Retinal and choroidal vascular density alterations in patients with schizophrenia: assessment via optical coherence tomography angiography

Abstract

Background

Previous studies have revealed the retinal and choroidal structural impairment in schizophrenia (SCZ). This study aimed to investigate alterations in retinal and choroidal vascular density in patients with SCZ using optical coherence tomography angiography (OCTA).​.

Methods

A total of 63 SCZ patients (63 eyes) and 61 healthy controls (61 eyes) were enrolled. Psychopathological symptoms in SCZ patients were assessed using the Positive and Negative Syndrome Scale (PANSS). OCTA images were acquired using an Optovue RTVue-XR system with a 3 × 3 mm scan size (304 × 304 pixels) to quantify the following parameters: retinal superficial vascular density (SVD), retinal deep vascular density (DVD), foveal avascular zone (FAZ), superficial FAZ (sFAZ), deep FAZ (dFAZ), choroidal vessel volume index (CVI), and choroidal vessel volume (CVV). Comparisons of retinal and choroidal vascular density were performed between the SCZ group and the control group. The diagnostic efficacy of these vascular density parameters in distinguishing SCZ patients from healthy controls was evaluated. Additionally, correlations between PANSS scores and retinal/choroidal vascular density parameters were analyzed in the SCZ group.​.

Results

Compared with the control group, the SCZ group exhibited significantly lower DVD (Z = -4.070, P < 0.001) and CVI (t = 2.204, P = 0.029), along with a significantly larger dFAZ (t = -3.953, P < 0.001). After adjusting for age, sex, and body mass index (BMI), DVD remained significantly associated with SCZ (β = -3.050, P = 0.011). Furthermore, the exploratory ROC analysis yielded an area under the curve (AUC) of 0.712 for DVD, indicating a modest level of discrimination between groups. Additionally, no significant correlations were observed between PANSS scores with DVD, dFAZ, or CVI in the SCZ group (all P > 0.05).

Conclusion

These results suggest that alterations in retinal and choroidal vascular parameters, particularly DVD, may be associated with SCZ and have potential value as a potential exploratory marker for distinguishing SCZ patients from healthy controls.

Clinical trial number

Not applicable.

Introduction

Schizophrenia (SCZ) is a complex psychiatric disorder characterized by an unclear etiology and heterogeneous pathological manifestations. Clinically, SCZ presents as a syndrome involving multiple symptom domains, including perceptual disturbances, cognitive impairments, emotional dysregulation, and behavioral abnormalities, along with mental incoordination [1]. According to the latest epidemiological data from China, the weighted lifetime prevalence of SCZ is 0.7% [2]. This disorder is associated with high relapse rates and significant disability, resulting in substantial psychological distress for patients and imposing considerable socioeconomic burdens on both families and society.

The precise pathogenesis of SCZ remains incompletely elucidated. Accumulating research indicates that vascular dysfunction may contribute to the pathophysiology of psychotic disorders, including SCZ [3]. Specifically, studies have documented microvascular alterations, impaired hemodynamic regulation, and genetic factors influencing cerebral circulation in individuals with SCZ [4]. Concurrently, neuroimaging investigations have consistently revealed cerebral blood flow abnormalities, a critical marker of cerebrovascular function, in both early and chronic stages of the disorder [5]. Furthermore, emerging evidence suggests that several risk factors for SCZ are linked to vascular dysregulation, such as aberrant angiogenesis, pathological vascular remodeling, and perturbations in hypoxia-related signaling pathways [6]. These microvascular anomalies may, in turn, induce structural and functional perturbations in neural networks, potentially serving as a mechanistic link between vascular pathology and the neurobiological substrates of SCZ.

The brain and retina share substantial embryological, structural, and functional homology, as both originate from neuroectodermal tissue and exhibit comparable neuronal cell types and laminar organization [7]. These central nervous system (CNS) structures maintain direct anatomical connectivity via the optic nerve. Notably, their vascular and neural development occur synchronously through neurovascular coupling mechanisms. Accumulating evidence demonstrates parallel anatomical and physiological alterations in both ocular and cerebral tissues across various neuropsychiatric disorders [8, 9]. This remarkable congruence, combined with the retina’s unique anatomical accessibility, positions retinal and choroidal (the vascular layer nourishing the outer retina) examination as a particularly valuable non-invasive approach for both biomarker discovery and pathophysiological investigation in neuropsychiatric conditions [10].

