Management of corneal neovascularization: Current and emerging therapeutic approaches

Abstract

Corneal neovascularization (CoNV) is a sight-threatening condition affecting an estimated 1.4 million people per year, and the incidence is expected to rise. It is a complication of corneal pathological diseases such as infective keratitis, chemical burn, corneal limbal stem cell deficiency, mechanical trauma, and immunological rejection after keratoplasties. CoNV occurs due to a disequilibrium in proangiogenic and antiangiogenic mediators, involving a complex system of molecular interactions. Treatment of CoNV is challenging, and no therapy thus far has been curative. Anti-inflammatory agents such as corticosteroids are the mainstay of treatment due to their accessibility and well-studied safety profile. However, they have limited effectiveness and are unable to regress more mature neovascularization. With the advent of advanced imaging modalities and an expanding understanding of its pathogenesis, contemporary treatments targeting a wide array of molecular mechanisms and surgical options are gaining traction. This review aims to summarize evidence regarding conventional and emerging therapeutic options for CoNV.

Corneal neovascularization (CoNV) is a sight-threatening condition that occurs after pathological insults such as microbial keratitis, herpes simplex keratitis, chemical burn, corneal limbal stem cell deficiency (LSCD), mechanical trauma, and immunological rejection following corneal transplantation. It is characterized by disruption of the cornea’s normal avascularity from the invasion of abnormal blood vessels in the stroma [Fig. 1]. These atypical vessels have increased vascular permeability and can cause chronic corneal edema, lipid and protein exudation, and scar formation leading to disarrangement of the regularly arranged lamellar collagen in the cornea. The loss of transparency can lead to visual impairment and blindness in severe cases. With an estimated annual incidence of 1.4 million people, CoNV represents a significant public health concern.[1]

Figure 1.

Clinical manifestations and etiologies of corneal neovascularization. SJS = Steven–Johnson syndrome, TEN = toxic epidermal necrolysis

Unfortunately, despite its prevalence, no curative therapy is available for this condition. Various treatments including anti-inflammatory agents, immunomodulatory drugs, fine needle diathermy (FND), and laser photocoagulation have been proposed to mitigate CoNV. However, these treatments often have limited effectiveness and may be associated with adverse effects. More recently, the introduction of anti-vascular endothelial growth factor (anti-VEGF) drugs has provided new therapeutic options. There is also growing interest in combining different treatment modalities in a synergistic manner to enhance the management of CoNV. This review aims to provide an overview of the existing treatment options for CoNV and shed light on emerging therapeutic approaches for this debilitating condition.

Methods

A literature search of the PubMed database was performed to include publications from inception to 2023. The key search terms included but were not limited to “corneal neovascularization,” “CoNV,” “anti-VEGF,” “anti-inflammatory,” “immunosuppressant,” “tyrosine kinase inhibitors,” “laser photocoagulation,” “fine needle diathermy,” “photodynamic therapy,” “gene therapy,” as well as other relevant treatment modalities in the titles and abstracts. In addition, an independent search of current literature and a thorough screening of previous reviews were also conducted to maintain a comprehensive search and article inclusion. Both animal and human clinical studies were included in this review.

Pathogenesis

The human cornea is an optically important structure. Together with the tear film, they contribute to 70% of the eye’s total refractive power. It is made up of six distinct layers – the epithelium, Bowman’s membrane, stroma, Dua’s layers, Descemet’s membrane, and the endothelium.[2,3] The stroma contains type I and V collagen fibrils interwoven in a regular fashion, forming layers of stacked lamellae to create an optically transparent media for light to pass through. CoNV is postulated to occur via a disruption in corneal angiogenic privilege due to a disequilibrium in proangiogenic and antiangiogenic factors in the cornea.

Proangiogenic factors involved in the pathogenesis of CoNV include vascular endothelial growth factor (VEGF), platelet-derived growth factor, matrix metalloproteinase (MMP), and inflammatory cytokines such as interleukin-1 (IL-1) and interleukin-6 (IL-6).[4] An extensively studied angiogenic mediator is VEGF-A, which is a member of a family of proteins including VEGF-A, VEGF-B, VEGF-C, VEGF-D, virally encoded VEGF-E, and placental growth factor.[5] VEGF-A is secreted by a wide variety of cells including macrophages, T-cells, fibroblasts, retinal pigment epithelial cells, and corneal cells upon inflammation and insult.[6] VEGF-A plays a dominant role in regulating angiogenesis by interacting with the tyrosine kinase receptors, VEGF-R1 and VEGF-R2.[7] VEGF-C and VEGF-D are secreted by the macrophages, and they bind to VEGF-R3 to stimulate lymphangiogenesis. MMPs are Zn²⁺-dependent enzymes categorized as either secreted MMPs or membrane-type MMPs. The first transmembrane-containing MMP to be identified was MMP-14. It has been linked to several molecular mechanisms involved in the selective binding and cleaving of VEGF-R1, which induces corneal angiogenesis.[8] MMP-2 and MMP-9 contribute to the degradation of extracellular matrix components and the upregulation of VEGF and tumor growth factor-β.[9] IL-1 and IL-6 are proinflammatory cytokines which upregulate the expression of chemokines and growth factors that lead to neovascularization.[10,11]

Antiangiogenic factors include endostatin and endostatin analogs (neostatin, arrestin, canstatin, and tumstatin), plasminogen inhibitors, serine protease inhibitors (angiostatin and pigment epithelial-derived factor), and soluble VEGF receptors.[12] Angiostatin binds to surface proteins in the vascular endothelial cells to hinder migration and tubule formation.[13] Soluble VEGF receptors block the downstream effects of VEGF ligands by depleting available VEGF molecules and preventing their attachment to membrane-bound VEGF receptors. Soluble VEGF-R1 has a high affinity to VEGF-A and is postulated to mediate corneal avascularity during development.[14] The presence of corneal insult or inflammation disrupts the equilibrium by upregulating proangiogenic factors and downregulating antiangiogenic agents, leading to corneal hemangiogenesis and lymphangiogenesis.

LSCD, a consequence of a dysfunctional limbal stem cell (LSC) population, can also lead to CoNV. LSC has the unique capacity to differentiate into normal corneal epithelial cells. Any damage in LSC leads to cellular apoptosis followed by replacement with abnormal conjunctival epithelial cells. Corneal stem cells play a vital role in maintaining the integrity of corneal surface. When these stem cells are damaged or deficient, the corneal surface becomes compromised, leading to the loss of barrier function that normally prevents blood vessels from growing into the cornea.[15] In LSCD, there is often an increase in inflammatory mediators such as cytokines and growth factors. These mediators create a proinflammatory environment in the cornea, promoting the infiltration of immune cells, which, in turn, release angiogenic factors promoting CoNV.[16] LSCD occurs when there are alterations in corneal tissue homeostasis due to depletion of LSC (such as in chemical injuries and mechanical trauma), absence of LSC (congenital aniridia), or in proinflammatory states (Steven–Johnson syndrome or contact lens–related hypoxia).[17]

Therapeutic approaches

Several treatment modalities have been derived based on the understanding of CoNV pathogenesis. Broadly, they can be categorized based on their mechanism of action. These include anti-inflammatory agents, anti-VEGF agents, and surgical vessel occlusion with laser or diathermy. Contemporary and experimental therapies such as gene therapy and nano-based drug therapy show promise in managing CoNV, but lack clinical studies to support their safety and efficacy. Treatment modalities studied in the literature are summarized in Tables 1–3. Key therapeutic approaches are evaluated below.

