Multimodal evaluation of 525 nm laser-induced retinal injury in rabbits: structural, functional and cytokine analyses Authors: Ling Guo, Kuijun Chen, Guanhua Li, Jian Zeng, Zhuangqing Fan, Congyi Li, Yangcheng Zou, Jie Xu, Jieyuan Zhang Published: 2026 ABSTRACT This study aimed to systematically study the characterize of retinal damage induced by 525 nm laser irradiation at different power densities through multimodal assessment methods, including structural, functional and cytokine analyses. A total of 24 rabbits were exposed to laser irradiation at 0.5, 200, 500, and 1000 mW/cm2 with a single-eye exposure duration of 0.25 s. Multimodal evaluations including ERG, OCT/OCTA, histopathology and aqueous humor cytokine microarray analysis were used to evaluate the characteristics of laser-induced injury in rabbit eyes. No visible functional or pathological damage was observed after low-power laser radiation (0.5 mW/cm2). Laser irradiation exceeding 200 mW/cm2 induced transient electroretinographic suppression and persistent laser spots in the fundus. Laser-induced retinal edema and hemorrhage were naturally absorbed within 7 days after injury, and cellular damage and structural changes persisted in the form of old plaques (14d). Notably, 1000 mW/cm2 laser exposure caused severe retinal detachment (1d), ultimately leading to a retinal hole (14d). There was no significant difference in the inflammatory array clustering analysis of anterior aqueous humor, but several individual cytokines showed statistically significant differences: ANG-1 was upregulated (500 and 1000 mW/cm2, p < 0.01) while VEGF-A was downregulated (500 mW/cm2, p < 0.05); MIP-1β was significantly increased (200 mW/cm2, p < 0.05), whereas IL-1RA was decreased (500 mW/cm2, p < 0.05). The 525 nm laser induced dose-dependent retinal injury, ranging from reversible functional inhibition to irreversible structural damage, and triggered the expression of cytokines related to vascular stability and inflammation regulation. Introduction The widespread use of handheld laser increases public health concerns about laser-induced eye injury [1, 2]. As a daily indicator, the use of a 525 nm green laser has led to the occurrence of retinal injury with children and adolescents accounting for 70–80% of patients [3, 4]. Although the U.S. Food and Drug Administration has established the policies to reduce the risk of laser use, such as the 5 mW limit on laser pointers [5]. The clinical case reports show that six laser pointers with a transmission power greater than 30 mW purchased online may still cause retinal damage, that well above the safety threshold[1]. Notably, laser beam intensity can be geometrically amplified when passing through optical systems, creating substantial ocular hazards. Therefore, systematic research on laser-induced retinal injury is urgently needed to improve the clinical management and prognosis of accidental laser eye exposure. As the main site of photothermal and photochemical damage, the retina is very sensitive to 450–700 nm laser exposure [6]. Ninel et al. [7] reported a case of accidental exposure to 450 nm laser, in which the patient had retinal hemorrhage, tears, and burn-like lesions at the initial diagnosis. Then, the patient’s visual acuity gradually decreased, with full-thickness macular holes and atrophy of the outer retina. Xu et al. [8] investigated the mouse models induced by 80mW, 100mW and 120mW lasers to confirm that higher energy lasers increased the incidence of retinal cystoid edema and intraretinal hemorrhage, and exacerbated the severity of exudative retinal detachment. Xu et al. [9] showed blue light exposure in retinal organoids could cause significant apoptosis of the photoreceptors. While current literature has discussed the effects of laser exposure on ocular tissues, there is still a lack of systematic multimodal assessment of the damage characteristics and outcomes of laser-induced retinal injury of different severity levels. Green laser wavelengths, typified by 525 nm radiation, have long been established in clinical ophthalmology [10, 11]. Argon lasers and similar green light sources are routinely used for therapeutic retinal photocoagulation, which relies on relatively high laser power to induce controlled retinal pigment epithelium (RPE) damage and form therapeutic scars [10]. The majority of current studies center on this therapeutic paradigm, emphasizing RPE-centered thermal damage under specific power condition [12]. In contrast, there is currently a lack of differentiated research on the characteristics of retinal damage caused by accidental laser irradiation under various power conditions, especially in addition to structural damage, multimodal comprehensive analysis of function and cytokines. Against this background, we propose