new research for glaucoma

Featured Topics

Featured series.

A series of random questions answered by Harvard experts.

Explore the Gazette

Read the latest.

Implantable device responds to opioid overdose

 Study detects ‘hidden consciousness’ in brain injury patients 

Examining new weight-loss drugs, pediatric bariatric patients

Researchers at HMS have successfully restored vision loss and reversed glaucoma-induced damage in mice.

Credit: Sinclair Lab/Harvard Medical School

Seeing clearly again

Ryan Jaslow

MEEI Communications

Harvard Medical School scientists reverse age-related vision loss, eye damage from glaucoma in mice

Harvard Medical School scientists report they have successfully restored vision in mice by turning back the clock on aged eye cells in the retina to recapture youthful gene function.

The team’s work, described Dec. 2 in the publication Nature, represents the first demonstration that it may be possible to safely reprogram complex tissues, such as the nerve cells of the eye, to an earlier age.   

In addition to resetting the cells’ aging clock, the researchers successfully reversed vision loss in animals with a condition mimicking human glaucoma, a leading cause of blindness around the world. 

The achievement represents the first successful attempt to reverse glaucoma-induced vision loss, rather than merely stem its progression, the team said. 

If replicated through further studies, the approach could pave the way for therapies to promote tissue repair across various organs and reverse aging and age-related diseases in humans.

“Our study demonstrates that it’s possible to safely reverse the age of complex tissues such as the retina and restore its youthful biological function,” said senior author David Sinclair, professor of genetics in the Blavatnik Institute at Harvard Medical School, co-director of the Paul F. Glenn Center for Biology of Aging Research at HMS and an expert on aging. 

Sinclair and colleagues caution that the findings remain to be replicated in further studies, including in different animal models, before any human experiments. Nonetheless, they add, the results offer a proof of concept and a pathway to designing treatments for a range of age-related human diseases.

“If affirmed through further studies, these findings could be transformative for the care of age-related vision diseases like glaucoma and to the fields of biology and medical therapeutics for disease at large,” Sinclair said.

“At the beginning of this project, many of our colleagues said our approach would fail or would be too dangerous to ever be used. Our results suggest this method is safe and could potentially revolutionize the treatment of the eye and many other organs affected by aging.” Yuancheng Lu, lead study author

For their work, the team used an adeno-associated virus (AAV) as a vehicle to deliver into the retinas of mice three youth-restoring genes — Oct4, Sox2, and Klf4 — that are normally switched on during embryonic development. The three genes, together with a fourth one, which was not used in this work, are collectively known as Yamanaka factors.

The treatment had multiple beneficial effects on the eye. First, it promoted nerve regeneration following optic-nerve injury in mice with damaged optic nerves. Second, it reversed vision loss in animals with a condition mimicking human glaucoma. And third, it reversed vision loss in aging animals without glaucoma.

The team’s approach is based on a new theory about why we age. Most cells in the body contain the same DNA molecules but have widely diverse functions. To achieve this degree of specialization, these cells must read only genes specific to their type. This regulatory function is the purview of the epigenome, a system of turning genes on and off in specific patterns without altering the basic underlying DNA sequence of the gene. 

This theory postulates that changes to the epigenome over time cause cells to read the wrong genes and malfunction — giving rise to diseases of aging. One of the most important changes to the epigenome is DNA methylation, a process by which methyl groups are tacked onto DNA. Patterns of DNA methylation are laid down during embryonic development to produce the various cell types. Over time, youthful patterns of DNA methylation are lost, and genes inside cells that should be switched on get turned off and vice versa, resulting in impaired cellular function. Some of these DNA methylation changes are predictable and have been used to determine the biologic age of a cell or tissue.

Yet, whether DNA methylation drives age-related changes inside cells has remained unclear. In the current study, the researchers hypothesized that if DNA methylation does, indeed, control aging, then erasing some of its footprints might reverse the age of cells inside living organisms and restore them to their earlier, more youthful state.