Table 1.

Summary of anti-inflammatory and anti-VEGF agents studied for corneal neovascularization

Agent		Level of Evidence		Mechanism of Action		Efficacy and Remarks		Studies
Anti-Inflammatory
Corticosteroids		Animal studies Prospective Clinical Studies		Inhibit phospholipase A2 Inhibit proliferation and migration of vascular endothelial cells		Low-cost, highly accessible Limited regression of mature CoNV More effective if started early after CoNV develops		Animal: Hos et al., 2011,[120] Hoffart et al., 2010[22] Human: Cursiefen et al., 2001[20]
						Long-term use associated with side effects such as infection and glaucoma
NSAIDs		Animal Studies		Inhibition of COX-1/2 enzymes		Differential effectiveness; COX-2 selective inhibitors (celecoxib, rofecoxib) have limited inhibition of VEGF production		Animal: Cooper et al., 1980,[121] Pakneshan et al., 2008[122]
						Long-term use is associated with side effects such as corneal ulceration and melting
						Generally ineffective in treating CoNV
Cyclosporine A		Animal Studies Randomized Controlled Clinical Trial		Calcineurin inhibitor, limits T-Cell response Inhibition of VEGF- induced primary endothelial angiogenesis		Topical, subconjunctival, and systemic administration has shown to be effective in inhibiting CoNV in experimental models Inhibitory effect of topical cyclosporine A was found to be significantly more effective than bevacizumab in an immune-mediated rabbit CoNV model		Animal: Bucak et al., 2013,[26] Ulusoy et al., 2020[123] Human: Bock et al., 2014[27]
						Limited inhibition of CoNV in post-keratoplasty patients
Tacrolimus		Animal Studies		Immunosuppressor, calcineurin inhibitor that suppresses T-cell/B-cell activation and proliferation		Subconjunctival tacrolimus was shown to be more effective than anti-rat- VEGF in an alkali-burn rat model Topical application is limited by its poor corneal permeability		Animal: Chen et al., 2018[31]
Methotrexate		Animal Studies		Anti-metabolite; Mechanism not well understood Anti-inflammatory response through reduction of chemokine- induced VCAM		Topical and subconjunctival administration effective in suture-induced CoNV rabbit models with associated reduction of VEGF and IL-6		Animal: Byun et al., 2011[124]
				Anti-angiogenesis associated with inhibition of VEGF and bFGF, similar to thalidomide
Thalidomide/Thalidomide Analogue (CC-3052)		Animal Studies		Inhibition of VEGF, FGF-2, and TNF-α		Effective inhibition of CoNV in experimental models, however, special caution should be exercised regarding dosage safety and efficacy		Animal: Kruse et al., 1998,[125] Abbas et al., 2002,[126] Lee et al., 2013[127]
						Side effects appear to be lower with thalidomide analogues (teratogenicity)
Interleukin Receptor Inhibitors (IL-1ra, IL-17 mAb, tocilizumab)		Animal Studies		Inhibition of pro- angiogenic activity of Interleukins Promote angiogenesis either directly – activation of endothelial cells, or indirectly – macrophage polarization and VEGF secretion		IL-1ra was associated with significant inhibition of CoNV in early initiation of therapy after induction of CoNV in a murine model Efficacy of CoNV inhibition in subconjunctival tocilizumab was comparable to that of bevacizumab in a suture- induced rabbit model		Animal: Dana et al., 1998,[128] Yoo et al., 2014[35]
TNF-α Inhibitors (Infliximab, Enterecept, adalimumab)		Animal Studies		Inhibition of TNF-α, a pro- inflammatory cytokine responsible for corneal lymphangiogenesis through the induction of VEGF-C production by macrophages TNF-α is also responsible for the expression of angiogenic factors		Experimental models displayed the inhibitory effects of TNF-α Inhibitors on CoNV Effectiveness of subconjunctival etanercept is enhanced when used in conjunction with bevacizumab in a rat silver nitrate cauterization model Subconjunctival adalimumab more effectively reduced VEGF expression than bevacizumab in a rat silver/potassium nitrate cauterization model		Animal: Ozdemir et al., 2013,[34] Özkaya et al., 2023[129]
Tyrosine Kinase Inhibitors (Sorafenib, Semaxanib, Rivoceranib, Regorafenib, Lapatinib, Axitinib, Dovitinib, Dasatinib, Sunitinib, Vatalanib, Nilotinib, Pazopanib, ZK261991, AG 1296)		Animal Studies Prospective Clinical Studies		Multitarget/Selective TKI in signaling pathways		Effectiveness comparable to corticosteroids and anti- VEGF Requires further evaluation of therapeutic effect and long-term safety in clinical studies		Animal: Onder et al., 2014[36] Human: Amparo et al., 2013[130]
Transforming growth factor-β-activated kinase 1 (5Z-7-oxozeaenol)		Animal Studies		Inhibits cell proliferation via suppression of cell cycle and DNA replication Inhibition of TNFα- mediated NFκB signaling		Topical administration of 5Z-7-oxozeaenol suppresses corneal neovascularization in a mouse model		Animal: Wang et al., 2022[131]
						Effect is enhanced when gelatin- nanoparticles- encapsulated 5Z-7- oxozeaenol are administered, likely due to the extended retention of the drug in the cornea
Anti-VEGF
Bevacizumab		Animal Studies		Humanized anti-VEGF-A monoclonal antibody		Prolonged topical use associated with loss of integrity of corneal epithelium		Animal: Manzano et al., 2007[43]
		Randomized Controlled Clinical Trials		Large molecular weight and size may impede corneal penetrance
								Human:
						Pooled meta-analysis suggests significant inhibitory effects of topical and subconjunctival bevacizumab on CoNV in both human and animal studies		Dastjerdi et al., 2010,[132] Petsoglou et al., 2013,[133] Papathanassiou et al., 2013[48]
		Meta-analysis
Ranibizumab		Animal Studies Prospective Clinical Studies		Recombinant humanized anti-VEGF-A monoclonal antibody Lower molecular weight and size allows greater drug penetration than bevacizumab		Effectiveness in inhibiting CoNV in comparison to bevacizumab remains inconclusive, but ranibizumab appears to induce neovascular reduction earlier than bevacizumab in animal models		Animal: Kim et al., 2014,[134] Liarakos et al., 2014[51] Human: Kim et al., 2013,[52] Ferrari et al., 2013[49]
						Topical ranibizumab was found to inhibit CoNV mostly through the reduction of vessel caliber rather than invasion area Comparative inhibitory effect of ranibizumab and bevacizumab in clinical studies remains to be further validated
Pegaptanib		Animal Studies		Oligonucleotide aptamer that selectively binds VEGF-165		Inconclusive treatment efficacy in both topical and subconjunctival administration in experimental animal models No inhibitory effect on corneal neovascularization was noted in a rabbit model when compared to topical prednisolone		Animal:
								Akar et al., 2013,[39] Andrade et al., 2021[53]
Aflibercept		Animal Studies Prospective Clinical Studies		Humanized recombinant fusion protein that binds VEGF and PlGF		Inconclusive treatment efficacy in both topical and subconjunctival administration in comparison to other anti- VEGFs in animal models No significant improvement of corneal neovascularization following subconjunctival administration of aflibercept in six patients		Animal: Oliveira et al., 2010,[135] Eski et al. 2022[136] Human: Sella et al., 2021[58]
Conbercept		Animal Studies Case Series Studies		Humanized recombinant fusion protein that binds specifically to VEGF-B, PlGF and various isoforms of VEGF-A		Topical, subconjunctival, and intrastromal conbercept was found to have statistically significant effect on CoNV regression in rabbit models		Animal: Liu et al., 2020,[137] Du et al., 2023[138] Human: Sun et al., 2023[139]
						Effectiveness of CoNV inhibition has yet to be compared to alternative anti-VEGFs
Laser Photocoagulation		Animal Studies Prospective Clinical Studies		Heat-induced photocoagulation of corneal vasculature		More commonly used in chronic CoNV, in part due to the limited efficacy of medical therapies on more established corneal vasculature Repeated treatments are often required due to recanalization of corneal vessels May cause damage of surrounding tissues		Animal: Cherry et al., 1976[140] Human: Gordon et al., 2002,[141] Kumar et al., 2016[77]
Photodynamic Therapy		Animal Studies Prospective Clinical Studies		Excitation of photosensitizers to occlude corneal vessels through thrombus formation		Relatively selective for corneal neovascular vessels with little damage to surrounding tissue Repeated treatments are often required for extensive disease High costs		Animal: Primbs et al., 1998[142] Human: Al-Torbak et al., 2012,[143] Yoon et al., 2007[144]
Fine Needle Diathermy		Animal Studies Prospective Clinical Studies		Coagulating current used to occlude corneal vessels		More effective than topical dexamethasone in mature CoNV and lymphatics regression in a sutured- induced CoNV mouse model Corneal vascularization secondary to thermal cautery may be reduced with combined anti-VEGF		Animal: Le et al., 2018[85] Human: Romano et al., 2016,[145] Pillai et al., 2000[80]

