a systemic marker in endometriosis

A total of 100 eyes from 100 female participants were analyzed, comprising 50 eyes from patients with laparoscopically or ultrasonographically confirmed pelvic endometriosis and 50 eyes from healthy controls. According to Table  1 , there were no statistically significant differences between the two groups in terms of age, spherical equivalent, intraocular pressure, central corneal thickness, central macular thickness, or parafoveal retinal thickness ( p  > 0.05 for all). Mean CA-125 levels in the endometriosis group were 43.7 ± 35.9 U/mL, and the average endometrioma size was 4.3 ± 2.1 cm. Based on laparoscopic staging data available for 21 patients, endometriosis severity was categorized according to the revised American Society for Reproductive Medicine (rASRM) classification. Of these, 6 patients (28.6%) had stage I, 7 (33.3%) had stage II, 5 (23.8%) had stage III, and 3 (14.3%) had stage IV disease. However, no statistically significant differences in OCTA parameters were observed among severity subgroups, although subgroup comparisons were limited by sample size. Table 1 Demographic and ocular parameters Parameter Endometriosis Group (Mean ± SD) Control Group (Mean ± SD) p -value Age (years) 36.2 ± 7.9 36.4 ± 4.3 0.773 Spherical Equivalent (D) –0.52 ± 0.78 –0.67 ± 1.01 0.469 Intraocular Pressure (mm Hg) 15.20 ± 2.32 14.22 ± 2.67 0.052 Central Corneal Thickness (µm) 552.36 ± 26.70 553.10 ± 25.02 0.855 Central Macular Thickness (µm) 218.76 ± 22.65 219.36 ± 20.12 0.738 Foveal Retinal Thickness (µm) 259.82 ± 16.50 260.28 ± 17.15 0.738 Parafoveal Retinal Thickness– Inferior (µm) 335.74 ± 13.65 339.52 ± 11.30 0.306 Parafoveal Retinal Thickness– Superior (µm) 339.18 ± 13.61 343.36 ± 12.07 0.107 Parafoveal Retinal Thickness– Nasal (µm) 340.70 ± 12.70 340.76 ± 13.15 0.982 Parafoveal Retinal Thickness– Temporal (µm) 323.44 ± 11.97 326.42 ± 12.00 0.217 CA-125 (U/mL) 43.7 ± 35.9 – – Endometrioma Size (cm) 4.3 ± 2.1 – – Demographic and ocular parameters As shown in Table  2 , the total vessel area density (VAD) in the superficial vascular complex (SVC) was significantly lower in the endometriosis group (34.55 ± 3.48%) compared to controls (37.14 ± 2.34%, p  < 0.001). Similarly, the FD-300 value, representing vessel density within a 300-µm annulus around the foveal avascular zone, was significantly reduced in the endometriosis group (22.5 ± 3.3%) versus controls (24.1 ± 3.1%, p  = 0.011). Segmental parafoveal VAD values—measured in the inferior, superior, nasal, and temporal quadrants—were also significantly lower in the endometriosis group ( p  < 0.05 for each comparison). No significant differences were observed in FAZ area or circularity index between the groups; however, the FAZ perimeter was significantly increased in the endometriosis group ( p  = 0.025). Table 2 Superficial vascular complex (SVC) OCTA parameters Parameter Endometriosis Group (Mean ± SD) Control Group (Mean ± SD) p -value Total VAD (%) 34.55 ± 3.48 37.14 ± 2.34 < 0.001 Foveal VAD (%) 4.70 ± 2.46 6.40 ± 4.42 0.188 Parafoveal VAD (%) 35.06 ± 4.08 38.08 ± 2.58 < 0.001 Inferior Parafoveal VAD (%) 35.72 ± 6.80 39.76 ± 2.58 < 0.001 Superior Parafoveal VAD (%) 36.50 ± 4.59 39.88 ± 3.31 < 0.001 Nasal Parafoveal VAD (%) 33.28 ± 4.48 36.20 ± 3.03 < 0.001 Temporal Parafoveal VAD (%) 35.14 ± 4.45 38.06 ± 3.48 0.001 FD-300 (%) 22.5 ± 3.3 24.1 ± 3.1 0.011 FAZ Area (mm²) 0.59 ± 0.14 0.53 ± 0.19 0.097 FAZ Perimeter 3.27 ± 0.48 3.01 ± 0.67 0.025 FAZ Circularity Index 0.68 ± 0.08 0.72 ± 0.08 0.055 VAD , Vessel Area Density; FD-300 , Vessel Area Density measured within a 300-µm annulus surrounding the FAZ; FAZ , Foveal Avascular Zone Superficial vascular complex (SVC) OCTA parameters VAD , Vessel Area Density; FD-300 , Vessel Area Density measured within a 300-µm annulus surrounding the FAZ; FAZ , Foveal Avascular Zone According to Table  3 , vessel area density