Retinopathy of Prematurity (ROP): An Overview of Biomarkers in Various Samples for Prediction, Diagnosis, and Prognosis

Introduction

Retinopathy of Prematurity (ROP) is a proliferative retinal vascular disease that affects preterm infants.^1,2 While ROP may regress spontaneously in some cases, it can progress from mild stages (ROP stages I and II) to more severe forms (ROP stages III, IV, and V) in others, potentially resulting in complete vision loss if not treated promptly.³ However, ROP is largely preventable, and reducing the incidence of blindness hinges on high-quality neonatal care, comprehensive ROP screening programs, and the expertise of skilled ophthalmologists.⁴ Although ROP is classified as a neovascular disease, the underlying molecular mechanisms remain unclear. It is hypothesized that multiple pathological risk factors, including low birth weight, gestational age, postnatal oxygen fluctuations, and inflammation, contribute to ROP progression by inducing retinal vascular occlusion and neovascularization.⁵ This, in turn, triggers oxidative stress and the release of pro-inflammatory and pro-angiogenic factors.⁶ These cytokines play a pivotal role in the onset and progression of ROP, prompting numerous studies to focus on detecting ROP-related factors for early prediction and intervention to mitigate the incidence and severity of ROP in preterm infants.

Currently, ROP screening by the neonatologist is based on birth weight and gestational age, with ROP staging based on binocular indirect ophthalmoscopy (BIO) and/or wide-field retinal imaging systems.^7,8 However, both methods require pupil dilation, the use of an eyelid speculum, and may involve scleral indentation, potentially causing side effects such as physiological stress, apnea, and cardiovascular instability.^9,10 Additionally, multiple examinations are often required to reach a clinical conclusion, with the decision to initiate treatment largely depending on the ophthalmologist’s clinical experience^11–13 or AI.¹⁴ In light of research on the pathogenesis of ROP, various factors have been measured in different sample types, including peripheral blood, umbilical cord blood, amniotic fluid, placenta, vitreous fluid, aqueous humor, urine, and tears, to elucidate their correlation with ROP pathobiology^15,16 (shown in Figure 1). This review aims to evaluate the existing ROP detection methods, with the goal of identifying more effective, safer, and simpler approaches that could reduce the frequency of ROP staging exams in preterm infants. Furthermore, it seeks to identify relevant biomarkers that could enable the development of more personalized ROP screening protocols, thereby improving the diagnosis and prognosis of ROP and alleviating the burden on patients, families, and healthcare providers.

Figure 1 Different detection methods of ROP, including peripheral blood, umbilical cord blood, amniotic fluid, placenta, vitreous humor, aqueous humor, urine and tear fluid.

Different Detection Methods

Peripheral Blood

Peripheral blood, composed of various cell types and compounds such as salts and proteins, has been extensively studied for its potential in screening, predicting, diagnosing, and prognosticating Retinopathy of Prematurity (ROP). Research by Villegas-Becerril et al has identified Vascular Endothelial Growth Factor (VEGF) and Insulin-like Growth Factor 1 (IGF-1) as important markers in predicting the risk of ROP.¹⁷ VEGF, a key molecule involved in pathological retinal vascular changes,^18,19 is downregulated during stage I ROP and upregulated in the Müller glial cells of the peripheral avascular retina during stage II.^20–22 Hellgren and Yenice discovered that preterm infants who later developed ROP had significantly lower circulating VEGF levels at birth.^23,24 Additionally, studies by Pieh and Yalin demonstrated that VEGF levels significantly increased in the blood of patients with late-stage or treatment-requiring ROP.^25,26 Thus, maintaining physiological levels of VEGF may prevent ROP progression, while reducing pathological VEGF levels in advanced ROP could help mitigate disease progression. These clinical findings are strongly supported by results from oxygen-induced retinopathy (OIR) animal models.^27–29 IGF-1, essential for retinal vascular formation during normal eye development,^30–32 stimulates VEGF synthesis.³³ When IGF-1 levels are low, VEGF is unable to activate the Akt signaling pathway during the first stage of ROP, leading to endothelial cell apoptosis. Several studies have confirmed that low blood IGF-1 levels are associated with the development and increased severity of ROP,^34–36 and early restoration of IGF-1 to normal levels can prevent the onset of ROP.³⁷