Legend: CoNV, Corneal Neovascularization; NSAID, Non-Steroidal Anti-Inflammatory Drug; COX; Cyclooxygenase, VEGF, Vascular Endothelial Growth Factor; IL, Interleukin; VCAM, Vascular Cell Adhesion Molecule; FGF, Fibroblast Growth Factor; TNF, Tumor Necrosis Factor; IL-1ra, Interleukin-1 Receptor Antagonist; IL-17 mAb, Interleukin-17 Monoclonal Antibody; TKI, Tyrosine Kinase Inhibitors; PlGF, Placental Growth Factor; Nuclear Factor-κB, NFκB

Anti-inflammatory agents

Inflammation is a key driver of corneal angiogenesis and is mediated by the release of proangiogenic cytokines such as IL-1, tumor necrosis factor-α (TNF-α), and VEGF.[18]

Corticosteroids are potent inhibitors of inflammation and remain widely accepted as first-line therapy for CoNV. By inhibiting vascular endothelial cells, steroids may exert a direct antiangiogenic effect through inhibition of VEGF expression from the vascular smooth muscle cells.[19] They also inhibit the migration of inflammatory cells, such as macrophages and lymphocytes, into the cornea. Inflammatory cells release angiogenic factors that promote neovascularization. By limiting the infiltration of these cells, steroids indirectly inhibit the growth of new blood vessels. Furthermore, steroids stabilize newly formed blood vessels, making them less fragile and less prone to leakage. This stabilization effect can prevent further expansion of the abnormal corneal vascular complex. However, their antiangiogenic effects are weak and insufficient to induce regression of long-term, mature CoNVs.[20]

The efficacy of steroids in suppressing angiogenesis is maximized if they are started immediately following corneal insult. Delays in administration may render them ineffective against subsequent development of CoNV.[21] In a chemical cauterization rat model, early and immediate application of dexamethasone resulted in a significant reduction of new vessel size compared to the control group.[22] Roshandel et al. have suggested prompt and aggressive treatment of CoNV associated with inflammatory conditions such as herpes simplex virus, immune stromal keratitis, and corneal graft rejection with early, frequent corticosteroids, before tapering them slowly based on clinical response.[23] However, the optimal duration of treatment remains controversial. Cursiefen et al.[20] showed that there was no difference in CoNV extent between post-keratoplasty patients who received either 6 months or 12 months of topical steroid therapy. Long-term corticosteroid usage is also associated with ocular and systemic side effects.[24] Despite this, corticosteroids remain the mainstay of treatment due to their low cost, accessibility, and anti-inflammatory properties.

Cyclosporine is an immunomodulatory agent which inhibits the migration of primary endothelial cells. The mechanism behind its antiangiogenic effects are not well characterized, but is thought to be related to an overexpression of HESR1 transcription factor and a downregulation of VEGF-R2.[25] The role of cyclosporine in the treatment of CoNV remains controversial as conflicting evidence on its efficacy have been reported. In one study, topical cyclosporine A 0.05% was found to be more effective than topical bevacizumab 0.5% (24.4% vs. 37.1%, P = 0.03) but less effective than topical dexamethasone 0.1% (24.4% vs. 5.9%, P = 0.02) in reducing the extent of neovascularization in a rabbit cornea model.[26] On the contrary, a randomized, prospective, multicenter clinical trial did not find any significant difference in the incidence of CoNV following high-risk keratoplasty among patients receiving either 5.13 or 7.7 mg subconjunctival cyclosporine A implant when compared to placebo.[27] This was postulated to be related to the strong proangiogenic stimulus after high-risk transplantation and poor drug penetration into the anterior chamber following local administration of cyclosporine.[28]

Tacrolimus is a calcineurin inhibitor which decreases T-cell activation and inflammatory cytokine release. It has a higher potency than cyclosporine, exhibiting similar effects at concentrations 100 times lower.[29] It reduces the production of angiogenic factors such as VEGF, epidermal growth factor, platelet-derived growth factor, prostaglandin E2, TNF-α, MMP-9, MMP-13, IL-1, IL-6, and hypoxia-inducible factor.[30] In a corneal alkali burn model, topical tacrolimus 0.05% administered four times daily was more effective than subconjunctival anti-rat-VEGF (0.5 mg/0.05 ml) in decreasing CoNV after 1 month of therapy.[31] Corneal epithelial defects and opacities were also significantly reduced by topical tacrolimus. However, the efficacy of topically instilled tacrolimus is limited by poor drug penetration across the cornea.[32] Cationic liposomal tacrolimus has been reported to be effective in circumventing this problem. With a surface potential of approximately +30 mV, this formulation binds to the negatively charged mucin layer of the ocular surface to increase the concentration of tacrolimus in the anterior chamber of mice by 20-fold at 90 min post-instillation.[32] Furthermore, this new drug delivery vehicle was more effective than topical tacrolimus in inhibiting CoNV, reducing corneal inflammation, and shortening the duration required for corneal epithelial recovery. However, adherence of topical tacrolimus may be an issue. Some patients may experience temporary discomfort, burning, or stinging in the eyes after applying tacrolimus ointment. Tacrolimus ointment might exacerbate dry eye symptoms in some individuals. Prolonged use of topical tacrolimus may also increase the risk of ocular infection because of local immunosuppression.[33]

Immunomodulatory biologic therapies including IL-6 receptor antagonist, tyrosine kinase inhibitors, and TNF-α inhibitors have been shown in experimental studies to be effective in treating CoNV.[34,35,36] Subconjunctival tocilizumab, an IL-6 receptor antagonist, was as effective as subconjunctival bevacizumab in reducing the extent of CoNV.[35] Subconjunctival sunitinib, a tyrosine kinase inhibitor, reduced neovascularization area in experimental rat corneas at a level comparable to subconjunctival bevacizumab.[37] Subconjunctival etanercept, a recombinant TNF receptor which antagonizes both TNF-α and TNF-β, demonstrated anti-inflammatory and antiangiogenic effects in an animal model, and the effects were further enhanced when given in conjunction with subconjunctival bevacizumab.[34] However, further human clinical studies are required to ascertain their clinical safety and utility.