Over the past decade, investigations into structural and functional changes in the retina have been conducted in psychotic spectrum disorders, particularly SCZ [11]. It has been reported that patients with SCZ exhibit significant thinning of the peripapillary retinal nerve fiber layer (RNFL), macular thinning, and reduced macular volume compared with healthy controls. These morphological features have been shown to correlate with the duration of SCZ [12]. Relative to healthy controls, individuals with SCZ and other psychotic disorders display retinal thinning, with a prominent reduction in subfoveal choroidal thickness [13]. Meta-analyses have further validated decreased macular thickness/volume, peripapillary RNFL thickness, and thickness of the macular ganglion cell-inner plexiform layer (GCIPL) in patients with SCZ compared to healthy subjects [14] Notably, retinal microvascular abnormalities have also been associated with the severity of psychotic symptoms and cognitive impairment [15]. However, research on choroidal structure in SCZ patients remains remarkably limited, despite its greater clinical significance compared to retinal examination. Notably, the choroid is more susceptible to systemic pathophysiological changes than the retina. For instance, the choroidal vasculature exhibits the highest blood flow density in the human body and represents the most frequent ocular site for metastatic lesions [16]. Additionally, neural regulation of choroidal blood flow can be perturbed by systemic disorders such as hypertension and diabetes mellitus [17]. Equally importantly, the choroid serves as the primary supplier of oxygen and nutrients to the retinal outer layers [18]; consequently, the structural and functional retinal impairments documented in previous SCZ studies may be mechanistically linked to reduced choroidal vascular density. Therefore, the present study focuses specifically on investigating alterations in choroidal vascular density in SCZ patients to elucidate its potential role in the pathophysiological mechanisms of the disorder.

Optical coherence tomography angiography (OCTA) has emerged as a state-of-the-art, non-invasive imaging modality for visualizing retinal microvasculature with high resolution [19, 20]. Utilizing a frequency division-based decorrelation algorithm, OCTA enables three-dimensional (3D) reconstruction of retinal microvascular networks [21]. Compared with conventional angiographic techniques, this advanced method offers several distinct advantages: (1) elimination of contrast agent administration requirements; (2) enhanced imaging speed for accurate blood flow detection; and (3) superior capability for depth-resolved 3D visualization of microvascular structures at different retinal layers [22, 23]. In light of these advantages, the present study employs OCTA to systematically assess alterations in retinal and choroidal vascular density in patients with SCZ.

In the current study, we aim to evaluate the association between retinal and choroidal vascular density with SCZ, hypothesizing that microvascular abnormalities in the eye may reflect systemic or neurovascular dysfunction associated with SCZ pathogenesis. By comparing OCTA-derived metrics between SCZ patients and healthy controls, this study seeks to identify potential ocular biomarkers linked to SCZ, which could offer insights into its underlying vascular pathophysiology and serve as a non-invasive tool for monitoring disease progression.

Materials and methods

Participants

This cross-sectional study was conducted at Chaohu Hospital of Anhui Medical University between March 2024 and May 2025. Participants were eligible for inclusion if they met the following criteria: (1) aged 18–65 years; (2) diagnosed with SCZ by at least two attending physicians in accordance with the Diagnostic and Statistical Manual of Mental Disorders, Fifth Edition (DSM-5) criteria; (3) drug-naïve status or discontinuation of all antipsychotic/psychotropic medications for ≥ 3 months (verified via medical records and caregiver report). Individuals were excluded if they fulfilled any of the following: (1) current or past diagnosis of depressive disorder, bipolar disorder, or other psychiatric conditions as per DSM-5; (2) systemic diseases (e.g., diabetes, cardiometabolic disease, endocrine/nervous system disorders) with potential impacts on the eye or optic nerve; (3) presence of retinopathy or choroidal disease; (4) history of ocular tumor, eye surgery, or ocular trauma; (5) conditions that could significantly interfere with fundus imaging quality; (6) pregnancy or lactation; (7) contraindications, allergies, or intolerance to local anesthetics or mydriatic agents. The control group consisted of subjects who underwent physical examinations at the Department of Ophthalmology, Chaohu Hospital of Anhui Medical University, during the same period. All control participants were confirmed to have no ocular abnormalities through clinical examinations and optical coherence tomography angiography (OCTA) imaging by an ophthalmologist from the medical center. This study was approved by the Ethics Committee of Chaohu Hospital of Anhui Medical University (registration number: KYXM-202403-001) and conducted in accordance with the principles of the Declaration of Helsinki. Written informed consent was obtained from all participants prior to enrollment.