A large body of research has also shown that various inflammatory factors in the blood, such as IL-6, IL-8, IL-18, and TNF-α, play significant roles in the occurrence and progression of ROP as regulators of angiogenesis. IL-6, which has both pro-inflammatory and anti-inflammatory effects,³⁸ can promote VEGF expression.³⁹ Multiple studies have reported elevated IL-6 levels in the plasma/serum of preterm infants before and after birth in those who eventually develop ROP, whether mild or severe.^34,40–42 IL-8, the first chemokine studied, has notable angiogenic activity and induces ocular inflammation.⁴³ Several studies have demonstrated a significant association between elevated IL-8 levels in the plasma/serum of preterm infants after birth and the development of severe ROP,^42,44,45 findings that have been further validated in animal experiments.⁴⁶ IL-18, a pro-inflammatory cytokine with immunomodulatory properties, exhibits both angiogenic and angiostatic effects.⁴⁷ Sood et al observed lower serum IL-18 levels in preterm infants with ROP compared to those without the disease, although IL-18 levels increased within three weeks after birth.⁴⁰ The biphasic pattern of IL-18 expression aligns with the two stages of ROP development,⁴⁸ underscoring its role as a time-sensitive angiogenic regulator.⁴⁰ TNF-α, primarily produced by monocytes or macrophages, is a major initiator of inflammation.⁴⁹ It enhances the production of other cytokines, such as IL-8, basic Fibroblast Growth Factor (bFGF), and MCP-1,⁵⁰ in perivascular retinal microglia through autocrine or paracrine mechanisms and plays a role in the formation of hypoxic retinal neovascularization.^51,52 Clinical studies by Hellgren and Holm found that elevated TNF-α levels in the serum/plasma of preterm infants are associated with the development and severity of ROP.^34,42,44 Moreover, animal studies have confirmed that inhibiting TNF-α significantly improves vascular restoration in the ischemic retina of the mouse OIR model and reduces pathological neovascularization.^53,54

In addition to cytokine detection, immune cell infiltration in peripheral blood has also been linked to ROP. Peripheral Blood Mononuclear Cells (PBMCs), which include lymphocytes (T cells, B cells, and Natural Killer cells), monocytes, and dendritic cells, are central to the immune response.⁵⁵ Kurtul and others demonstrated that lymphocyte count is negatively correlated with ROP, indicating its independent predictive value.⁵⁶ Further studies by Zhou and Li on PBMCs in ROP patients requiring treatment revealed significant changes in non-coding RNAs (tsRNA, miRNA, circRNA), with hsa_circRNA_061346, hsa_circRNA_092369, and hsa_circRNA_103554 emerging as promising biomarkers and molecular targets for diagnosing treatment-required ROP.^57,58

Moreover, research on the genetic polymorphisms of peripheral blood has identified numerous genes involved in pathways related to fetal retinal growth and development, angiogenesis, inflammation, neurodegeneration, and oxidative stress, which are associated with ROP.⁵⁹ Rathi et al found that gene variants in the complement pathway (CFH, CFB, C3), ECM remodeling (FBLN5, MMP9), leukocyte transendothelial migration and activation (CXCR4), HIF1A signaling and angiogenesis (ANGPT2, H2AFX, and VEGF), and developmental processes (TGFb1, IHH) are closely linked to the pathogenesis of ROP.⁵⁹ Fevereiro-Martins and others discovered that polymorphisms in genes involved in the WNT signaling pathway, VEGFA gene, eNOS gene, and Brain-Derived Neurotrophic Factor (BDNF) gene are also associated with ROP development.⁶⁰ Xu et al highlighted the role of lactate metabolism in ROP development, finding that a large number of lactate metabolism-related genes (LMRGs) are significantly associated with ROP, confirming the involvement of these genes in the disease’s pathogenesis.⁶¹

Peripheral blood, typically collected from peripheral veins, is easily accessible and represents a minimally invasive and routine method for preterm infants, allowing for repeated analysis. Blood tests reflect cellular activity throughout the body and have been widely applied across various diseases.