Anti-vascular endothelial growth factor

Anti-VEGF agents such as pegaptanib, bevacizumab, ranibizumab, and aflibercept have been administered off-label, via topical, subconjunctival, and intrastromal routes, to treat CoNV.[38,39,40,41] Among the anti-VEGF agents, bevacizumab is most commonly used to treat neovascular age-related macular degeneration and diabetic macular edema due to its widespread availability and cost-effectiveness. Bevacizumab is a large, full-length humanized monoclonal antibody that binds to all isoforms of VEGF-A. Owing to its large molecular weight of 149 kD, it is typically unable to penetrate the cornea. However, in experimental animal models, the pathologically vascularized stroma permits penetration of large immunoglobulin molecules and increases the viability of topical administration.[42] In a rat model of chemically induced CoNV, topical bevacizumab 4 mg/ml drops administered twice daily for 1 week was demonstrably more effective than placebo in reducing the area of CoNV (38.2% vs. 63.5%, P < 0.02).[43] A prospective clinical trial of short-term topical bevacizumab 10 mg/ml eyedrops administered two to four times daily for 3 weeks successfully reduced the area of neovascularization (47.5%, P < 0.001) and diameter of neovascular vessels (36.2%, P = 0.003) at 6 months.[44] Although no adverse ophthalmic or systemic events were reported, this study excluded patients with poor corneal epithelial integrity (either existing epithelial defects or previous corneal surgeries) and utilized a shorter treatment course with a lower concentration of topical bevacizumab, compared to other studies. In contrast, separate studies have reported corneal epitheliopathy, approaching 60% of the eyes which received topical bevacizumab 12.5 mg/ml over the course of 3 months, and development of epithelial defects in 16.7% of eyes after administration of topical bevacizumab 5 mg/ml for up to 1 year of therapy.[45,46] Both studies reported improvements in CoNV parameters. Despite improvements in neovascularization, there was no statistically significant improvement in visual acuity.[43,45,46]

Subconjunctival bevacizumab (2.5 mg/0.1 ml) administered monthly for up to 3 months in patients with CoNV has been demonstrated to reduce the extent and density of neovascular vessels (6.0 ± 1.2 clock hours before and 4.6 ± 1.0 clock hours after bevacizumab injection, P = 0.008).[47] A meta-analysis of seven clinical trials of subconjunctival bevacizumab showed a pooled reduction of 32% (95% confidence interval [CI]: 10%–54%) in the neovascularization area.[48] The pooled mean change in best corrected visual acuity in three eligible studies demonstrated a mean improvement of 0.04 log of minimum angle of resolution (logMAR; 95% CI: -0.01–0.09), but this improvement was not statistically significant. The analyses were limited by substantial heterogeneity among the studies (studies for neovascularization area: I² = 92.8%, best correct visual acuity: I² = 73%). Interestingly, this meta-analysis demonstrated that the efficacy of subconjunctival bevacizumab was comparable to that of topical bevacizumab, although this was limited by the small population and large heterogeneity between the studies.

Ranibizumab is a recombinant humanized monoclonal antibody that is affinity-matured to optimize VEGF-A binding potential. It has a smaller molecular weight of 48 kD, which theoretically allows better penetration of the cornea than bevacizumab. Topical administration of ranibizumab 10 mg/ml four times daily for 3 weeks has been reported to reduce the neovascularization extent (55.3%, P < 0.001) and the diameter of neovascular vessels (59%, P < 0.001) at 4 months.[49] Similar to studies on topical bevacizumab, there were no statistically significant changes in invasion area or visual acuity. No adverse ophthalmic or systemic events were reported, though patients with previous ocular surgery or corneal epithelial defects were excluded from the study. Comparison of the bevacizumab and ranibizumab studies with similar study population characteristics by Cheng et al. and Ferrari et al. demonstrated earlier improvements in vessel caliber and neovascularization area in the topical ranibizumab group, likely attributable to the smaller molecular size.[50] However, there were no statistically significant differences in neovascularization area, vessel caliber, or invasion area between the two medications at the study endpoint.

Subconjunctival and intrastromal administration of ranibizumab has also been investigated as a potential therapeutic option. In rabbit cornea models, subconjunctival ranibizumab (1 mg/0.1 ml) administered 1 h after alkali burns resulted in smaller vascularization areas at 2 weeks (control: 22.63 ± 4.03% vs. treated: 6.13 ± 4.22%, P = 0.001).[51] VEGF levels were also significantly lower in all the sampled anterior segment tissues including the cornea, iris, aqueous humor, and conjunctiva. The extent of cornea scarring area was, however, similar to the ranibizumab-treated and control corneas. A clinical randomized prospective study of 16 patients with chronic CoNV demonstrated that combined subconjunctival and intrastromal bevacizumab (2.5 mg/0.1 ml) and ranibizumab (1 mg/0.1 ml) both reduced the neovascularization area at 1 month.[52] Notably, bevacizumab resulted in greater reduction in neovascularization area compared to ranibizumab (28.4 ± 9.01% vs. 4.51 ± 11.64%, P = 0.001). Further prospective randomized controlled trials with larger study populations will be required to provide evidence on the long-term efficacy of bevacizumab versus ranibizumab.

Pegaptanib is an oligonucleotide aptamer that selectively binds to VEGF-165. It has shown variable efficacy in treating CoNV in animal models. Topical 0.5% and 1.0% pegaptanib sodium diluted in 15 ml of 0.5% carboxymethylcellulose sodium were both less effective than topical 1.0% prednisolone acetate in reducing neovascularization area, total vascular length, and number of neovascular vessels in a rabbit alkali burn model.[53] Another experimental rat model investigating the efficacy of subconjunctival bevacizumab, ranibizumab, and pegaptanib found that all three agents reduced the area of neovascularization, but bevacizumab produced a larger area of reduction compared to ranibizumab and pegaptanib.[39]

Aflibercept is a soluble fusion protein of VEGF receptors 1 and 2.[54] It is a relatively newer anti-VEGF compared to bevacizumab and ranibizumab, with a wider range of target molecules including VEGF-A, VEGF-B, and placental growth factor.[55] While an animal study showed that topical aflibercept 2 mg/0.5 ml and 2 mg/5 ml were effective in reducing neovascularization surface area, the effect was comparable to topical bevacizumab 2.5 mg/1 ml.[40] In another rat model of chemically induced CoNV, subconjunctival administration of bevacizumab (1.25 mg/0.05 ml) or ranibizumab (0.5 mg/0.05 ml) or aflibercept (1.25 mg/0.05 ml) was effective in inhibiting CoNV, inflammation, fibroblast activity, and reduced VEGF levels in neovascularized arteries.[56] No significant differences in efficacy were noted among the three medications. A comparison of topical aflibercept and topical bevacizumab in corneal chemical burn rat models demonstrated smaller areas of neovascularization at day 7 and day 10 in rat corneas receiving topical aflibercept compared to topical bevacizumab (aflibercept: 21.73 ± 14.59% on day 7 and 31.0 ± 23.61% on day 10 vs. bevacizumab: 51.27 ± 15.50% on day 7 and 54.4 ± 11.33% on day 10, P < 0.001).[57] Clinically, a single subconjunctival injection with 0.08 ml of aflibercept (25 mg/ml) in six eyes with CoNV did not demonstrate significant changes in the extent, density, or centricity of corneal blood vessels up to 3 months after treatment.[58] Further studies are required to ascertain the effectiveness of pegaptanib and aflibercept in treating CoNV.