values within the deep vascular complex (DVC) were similarly reduced in the endometriosis group. Total DVC VAD was significantly lower in endometriosis (36.99 ± 3.53%) versus controls (39.07 ± 2.89%, p  = 0.001). The FD-300 value for the DVC layer was also significantly lower in the endometriosis group (28.6 ± 4.5%) than in controls (31.2 ± 3.8%, p  = 0.002). All segmental parafoveal VADs within the DVC (inferior, superior, nasal, temporal) were significantly lower in the endometriosis group. However, FAZ metrics—area, perimeter, and circularity index—did not differ significantly between groups in the deep layer. Table 3 Deep vascular complex (DVC) OCTA parameters Parameter Endometriosis Group (Mean ± SD) Control Group (Mean ± SD) p -value Total VAD (%) 36.99 ± 3.53 39.07 ± 2.89 0.001 Foveal VAD (%) 15.28 ± 3.54 17.30 ± 4.67 0.017 Parafoveal VAD (%) 38.08 ± 3.77 40.56 ± 2.84 < 0.001 Inferior Parafoveal VAD (%) 36.48 ± 7.58 39.68 ± 3.50 0.036 Superior Parafoveal VAD (%) 38.40 ± 4.36 40.64 ± 4.95 0.001 Nasal Parafoveal VAD (%) 37.96 ± 4.62 40.32 ± 2.95 0.003 Temporal Parafoveal VAD (%) 38.66 ± 3.82 41.08 ± 3.06 < 0.001 FD-300 (%) 28.6 ± 4.5 31.2 ± 3.8 0.002 FAZ Area (mm²) 0.33 ± 0.11 0.32 ± 0.12 0.544 FAZ Perimeter 2.25 ± 0.39 2.19 ± 0.46 0.513 FAZ Circularity Index 0.78 ± 0.06 0.80 ± 0.06 0.106 VAD , Vessel Area Density; FD-300 , Vessel Area Density measured within a 300-µm annulus surrounding the FAZ; FAZ , Foveal Avascular Zone Deep vascular complex (DVC) OCTA parameters VAD , Vessel Area Density; FD-300 , Vessel Area Density measured within a 300-µm annulus surrounding the FAZ; FAZ , Foveal Avascular Zone

This study was designed as a retrospective cross-sectional analysis conducted at the Departments of Obstetrics and Gynecology and Ophthalmology of Duzce University. Ethical approval was obtained from the Non-Interventional Clinical Research Ethics Committee of Duzce University on May 26, 2025 (Approval No: 2025/142). The study was conducted in accordance with the ethical principles outlined in the Declaration of Helsinki. All data were retrieved retrospectively from institutional electronic medical records and anonymized prior to analysis. Written informed consent had been obtained from all participants as part of routine clinical protocols. The study was reported in accordance with the CONSORT checklist for observational studies where applicable. The study population consisted of 100 eyes from female participants aged 18 to 45 years. Fifty eyes were selected from patients diagnosed with pelvic endometriosis between 2023 and 2024, and fifty eyes from age-matched healthy women served as the control group. The diagnosis of pelvic endometriosis was established either through laparoscopic visualization with histopathological confirmation or based on transvaginal ultrasonographic findings consistent with ovarian endometriomas, such as homogenous low-level internal echoes and ground-glass appearance. The diagnostic approach was determined according to routine clinical protocols and performed by experienced gynecologists. Staging of endometriosis based on the revised American Society for Reproductive Medicine (rASRM) classification was available for a subset of patients diagnosed via laparoscopy; however, due to incomplete data across all participants, disease staging was not included in the main statistical analysis. Inclusion criteria for both groups included female sex, age between 18 and 45 years, spherical equivalent refractive error within ± 3.00 diopters, best-corrected visual acuity (BCVA) of 20/20 or better in the study eye, and absence of any retinal or optic nerve pathology. Exclusion criteria were as follows: prior ocular surgery or trauma, presence of glaucoma, diabetic retinopathy, uveitis, or other ocular diseases affecting the retinal vasculature; systemic diseases known to impair microvascular circulation such as diabetes