the scientific hypothesis that 525 nm continuous-wave laser irradiation of the eye with different power densities will cause significantly distinct damage responses. In this study, we adopted a multimodal evaluation to systematically characterize the retinal damage induced by 525 nm laser at graded power densities in a rabbit model. The methodologies employed include fundoscopic examination, OCT, ERG, histopathological analysis, inflammation factor profiling, and observations continued until 14 days after injury. This multifaceted approach allowed for a thorough evaluation of both the structural and functional ramifications of laser exposure on retinal integrity. Our work complements the current literature dominated by RPE-centered thermal injury studies, and provides novel insights into understanding, diagnosing and managing early accidental laser-induced retinal damage. Materials and methods Experimental setup The laser system is a laser generator from Jiangsu Shuguang Optoelectronics (Class 4 laser). The system specifications are as follows: a green fiber-coupled semiconductor laser operating continuously at 525 nm with a total power of 3.87 W. The beam divergence (divergence angle: 0.58°—3.39°). By adjusting the zoom emission lens, the estimated retinal spot diameter was 22 mm; the number of exposures per eye was 1; the distance from the laser source to the eye was 100 cm; and the beam mode was a divergent beam (divergent angle 1°). Throughout the entire experiment, all the above parameters remain unchanged. Only the laser power density is adjusted by adjusting the laser drive current: 0.19 A for 0.5 mW/cm2, 0.50 A for 200 mW/cm2, 0.95 A for 500 mW/cm2 and 1.89 A for 1000 mW/cm2.And the laser irradiation time was precisely controlled by an electronic shutter. A laser power meter (head PM100D, probe 121 C) measures the power density at the target. These components are illustrated in the schematic diagram of Fig. 1.Fig. 1Schematic drawing of the laser exposure setup for establishing rabbit retinal injury model at 525 nm Schematic drawing of the laser exposure setup for establishing rabbit retinal injury model at 525 nm Animal subjects The experimental chinchilla rabbits were male, weighing 2.34 ± 0.17 kg. Before exposure, the chinchilla rabbits were examined with a slit-lamp examination (Topcon, Tokyo, Japan) and fundus photography to confirm normal ocular tissues. In this experiment, the left eye of each rabbit was designated as the target for laser exposure, and the right eye served as the control. The rabbits were randomly allocated to 4 groups: 0.5 mW/cm2, 200 mW/cm2, 500 mW/cm2, and 1000 mW/cm2, with 6 rabbits per group. General anesthesia was induced via intramuscular injection of a compound anesthetic (Suta50 and Subexin) at a dosage of 10 mg/kg body weight. Following anesthesia, bilateral pupil dilation was achieved by instilling compound tropicamide eye drops (Santen Pharmaceutical Co., LTD, Shiga Plant, Japan). After securing the animals, eyelid specula were used to retract the eyelids and prevent blinking reflexes. Laser irradiation was delivered perpendicularly to the left ocular surface using parameters corresponding to the respective power density groups. Exposure duration was controlled at 0.25 s using a GCI-73 M precision electronic timer (Hengyang Optics, Guangzhou, China). Electroretinography(ERG) recording Circular corneal electrodes were placed on the bilateral corneal surfaces of rabbits, with a silver needle reference electrode inserted subcutaneously into the scalp and the ground electrode clamped to the tail. ERG signals were amplified at 128 Hz using a flash stimulator (IRC Co., LTD, China) and bandpass-filtered between 1 and 1000 Hz. Light stimulation was delivered via the Bright Adaptation 3.0 software (30 cd/m2 background illumination, 3.0 cd·s/m2 light stimuli, and 10 repetitions). Waveforms were recorded at baseline (before laser exposure) and at 15 s, 30 s, 45 s, 60 s, 90 s, 120 s, 5 min, 10 min, 1 h, 7 days, and 14 days post-exposure. The rabbit’s cornea was lubricated with sodium hyaluronate eye drops (Hyco-Comod, Germany) during recordings. The ERG waveform primarily consisted of a negative-phase a-wave followed by a positive-phase b-wave to form a biphasic response. The a-b wave amplitude was measured simultaneously in both eyes using an ERG electrogram. Since the left eye served as the injured eye and the right eye as the control, the a-b wave amplitude ratio was calculated as (left eye a-b wave amplitude)/(right eye a-b wave amplitude). Fundus examination The experimental animals were examined 1 day, 7 days, and 14 days post-laser exposure. The animals were anesthetized and their eyes were dilated with the above method. The operator gently lifted the eyelids with the index finger and thumb to fully