Intrastromal anti-VEGF therapy was proposed as a more effective drug administration route for CoNV due to the delayed drug elimination from the avascular cornea. Multiple case series and clinical studies have demonstrated the efficacy of intrastromal injections in regression of neovascularization and sustained reduction of VEGF levels.[59,60,61,62] In a suture-induced rabbit cornea neovascularization model, intrastromal bevacizumab (1.25 mg/0.05 ml) and intrastromal aflibercept (2 mg/0.05 ml) reduced CoNV more effectively at 2 weeks (88.1% and 82.5%, respectively) compared to subconjunctival bevacizumab (1.25 mg/0.05 ml) or subconjunctival aflibercept (2 mg/0.05 ml) (64.5% and 69.9%, respectively).[41] However, intrastromal injections may be associated with procedural complications such as Descemet’s membrane perforation or detachment.

Complications associated with topical anti-VEGF therapy include corneal wound healing impairment that manifests as persistent epithelial defects and corneal stromal thinning.[45,46,63] A theoretical side effect is neurotrophic keratopathy due to inhibition of VEGF-mediated neural growth.[64] However, this has yet to be observed clinically. Risk factors for development of corneal epitheliopathy include the use of high concentrations of bevacizumab (>10 mg/ml), extended duration of therapy, pre-existing corneal epithelial defects, and previous corneal surgeries. No systemic complications of topical anti-VEGF, such as cerebrovascular accidents and thromboembolic events, have been reported in the limited number of clinical studies published to date. Nonetheless, it is prudent to take precautions to reduce systemic absorption of the drug, such as performing punctal occlusion or applying punctal plugs before topical administration.[50] Subconjunctival injection of anti-VEGF is safe and well tolerated, with no significant side effects reported in human or animal studies. Common adverse events include subconjunctival hemorrhage (28%), pain (31%), anterior chamber reaction (8%), and irritation (3%).[48,65,66,67]

Although current studies show promise in treating CoNV, anti-VEGF agents only produce a partial and temporary reduction in neovascularization, with most effects occurring within a month of commencing therapy with progressive waning over the subsequent 2 months.[68] Neovascularization may also recur after successful treatment with subconjunctival bevacizumab, necessitating repeated injections.[69] While there are anecdotal reports of subconjunctival bevacizumab successfully treating CoNV, these reports are rare and confounded by the concomitant or preceding administration of other therapeutics such as corticosteroids.[70,71,72] In addition, there is a paucity of long-term clinical outcomes, with most studies reporting follow-up of up to 3 months. The advent of new anti-VEGF drugs, such as brolucizumab and faricimab, further expands the armamentarium of therapies available. Therefore, further large-scale studies and long-term evaluation of anti-VEGF agents are required to ascertain their effectiveness and safety in treating CoNV.

Physical ablation of neovascular vessels

Medical therapy may have limited effectiveness in treating chronic CoNV with mature and well-established vascularization. In such cases, physical occlusion of neovascular vessels can be considered with laser photocoagulation, photodynamic therapy, or FND.

Laser photocoagulation using argon, yellow dye, and Nd:YAG neodymium:yttrium-aluminium-garnet laser has been shown to be effective in treating CoNV in animal[73,74,75] and clinical studies.[76,77] Although laser photocoagulation can obliterate efferent neovascular vessels, as they are wider and have relatively slower blood flow, it is often unable to cauterize narrower and deeper afferent vessels with faster blood flow. A single prospective study of 40 patients with CoNV receiving double-frequency Nd:YAG laser photocoagulation demonstrated improvements in the mean area of CoNV from 31.9% to 17.6% after 3 months of laser treatment.[77] Of 185 corneal vessels, 53.5% were completely occluded, 9.2% were partially occluded, and 37.3% were recanalized at the end of 3 months. Hence, multiple laser photocoagulation sessions may be required for patients with extensive CoNV in view of the possibility of recanalization. Side effects associated with laser photocoagulation include corneal endothelium or crystalline lens damage, corneal hemorrhage, corneal thinning, and iris atrophy.[77]

Photodynamic therapy achieves vessel occlusion via reactive oxygen species generation, which arise from the interaction between light energy and photosensitizers. In the treatment of CoNV, photodynamic therapy with intravenous or intrastromal verteporfin was effective in inducing angiogenic and lymphogenic regression in a rodent suture-induced CoNV model.[78,79] Long-term allograft survival was also improved significantly in photodynamic therapy-treated eyes when compared to controls.[78] A potential advantage of using photodynamic therapy is its selectivity for neovascular vessels, allowing it to induce vascular regression with minimal damage to surrounding tissues.

FND involves the use of high-frequency electrical currents to generate heat. The tip of the fine needle is heated using diathermy and subsequently applied in close proximity to the abnormal blood vessels. The heat generated coagulates the blood vessels and stops blood flow, leading to their closure.[80] Subsequent studies demonstrated that occlusion of corneal vessels using FND among pre-keratoplasty patients and patients with lipid keratopathy was successful in reducing lipid deposition in 82.3% of treated eyes, with 84.6% of high-risk corneal grafts surviving beyond 1 year.[81] Unfortunately, corneal diathermy may induce further CoNV due to thermal damage. In a mouse suture-induced CoNV model, FND monotherapy inhibited angiogenesis at early time points, but was associated with secondary corneal angiogenesis, with elevated proangiogenic factors such as VEGF-A, VEGF-C, and VEGF-D.[82] Hence, combination therapy with anti-VEGF agents has been proposed to complement its usage. In the same study, addition of anti-VEGF agents with FND significantly regressed corneal blood and lymphatic vessels at 1 week and prevented the undesirable effect of angiogenesis observed in the monotherapy group.[82] Clinically, FND combined with subconjunctival bevacizumab treatment before penetrating keratoplasty (PK) in patients with at least one quadrant of CoNV resulted in rejection-free graft survival rates of more than 90% after 1 year and 78% after 3 years.[83] A retrospective review of children who received FND with subconjunctival bevacizumab reported complete CoNV resolution in 88.9% within 1 month of treatment and an improvement in mean corrected distance visual acuity from 0.66 ± 0.31 to 0.50 ± 0.37 logMAR (P = 0.02).[84] Combination of FND and corticosteroids has also been shown to regress both blood vessels and lymphatic vessels in animal models.[85] However, no clinical studies are available to support the efficacy of this combination. Another strategy to reduce excessive thermal damage is to selectively cauterize afferent arterioles, which comprise < 1% of the neovascular vessels.[86] Spiteri et al. identified afferent neovascular vessels with fluorescein and indocyanine green corneal angiography and performed selective FND. They showed that the number of afferent vessels ranged between one and three and that the vessels had a mean diameter of 40 µm. After selective FND, the mean area of neovascularization was reduced by 1.80 ± 1.40 mm² (P < 0.01) up to 12 weeks postoperatively, demonstrating the viability of this protocol in reducing CoNV while minimizing thermal damage.[87]