mellitus or uncontrolled hypertension; pregnancy or lactation; and inadequate OCTA image quality (signal strength index < 30 dB or segmentation artifacts). The rationale for using a 30 dB signal strength threshold was based on previous technical recommendations to ensure image quality and reliability [ 11 ]. Axial length measurements were not available in this retrospective dataset and represent a methodological limitation. In addition, hormonal therapy status (e.g., GnRH analogs, oral contraceptives, progestins) was not systematically recorded and could not be evaluated as a confounding variable. All macular OCTA images were obtained during the early follicular phase (days 1–5) of the menstrual cycle to minimize hormonal influence on vascular parameters. OCTA imaging was performed using the Heidelberg Spectralis system with a 10 × 10° high-resolution scan centered on the fovea, covering a 3 × 3 mm area. Automated segmentation was used to delineate the superficial vascular complex (SVC) and deep vascular complex (DVC), based on standard anatomical boundaries. Only high-quality images without motion or segmentation artifacts were included. The Early Treatment Diabetic Retinopathy Study (ETDRS) grid was not explicitly used; however, the parafoveal region was operationally defined as the annular area between 1 mm and 3 mm diameters centered on the fovea. Quantitative analysis was conducted using the OCTA open-source software. Extracted parameters included total and parafoveal vessel area density (VAD), foveal avascular zone (FAZ) area, perimeter, and circularity, as well as the FD-300 metric, which quantifies the vessel density in a 300-µm wide ring surrounding the FAZ and reflects perifoveal capillary perfusion. Segmental parafoveal VAD values were also calculated for the nasal, temporal, superior, and inferior quadrants. All image analyses were performed by a single masked examiner trained in OCTA interpretation, which minimized subjective bias and ensured consistency [ 11 ]. The processing pipeline and graphical interface of the OCTAVA software used for quantitative OCTA analysis are illustrated in Fig.  1 . Fig. 1 Screenshot from the OCTAVA software used for quantitative OCTA image analysis Screenshot from the OCTAVA software used for quantitative OCTA image analysis Statistical analyses were performed using SPSS Statistics for Mac, version 30 (IBM Corp., Armonk, NY, USA). The normality of distribution for continuous variables was assessed using the Kolmogorov–Smirnov test. Data were expressed as mean ± standard deviation (SD). Between-group comparisons were performed using the independent samples t-test or the Mann–Whitney U test, depending on data distribution. Categorical variables were compared using the Chi-square or Fisher’s exact test, as appropriate. A two-tailed p -value of < 0.05 was considered statistically significant. To account for multiple comparisons across OCTA-derived parameters, the Benjamini-Hochberg false discovery rate (FDR) correction was applied to control for Type I error. Although axial length and hormonal therapy status were not available in this retrospective dataset, age and refractive error were assessed and found to be comparable between groups. No multivariate adjustment could be performed due to the lack of additional covariate data. Observed standard deviation differences between groups were acknowledged as a source of potential variability and were considered during interpretation. Prior to data collection, a priori sample size estimation was performed using G*Power (version 3.1) software. Assuming a two-tailed α of 0.05, power (1–β) of 0.95, and a medium effect size (Cohen’s d = 0.5), the minimum required sample size was calculated as 36 participants. With a total of 100 eyes included in the final analysis, the study was adequately powered to detect statistically significant differences between groups [ 12 ].