expose the eyeballs. The laser scanning ophthalmoscope Daytona P200T(Opal, UK) was used for fundus scanning and photography. Optical coherence tomography(OCT) imaging High-resolution OCT and optical coherence tomography angiography (OCTA) were performed on 1 day, 7 days and 14 days after laser exposure, respectively. The animals were anesthetized and dilated with the above methods. The operator gently lifted the eyelids with the index finger and thumb to fully expose the eyeball. The retinal OCT and OCTA scans were performed using the optical microSS-OCTA (VG100, software version V2.1.01, China) and the laser scanning ophthalmoscope Daytona P200T (Ouburg, UK). The scanner utilized the following settings: Angio 6 × 6, ONH Angio 6 × 6, 512 × 512 and line scan 64. Central retinal thickness (CRT) was measured on OCT images. On OCTA, vessel density were automatically calculated in the capillary plexuses using the built-in software. Micropathological staining To analyze the morphological changes of the retinal tissue, the following procedures were strictly performed: At each designated time point, animals were euthanized and the eyeballs were promptly enucleated and placed on ice. Subsequent microdissection was carried out to precisely isolate the target tissues. Briefly, the eyeballs were sectioned along the corneal limbus under an ophthalmic microscope and the anterior segments were removed. The vitreous body was gently scraped away to fully expose the retina. The posterior polar eyecup tissues were then carefully collected and fixed in 4% paraformaldehyde solution. Finally, the samples were processed through paraffin embedding, hematoxylin–eosin (HE) staining and systematically examined for microstructural observations under a microscope. Protein microarray of inflammatory cytokines and chemokines The aqueous humor supernatant collected on day 7 post-laser exposure was frozen at −80℃ until use. 100 μL of aqueous humor supernatant (per rabbit) was used for cytokine and chemokine detection and quantification by a Quantibody Inflammation Array I kit (RayBiotech, Inc., Norcross, GA; Cat. No. QAL-CAA-30) according to the manufacturer’s instructions. The array was designed to quantitatively detect 30 cytokines simultaneously (Table 1). The signals (green fluorescence, Cy3 channel, 532nm excitation) were captured using InnoScan 300 laser scanner (Innopsys, France). Quantitation for four replicate data was performed using RayBiotech QAL-CYT-1 software. Sample concentrations (pg/ml) were determined from mean fluorescence intensities (median values) compared against a five-parameter linear regression standard curve generated from standards provided by the manufacturer. A series of isochronal maps reflecting the concentrations of the cytokine detected by the microarrays was constructed by Matlab software (R2009a, MathWorks, Inc., Natick, MA) to transform numerical data into simplex graphical patterns.Table 1Quantibody® inflammation array I kitProinflammatory factorTNF-a、MIP-1β、IL-1α、IL-1β、IL-17A、Leptin、IL-17F、TWEAK RAnti-inflammatory cytokineIL-13、IL-1RA、IL-21、Matrilin-3、ChemokinesIL-8、RANTES、BLC、TSP-2、6Ckine、SCFAdhesion factorNCAM-1、Contactin-1Growth factorVEGF-α、ANGPTL3、ANG-1Other factorsFLRG、SREC-II、Notch-3、BLMH、AMSH、CMASP3、MMP-9 Table 1Quantibody® inflammation array I kitProinflammatory factorTNF-a、MIP-1β、IL-1α、IL-1β、IL-17A、Leptin、IL-17F、TWEAK RAnti-inflammatory cytokineIL-13、IL-1RA、IL-21、Matrilin-3、ChemokinesIL-8、RANTES、BLC、TSP-2、6Ckine、SCFAdhesion factorNCAM-1、Contactin-1Growth factorVEGF-α、ANGPTL3、ANG-1Other factorsFLRG、SREC-II、Notch-3、BLMH、AMSH、CMASP3、MMP-9 Quantibody® inflammation array I kit Statistical analysis The results were analyzed using a statistical software package (GraphPad Prism). Columns and error bars indicated means and standard errors, respectively. The Shapiro–Wilk test was used to evaluate the normality of data distribution. Normally distributed data were analyzed by one-way analysis of variance (ANOVA) and Bonferroni correction was applied for multiple comparisons. Non-normally distributed data were analyzed using nonparametric tests. P values less than 0.05 were considered to be statistically significant. Results The characteristics of ERG changes in different power density laser ERG examination is used to evaluate visual function by measuring the bioelectric activity of the visual system. The a-b amplitude ratio of normal animals was 1.07 ± 0.145. At 15 s after exposure to the lasers, the a-b wave amplitude ratio were 1.01 ± 0.27 (0.5mW/cm2 group, Fig. 2A, p > 0.05), 0.88 ± 0.08 (200mW/cm2 group, Fig. 2B, p < 0.05), 0.71 ± 0.09 (500mW/cm2 group, Fig. 2C, p < 0.05) and 0.43 ± 0.07 (0.5mW/cm2 group, Fig. 2D, p < 0.05), which produced varying degrees of glare, which impacted visual function.Fig. 2The change of ERG in rabbits