Umbilical Cord Blood

Umbilical cord blood, collected from the umbilical vein at the time of delivery, has been extensively studied for its relationship with ROP, particularly in the context of VEGF and IGF-1 levels. Research consistently shows that elevated VEGF levels in the cord blood of ROP patients and low serum VEGF content in umbilical cord blood are independent risk factors for the development of ROP in preterm newborns.^15,24,35 IGF-1 levels are also decreased, with low levels of IGF-1 in the cord blood of very preterm infants negatively correlating with severe ROP,^15,35,62,63 positioning cord blood IGF-1 as a biomarker for severe ROP risk. Cekmez et al found that levels of Apelin, a vascular endothelial growth factor necessary for normal vascular growth and endothelial cell proliferation,^64,65 are lower in the umbilical cord blood of ROP patients at birth, with changes in Apelin levels positively correlating with changes in IGF-1 levels.⁶² Apelin is implicated in cell proliferation and angiogenesis.⁶⁶

Furthermore, elevated expression of numerous inflammatory factors in umbilical cord blood has been linked to ROP development and severity, with the expression of inflammatory factors (MIP-1β, MCP-1) inversely correlated with gestational age (GA) and birth weight (BW).⁶⁷ Yu et al reported that elevated levels of Monocyte Chemoattractant Protein-1 (MCP-1), Macrophage Inflammatory Proteins (MIP-1α, MIP-1β), and IL-7 in umbilical cord serum predict ROP risk, with MIP-1β associated with ROP severity.⁶⁷ Park et al demonstrated that high concentrations of inflammatory mediators (IL-6 and C5a) in umbilical cord blood collected at birth are significantly associated with an increased risk of severe ROP and Type 1 ROP.⁶⁸ Elevated IL-6 levels in umbilical cord plasma can predict ROP severity and serve as an independent marker for severe ROP, while increased C5a concentrations can assess ROP severity and the need for laser treatment.⁶⁸ Thus, the combined analysis of various inflammatory factors offers a more accurate prediction of ROP development.

In addition to cytokines, protein level detection in umbilical cord blood has also been closely linked to ROP. TGFBI, an extracellular matrix protein associated with angiogenesis, inflammation, and embryonic development, plays a key role in regulating immune and inflammatory responses.⁶⁹ Endoglin, critical in post-occlusive reperfusion and neovascular diseases, is involved in angiogenesis, neovascular formation, and vascular remodeling.⁷⁰ Song et al found that reduced TGFBI levels in cord blood are significantly associated with severe ROP and Type 1 ROP, with low Endoglin levels serving as a predictor for Type 1 ROP.⁷⁰ A combined predictive model based on TGFBI, Endoglin levels, and weight data can serve as a reliable birth indicator for neonatal ROP risk.⁷⁰ Additionally, Madan et al found higher levels of deamidated globulin chains in the cord blood of preterm infants with severe ROP,⁷¹ where deamidation is a form of protein damage associated with various pathological conditions.^72,73

Umbilical cord blood at birth directly reflects the intrauterine environment’s impact on the fetus, including hypoxia, stress, injury, and infection/inflammation.⁶⁸ Compared to peripheral blood, it is collected at an earlier stage, allowing for a more timely reflection of the pathological processes involved in ROP.

Amniotic Fluid

Amniotic fluid, collected through direct puncture during delivery, has been examined in preterm infants to assess its relationship with Retinopathy of Prematurity (ROP). Studies, such as those by Woo et al, have identified significant correlations between elevated levels of inflammatory mediators (eg, IL-6 and IL-8) and angiogenic mediators (eg, Endoglin, Endostatin, and IGFBP-2) in the amniotic fluid and the severity of ROP.⁷⁴ These biomarkers, combined with prenatal factors like gestational age and birth weight, can predict ROP’s occurrence and progression.⁷⁴ Endostatin, known for inhibiting angiogenesis, may counteract the effects of VEGF and other growth factors.⁷⁴ IGFBP-2, which plays a role in cell proliferation and angiogenesis, is also expressed in fetal and placental tissues.^33,75 Additionally, Jang et al found that lower levels of IL-10 and TNF-α in amniotic fluid were significantly associated with ROP, and increased MMP-2 levels were identified as a risk factor for the disease.⁷⁶ MMP-2, a protease involved in extracellular matrix degradation, is linked to pathological retinal neovascularization.^77,78

Because amniotic fluid is collected early, it allows for the detection of factors that may reflect the initial stages of ROP, providing a timely insight into the disease’s development.