Past work had achieved this feat in cells grown in laboratory dishes but fell short of demonstrating the effect in living organisms.

The new findings demonstrate that the approach could be used in animals as well.  

Overcoming an important hurdle 

Lead study author, Yuancheng Lu , research fellow in genetics at HMS and a former doctoral student in Sinclair’s lab, developed a gene therapy that could safely reverse the age of cells in a living animal. 

Lu’s work builds on the Nobel Prize winning discovery of Shinya Yamanaka, who identified the four transcription factors, Oct4, Sox2, Klf4, c-Myc, that could erase epigenetics markers on cells and return these cells to their primitive embryonic state from which they can develop into any other type of cell.   

Subsequent studies, however, showed two important setbacks. First, when used in adult mice, the four Yamanaka factors could also induce tumor growth, rendering the approach unsafe. Second, the factors could reset the cellular state to the most primitive cell state, thus completely erasing a cell’s identity.

Lu and colleagues circumvented these hurdles by slightly modifying the approach. They dropped the gene c-Myc and delivered only the remaining three Yamanaka genes, Oct4, Sox2, and Klf4.  The modified approach successfully reversed cellular aging without fueling tumor growth or losing their identity. 

Gene therapy applied to optic nerve regeneration

  In the current study, the researchers targeted cells in the central nervous system because it is the first part of the body affected by aging. After birth, the ability of the central nervous system to regenerate declines rapidly. 

To test whether the regenerative capacity of young animals could be imparted to adult mice, the researchers delivered the modified three-gene combination via an AAV into retinal ganglion cells of adult mice with optic nerve injury.  

For the work, Lu and Sinclair partnered with Zhigang He , HMS professor of neurology and of ophthalmology at Boston Children’s Hospital, who studies optic nerve and spinal cord neuro-regeneration.

The treatment resulted in a two-fold increase in the number of surviving retinal ganglion cells after the injury and a five-fold increase in nerve regrowth. 

“At the beginning of this project, many of our colleagues said our approach would fail or would be too dangerous to ever be used,” said Lu. “Our results suggest this method is safe and could potentially revolutionize the treatment of the eye and many other organs affected by aging.”

Reversal of glaucoma and age-related vision loss 

More like this.

Focusing on the fovea

Linking sight and movement

Making sense of how the blind ‘see’ color

Following the encouraging findings in mice with optic nerve injuries, the team partnered with colleagues at Schepens Eye Research Institute of Massachusetts Eye and Ear Bruce Ksander , HMS associate professor of ophthalmology, and Meredith Gregory-Ksander , HMS assistant professor of ophthalmology. They planned two sets of experiments: one to test whether the three-gene cocktail could restore vision loss due to glaucoma and another to see whether the approach could reverse vision loss stemming from normal aging.

In a mouse model of glaucoma, the treatment led to increased nerve cell electrical activity and a notable increase in visual acuity, as measured by the animals’ ability to see moving vertical lines on a screen. Remarkably, it did so after the glaucoma-induced vision loss had already occurred.

“Regaining visual function after the injury occurred has rarely been demonstrated by scientists,” Ksander said. “This new approach, which successfully reverses multiple causes of vision loss in mice without the need for a retinal transplant, represents a new treatment modality in regenerative medicine.” 

The treatment worked similarly well in elderly, 12-month-old mice with diminishing vision due to normal aging. Following treatment of the elderly mice, the gene expression patterns and electrical signals of the optic nerve cells were similar to young mice, and vision was restored. When the researchers analyzed molecular changes in treated cells, they found reversed patterns of DNA methylation — an observation suggesting that DNA methylation is not a mere marker or a bystander in the aging process, but rather an active agent driving it.

“What this tells us is the clock doesn’t just represent time — it is time,” said Sinclair. “If you wind the hands of the clock back, time also goes backward.” 