before and after laser exposure.Sample waveforms of ERG before laser exposure and 15 s post-laser exposure in the 0.5 mW/cm2 A, 200 mW/cm2 B, 500 mW/cm2 C and 1000 mW/cm.2 D groups. E Comparison of a-b wave amplitudes in eyes before laser exposure and 15 s post-laser exposure in each group. F Comparison of a-b wave amplitudes in eyes before laser exposure, 15 s, 30 s, 45 s, 60 s, 90 s, 120 s, 5 min, 10 min, 1 h, 7 and 14 days post-laser exposure. Data were presented as Mean ± SEM. *p < 0.05, **p < 0.01, ***p < 0.001 (n = 6) The change of ERG in rabbits before and after laser exposure.Sample waveforms of ERG before laser exposure and 15 s post-laser exposure in the 0.5 mW/cm2 A, 200 mW/cm2 B, 500 mW/cm2 C and 1000 mW/cm.2 D groups. E Comparison of a-b wave amplitudes in eyes before laser exposure and 15 s post-laser exposure in each group. F Comparison of a-b wave amplitudes in eyes before laser exposure, 15 s, 30 s, 45 s, 60 s, 90 s, 120 s, 5 min, 10 min, 1 h, 7 and 14 days post-laser exposure. Data were presented as Mean ± SEM. *p < 0.05, **p < 0.01, ***p < 0.001 (n = 6) The a-b wave amplitude measurements reflected the injury effects induced by different laser energies (Fig. 2F). In the 0.5 mW/cm2 group, the amplitude ratio exhibited only a transient decrease and recovered to 94.39% of the pre-injury level within 15 s. With increasing energy, the 200 mW/cm2 and 500 mW/cm2 groups demonstrated more substantial reductions (approximately 17.43% and 33.64%, respectively), and their recovery times were correspondingly prolonged to 30 s and 90 s. Of note, the 1000 mW/cm2 group produced the most severe response, where the amplitude ratio plummeted by 59.81% and the recovery process was delayed until 10 min. Taken together, these data establish two key findings: first, a higher laser energy resulted in a more profound initial suppression of the retinal light response and more critically, the time required for recovery was also significantly extended. The lesioned spot in fundus images in different power density laser Fundoscopic examination was employed to detect and characterize abnormal laser spots in the retina. At one day post-exposure, no lesions were readily detectable in the 0.5 mW/cm2 group. In contrast, a grade I laser spot (a pale green fluorescent spot with an indistinct edge) was observed in the 200 mW/cm2 group. The 500 mW/cm2 group presented a grade II spot (a green fluorescent spot with a clear edge), while the most severe presentation, a grade III-IV spot (a dark green fluorescent spot surrounded by a pale-green ring), was noted in the 1000 mW/cm2 group (Fig. 3). Overall, the fundoscopic findings revealed a positive correlation between laser power density and lesion severity, with higher energies producing spots that were both brighter and larger in area [13].Fig. 3The fundus photograph of the rabbit retina under different power density laser. The red arrow indicated the retinal lesions The fundus photograph of the rabbit retina under different power density laser. The red arrow indicated the retinal lesions During the 7- to 14-day observation period, the color of the fundus laser spots faded spontaneously without any therapeutic intervention. In the 200 mW/cm2 group, the bright laser spots were no longer detectable by day 14 post-injury. In the 500 mW/cm2 group, the spot color gradually diminished, slowly transitioning into a gray, aged lesion surrounded by a faint halo, indicative of a protracted repair process. In contrast, the high-reflection bright plaques in the 1000 mW/cm2 group exhibited marked persistence, with little fading observed, revealing more severe and potentially irreversible tissue damage (Fig. 3). OCT retinal disruption features in different power density laser The results of the OCT scan (1 day post-exposed) showed that no laser damage spot was visible in the fundus in the 0.5 mW/cm2 group. In the 200 mW/cm2 group, the neuroepithelial layer exhibited enhanced reflectivity, increased thickness, and disorganized structure. In the 500 mW/cm2 group, the laser injury affected the entire retina and choroid layers. The neuroepithelial layer's reflectivity was enhanced and its size was larger; the structure had completely disappeared with cystoid edema (non-reflective dark area). In the 1000 mW/cm2 group, the laser-induced retinal damage was more severe and spread to the choroid, accompanied by the separation of the neuroepithelial layer and the pigment epithelial layer (non-reflective dark area) around the laser spot. 7–14 days after injury, the thickness of the neuroepithelial layer at the laser spot has basically recovered (200 mW/cm2 group), indicating that the neuroepithelial edema caused by the laser can be gradually absorbed. The large-area structural disorder of the neuroepithelial layer is difficult