Placenta

The placenta serves as the critical interface between mother and fetus, facilitating the exchange of nutrients and oxygen. Research has shown that placental pathology significantly impacts postnatal conditions in preterm infants, such as bronchopulmonary dysplasia, necrotizing enterocolitis, and neurocognitive and neurovascular development.^79–85 Histological examinations of the placenta can provide valuable insights into the risks of adverse neonatal outcomes, particularly in extremely preterm infants (EPT) born before 28 weeks of gestation.⁸⁶

Placental function is directly linked to the underlying mechanisms of ROP. For instance, Leviton et al demonstrated that preterm infants with impaired placental implantation exhibited significantly lower IGF-1 levels in peripheral blood on the first day post-birth.⁸⁷ However, the relationship between placental infection, inflammation, and ROP remains controversial. Some studies suggest a strong association between placental inflammation (eg, chorioamnionitis, funisitis) and an increased risk of ROP.^88–92 Inflammatory conditions in the placenta may trigger cytokine release into the fetal circulation, contributing to ROP development. Chen et al, for example, found that co-existing placental bacterial infection and histological inflammation were linked to an increased risk of Zone I ROP but not severe ROP.⁸⁸ Conversely, other studies have suggested no significant correlation and even a protective effect of placental inflammation against ROP.^93–97 Owen et al posited a protective relationship between acute placental inflammation and ROP risk,¹⁶ a finding supported by Park et al, who observed a decreased ROP risk with the progression of acute histological chorioamnionitis (HCA).⁹⁷ These contradictory findings indicate a complex relationship between placental function, inflammation, and ROP, warranting further investigation.

Moreover, epigenetic investigations by Bulka et al on placental tissue unveiled that placental DNA methylation at 16 CpG sites in 8 genes linked to mothers of extremely low gestational age newborns (ELGAN) is associated with early ROP onset risk.⁹⁸ Methylation patterns in genes like Serum Amyloid A (SAA1 and SAA2), Myeloperoxidase (MPO), C-Reactive Protein (CRP), and Tumor Necrosis Factor Receptor Superfamily Member 1B (TNFRSF1B) exhibit a negative correlation with early ROP onset risk.⁹⁸ Conversely, methylation at three sites in the Tumor Necrosis Factor Receptor Superfamily Member 1A (TNFRSF1A) gene is positively linked to early ROP risk. Furthermore, methylation at two sites in the Brain-Derived Neurotrophic Factor (BDNF) and Angiopoietin 1 (ANGPT1) genes respectively correlates positively/negatively with early ROP risk.⁹⁸ Assessment of CpG methylation sites across multiple genes can serve as valuable predictive biomarkers for ROP, offering avenues for early intervention to mitigate disease severity.

Studies on other placental elements, such as chemical composition, have also been insightful. For example, Deev et al found altered levels of elements like nitrogen and potassium in the placentas of ROP patients, with nitrogen content potentially serving as a preclinical biomarker for the disease.⁹⁹ Research into placental nutrients has similarly shown correlations between omega-3 receptor expression and ROP occurrence,¹⁰⁰ emphasizing the role of omega-3 in retinal vascularization and reducing the severity of pathological neovascularization.^54,101

Histological examination of placental tissue offers early insights into the pathological processes of ROP, enhancing our understanding of the disease’s pathogenesis and providing clinical insights for early prediction and prevention of ROP onset and progression.

Vitreous Fluid

The vitreous humor, which constitutes 80% of the eye’s volume, is predominantly composed of water (98–99.7%).¹⁰² It forms a transparent gel structure containing a myriad of diluted structural macromolecules (such as hyaluronic acid (HA), proteoglycans, collagen, and non-collagen proteins), non-structural proteins (serum), and a small population of cells (hyalocytes).¹⁰³ Given its proximity to the retina, particularly when the blood-retinal barrier (BRB) is compromised, the vitreous can accumulate secretory products from the retina.¹⁰⁴ Consequently, in vitreoretinal diseases, the composition of the vitreous undergoes alterations due to the varied expression of proteins under distinct pathological conditions.¹⁰³

Studies have identified elevated levels of angiogenic factors like VEGF,^59,105–107 erythropoietin,¹⁰⁸ angiopoietin-1, and angiopoietin-2 in the vitreous of ROP patients,¹⁰⁹ while anti-angiogenic factors such as pigment epithelium-derived factor and VEGF165b are decreased.^106,110 Angiopoietins work in conjunction with VEGF, contributing to both physiological and pathological neovascularization.¹⁰⁹ Higher levels of Ang-1 and Ang-2 have been observed in the vitreous of infants with moderate to late-stage ROP.¹¹¹ Although intravitreal anti-VEGF injections are a common treatment for ROP, particularly aggressive posterior ROP.^112,113 Nonetheless, some side effects have been observed in neonates treated with anti-VEGF for ROP, including neovascular recurrence,¹¹⁴ enduring impairments in photoreceptor integrity,¹¹⁵ and compromised eye and organ development.^116–118 Studies also indicate resistance to anti-VEGF therapy in some patients,¹¹⁹ underscoring the necessity of identifying alternative angiogenic or anti-angiogenic cytokines implicated in ROP pathogenesis.