The researchers said that if their findings are confirmed in further animal work, they could initiate clinical trials within two years to test the efficacy of the approach in people with glaucoma. Thus far, the findings are encouraging, researchers said. In the current study, a one-year, whole-body treatment of mice with the three-gene approach showed no negative side effects.

Other authors on the paper include Benedikt Brommer, Xiao Tian, Anitha Krishnan, Margarita Meer, Chen Wang, Daniel Vera, Qiurui Zeng, Doudou Yu, Michael Bonkowski, Jae-Hyun Yang, Songlin Zhou, Emma Hoffmann, Margarete Karg, Michael Schultz, Alice Kane, Noah Davidsohn, Ekaterina Korobkina, Karolina Chwalek, Luis Rajman, George Church, Konrad Hochedlinger, Vadim Gladyshev, Steve Horvath, and Morgan Levine.

This work was supported in part by a Harvard Medical School Epigenetics Seed Grant and Development Grant, The Glenn Foundation for Medical Research, Edward Schulak, the National Institutes of Health (grants R01AG019719,R37AG028730, R01EY026939, R01EY021526, R01AG067782, R01GM065204, R01AG065403, R01EY025794, R24EY028767 and R21EY030276), and the St. Vincent de Paul Foundation.

Share this article

You might like.

Without assistance, it allows for precise administration of naloxone at the moment it is needed

25% of participants with severe brain injury followed instructions covertly

Researcher says study found variation in practices, discusses safety concerns overall for younger users

John Manning named next provost

His seven-year tenure as Law School dean noted for commitments to academic excellence, innovation, collaboration, and culture of free, open, and respectful discourse

Loving your pup may be a many splendored thing

New research suggests having connection to your dog may lower depression, anxiety  

Good genes are nice, but joy is better

Harvard study, almost 80 years old, has proved that embracing community helps us live longer, and be happier

An official website of the United States government

The .gov means it’s official. Federal government websites often end in .gov or .mil. Before sharing sensitive information, make sure you’re on a federal government site.

The site is secure. The https:// ensures that you are connecting to the official website and that any information you provide is encrypted and transmitted securely.

• Publications
• Account settings

Preview improvements coming to the PMC website in October 2024. Learn More or Try it out now .

• Advanced Search
• Journal List
• Mayo Clin Proc Innov Qual Outcomes
• v.6(6); 2022 Dec

Updates on the Diagnosis and Management of Glaucoma

Glaucoma is the leading cause of blindness throughout the world (after cataracts); therefore, general physicians should be familiar with the diagnosis and management of affected patients. Glaucomas are usually categorized by the anatomy of the anterior chamber angle (open vs narrow/closed), rapidity of onset (acute vs chronic), and major etiology (primary vs secondary). Most glaucomas are primary (ie, without a contributing comorbidity); however, several coexisting ophthalmic conditions may serve as the underlying etiologies of secondary glaucomas. Chronic glaucoma occurs most commonly; thus, regular eye examinations should be performed in at-risk patients to prevent the insidious loss of vision that can develop before diagnosis. Glaucoma damages the optic nerve and retinal nerve fiber layer, leading to peripheral and central visual field defects. Elevated intraocular pressure (IOP), a crucial determinant of disease progression, remains the only modifiable risk factor; thus, all current treatments (medications, lasers, and operations) aim to reduce the IOP. Pharmacotherapy is the usual first-line therapy, but noncompliance, undesirable adverse effects, and cost limit effectiveness. Laser and surgical treatments may lower IOP significantly over long periods and may be more cost effective than pharmacotherapy, but they are plagued by greater procedural risks and frequent treatment failures. Traditional incisional procedures have recently been replaced by several novel, minimally invasive glaucoma surgeries with improved safety profiles and only minimal decreases in efficacy. Minimally invasive glaucoma surgeries have dramatically transformed the surgical management of glaucoma; nevertheless, large, randomized trials are required to assess their long-term efficacy.