In the present study, a significant reduction in retinal microvascular density was observed in women with pelvic endometriosis compared to healthy controls, particularly in the FD-300 metric, which quantifies vessel density in the 300-µm-wide annulus surrounding the foveal avascular zone (FAZ). FD-300 is a highly sensitive marker of perifoveal perfusion and is less affected by segmentation errors compared to whole parafoveal measures [ 13 ]. Its consistent reduction in both superficial and deep retinal plexuses suggests early microvascular compromise in endometriosis patients, despite preserved visual function and normal ophthalmic examination. According to our data, these findings support the emerging view that endometriosis extends beyond the reproductive tract and may reflect a systemic disease with vascular involvement. Previous epidemiological studies have shown an increased risk of cardiovascular events, including myocardial infarction, stroke, and accelerated atherosclerosis, in women with endometriosis [ 8 , 14 ]. Such associations are attributed to chronic inflammation, endothelial dysfunction, and oxidative stress. In line with this, recent studies have demonstrated impaired endothelial-dependent vasodilation in young women with endometriosis, even in the absence of traditional cardiovascular risk factors, as evidenced by reduced reactive hyperemia index (RHI) values [ 15 , 16 ]. The inflammatory microenvironment of endometriosis is characterized by elevated levels of cytokines such as IL-1β, IL-6, and TNF-α in both the serum and peritoneal fluid [ 17 ]. These pro-inflammatory mediators promote oxidative stress, reduce nitric oxide bioavailability, and impair endothelial integrity, leading to microvascular dysfunction [ 18 ]. Over time, chronic endothelial activation may culminate in microvascular rarefaction, which is increasingly recognized in inflammatory and autoimmune diseases [ 19 , 20 ]. Consistent with our findings, similar OCTA-based reductions in retinal vascular density have been reported in systemic lupus erythematosus and Behçet’s disease [ 21 ]. Importantly, none of the participants in this study exhibited clinical retinal pathology, suggesting that these vascular alterations are subclinical and potentially reversible. Moreover, endometriosis has been linked with other microvascular syndromes such as migraine and Raynaud’s phenomenon [ 22 ], supporting a broader systemic vascular phenotype. These conditions may share overlapping mechanisms involving vascular tone dysregulation, inflammation, and autonomic imbalance. Interestingly, despite endometriosis being a hyperangiogenic disease with upregulated vascular endothelial growth factor (VEGF) expression [ 4 ], our study revealed a paradoxical reduction in retinal capillary density. This discrepancy may be explained by the compartmentalized nature of angiogenesis in endometriosis, where local VEGF secretion sustains ectopic lesions without significant systemic spillover. Moreover, the retina, protected by the blood-retinal barrier and governed by strict autoregulation, may be less responsive to systemic VEGF elevations. This concept aligns with similar observations in diabetic retinopathy, where high intraocular VEGF levels co-exist with capillary dropout [ 23 , 24 ]. n some studies, anti-VEGF treatment did not significantly alter retinal vessel density, suggesting that VEGF alone does not fully account for vascular integrity [ 25 ]. A plausible hypothesis is that ectopic lesions in endometriosis act as an “angiogenic sink,” consuming VEGF locally and depriving systemic circulation, including the retina, of adequate angiogenic support [ 26 ]. Alternatively, chronic inflammation may override angiogenic signaling, leading to microvascular constriction or regression. Such findings stand in contrast to the more consistent reduction observed in FD-300 values, which reflect microvascular density in a narrow annulus surrounding the foveal avascular zone. This suggests that perifoveal capillary compromise may be more sensitive to systemic influences than other retinal regions. However, not all vascular parameters followed a uniform pattern of alteration. For instance, no significant differences were observed in the FAZ area and circularity index between groups, which may appear contradictory to the hypothesis of generalized microvascular rarefaction. This inconsistency might reflect the relatively preserved architecture of the foveal avascular zone in the absence of overt retinal pathology or the limitations of cross-sectional OCTA in detecting subtle morphological changes. Additionally, quadrant-specific variations in parafoveal VAD were noted (e.g., greater reductions in the inferior and nasal sectors), yet these differences were not statistically homogeneous across all regions. Such spatial heterogeneity may reflect localized autoregulatory differences or technical variability in image acquisition and segmentation. These regional discrepancies should be interpreted with caution, and future studies incorporating perfusion mapping and stratified analyses are warranted to clarify their relevance. Hormonal factors may also play a role in retinal vascular homeostasis. Estrogen generally exerts vasodilatory and anti-inflammatory effects, but in endometriosis, resistance to progesterone and the presence of chronic inflammation may shift estrogen’s effects toward maladaptive pathways [ 27 – 31 ]. Although all patients were imaged in the early follicular phase to control for hormonal