Beyond angiogenic factors, studies have documented escalated levels of numerous inflammatory factors, complement factors, and chemokines in the vitreous of ROP patients. Sato and Velez-Montoya, for example, found increased levels of erythropoietin, VEGF, IL-6, IL-7, IL-15, Eotaxin, G-CSF, IP-10, and RANTES in the vitreous.¹²⁰ Complement factors like CFH, C3, and C4,⁵⁹ typically downregulated in normal preterm infants due to immune immaturity,^121,122 are upregulated in ROP patients, suggesting a significant role for the complement pathway in ROP development.⁵⁹ The chemokine receptor CXCR4, which facilitates lymphocyte and monocyte migration,¹²³ has been shown to influence retinal vascular sprouting,¹²⁴ and RANTES, a chemokine critical for innate immunity in neonates, has demonstrated complex roles in ROP.¹²⁵ Sato et al have found that RANTES exhibits significantly higher vitreous levels in both active and inactive ROP patients compared to non-ROP individuals.¹²⁰ However, another study reported lower RANTES concentrations in the vitreous of preterm infants with severe ROP, with higher RANTES levels associated with reduced ROP risk,⁴⁴ suggesting a protective role of RANTES.

Additionally, examinations of proteins in the vitreous of ROP patients, such as Opticin, a glycoprotein abundant in the human vitreous extracellular matrix (ECM) and an endogenous anti-angiogenic factor,^126,127 have shown reduced expression of the anti-angiogenic protein Opticin in the vitreous of ROP patients.³ Intravitreal administration of Opticin has demonstrated efficacy in preventing retinal neovascularization development, with animal studies corroborating Opticin’s ability to promote regression of established retinal neovascularization in a mouse model of oxygen-induced retinopathy.³

However, due to the invasive nature of vitreous sampling, this method of “screening” or of prophylactic treatment, is not currently practical. Current knowledge primarily comes from animal studies, which provide valuable insights into retinal conditions and help guide clinical research on ROP’s pathogenesis.

Aqueous Humor

The aqueous humor is secreted by the ciliary epithelium, extending to the serrated margin and reaching the peripheral retina. Therefore, the cytokine levels in the aqueous humor of ROP patients are likely to reflect retinal-related cytokine levels, similar to those in the vitreous.¹⁰⁷ Studies by Nonobe and Velez-Montoya have found elevated levels of VEGF in the aqueous humor of late-stage ROP patients, with a significant reduction in VEGF concentration after intravitreal injection of anti-VEGF (bevacizumab).^107,128 Liang et al further observed that VEGF levels in the aqueous humor of preterm infants with severe ROP (aggressive retinopathy of prematurity, A-ROP) were higher than those with milder forms (threshold ROP and type 1 pre-threshold ROP).¹²⁹ The VEGF levels were negatively correlated with the lesion area and stage of ROP but positively correlated with the degree of venous tortuosity, while unrelated to arterial tortuosity.¹²⁹

Moreover, Lyu et al detected elevated levels of VEGF, interferon-γ (IFN-γ), IL-10, and IL-12 in the aqueous humor of ROP patients, correlating with the severity of ROP.¹³⁰ Notably, higher levels of VEGF and MIP-1β were independently associated with the need for ROP retreatment.¹³⁰

In addition to VEGF and cytokines, other anti-angiogenic factors such as Col1a1, tRF-1001, and 16K-PRL have been detected in the aqueous humor of ROP patients. Animal studies have shown that Col1a1 can reduce retinal neovascularization and ischemic areas, making it a promising therapeutic target.¹³¹ Clinical studies by Xia et al also observed upregulated expression of Col1a1 in the aqueous humor of ROP patients.¹³¹ tRF-1001, a novel class of non-coding RNA transcripts with anti-angiogenic properties, has shown downregulation in the retina of an oxygen-induced retinopathy (OIR) animal model and in the aqueous humor of patients with age-related macular degeneration (AMD).¹³² The upregulation of tRF-1001 could inhibit pathological angiogenesis, offering potential therapeutic avenues.¹³² 16K-PRL, a potent angiogenesis inhibitor produced locally in the eyes of ROP patients, has been linked to the regression of intraocular vessels, with Dueñas et al reporting significantly elevated PRL levels in the aqueous humor of ROP patients.¹³³

Aqueous humor is generally collected using a 30-gauge needle through a clear corneal puncture, obtaining 20–30 μL of undiluted aqueous humor from each eye, is somewhat invasive and may pose severe complications. Current research on aqueous humor predominantly focuses on glaucoma and certain fundus diseases post-intravitreal drug injection, collecting aqueous humor for relevant research to provide insights into ROP pathogenesis-related mechanisms.