Article Highlights

• • Glaucoma, a leading cause of blindness throughout the world, presents with an open or closed anterior chamber angle, structural damage to the optic nerve (seen in all stages), and visual field defects (seen in later stages). Glaucoma may be asymptomatic until the late stages, thereby emphasizing the need for general physicians to understand important diagnostic criteria and management options.
• • The progression of glaucoma is mitigated by lowering the intraocular pressure, which is done with topical medications, laser procedures, or incisional operations.
• • Minimally invasive glaucoma surgery, with a favorable safety profile and efficacy that rivals traditional incisional procedures, has transformed glaucoma care.

Glaucoma can be defined as a progressive optic neuropathy that induces optic disc cupping and retinal ganglion cell apoptosis. 1 As the world’s leading cause of irreversible blindness, the disease currently affects 3.5% of individuals aged between 40 and 80 years. The incidence of glaucoma is increasing, together with life expectancies, in resource-limited countries, and nearly 112 million people are expected to be affected by 2040. 1 , 2 Early detection can slow disease progression, but because visual field loss may be asymptomatic until the late stages, a timely diagnosis is frequently delayed. 3 Common risk factors for glaucoma include advancing age, a positive family history, race (non-Caucasian), and elevated intraocular pressure (IOP). 4 , 5 Once diagnosed with glaucoma, most patients require lifelong care.

Aqueous humor is produced by the ciliary body, and after percolating through the posterior chamber, around the lens, and through the pupil, it exits the eye through the semiporous trabecular meshwork (TM) in the iridocorneal angle of the anterior chamber. Aqueous humor then flows into the circumferential vascular collection duct (Schlemm canal) and leaves the eye through the distal collector channels that drain into the episcleral venous system. 6 , 7 , 8 A detailed anatomical view of the anterior eye segment and the aqueous outflow pathway is displayed in Figure 1 . The pathogenesis of glaucoma includes inadequate drainage of aqueous humor because of increased resistance through the meshwork 7 or occlusion of the angle, 9 both of which elevate the IOP. Elevated IOP contributes to an irreversible, progressive ocular neuropathy characterized by retinal ganglion cell apoptosis. 1 Patients with elevated IOP without other signs of glaucoma are said to have ocular hypertension, and those with optic disc enlargement but normal IOP and no other signs of glaucoma are classified as glaucoma suspects.

Ultrasound biomicroscopy (UBM) of the anterior eye segment. A, UBM shows the ciliary zonules (CZ), ciliary body (CB), sclera (S), cornea (C), anterior chamber (AC), posterior chamber (PC), and lens (L). The anterior chamber angle (ACA) is indicated by the arrow. B, Magnified UBM of the ACA shows the trabecular meshwork (TM), Schlemm canal (SC), and collector channels (CCs).

The risk factors and pathogenesis that underly glaucoma have been well described in the literature; however, the biological basis of the disease remains incompletely understood. The biomechanical and vascular theories of glaucoma propose that elevated IOP compromises axonal integrity at the optic nerve head (ONH), which leads to ganglion cell apoptosis. 5 The biomechanical theory posits that abnormally narrow scleral fenestrations at the ONH limit axoplasmic flow, 5 , 7 , 10 whereas the vascular theory states that decreased perfusion pressure leads to hypoxia and ischemic damage of the ONH. 5 , 7 , 11 Both theories include IOP as a risk factor; however, one-third of patients with glaucoma have normal IOPs (normal tension glaucoma). 5 Glaucoma has been associated with Alzheimer disease 12 and a loss of cognitive function, 13 which suggests that neurodegeneration may contribute to the pathogenesis. 5 However, despite the different pathogenetic theories, elevated IOP consistently contributes to disease progression and remains the only treatable risk factor. 5 , 7

The goal of glaucoma treatment is to lower IOP with medications, laser procedures, and/or operation. First-line therapy is usually pharmacotherapy, with laser and surgical procedures added for further IOP reduction in eyes with inadequate initial responses. Incisional operations consist of filtration procedures (eg, trabeculectomy) or tube shunt implantation, both of which reroute aqueous humor flow past the damaged angle into the subconjunctival space forming a filtration bleb. 14