OCTA imaging

Optical coherence tomography angiography (OCTA) was performed using the RTVue-XR Avanti system (Optovue, Inc., Fremont, CA, USA) integrated with AngioVue OCTA software (Version 2018.1.0.43). A nominal scan field of 3 × 3 mm (sampling density: 304 × 304 pixels) was employed, with the scan centered on the foveal region. The OCT angiograms centered on the fovea were automatically segmented to define the superficial plexus from 3 μm below the internal limiting membrane to 15 μm below the inner plexiform layer and the deep plexus from 15 to 70 mum below the inner plexiform layer. The OCTA system was equipped with a light source featuring a central wavelength of 840 nm and a bandwidth of 50 nm. OCTA images with significant artifacts and decreased image quality were excluded, including those with (1) a quality signal strength index < 40, (2) definite residual motion artifacts, (3) indiscrete segmentation of each retinal layers or slabs, (4) poor centering, or (5) signal loss due to blinking. To determine the measurement repeatability of the OCTA parameters, we calculated the intraclass correlation coefficient (ICC). An ICC (ratio of the subject variance to the total variance) close to 1 indicated low variance between the two examinations (poor: ICC ≤ 0.40; fair: 0.40 < ICC ≤ 0.59; good: 0.60 < ICC ≤ 0.74; excellent: 0.75 < ICC ≤ 1.00). All parameters included in the final analysis exhibited at least fair-to-good reliability (ICC > 0.40).

Macular retinal microvascular density was analyzed using OCTA images of both the deep and superficial retinal layers. Specifically, the superficial vascular density (SVD) was defined as the vascular density within the region extending from 3 μm below the internal limiting membrane (ILM) to 15 μm below the inner plexiform layer (IPL). Conversely, the deep vascular density (DVD) was assessed in the region spanning 15–70 μm below the IPL. The area of the foveal avascular zone (FAZ) was quantified using ImageJ software, with measurements performed on en face OCTA images to evaluate both the superficial FAZ (sFAZ) and deep FAZ (dFAZ).

Choroidal blood flow parameters including choroidal vessel volume index (CVI) and choroidal vessel volume (CVV) were analyzed using built-in software. CVI was defined as the ratio of choroidal vessel volume to total choroidal volume within a specified three-dimensional region, representing a three-dimensional vessel density and expressed as a percentage (higher values indicate greater vascular density). Building upon prior research [24], the present study employed the ResUnet model for segmentation and the Phansalkar thresholding method for binarization to quantify choroidal thickness and the CVI across the entire choroidal volume. CVV is defined as the volume of choroidal vessels per unit area (volume/area), expressed in microns. The unit “microns (µm)” is methodologically justified: volume is quantified in cubic microns (µm³) and area in square microns (µm²), yielding a normalized metric with dimensions of length (µm³/µm² = µm). This definition and unit align with recent OCTA-based choroidal vascular research [25, 26] and directly reflect the software’s automated calculation output, ensuring consistency with established analytical pipelines.

Sample calculation

The sample size for this study was determined using G*Power software version 3.1.9.6. For an independent t-test (two-tailed) with a medium effect size of 0.5, a significance level of 0.05, and a power of 0.80, the minimum required sample size was calculated to be 64 for each group, resulting in a total of 128 participants. However, during the screening process, 3 subjects in the control group and 1 subject in the patient group were excluded based on the inclusion and exclusion criteria, leaving a final sample size of 124 participants (61 controls and 63 patients). For all participants (63 patients with SCZ and 61 healthy controls), the right eye was prioritized for OCTA imaging.

Statistical analysis

Data analysis was performed using SPSS version 17.0 (IBM Corporation, Armonk, NY, USA). The Kolmogorov-Smirnov test was applied to assess the normality of data distributions. Descriptive statistics were computed, with normally distributed quantitative variables expressed as mean ± standard deviation (SD), and non-normally distributed quantitative variables presented as median and interquartile range (IQR; 25th -75th percentiles). For between-group comparisons, the independent samples t-test was used to analyze normally distributed variables, while the Mann-Whitney U test was employed for non-normally distributed variables. Multiple linear regression models were constructed to compare retinal and choroidal vascular density between the two groups after adjusting for potential confounding factors. As an exploratory analysis, receiver operating characteristic (ROC) curves and their corresponding area under the curve (AUC) values were generated to provide a preliminary assessment of the potential for selected parameters to differentiate SCZ patients from controls. Pearson’s correlation analysis was performed to explore potential associations between variables. A two-tailed P-value < 0.05 was considered statistically significant.