to recover. After the edema subsides, part of the outer structure is retained to form irreversible scar tissue (500 mW/cm2 group). However, in the 1000 mW/cm2 group, retinal tissue was interrupted with edge curl (Fig. 4A).Fig. 4Retinal structural and thickness changes following laser irradiation. A Representative OCT cross-sectional images of the retina in different power density laser group at different time points. The green line represented the en-face location of the OCT image. The red arrow indicated the position of the laser spot. B Quantitative analysis of central retinal thickness across different laser power density groups. Data were presented as Mean ± SEM. *p < 0.05, **p < 0.01, ***p < 0.001. vs. the 0.5 mW/cm.2 group (n = 3) Retinal structural and thickness changes following laser irradiation. A Representative OCT cross-sectional images of the retina in different power density laser group at different time points. The green line represented the en-face location of the OCT image. The red arrow indicated the position of the laser spot. B Quantitative analysis of central retinal thickness across different laser power density groups. Data were presented as Mean ± SEM. *p < 0.05, **p < 0.01, ***p < 0.001. vs. the 0.5 mW/cm.2 group (n = 3) To evaluate the structural damage induced by laser irradiation, we quantified the dynamic changes in CRT across all groups over the 14-day follow-up period. In the 0.5 mW/cm2 group, CRT remained stable at baseline levels throughout the observation period (p > 0.05 at all time points). In the 200 mW/cm2 group, CRT was increased on day 1, then gradually decreased over time, returning to near-baseline levels by day 14. In the 500 mW/cm2 group, a marked increase in CRT was observed on day 1 (p < 0.01 vs. 0.5 mW/cm2), thickness returned to a level not significantly different from the 0.5 mW/cm2 control group by day 7. The 1000 mW/cm2 group exhibited the most severe and persistent structural changes: CRT peaked on day 1 (p < 0.001 vs. 0.5 mW/cm2), decreased progressively thereafter, and was significantly lower than the 0.5 mW/cm2 group by day 14 (p < 0.01) (Fig. 4B). Retinal local blood flow density features in different power density laser. To assess laser-induced changes in fundus blood flow, we then performed OCTA on the experimental animals. In the 200 mW/cm2 group, no significant alteration in local blood flow density around the laser spot was observed post-exposure. However, in both the 500 and 1000 mW/cm2 groups, cross-sectional OCTA on day 1 revealed capillary leakage and bleeding at the lesion site, which resolved within 7 to 14 days. Concurrently, while these two groups exhibited high local blood flow density at the spot on day 1, a shift occurred from day 7 to 14: the central density peak diminished and the high-density area began to diffuse. Notably, by day 14, the high-density shadow in the 500 mW/cm2 group showed signs of dissipation, whereas it continued to spread in the 1000 mW/cm2 group(Fig. 5A).Fig. 5Retinal microvascular changes following laser irradiation assessed by OCT angiography. A Representative OCTA and corresponding cross-sectional perfusion images of the retinal superficial capillary plexus different power density laser group at different time points. Red arrows indicate the laser-irradiated area. B Quantitative analysis of superficial retinal vessel density across different laser power density groups over time. Data were presented as Mean ± SEM. *p < 0.05, **p < 0.01, ***p < 0.001. vs. the 200 mW/cm.2 group (n = 4) Retinal microvascular changes following laser irradiation assessed by OCT angiography. A Representative OCTA and corresponding cross-sectional perfusion images of the retinal superficial capillary plexus different power density laser group at different time points. Red arrows indicate the laser-irradiated area. B Quantitative analysis of superficial retinal vessel density across different laser power density groups over time. Data were presented as Mean ± SEM. *p < 0.05, **p < 0.01, ***p < 0.001. vs. the 200 mW/cm.2 group (n = 4) In the 200 mW/cm2 group, vessel density was 46.29 ± 9.73% on day 1, decreased significantly to 12.21 ± 3.21% on day 7, and remained at 17.33 ± 3.80% by day 14. In the 500 mW/cm2 group, vessel density exhibited a biphasic pattern: it was 69.08 ± 7.85% on day 1, increased to a peak of 76.12 ± 6.38% on day 7 (significantly higher than the 200 mW/cm2 group, p < 0.001), and subsequently decreased to 43.25 ± 5.68% on day 14(significantly higher than the 200 mW/cm2 group, p < 0.05). In the 1000 mW/cm2 group, vessel density was 67.94 ± 0.93% on day 1, reached a peak of 81.17 ± 4.40% on day 7 (significantly higher than the 200 mW/cm2 group, p < 0.001), and remained at 71.53 ± 5.25% on day 14. At day 14, vessel density in the 1000 mW/cm2 group was