LSC transplantation

LSCD is an ocular surface disease caused by abnormal LSCs and can lead to CoNV. Management is challenging and medical therapy is only effective for early or partial LSCD.[88] Surgical management of LSCD can be categorized into allograft, autograft, and cultivated LSC transplantation. LSC grafting techniques include autologous limbal transplant, allogenic limbal transplant, keratolimbal allograft, and simple limbal epithelial transplantation (SLET). A meta-analysis of 40 studies with 2202 eyes demonstrated that autologous limbal transplantations were associated with greater visual acuity improvement (76%) than allogenic limbal transplants (52.3%), cultivated autologous LSC transplants (56.4%), and cultivated allogenic LSC transplants (43.3%) (P < 0.001).[89] SLET utilizes a small limbal tissue autograft that is transplanted from the healthy eye to the diseased eye.[90] This is accomplished by dividing the strip of donor limbal tissue into small pieces and distributing them over an amniotic membrane placed over the diseased cornea. The outcomes of SLET have been favorable and are likely comparable to those of cultivated LSCs – 83% of reported SLET operations in the literature restored the defective corneal epithelium and 69% resulted in improvements in visual acuity.[91] Common complications reported in these studies following SLET include focal recurrence of LSCD, progression of conjunctivalization, progressive symblepharon, and keratitis.[91] No large-scale head-to-head comparison studies of SLET and other LSC transplantation techniques have been performed.

Mitomycin intravascular chemoembolization

Mitomycin C (MMC) intravascular chemoembolization (MICE) is an emerging technique used to treat CoNV and lipid keratopathy. MMC inhibits vascular endothelial cell proliferation and hinders neovascular vessel growth and repair.[92] Mimouni and Ouano described a case series of three patients who received 0.05 ml of MMC (0.4 mg/ml) injected directly into a corneal neovascular vessel.[93] All three patients experienced a regression of neovascularization and lipid keratopathy over a follow-up period of 1 year, with no intraoperative or postoperative complications identified. To avoid possible unwanted loss of limbal vascularization, MMC should not be injected in the direction of the limbus. Another case series reported success in inducing regression of neovascularization in two cases of MICE performed in patients with failed PK.[94] Further clinical studies are warranted to elucidate the long-term outcomes and safety of this treatment.

Corneal cross linking

Corneal cross-linking (CXL) is a procedure that utilizes ultraviolet-A (UVA) and riboflavin (vitamin B2) to treat corneal ectasias.[95] CXL was shown in an in vivo rat model to be able to temporarily suppress suture-induced corneal hemangiogenesis and lymphangiogenesis.[96] This is postulated to occur via the combined effects of UVA and riboflavin – UVA irradiation has been suggested to produce an anti-lymphangiogenic effect at the cellular level, while riboflavin reduces the lipopolysaccharide-induced synthesis of inflammatory cytokines TNF-α, IL-1, and IL-6.[97,98] UVA activation of riboflavin further releases reactive oxygen species, which is theorized to reduce neovascular vessel formation.[99] Further short-term observations of corneal parameters in animal studies displayed a partial effectiveness in treating CoNV. Xu et al. studied 60 rabbits with alkali burn-induced CoNV, which were divided into six groups.[100] Group A (control) was not given any treatment, while groups B, C, and D received 30, 15, and 45 min of CXL immediately after injury, respectively. Groups E and F received CXL for 30 min on day 1 and 3 after the injury, respectively. CXL was performed with riboflavin 0.1% drops dissolved in dextran 20% applied every 2–5 min for 30 min, and with continual administration during UVA irradiation once every 3 min. Irradiation with UVA was performed at an irradiance of 3 mW/cm² and a surface dose of 5.4 J/cm². Over the course of 2 weeks, the extent of CoNV increased in a time-dependent manner across all groups. All corneas treated with CXL displayed less neovascularization than the control group, with the most significant reduction observed in corneas receiving 45 min of UVA irradiation. In a separate rabbit alkali burn model, comparison of CXL using Dresden protocol with matrix-regenerating agent also showed improvement in epithelial loss, stromal edema, corneal vascularization, and leukocytic infiltration in both groups.[101] CXL was, however, not able to reduce corneal opacification from alkali burns in animal models.[101,102,103] In a retrospective case series involving five patients, CXL was performed before or in combination with high-risk PK in patients with thin or fragile corneas. While these results may be affected by the accompanying PK, reduction of CoNV was observed in all patients (mean 70.5% ± 22.7%) and revascularization was not observed over a mean follow-up of 16.4 weeks.[104] CXL remains an experimental treatment for CoNV, and additional studies relating to the optimal riboflavin dose and irradiation regimen are required.

Gene therapy

Gene therapy refers to the transfer of genetic material to facilitate genomic augmentation, suppression, or repair, with the aim of providing durable, long-term therapeutic expression or suppression of the edited gene.[105] This attempts to overcome the drawback of conventional pharmacological approaches, where the effects are temporary and require repeated treatment. Gene therapy has been studied to treat various anterior segment diseases such as dry eye disease, corneal and conjunctival scarring, epithelial wound healing, and improve corneal graft survival.[106]

Multiple experimental animal models targeting the angiogenic VEGF pathway via gene transfer of VEGF receptors, Flt-1 and Flk-1, have shown variable success in regressing CoNV. Gene transfer of Flt-1 using recombinant adeno-associated virus and non-viral micellar nano vectors, as well as recombinant adenovirus-mediated delivery of soluble Flk-1 produced favorable inhibition of CoNV.[107,108,109] Subconjunctival administration of lipoplexes carrying the VEGF and roundabout4 transcription factor, GA-binding protein gene, produced temporary improvement on CoNV.[110] Downregulation of VEGF via recombinant adenovirus-driven antisense VEGF RNA and nanoparticle-mediated delivery of shRNA VEGF-A plasmids reduced CoNV.[111,112] Techniques such as adeno-associated virus and equine infectious anemia virus-based gene therapies to upregulate the antiangiogenic factors endostatin and angiostatin to inhibit VEGF and FGF-2 were effective in suppressing CoNV, immune cell infiltration, and corneal opacification.[113,114] More recently, downregulation of MMP-9 using lipid nanoparticles and the use of cholesterol-modified small interfering RNA targeting stromal cell-derived factor-1 have shown promise in treating CoNV.[115,116] Ocular delivery of small interfering RNA via pH-sensitive vehicles had comparable efficacy to that of ranibizumab.[117] The only Phase III clinical gene therapy study is that of aganirsen, an antisense oligonucleotide preventing insulin receptor substrate-1 expression, which has been shown to inhibit CoNV and reduce the need for transplantation in patients with keratitis.[118] Aganirsen has since been granted an Orphan Drug Designation by the US Food and Drug Administration for the prevention of corneal graft rejection in 2016.

Despite advancements, more research is required to address the challenges associated with gene therapy in CoNV. Gene therapy has very limited clinical evidence available regarding its efficacy and side effects. The prominent concern relates to the risk of developing host immune responses that may lead to the generation of neutralizing antibodies against vectors carrying the therapeutic gene, which may reduce therapeutic efficacy. Another concern is the development of unwanted, “off-target” effects both locally and systemically. However, these side effects theoretically carry a lower risk in the case of corneal diseases, due to its inherent immune privilege and avascular environment. Finally, the high cost of research and development associated with gene therapy represents a major hurdle before preclinical studies can progress to clinical trials.