Urine

Urinary VEGF levels are easily measurable and can be used for the early diagnosis of ROP. Studies by Levesque and others have found that low levels of VEGF in urine during the first month after birth are associated with the development of ROP.¹³⁴ Early research by Kwinta and subsequent studies by Levesque and Yenice found that infants with ROP requiring active intervention exhibited decreased urinary VEGF levels.^24,134,135

In addition to VEGF, other inflammatory markers such as IL-8 and IL-6 can also be detected in the urine of ROP patients. IL-8 plays a significant role in the recruitment of neutrophils in inflammation, cell adhesion, tumor growth, angiogenesis, neuronal protection, and brain development.⁴³ Studies have shown similar levels of IL-6, IL-8, and VEGF in the urine of both ROP and non-ROP groups, but there is a positive correlation between the levels of IL-8, IL-6, and VEGF in urine, indicating that the level of IL-6 increases with the levels of IL-8 and VEGF in urine, and the level of VEGF also increases with the level of IL-8 in urine.⁴⁵

Beyond VEGF and inflammatory markers, researchers have also explored the relationship between urine metabolites and ROP, including oxidative products like 8-hydroxy-2ʹ-deoxyguanosine (8-OHdG) and N-terminal pro B-type natriuretic peptide (NTproBNP). 8-OHdG, a sensitive biomarker of oxidative DNA damage,¹³⁶ has been found at significantly higher levels in the urine of ROP patients, suggesting its potential as a screening indicator for ROP.¹³⁷ NTproBNP, widely used to assess conditions such as heart failure and bronchopulmonary dysplasia,^138,139 was also found to be elevated in the first month after birth in preterm infants under 30 weeks of gestation, correlating with a higher risk of severe ROP.¹⁴⁰ Czernik et al identified a significant increase in the urine NTproBNP/creatinine ratio (UNBCR) in preterm infants with severe ROP, further underscoring its predictive value.¹⁴¹

However, urine collection in newborns presents challenges due to the difficulty in obtaining samples on demand. Typically, urine is collected using diapers or pads, which may lead to contamination and lower accuracy, limiting its utility as a diagnostic method.

Tears

Tears, a complex mixture of proteins, lipids, mucins, water, and salts, have recently been shown to contain 1526 proteins through proteomic analysis,¹⁴² making them less complex than serum or plasma. Tears are ideal for assessing biomarkers and molecular characteristics related to ocular vascular changes in both health and disease. Numerous studies have explored the relationship between ocular diseases and tear composition, including dry eye disease (DED), vernal conjunctivitis, diabetic retinopathy, Graves’ orbitopathy, ocular tumors, and glaucoma.^142–147

Vinekar et al analyzed pro-angiogenic factors in the tears of preterm infants with ROP, finding lower VEGF levels in those with ROP, especially in progressing cases.¹⁴⁸ Conversely, angiogenin levels were higher, and the ratio of angiogenin to birth weight, gestational age, and/or VEGF could serve as a potential non-invasive screening biomarker for ROP.¹⁴⁸ Magnani and others further confirmed that tear VEGF levels negatively correlate with the severity of ROP, with significantly lower levels in late-stage (stage 3) ROP patients compared to those in earlier stages or those with regressed or resolving ROP.¹¹¹ In contrast, Ang-1 and Ang-2 levels were positively correlated with ROP severity, with late-stage (stage 2) ROP patients showing significantly higher levels of these factors.¹¹¹ However, VEGF, Ang-1, and Ang-2 levels in tears alone are not reliable biomarkers for assessing ROP severity or treatment needs.¹¹¹

Additionally, studies have shown an upregulation of numerous inflammatory factors in tears, such as CCL2, RANTES-CCL5, and IL-8. Vinekar et al detected higher levels of RANTES-CCL5 and IL-8, along with lower levels of IL-6 and sL-selectin, in the tears of infants with ROP, though MCP1-CCL2 levels remained unchanged.¹⁴⁹ Baba et al found significantly elevated CCL2 levels in the tears of severe ROP patients, suggesting that tear CCL2 could serve as a useful biomarker for assessing ROP severity.¹⁵⁰ CCL2, a chemokine involved in monocyte and macrophage recruitment to inflammation sites, plays a role in the pathogenesis of various retinal vascular proliferative diseases.¹⁵¹