significantly higher than that in the 500 mW/cm2 group (p < 0.05) (Fig. 5B). The severity of micropathological damage in different power density laser. Histopathological examination of retinal damage in rabbits was performed 1,7 and 14 days after irradiation. At 1 day post-injury, the 0.5 mW/cm2 group exhibited essentially normal architecture across all retinal layers, without evident cellular edema, necrosis or inflammatory cell infiltration. The retinal pigment epithelium (RPE) layer remained continuous and the photoreceptor layer displayed an intact morphology. In the 200 mW/cm2 group, prominent edema was observed in the outer retinal layers, including the outer plexiform layer, outer nuclear layer and photoreceptor layer, along with enlarged intercellular spaces. In some regions, the photoreceptor layer showed irregular morphological changes, such as disorganization and swelling. The 500 mW/cm2 group showed marked pan-retinal edema, extensive vacuolar degeneration was observed in cells of the outer nuclear layer and photoreceptor layer, with cellular necrosis and disintegration present in certain regions. Furthermore, the RPE layer exhibited disrupted continuity with focal detachment observed. The 1000 mW/cm2 group displayed severe necrosis and disintegration throughout all retinal layers at the laser spot with destruction of cellular structures and formation of abundant cellular debris. Widespread RPE layer detachment was observed and hemorrhage or exudation was noted in the choroidal layer. 7–14 days after irradiation, in the 0.5 mW/cm2 group, the cellular morphology and arrangement across all retinal layers were indistinguishable from those of normal tissues. In the 200 mW/cm2 group, the edema had partially subsided and the cells in the outer nuclear layer and photoreceptor layer gradually became regularly arranged, though a few cells still showed vacuolar changes on the 7th day. On day 14, the retinal layered structure was largely restored with only slight irregularities in cellular arrangement remaining. On day 7, in the 500 mW/cm2 group, necrotic cells began to be cleared and areas of tissue lysis appeared. The number of cells in the inner nuclear layer and outer nuclear layer decreased and the photoreceptor layer structure was disorganized. On day 14, the retinal laminar structure partially recovered. However, significant cell loss remained in the outer nuclear layer and photoreceptor layer and the RPE layer showed incomplete repair. On day 7, in the 1000 mW/cm2 group, the necrotic tissue further underwent lysis, forming large areas lacking identifiable retinal structure. Only a few cells remained in the inner nuclear layer and outer nuclear layer and a vigorous inflammatory response was observed, accompanied by inflammatory cell infiltration. On day 14, severe destruction of the retinal structure was still evident with only partial residual cells in the inner nuclear layer and outer nuclear layer. The photoreceptor layer had almost completely disappeared and fibrotic scar tissue formed in the RPE layer and choroid (Fig. 6).Fig. 6Pathological observations of retinal damage in rabbit eyes post-exposure with 525 nm laser. Red arrow: enlarged intercellular spaces, with irregular morphology of photoreceptor layer in some areas; Light yellow arrow: cell structure destruction, with a large number of cell necrosis; Yellow arrow: visible hemorrhage or exudation; Green arrow: full-thickness edema; Pink arrow: photoreceptor layer almost disappeared, Scale bar = 50 μm Pathological observations of retinal damage in rabbit eyes post-exposure with 525 nm laser. Red arrow: enlarged intercellular spaces, with irregular morphology of photoreceptor layer in some areas; Light yellow arrow: cell structure destruction, with a large number of cell necrosis; Yellow arrow: visible hemorrhage or exudation; Green arrow: full-thickness edema; Pink arrow: photoreceptor layer almost disappeared, Scale bar = 50 μm Acute laser damage caused a sterile inflammatory response in the fundus of the eyes. We collected anterior aqueous samples from each group at 7 days and measured the level of 30 cytokines using the RayBio QAL-CAA-30 Inflammation Array. Cluster and principal component analysis (Fig. 7. A, B) revealed no significant dose-dependent differences in the inflammatory array between laser-treated groups and controls. Notably, one sample in the 200 mW/cm2 group exhibited markedly increased levels of NCAM-1, Matrilin-3, and NOTCH3.Fig. 7Inflammatory cytokines/chemokines concentration of different anterior aqueous humor groups. A. Cluster diagram, B. PCA diagram, C. Bar chart. The data were presented quantitatively as mean ± standard deviation (n = 4) Inflammatory cytokines/chemokines concentration of different anterior aqueous humor groups. A. Cluster