variation, hormonal therapies and fluctuations were not systematically recorded, representing a limitation. This study has several limitations that should be considered when interpreting the findings. First, although we hypothesized potential mechanisms such as systemic inflammation, endothelial dysfunction, and dysregulated angiogenesis, the study did not include laboratory biomarkers (e.g., CRP, IL-6, VEGF) or vascular function assessments (e.g., flow-mediated dilation or reactive hyperemia index) that could directly support these hypotheses. As a result, the pathophysiological links between endometriosis and retinal microvascular rarefaction remain speculative. Second, axial length measurements were not performed. Since axial length can influence retinal image magnification and vascular density estimations in OCTA analysis, this may have introduced a minor measurement bias, although both groups were age- and refraction-matched. Third, although OCTA scans were timed during the early follicular phase to minimize hormonal variability, the use of hormonal therapies such as oral contraceptives or GnRH analogues was not systematically evaluated. These agents may have influenced microvascular parameters and should be controlled for in future studies. Fourth, the study population was limited to women aged 18 to 45 years without systemic comorbidities. While this improved internal validity and reduced confounding, it may limit the generalizability of the results to broader populations, including older individuals or those with established cardiovascular disease. Finally, the cross-sectional nature of the study precludes any causal inferences. Longitudinal research is warranted to evaluate whether retinal microvascular changes progress with disease severity or respond to therapeutic interventions. In conclusion, women with pelvic endometriosis demonstrated significantly reduced retinal capillary density compared to age-matched healthy controls, as measured by OCTA. Notably, FD-300 and parafoveal VAD values—key indicators of perifoveal microvascular integrity—were consistently lower in both the superficial and deep retinal layers of endometriosis patients. These findings suggest that endometriosis may be associated with subclinical systemic microvascular compromise, even in the absence of overt cardiovascular comorbidities or retinal pathology. While causality cannot be inferred from this design, the results support the growing concept of endometriosis as a multisystem disorder with vascular and inflammatory implications that extend beyond the pelvic cavity. OCTA-derived retinal microvascular parameters may serve as accessible, non-invasive surrogate markers for systemic endothelial health in this population. Future longitudinal and mechanistic studies incorporating systemic biomarkers, disease severity indices, and hormonal modulation status are warranted to validate and expand upon these observations.

Endometriosis is a chronic, estrogen-dependent inflammatory disorder characterized by the presence of endometrial-like tissue outside the uterus. Affecting approximately 10% of women of reproductive age, it is a leading cause of pelvic pain and infertility. The pathogenesis of endometriosis involves multiple complex mechanisms, including retrograde menstruation, immune dysfunction, and neovascularization [ 1 – 3 ]. The vascular component of endometriosis is particularly important, as the survival and progression of ectopic lesions depend on neovascularization. Increased expression of angiogenic mediators such as vascular endothelial growth factor (VEGF) is well documented in endometriosis and correlates with disease severity [ 4 , 5 ]. This aberrant angiogenic profile has prompted investigations into whether endometriosis may have systemic microvascular consequences [ 6 ]. Indeed, women with endometriosis have been shown to exhibit a higher prevalence of systemic vascular disorders, including cardiovascular disease, hypertension, stroke, migraine, and Raynaud’s phenomenon. These associations are thought to be driven by chronic inflammation-induced endothelial dysfunction and microvascular injury [ 7 , 8 ]. However, the specific vascular beds involved and the clinical utility of these associations remain unclear. The retinal microvasculature is considered a sensitive biomarker of systemic vascular and inflammatory conditions. Optical coherence tomography angiography (OCTA) is a novel, non-invasive imaging technique that allows for the quantitative assessment of retinal capillary networks. OCTA has proven effective in detecting subclinical retinal microvascular alterations in systemic diseases such as diabetes, systemic lupus erythematosus, and hypertension [ 9 , 10 ]. Despite the established microvascular burden in endometriosis, there is currently no data exploring whether retinal vascular alterations are present in this population. The retina offers a unique, non-invasive opportunity to explore systemic microvascular involvement, and may serve as a surrogate site for subclinical endothelial dysfunction. Therefore, this study aims to quantitatively assess retinal vessel density in women with pelvic endometriosis using OCTA, and to explore whether these findings may reflect subclinical microvascular alterations potentially associated with the systemic nature of the disease. Rather than proposing direct clinical application, this work is intended as an exploratory analysis that may inform future investigations into retinal biomarkers in systemic gynecologic disorders.