Tear protein concentrations are relatively high and can be easily obtained through minimally invasive methods.¹⁵² Tear proteomics has identified potential biomarkers for other ocular diseases, including growth factors and VEGF,^153–155 with functions spanning angiogenesis, immune or inflammatory responses, visual perception, and metabolism.¹⁵² Shipton and others validated the feasibility of using tear proteins to assess ROP risk, finding an increase in LDH-B chains in infants at higher risk for ROP and a rise in immunoglobulin concentrations with postmenstrual and postnatal age.¹⁰ Rathi and Patnaik further found that MMP2 and MMP-9 expression was significantly higher in the tears of severe ROP patients compared to those with mild ROP or without the disease.^3,59 These findings highlight the potential of tear MMPs as early predictive biomarkers for ROP, with MMP levels increasing with disease severity and showing no response to laser treatment, underscoring their diagnostic value.

Tear collection is relatively non-invasive and straightforward, most commonly done using sterile Schirmer strips (5 x 35 mm) to collect tears or conjunctival secretions. Other methods include glass capillaries, surgical sponges, glass rods and scrapers, and phenol red thread tear tests. Tears can be safely collected, reliably analyzed repeatedly, and associated with disease staging and progression, making them a valuable tool for predicting the onset and progression of ROP in preterm infants.

Comparison of Different Detection Methods and Factors

Comparison of Different Detection Methods

Research on ROP has revealed that various detection methods possess distinct advantages and disadvantages. The collection of vitreous and aqueous humor, while invasive and primarily used for laboratory research, offers a more direct insight into the retinal condition and aids in clarifying the pathogenesis of ROP. Sampling of amniotic fluid, umbilical cord blood, and placental tissue occurs relatively early, allowing for an assessment of the in utero environment, which can provide early predictions regarding the occurrence and progression of ROP. Peripheral blood and urine samples are easier to obtain; however, their collection typically occurs later and is susceptible to influence from a variety of systemic factors. Tears, on the other hand, can be collected safely at multiple time points, allowing for the dynamic monitoring of disease progression. This makes them an ideal source for identifying and validating molecular markers closely associated with ocular diseases, including ROP.¹⁵⁶

Comparison of Different Detection Factors

VEGF has emerged as the most extensively studied and modifiable factor in ROP. The pathogenesis of ROP is divided into two phases: the first phase involves the downregulation of VEGF levels due to high environmental oxygen, which inhibits retinal vascular growth. The second phase is marked by accelerated retinal development and metabolism, leading to relative retinal hypoxia. This hypoxia stimulates increased VEGF production in the avascular retina, promoting extensive retinal neovascularization.^20,157 Numerous studies have examined VEGF levels across various sample types in ROP patients, with most indicating elevated VEGF levels. Showing differences in severity in peripheral blood, ie, decreased VEGF in mild ROP patients and increased VEGF in severe ROP patients, consistent with the two development stages of ROP disease. Notably, VEGF levels in tears and urine are decreased in ROP patients. Lower tear VEGF levels may reflect systemic VEGF levels; Vinekar et al reported that infants with higher tear VEGF levels exhibited more systemic angiogenesis, likely as a response to prematurity-related inflammatory complications, independent of ROP status.¹⁴⁸

Insulin-like growth factor-1 (IGF-1) is a critical peptide hormone with structural similarity to insulin, playing a fundamental role in fetal and neonatal growth and development.¹⁵⁸ It promotes cell proliferation and differentiation, inhibits apoptosis, and has metabolic and immunomodulatory effects akin to insulin.¹⁵⁹ IGF-1 is considered a positive regulator of VEGF, effectively facilitating retinal neovascularization.¹⁶⁰ Recent studies have increasingly elucidated the relationship between IGF-1 and ROP, with multiple investigations into peripheral and umbilical cord blood suggesting that low IGF-1 levels are crucial in the onset and progression of ROP.