diagram, B. PCA diagram, C. Bar chart. The data were presented quantitatively as mean ± standard deviation (n = 4) Statistical analysis was performed on the cytokine levels across all experimental groups. Regarding inflammatory factors, several significant alterations were observed: the proinflammatory macrophage inflammatory protein MIP-1β was significantly elevated in the 200 mW/cm2 group (p = 0.046), whereas the anti-inflammatory factor IL-1RA was markedly reduced in the 500 mW/cm2 group (p = 0.043). Among the chemotactic and adhesion factors, BLC showed a significant decrease in the 1000 mW/cm2 group (p = 0.031). Similarly, the level of 6Ckine was significantly reduced in both the 0.5 mW/cm2 (p = 0.045) and 200 mW/cm2 groups (p = 0.043). A notable trend was observed with growth factors: Angiopoietin-1 (ANG-1) was significantly upregulated in both the 500 mW/cm2 (p = 0.008) and 1000 mW/cm2 (p = 0.039) groups. In contrast, Vascular Endothelial Growth Factor A (VEGF-A) was significantly down-regulated specifically in the 500 mW/cm2 group (p = 0.039) (Fig. 7C). Discussion This study systematically investigated retinal damage and recovery characteristics induced by 525 nm laser irradiation at graded power densities (0.5, 200, 500, and 1000 mW/cm2) in chinchilla rabbits. Our results demonstrated that laser exposure triggered a series of dose-dependent alterations in retinal function, structure and secondary cytokine responses. Specifically, low-power laser irradiation (0.5 mW/cm2) did not induce obvious functional or pathological retinal damage. By contrast, high-power laser exposure (200, 500, and 1000 mW/cm2) led to significantly reduced ERG amplitudes with slow recovery, formation of permanent laser lesions of varying severity involving both the retina and choroid, and severe injury (500 and 1000 mW/cm2) characterized by retinal detachment, hemorrhage and dynamic cytokine fluctuations. A key strength of our multimodal design is that it explicitly links these dimensions —for example, ERG reduction corresponded to increased central retinal thickness on OCT, and reciprocal ANG-1/VEGF-A changes explained OCTA-detected vascular leakage. Collectively, these findings indicate that laser-induced retinal injury is closely associated with power density: a higher laser power was associated with more severe retinal injury and slower recovery. In this study, 525 nm laser radiation induced a series of typical pathological changes in rabbit retinas, including retinal edema, hemorrhage, detachment, and holes, which are consistent with the clinical features of laser retinopathy induced by handheld laser devices in children reported by Neffendorf [14]. Joglekar et al. also reported 51 cases of ocular injury caused by recreational laser irradiation, with favorable outcomes observed in patients with laser-induced premacular hemorrhage and decreased visual acuity after 3-month follow-up [15]. The chinchilla rabbit model used in this study recapitulates key features of human laser retinal injury well, owing to its similar ocular pigmentation to humans [16]. Therefore, the injury characteristics and natural recovery patterns induced by different-energy 525 nm lasers observed in this study can provide important evidence for the development of clinical retinal repair strategies. Analysis of inflammatory cytokines in anterior aqueous humor revealed no significant separation among groups in clustering analysis, whereas several key individual cytokines showed statistically significant differences. This discrepancy is explained by the fact that clustering analysis reflects the global expression profile of all detected factors, whereas individual cytokine analysis focuses on functionally critical molecules that may undergo significant regulation even without prominent global shifts [17, 18]. ANG-1 was significantly upregulated, while VEGF-A was significantly downregulated in the high-power laser groups. VEGF-A is a highly specific pro-angiogenic factor, whereas ANG-1 primarily participates in vascular maturation and integrity maintenance. These molecular changes correspond well with the reduced vessel density and perfusion index observed on OCTA. The reciprocal changes in these two factors likely reflect a compensatory homeostatic response aimed at preserving vascular stability after retinal vascular injury. This ANG-1/VEGF-A signaling might represent a therapeutic strategy to stabilize retinal vasculature after accidental laser injury. However, Kakimoto et al. reported marked upregulation of multiple pro-inflammatory cytokines (including IL-6 and MCP-1) after 200 mW/cm2 laser exposure in pigmented male Rex rabbits [19]. Although our cytokine chip does not contain these two types of pro-inflammatory factors, the proinflammatory macrophage inflammatory protein MIP-1β