Results

Characteristics of study participants

As presented in Table 1, compared with the control group, the SCZ group exhibited significantly higher age (Z = -4.952, P < 0.001) and body mass index (BMI) (t = -2.444, P = 0.016). Additionally, there was a significant difference in sex distribution between the two groups (χ² = 9.299, P = 0.002). No statistically significant differences were observed between the control and SCZ groups with respect to smoking (χ² = 0.806, P = 0.369), systolic blood pressure (SBP) (t = -0.410, P = 0.682), diastolic blood pressure (DBP) (t = -1.647, P = 0.100), best-corrected visual acuity (BCVA) (t = -0.669, P = 0.504), and intraocular pressure (IOP) (Z = 0.916, P = 0.360).

Table 1.

Comparison of demographic characteristics and clinical profiles between control group and SCZ group

Age (years)	Control group	SCZ group	t/Z/χ2	P
	30 (25, 44.5)	49 (37, 54)	−4.952	< 0.001
Female, n (%)	38 (62.3%)	22 (34.9%)	9.299	0.002
Smoking, n (%)	8 (13.1%)	12 (19.4%)	0.806	0.369
BMI (kg/m2)	22.93 ± 2.56)	24.42 ± 4.06	−2.444	0.016
SBP (mmHg)	124.48 ± 4.50	124.98 ± 8.71	−0.410	0.682
DBP (mmHg)	77 (75, 80)	80 (76, 82)	−1.647	0.100
BCVA (log MAR)	1 (1, 1)	1 (0.8, 1)	−0.669	0.504
IOP (mmHg)	15 (13, 16)	15 (14, 17)	0.916	0.360

BMI: body mass index; SBP: systolic blood pressure; DBP: diastolic blood pressure; BCVA: best-corrected visual acuity; log MAR: logarithm of the minimum angle of resolution; IOP: intraocular pressure

Comparison of retinal and choroidal vascular parameters between the control group and the SCZ group

As shown in Table 2, compared with the control group, the SCZ group demonstrated significantly lower DVD (Z = -4.070, P < 0.001) and CVI (t = 2.204, P = 0.029). In contrast, the SCZ group exhibited a significantly larger dFAZ (t = -3.953, P < 0.001) compared to the control group. No statistically significant differences were observed between the two groups in terms of SVD (t = 0.612, P = 0.541), sFAZ (Z = -0.983, P = 0.326), or CVV (t = -1.678, P = 0.096).

Table 2.

Comparison of retinal and choroidal vascular density between control group and SCZ group

SVD (%)	Control group	SCZ group	t/Z	P
	33.84 ± 3.59	33.46 ± 3.24	0.612	0.541
DVD (%)	36 (32, 40)	29 (27, 36)	−4.070	< 0.001
sFAZ (mm2)	0.18 (0.12, 0.23)	0.19 (0.14, 0.27)	−0.983	0.326
dFAZ (mm2)	0.48 ± 0.26	0.69 ± 0.33	−3.953	< 0.001
CVV (µm)	107.59 ± 40.77	120.24 ± 42.81	−1.678	0.096
CVI (%)	45.05 ± 5.27	42.92 ± 5.48	2.204	0.029

SVD: retinal superficial vascular density; DVD: retinal deep vascular density; FAZ: foveal avascular zone; sFAZ: superficial FAZ; dFAZ: deep FAZ; CVV: choroidal vessel volume; CVI: choroidal vessel volume index

After adjusting for age, sex, and BMI, multiple linear regression analysis revealed that DVD was significantly associated with SCZ (β = -3.050, P = 0.011; Table 3). However, no significant associations were found between dFAZ or CVI and SCZ following adjustment for the same confounding factors (both P > 0.05).

Table 3.