A wide range of inflammatory factors, including TNF-α, IL-6, IL-8, IL-18, IL-10, MCP-1, MIP-1α/β, and CCL2, have been found to be elevated in various sample types from ROP patients and are believed to interact synergistically with VEGF. TNF-α, a key cytokine in inflammation with paradoxical anti-inflammatory properties, has been shown to influence retinal angiogenesis. Its expression is upregulated in animal models of retinal neovascularization and human proliferative eye diseases.¹⁶¹ Blocking TNF-α can enhance physiological angiogenesis while reducing pathological neovascularization in mouse models of oxygen-induced retinopathy.^53,54 IL-6, known to promote pathological angiogenesis, including retinal angiogenesis, increases in diabetic retinopathy.^162–164 IL-8, another inflammatory cytokine, plays a crucial role in promoting angiogenesis and is elevated under conditions of retinal neovascularization.^162,163,165 However, some studies report conflicting results or no significant differences for these inflammatory markers. Additionally, certain studies have detected immune cells (PBMCs), complement factors (C3, C4), and angiopoietins (Ang-1 and Ang-2) as contributors to the development and progression of ROP.

Furthermore, specific factors demonstrate unique expression patterns in particular samples. For instance, reduced TGFBI levels in umbilical cord blood proteins are significantly associated with severe ROP. Increased MMP levels in amniotic fluid and tears are identified as risk factors for the onset and independent risk factors for the progression of ROP. Elevated opticin levels in the vitreous have been shown to prevent retinal neovascularization. Additionally, Col1a1 and 16K-PRL are significantly upregulated in the aqueous humor of ROP patients. Placental function and inflammation also exhibit a complex association with ROP, and significant elevations in 8-OHdG and NTproBNP have been observed in the urine of ROP patients.

Common influencing factors and their related mechanisms are shown in Table 1. The detection results of different factors across various, and even identical, sample types are largely consistent, though some contradictory findings may be attributed to differences in the stage of ROP at the time of sampling, variations in corrected gestational age, and the overall condition of the patients. Further research is needed to examine related factors in the same or different sample types at consistent stages to enhance our understanding of ROP pathogenesis and to provide more effective recommendations for early prevention and treatment in clinical practice.

Table 1 Predominant Cytokines Implicated in Retinopathy of Prematurity and Their Mechanistic Roles

Conclusion

Retinopathy of Prematurity (ROP), characterized by abnormal and uncontrolled vascular growth in the immature retina of preterm infants, remains a leading cause of potentially preventable blindness in children globally. Various detection methods for ROP, including peripheral blood, umbilical cord blood, amniotic fluid, placenta, vitreous humor, aqueous humor, urine, and tears, offer distinct advantages and limitations. Among these, the simplicity, safety, and speed of tear collection make it a particularly promising method for use in preterm infants. Tears can be safely and repeatedly collected, and they are associated with disease staging and progression. Previous studies have investigated relevant cytokines in the tears of premature infants with retinopathy, such as angiogenic factors (VEGF, Ang-1, and Ang-2), inflammatory factors (CCL2, RANTES-CCL5, IL-8), and MMPs (MMP-2 and MMP-9). It was found that the levels of MMPs (MMP-2 and MMP-9) increased with the severity of the disease. The establishment of tear protein expression profiles has also played a significant role in the study of ocular diseases, facilitating patient compliance and longitudinal monitoring. Identifying biomarkers that accurately predict the risk and severity of ROP is critical for many infants, their families, and the healthcare system. As neonatal medicine continues to evolve, the clinical and practical challenges associated with ROP are likely to grow, underscoring the need for further investigation into potential tear biomarkers for ROP. Such research could lead to more effective early detection and intervention strategies, ultimately reducing the incidence of ROP-related blindness.

Retinal vascular disorders such as Familial Exudative Vitreoretinopathy (FEVR), Incontinentia Pigmenti (IP), and Coats Disease severely impair infant vision by damaging retinal vasculature. Previous studies have identified disease-specific biomarkers, such as Okamoto et al¹⁶⁶ reported that the combination of V-shaped vascular notch, brushy vascular ends, and csAR serves as a biomarker for autosomal dominant FEVR (AD-FEVR) patients with pathogenic variants in the Norrin/β-catenin genes. Woffendin et al¹⁶⁷ demonstrated that marked skewing of X–inactivation patterns is a hallmark of IP. Zhang et al¹⁶⁸ proposed angiogenin as a potential biomarker for retinal vascular abnormalities, with VEGF and MCP-1 concentrations positively correlating with the severity of retinal exudation. These disorders, along with Retinopathy of Prematurity (ROP), share common features of retinal vascular dysfunction. The biomarkers discussed in this study may provide a cross-disease reference framework for early prediction and severity assessment of such retinal vascular abnormalities.