Nanoparticle-based therapy

Nanomedicine and nano-based drug delivery therapy is a new but rapidly developing field, where materials of the nanoscale are employed to enhance drug delivery to provide better therapeutic efficacy. The potential functions of nanoparticles in the treatment of CoNV are wide ranging, but can be broadly classified into three groups: (1) enhancing drug delivery via nanocarriers, (2) acting as vehicles for gene therapy, and (3) functioning as therapeutic agents.[119] Nanoparticle-based therapies studied in the literature are summarized in Table 2. Notably, these therapies are still in the preclinical research stage and their clinical safety and efficacy remains to be determined. Further work is required to determine the potential toxicity of nanoparticles as well as to overcome the high cost of production associated with the large-scale production for clinical use. Nonetheless, nanoparticle-based therapy remains a promising and viable method to treat CoNV.

Table 2.

Summary of surgical, gene-based and nano-based delivery therapy modalities studied for corneal neovascularization

Agent		Level of Evidence		Mechanism of Action		Efficacy and Remarks		Studies
LSC Transplantation		Animal Studies Randomized Controlled Clinical Trial Meta-analysis		Transplantation of allograft, autograft or cultivated LSC		In a LSC deficient rabbit model, LSC with limbal stromal stem cell transplantation was significantly more effective than LSC transplantation at CoNV regression Comparative studies on the effectiveness of different graft types and transplant procedures remain inconclusive		Animal: Zhu et al., 2022[146] Human: Zakaria et al., 2014,[147] Le et al., 2020[148]
MICE		Case Series		Ablation of CNV with endothelial cell proliferation inhibition by MMC		Effective in reducing corneal neovascularization in the described clinical case series Further assessment is required for the safety and efficacy of MMC		Human: Mimouni et al., 2022,[93] Addeen et al., 2023[94]
Corneal Cross- Linking		Animal Studies Case Series		Riboflavin inhibition of TNF-α, IL-1, and IL-6 synthesis ROS generation associated with riboflavin activation by UV-A irradiation opposes angiogenesis and lymph angiogenesis		Effectiveness of CoNV inhibition observed in animal models and even in patients with progressive CoNV in need of high risk-keratoplasty Further studies on optimal dosage and regimen required		Animal: Xu et al., 2018,[149] Kesim et al., 2021,[150] Colombo- Barboza et al., 2014[103] Human: Schaub et al., 2021[104]
Gene Therapy
Aganirsen		Randomized Controlled Clinical Trial		IRS-1 mRNA expression inhibiting antisense oligonucleotide		Phase III study reported safety and efficacy of aganirsen Reduced need for transplantation in viral keratitis and central corneal neovascularization		Human: Cursiefen et al., 2014[118]
Nano-based Delivery Therapy
Nanocomposite Hydrogels		Animal Studies		Different NPs embedded in hydrogels to capitalize on biomaterial properties that improve drug solubility, penetration and inhibitory effect on CoNV		Subconjunctival injection of mesoporous silica nanoparticles integrated thermogel loaded with bevacizumab and cyclosporin showed significant inhibition of corneal neovascularization as compared to individual bevacizumab and cyclosporin therapies		Animal: Lyu et al., 2021[151]
Liposomal NP		Animal Studies		Phospholipid bilayer membranes capable of carrying both hydrophilic and hydrophobic molecules Enhances drug penetration and retention to improve therapeutic efficacy of vehicle drug		Therapeutic effect and drug retention time of cationic liposomes encapsulating tacrolimus were better than commercial tacrolimus eyedrops and the free drug at CoNV inhibition		Animal: Lin et al., 2021[32]
Polymeric Micelles		Animal Studies		Hydrophobic fatty acid core with a hydrophilic outer layer Smaller volume than liposomes may enhance corneal drug penetrance		In a corneal transplantation rat model, CoNV inhibitory effect of topical cyclosporine A micelle formulation was comparable to systemic cyclosporine A and significantly greater than the control group		Animal: Di Tommaso et al., 2012[152]
Natural/Synthetic Polymeric NP		Animal Studies		Sub-micro colloidal particles formulated for enhanced drug penetration, bioavailability, biodegradability, and non-toxicity		Kaempferol-loaded gelatin NPs was significantly more effective than kaempferol solution (poorly soluble flavonoid) at inhibiting corneal vessels formation in a mouse model In chemically induced cornea neovascularization, mice corneas treated with gp91 peptide-encapsulated gelatin nanoparticles –an anti- angiogenic nicotinamide adenine dinucleotide phosphate oxidases		Animal: Chuang et al., 2019,[153] Chu et al., 2023[154] Zhang et al., 2018[155]
						inhibitor, had a significant reduction in neovascular vessel area (8%) when compared to phosphate-buffered solution (34%) and gp91 (20%) Subconjunctival bevacizumab- encapsulated polylactide-co-glycolide had significantly greater inhibitory effect on corneal neovascularization than bevacizumab in an alkali burn mouse model
Inorganic Nanoparticles		Animal Studies		Submicron inorganic particles utilized for drug delivery Downregulation of VEGFR-2 expression in Gold NPs Anti-inflammatory action by the ROS scavenging properties of Ce NPs		Gold NP resulted in significant reduction in corneal neovascularization in a mouse model compared to control Ce NPs had comparable effectiveness in corneal neovascularization reduction to dexamethasone in an alkali mouse model Potential toxicity due to non- biodegradable nature		Animal: Cho et al., 2015,[156] Zheng et al., 2019[157]
Nanoemulsions		Animal Studies		Immiscible oil and water layers with accompanying active surfactants that allow for enhanced stability and sustained delivery of hydrophobic drugs		In an alkali-burn mouse model, topical application of eNano-Ro5 significantly reduced corneal neovascularization as compared to the control group, but was slightly less effective than topical dexamethasone Multicomponent nature requires extensive research and complicates manufacturing processes		Animal: Delgado- Tirado et al., 2022[158]
Dendrimers		Animal Studies		Polymeric NPs with a regular branched structure where surface groups can increase drug solubility and stability Modifiable chemical parameters by altering surface groups allow for further drug development		Subconjunctival injectable dendrimer- dexamethasone gel significantly limited the extent of corneal neovascularization as compared to free dexamethasone in an alkali burn rat model		Animal: Soiberman et al., 2017[159]
Nanowafers		Animal Studies		Nano-sized polymer with arrays of drug-filled nano reservoir to enable sustained drug release and absorption		Axitinib nanowafer was more effective than topical axitinib in inhibiting corneal neovascularization in a murine ocular burn model		Animal: Yuan et al., 2015[160]

LSC, Limbal Stem Cell; MMC, Mitomycin Intravascular Chemoembolization; MMC, Mitomycin; TNF-α, Tumour Necrosis Factor α; IL, Interleukin; ROS, Reactive; UV-A, Ultraviolet A; IRS-1, Insulin Receptor Substrate 1; mRNA, messenger Ribonucleic Acid; NP, Nanoparticle; VEGFR, Vascular Endothelial Growth Factor Receptor; Ce, Cerium oxide; eNano-Ro5, Nanoemulsion containing Ro5-3335

Table 3.