Comparison of DVD, dFAZ, and CVI between control group and SCZ group after adjusting age, sex, and BMI

DVD (%)	β	95%CI	P
	−3.050	−5.390 to−0.709	0.011
dFAZ (mm2)	−0.009	−0.052 to 0.034	−0.680
CVI (%)	−1.750	−3.993 to 0.492	0.125

DVD: retinal deep vascular density; FAZ: foveal avascular zone; dFAZ: deep FAZ; CVI: choroidal vessel volume index

Diagnostic value of DVD for distinguishing SCZ patients from healthy controls

The ability of DVD to differentiate patients with SCZ from healthy controls was explored using ROC curve analysis (Fig. 1). The AUC was 0.712 (95% CI: 0.625–0.809) with an 86.6% positive predictive value (PPV) and a 74.5% negative predictive value (NPV), providing preliminary evidence of modest discriminatory capacity within this cohort. To further verify the robustness of the model’s diagnostic performance, we performed 10 iterations of random dataset splitting (70% training/30% validation) and computed the AUC for each run. Specifically, the training set yielded AUC values ranging from 0.701 to 0.734 (mean ± SD: 0.717 ± 0.016), whereas the validation set exhibited an AUC range of 0.698 to 0.723 (mean ± SD: 0.710 ± 0.012). These consistent results across multiple splitting iterations collectively confirm the stable and reliable diagnostic performance of the model. While this value exceeds a commonly cited threshold of 0.7 for minimal acceptable discrimination, its clinical utility cannot be established from this single-site, cross-sectional study and requires external validation, calibration, and pre-specified threshold determination.

Fig. 1.

ROC curve of DVD in identification of patients with SCZ from healthy volunteers

Associations between retinal/choroidal vascular parameters and disease activity in patients with SCZ

The severity of symptoms in patients was evaluated using the Positive and Negative Syndrome Scale (PANSS), a validated rating instrument comprising 7 items assessing positive symptoms (PANSS-P), 7 items measuring negative symptoms (PANSS-N), and 16 items evaluating general psychopathology (PANSS-G) in SCZ [27]. The correlations between PANSS scores and DVD, dFAZ, and CVI in SCZ patients were summarized in Fig. 2. No significant correlations were observed between PANSS scores (including total, PANSS-P, PANSS-N, and PANSS-G) and DVD, dFAZ, or CVI in the SCZ group (all P > 0.05).

Fig. 2.

Correlation between the PANSS scores and DVD, dFAZ, and CVI in patients with SCZ. ×: no significance

Discussion

This study investigated retinal and choroidal vascular density alterations in patients with SCZ using OCTA and explored their associations with disease activity. Four key findings were observed. Firstly, the SCZ group demonstrated significantly lower DVD and CVI, as well as a significantly larger dFAZ compared with the control group. Secondly, after adjusting for age, sex, and BMI, DVD remained significantly associated with SCZ. Thirdly, ROC analysis revealed that DVD exhibited moderate discriminative capacity in differentiating SCZ patients from controls, with an AUC of 0.712. Fourthly, no significant correlations were found between PANSS scores and DVD, dFAZ, or CVI in the SCZ group.

OCTA is widely used to evaluate vascular structures in various ocular conditions, including age-related macular degeneration, diabetic retinopathy, uveitis, and glaucoma [23, 28]. By detecting impaired perfusion sites in both superficial and deep capillary plexuses, OCTA provides valuable depth-resolved information regarding the severity of perfusion deficits [29]. Consequently, it has been extensively applied in the diagnosis of multiple disorders, such as multiple sclerosis, Alzheimer’s disease, and diabetic retinopathy [30–32]. Given its noninvasive nature and clinical feasibility, OCTA was selected as the preferred retinal imaging modality over fluorescein fundus angiography (FFA) in the present study to assess alterations in retinal and choroidal vascular density, in patients with SCZ.