was significantly elevated in the 200 mW/cm2 group (p = 0.046), whereas the anti-inflammatory factor IL-1RA was markedly reduced in the 500 mW/cm2 group (p = 0.043). The finding also reflects a similar trend of inflammatory changes. In our study, laser lesions were targeted to the posterior pole of the retina, thereby avoiding direct injury to large retinal vessels and resulting in relatively mild cytokine secretion and moderate vascular stability responses [20, 21]. Conversely, laser irradiation directed at highly vascularized areas such as the optic disc may cause more extensive vascular damage and trigger stronger downstream cytokine release [22, 23]. Notably, one sample in the 200 mW/cm2 group showed markedly elevated levels of NCAM-1, Matrilin-3 and Notch3, all of which were associated with cell adhesion, neural development, and injury repair. NCAM-1 plays a crucial role in regulating neurogenesis, neurite growth, cell proliferation, and migration [24]. Matrilin-3 inhibits astrocyte-mediated neuroinflammation following ischemic stroke through both intracellular and extracellular mechanisms [25]. Notch3, a transmembrane developmental signaling receptor, maintains the resting state of neural progenitor cells [26]. The results of this sample suggested that laser irradiation of the fundus may have a potential risk of neurological damage, highlighting laser-induced neuro-retinal injury as a meaningful direction for future investigation. The 525 nm green laser primarily targets the RPE and induces controllable retinal damage mainly through photothermal injury, oxidative stress, and subsequent acute inflammatory cascades [27–29]. In this study, all optical parameters were strictly controlled to ensure reproducible retinal irradiation. When the laser power density exceeded approximately 200 mW/cm2, obvious RPE damage and scar formation occurred, further triggering robust secondary immune and inflammatory reactions. Higher laser power resulted in more severe RPE disruption, stronger inflammatory activation, and more prominent vascular dysregulation, consistent with our changes in CRT, vessel density and cytokine profiles. These structural and functional alterations observed via OCT and histology were closely correlated with molecular changes. This experiment selected aqueous humor for cytokine detection considering that aqueous humor samples can reflect the overall inflammatory response of the eye, which offers continuous monitoring, is minimally invasive and is clinically feasibility. However, retinal-derived cytokines may be diluted or delayed due to the "vitreous to aqueous humor" conduction process. This may be one of the reasons why the cluster analysis of cytokines in this paper shows no significant difference. It must be acknowledged that vitreous sampling can more directly reflect the local inflammatory response of the retina [10]. Future research should supplement these findings of local inflammatory reaction of the retina after laser irradiation. Meanwhile, the Chinchilla rabbit lacks a critical visual region—the macula, which limits the direct translation of our findings to human macular damage. In the future, non-human primates (such as macaques) with eye structures more similar to humans will be selected as animal models for laser-induced ocular injury to address the limitations of this study. In summary, this work provides a comprehensive, multimodal characterization of the structural, functional and cytokine features, as well as the clinical consequences of laser-induced retinal damage at varying severity levels. We established a theoretical model capable of explaining both the immediate injury and the subsequent long-term repair, thereby laying a vital foundation for future efforts to clarify therapeutic approaches, refine treatment workflows and improve patient prognosis in cases of laser-induced retinal injury. Conclusions In conclusion, this study provides a multimodal, structure–function-cytokine evaluation of 525 nm at graded power densities laser-induced retinal injury in rabbits. Low-power (0.5 mW/cm2) exposure caused no detectable damage. In contrast, power densities ≥ 200 mW/cm2 induced transient ERG suppression, persistent fundus lesions, and dose-dependent structural disruptions ranging from localized necrosis (200 mW/cm2) to full‑thickness retinal perforation (1000 mW/cm2). Importantly, our cytokine analysis revealed a compensatory ANG-1↑/VEGF-A↓ response in high-power groups, which correlated with OCTA-detected vascular changes and histologically observed RPE injury. The integrated structure–function-cytokine framework not only clarifies the pathophysiology of laser-induced retinal injury but also establishes a practical template for future preclinical evaluation of retinal protective or regenerative therapies.