Summary of other treatment modalities studied for corneal neovascularization

Agent		Level of Evidence		Mechanism of Action		Efficacy and Remarks		Studies
Combination Therapy
Corticosteroids and Heparin		Animal Studies Case Reports		Corticosteroid inhibition of phospholipase A2 and subsequent inflammatory response		Effectiveness of combination therapy remains to be further evaluated in comparative clinical studies		Animal: Nikolic et al., 1986,[161] Aydin et al., 2008[162]
				Mechanism of heparin in reducing neovascularization not well understood; possibly via direct modulatory effect on angiogenic growth factors, or enhancing uptake of				Human: Tommila et al., 1987[163]
				corticosteroids into endothelial cells
Corticosteroids and Doxycycline		Animal Studies		Corticosteroid inhibits phospholipase A2 and subsequent inflammatory response Doxycycline inhibits MMP		Combination of triamcinolone acetonide and doxycycline was more effective at CoNV inhibition than the individual therapies		Animal: Aydin et al., 2008[162]
Corticosteroids and anti- VEGF		Animal Studies		Inhibition of angiogenic factors and proinflammatory cytokines like VEGF and IL-6		Effectiveness of the combination of corticosteroids and anti- VEGF did not differ significantly from individual therapeutics Remains to be further evaluated in studies with larger sample sizes and longer follow-up time		Animal: Kang et al., 2010,[164] Hoffart et al., 2010[22]
NSAID and anti-VEGF (Diclofenac and Bevacizumab via thermosensitive hydrogel)		Animal Studies		Dual delivery of diclofenac and bevacizumab by the thermosensitive hydrogel (Poly (dl-lactide)-poly (ethylene glycol)-poly (dl-lactide)		Proposed hydrogel caused minimal toxicity to tested cell lines at concentrations ranging from 0.05 mg/ml to 0.8 mg/ml with good ocular biocompatibility after a single subconjunctival injection		Animal: Shi et al., 2022[165]
						Superior anti-angiogenic effects as compared to subconjunctival bevacizumab injection alone
TNF-α inhibitor and anti- VEGF		Animal Studies		Inhibition of TNF-α, a pro- inflammatory cytokine responsible for corneal lymphangiogenesis through the induction of VEGF-C production by macrophages		Simultaneous inhibition of TNF-α and VEGF was more effective in CoNV inhibition than monotherapies		Animal: Ozdemir et al.,2013,[34] Zhou et al., 2023[166]
				TNF-α is also responsible for the expression of angiogenic factors
Laser Photocoagulation and anti-VEGF		Case Reports		Heat-induced photocoagulation of corneal vasculature Inhibition of VEGF		Clinical effectiveness of the combination therapy exhibited in multiple case reports Effectiveness yet to be established in a comparative clinical study		Human: Gerten et al., 2008,[167] Anand et al., 2018,[168] Donato et al., 2021[169]
PDT and anti-VEGF		Animal Studies Case Series Studies		Excitation of photosensitizers to occlude corneal vessels through thrombus formation Inhibition of VEGF		Combination therapy was more effective than Individual monotherapies in a suture- induced rabbit model and preliminary clinical studies Anti-VEGF agents suppress VEGF release post-PDT, reducing the risk of angiogenesis and recurrence		Animal: Kim et al., 2016[170] Human: Hamdan et al., 2015,[171] Yoon et al., 2019[172]
FND and anti-VEGF		Animal		Coagulating current used to		FND and bevacizumab		Animal:
		Studies		occlude corneal vessels		combination successfully		Le et al.,
						regressed mature CoNV in		2020[82]
		Case		Inhibition of VEGF		patients previously
		Series				unresponsive to		Human:
		Studies				corticosteroids		Koenig et al.,
								2012,[173]
								Elbaz et al.,
								2015,[84]
								Mestanoglu et al.,
								2022[174]
Other Therapies
Descemet Membrane		Case		Selective removal and		Single study revealed a		Human:
Endothelial Keratoplasty		Series		transplantation of Descemet		significant regression in		Hayashi et al.,
		Studies		membrane and endothelium		corneal neovascularization		2021[175]
						postoperatively
Mesenchymal Stem Cell Therapy		Animal		Source of progenitor cells		Efficacy, bio-variability, and		Animal:
		Studies		Anti-inflammatory properties		safety to be further evaluated		Yao et al.,
						in clinical studies		2012,[176]
								Song et al.,
								2018[177]
Tetracyclines (Doxycycline, Minocycline)		Animal Studies Case Series Studies		MMP and PI3K/Akt-eNOS pathway inhibition MMP, ERK1/2 and Akt pathways inhibition		Lesser side effects than Anti- VEGF and corticosteroids		Animal: Aydin et al., 2008,[162] Xiao et al., 2012[178]
Other Antimicrobials (Tigecycline, Itraconazole, Dihydroartemisinin)		Animal Studies		Tigecycline inhibits MMP Itraconazole inhibits cholesterol biosynthesis Dihydroartemisinin inhibits VEGF, ERK1/2 and p38 expression		Effectiveness in CoNV inhibition displayed in experimental models Requires further evaluation of therapeutic effect in clinical studies		Animal: Zhong et al., 2011,[179] Goktas et al., 2014[180]
Flavonoids		Animal		Anti-angiogenic properties:		Variable effectiveness of		Animal:
(Epigallocatechin		Studies		Mechanism not well		individual flavonoids in		Li et al,
Gallate, Kaempferol,				understood, may be		different experimental		2018,[181]
Fisetin, Luteolin,				associated to corticosteroid-		models of CoNV		Chuang et al.,
Genistein, Naringenin,				like molecular structure				2019[153]
Quercetin, Coumarin
Esculetin)
Non-Flavonoid		Animal		Anti-angiogenic properties		Variable bioavailability may		Animal:
Phytochemicals		Studies		Mechanism not well		limit the effectiveness		Bian et al.,
(Curcumin, Resveratrol,				understood				2008,[182] Kim
Withaferin A, Xanthatin,								et al. 2010,[183]
Triptolide,
Thymoquinone,
Glycyrrhizin)
Vitamins		Animal		Suppresses angiogenic factors such as VEGF and MMP		Requires further evaluation of therapeutic effect in clinical studies		Animal:
(Ascorbic Acid,		Studies						Suzuki et al.,
Riboflavin, 1α,25-								2000,[184] Lee
dihydroxyvitamin D3)				Mechanism not well				et al., 2012[185]
				understood

MMP, Matrix Metalloproteinase; VEGF, Vascular Endothelial Growth Factor; NSAID, Non-Steroidal Anti-Inflammatory Drug; IL, Interleukin; TNF-α, Tumor Necrosis Factor α; PDT, Photodynamic Therapy; FND, Fine Needle Diathermy; PI3K/Akt-eNOS, Phosphatidylinositol 3-Kinase/Akt-endothelial Nitric Oxide Synthase; ERK1/2, Extracellular Signal-Regulated Kinase 1/2

Conclusion

CoNV is a common complication of severe corneal pathologies that can result in visually significant disabilities. Recent advances in our understanding of its pathogenesis, aided by developments in imaging and pharmacological approaches, have led to the development of a range of treatment strategies. While conventional therapy has focused on reducing inflammation and targeting the VEGF pathway, an expanding array of pharmacological agents, such as tyrosine kinase inhibitors, biologic immunomodulators, and gene therapy, may be useful in improving treatment efficacy. Moreover, targeting multiple angiogenic pathways in the form of combination therapy could produce a synergistic effect and should be considered in future drug and clinical trials. However, further research, rigorous clinical trials, and long-term follow up studies are essential to refine these protocols and ensure their safety and efficacy. As our knowledge continues to expand and innovative therapies emerge, the outlook for patients with CoNV is increasingly promising.

Financial support and sponsorship

Nil.

Conflicts of interest

There are no conflicts of interest.