The retina has been proposed as an early indicator of microangiopathies, where neurodegeneration and neuroinflammation can be noninvasively detected in both neurological and systemic disorders [33]. Moreover, sharing the same embryonic origin as the brain, the retina is regarded as a “window” that reflects cerebral pathological changes [34]. Based on these characteristics, recent research has focused on retinal imaging studies to identify potential pathophysiological markers of SCZ. For instance, compared with healthy controls and patients with acute SCZ, those with chronic or long-term chronic SCZ exhibit significantly reduced overall retinal nerve fiber layer (RNFL) thickness [35]. A meta-analysis involving 2079 eyes from patients with SCZ and 1571 eyes from controls further demonstrated that patients had significantly thinner average peripapillary RNFL (pRNFL), average macular thickness (MT), and macular ganglion cell layer-inner plexiform layer (GCL-IPL) than controls [36]. While these studies have primarily focused on alterations in retinal thickness as potential biomarkers for SCZ, research on retinal vascular density remains relatively limited. A critical distinction between vascular density changes (as measured by OCTA) and structural layer thinning (as detected by conventional OCT) merits focused discussion, as these modalities capture complementary aspects of retinal pathology. Structural OCT studies in SCZ have predominantly highlighted reductions in inner retinal layer thickness, attributed to neuronal loss or impaired synaptic integrity [37]. In contrast, our OCTA findings provide direct evidence of vascular compromise in the deep plexus, which precedes or accompanies structural thinning in neurodevelopmental or neurodegenerative disorders. This aligns with evidence that inner retinal structures are affected in schizophrenia spectrum disorders, while structural OCT studies often show layer-specific thinning. Vascular density loss may contribute to structural deterioration by reducing oxygen and nutrient delivery to metabolically demanding inner retinal cells, creating a vicious cycle of ischemia-induced neuronal dysfunction and further vascular regression [38]. Conversely, structural changes (e.g., synaptic loss) could disrupt the neuro-glio-vascular coupling that regulates vascular perfusion, leading to reduced vascular density [39]. This bidirectional relationship suggests that combining OCTA-derived vascular metrics (e.g., DVD) with conventional OCT structural measures may enhance the sensitivity of retinal imaging for detecting SCZ-related pathology, as vascular changes may emerge earlier than overt structural thinning. The retinal microvasculature, which shares similar structural and functional characteristics with cerebral microvessels, could provide additional insights into the microangiopathic and neuroinflammatory processes underlying SCZ. In the present study, the SCZ group exhibited significantly reduced DVD compared to controls. This association remained statistically significant after adjusting for potential confounders including age, sex, and BMI. Our findings align with previous reports suggesting greater vulnerability of the deep retinal layers to progressive vascular impairment relative to superficial layers, potentially attributable to hemodynamic dysfunction or disrupted neuro-glio-vascular interactions [39]. Subsequently, the diagnostic performance of DVD in differentiating patients with SCZ from healthy controls was evaluated. ROC analysis revealed that DVD exhibited moderate discriminative ability, yielding an AUC of 0.712. This finding suggests that DVD may serve as a potential imaging biomarker to assist in the identification of SCZ, though further validation in larger and more diverse cohorts is warranted to confirm its clinical utility.

FAZ, a capillary-free region surrounded by a ring of capillaries in the central fovea, is critical for central visual function. Pathological changes affecting the FAZ can lead to varying degrees of visual impairment [40]. It has been established that an enlarged FAZ area is associated with more severe capillary occlusion and nonperfusion in the macular region [41]. In diabetic retinopathy, FAZ enlargement is recognized as one of the earliest signs, showing a significant correlation with visual acuity and disease severity [42]. Additionally, a recent study reported that FAZ, particularly the dFAZ, is positively correlated with disease activity in systemic lupus erythematosus (SLE) [43]. Regarding schizophrenia spectrum disorders (SSDs), a meta-analysis found no significant differences in FAZ parameters between SSD participants and controls [44]. Consistently, in the present study, although the SCZ group exhibited a significantly larger dFAZ, no significant association between dFAZ and SCZ was observed after adjusting for the confounding factors. Taken together, these results do not support the utility of FAZ parameters, including dFAZ, as reliable neurobiological indicators for SCZ. Future studies incorporating more detailed ocular imaging protocols and comprehensive clinical phenotyping may help elucidate the potential relevance of such retinal microvascular changes in this disorder.

The choroid is structured to contain a highly complex network of small blood vessels with a large surface area, which is presumably designed to facilitate the exchange of oxygen and other essential nutrients. This network lies in close proximity to the outer retinal layers, and the proper structure and function of the choroidal vasculature are critical for optimal retinal function [45], making it a promising target for investigating microvascular pathology. It has been reported that the choriocapillaris flow area (CCFA) is significantly reduced in patients with diabetic macular ischemia (DMI) and is strongly correlated with retinal capillaries [46]. Another study indicated that choriocapillary vascular density (CCVD) negatively correlates with disease activity in SLE. CVI is a relatively new quantitative measure of choroidal vascularity, providing objective data on the proportion of choroidal luminal and stromal areas. Defined as the ratio of luminal choroidal area (LCA) to total choroidal area (TCA), CVI is a more stable parameter than choroidal thickness for assessing choroidal structural alterations in various diseases [47]. It has been reported that choroidal vessel density, including CVI, is lower in patients with major depressive disorder (MDD) compared with healthy controls [26]; however, no significant correlation was observed between CVI and MDD after adjusting for age and sex. Consistently, in the present study, although the SCZ group exhibited a significantly lower CVI, no significant association between CVI and SCZ was found after adjusting for confounding factors. Notably, our results differ from those in early-stage psychosis populations (ultra-high risk for psychosis [UHR-P] and first-episode psychosis [FEP]), where CVI was increased in FEP compared to UHR-P and healthy controls, and higher in UHR-P than controls after outlier exclusion [10]. This discrepancy is likely driven by two key factors: (1) disease stage: early psychosis may involve transient choroidal hyperperfusion linked to acute neuroinflammation or metabolic remodeling [10], whereas established SCZ (the focus of our study) may be characterized by progressive vascular dysfunction or hypoperfusion due to long-term neurobiological changes; and (2) medication exposure: our SCZ patients were medication-free for ≥ 3 months, eliminating antipsychotic-related vascular effects, whereas FEP patients in prior studies may have been acutely medicated. Additionally, factors such as smoking and cardiometabolic status, which were not fully addressed in previous comparisons, may also contribute to the observed differences and warrant further exploration as potential confounders. These findings suggest that while choroidal vascular alterations may be present in certain neuropsychiatric conditions, their association with specific disorders such as SCZ may be influenced by confounding variables and context-dependent factors including disease trajectory, methodological approach, and clinical characteristics (e.g., medication status). Therefore, further investigation with larger, well-controlled cohorts and rigorously controlling for cardiometabolic confounders is warranted to clarify their role in SCZ pathophysiology. Future longitudinal studies should also track choroidal parameter changes from UHR-P to chronic SCZ, to determine their utility as biomarkers of disease progression or treatment response, which will help establish a more comprehensive understanding of the dynamic relationship between choroidal vascular alterations and the course of psychotic disorders.

This study has several limitations that warrant consideration. Firstly, the single-center design, which relied exclusively on data from a single hospital, may introduce selection bias, thereby limiting the generalizability of the findings. Secondly, the relatively small sample size used in the current analysis underscores the need for future large-scale, multi-center studies with expanded sample sizes to validate and strengthen these results. Thirdly, as a cross-sectional study, this investigation is inherently limited in its ability to establish a causal relationship between retinal/choroidal vascular density alterations and SCZ. Causal inferences regarding the directionality of such associations would require validation through longitudinal or interventional studies. Lastly, notable disparities in sex, age, and BMI were observed between the study groups. These demographic imbalances may introduce confounding effects, potentially influencing the interpretation of the results and thus constituting an additional limitation of the current analysis.

Conclusion

In summary, OCTA provides a non-invasive modality for assessing retinal and choroidal vascular density alterations in individuals with SCZ. Moreover, DVD holds potential as a diagnostic biomarker for SCZ. To validate the utility of retinal vascular density as novel biomarkers for SCZ, large-scale multicenter studies are imperative for confirmation. Additionally, longitudinal studies are warranted to investigate the causal relationship between changes in retinal vascular density and disease activity, which may shed light on the underlying pathophysiological mechanisms of SCZ.

Author contributions

Jun Liu, Hui Zhang, Dachuan Zhang, and Yong Wang were involved in the conception and design of the study. Jun Liu, Hui Zhang, Guangyuan Wang, Suqi Song, Kai Zhang, Ruimei Ni, Le Ren, and Yayun Xu were responsible for data collection and statistical analysis. Jun Liu constructed a draft of the manuscript. Yong Wang and Zhengxuan Jiang have provided relevant feedback and critical revisions of the manuscript. The authors read and approved the final manuscript.

Funding

This work was financially supported by grants from the Research Fund of Anhui Institute of Translational Medicine (2023zhyx-C96) and Hefei Health Science and Technology Project (Hwk2025jqz007).

Data availability

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

Declarations

Ethics approval and consent to participate

This study was approved by the Ethics Committee of Chaohu Hospital of Anhui Medical University (registration number: KYXM-202403-001) and conducted in accordance with the principles of the Declaration of Helsinki. Written informed consent was obtained from all participants prior to enrollment.

Competing interests

The authors declare no competing interests.