Traditional incisional operations lower the IOP effectively; however, complication rates, including scar tissue proliferation, endophthalmitis, and conjunctival hemorrhage, are high. The IOP-lowering effect often decreases over time, which results in high 5-year reoperation rates (trabeculectomy, 15.1%; tube shunt implantation, 14.0%; EX-PRESS shunt, 18.3%). 15 , 16 , 17 These high reoperation rates speak to the need for procedures that increase conventional aqueous outflow while protecting the conjunctiva from surgical manipulation. This has led to the development of several conjunctival sparing, minimally invasive glaucoma surgeries (MIGSs) for the treatment of primary open-angle glaucoma (POAG). Minimally invasive glaucoma surgeries do not reduce IOP as well as traditional filtering procedures, but they have excellent safety profiles. 18

We believe that because of the expanding treatment options and increasing worldwide prevalence of glaucoma, an updated commentary on glaucoma and its treatment options is important for medical physicians. In this article, we provide a comprehensive updated review of the diagnosis and management of adult glaucoma through 2022.

A broad literature search with no time frame was carried out in PubMed with the following key words: “glaucoma prevalence,” “glaucoma risk factors,” “glaucoma diagnosis,” “glaucoma management,” “open-angle glaucoma,” angle-closure glaucoma,” “secondary glaucoma,” “tonometry,” “glaucoma medication,” “conventional aqueous outflow,” “unconventional aqueous outflow,” “glaucoma laser procedures,” “trabeculectomy,” “glaucoma tube shunt surgery,” and “minimally invasive glaucoma surgery.” Identified articles and their references were scrutinized, and those relevant to the subject matter were selected.

Diagnosis of Glaucoma

Types of glaucoma.

Glaucoma may be broadly categorized as open-angle glaucoma (OAG) and angle-closure glaucoma (ACG). Primary OAG and primary ACG are seen most frequently; however, several ocular conditions cause secondary glaucomas ( Table 1 ).

Table 1

Common Glaucoma Types are Listed According to Whether the Anterior Chamber Angle is Open or Closed a

Most eyes with glaucoma have diminished conventional aqueous outflow despite a normal gonioscopic appearance of the iridocorneal angle. These OAGs are slowly progressive optic neuropathies in which ONH cupping gradual increases and peripheral visual field loss develops. 15 , 19 The most common type of glaucoma—the POAG—affects 74% of patients with glaucoma. 20 Outflow resistance may be modulated by hydrodynamic pore-substrate interactions within the inner wall of the Schlemm canal, and patients with POAG have been found to have reduced pore density. 21

Several types of secondary OAG occur much less frequently than POAG. Pigmentary glaucoma occurs when friction between the lens zonules and iris pigment epithelium releases pigment granules that lodge in the TM and increase outflow resistance. 22 , 23 Exfoliative glaucoma, the most common form of secondary OAG, occurs when microscopic clumps of protein fibers are synthesized within the eye and clog the TM. 24 Exfoliation material has also been found in the heart, kidney, liver, and lungs. 24 , 25 Other forms of secondary OAG include uveitic and traumatic glaucomas, 26 , 27 , 28 use of ocular or systemic corticosteroids, 29 and antineoplastic drugs. 30 Increased episcleral venous pressure due to conditions such as carotid-cavernous sinus fistulas may cause OAG. 31

Angle-closure glaucomas are rapidly progressive ocular neuropathies characterized by the occlusion of at least 270° of the iridocorneal angle. 3 Angle-closure glaucomas are only one-third as common as OAGs; however, they are responsible for approximately 50% of all glaucoma-induced blindness. Primary ACG, which arises from pupillary block (appositional closure of the iridocorneal angle that results from an increasing pressure differential between the anterior and posterior chambers of the eye 32 ) or plateau iris (an anteriorly positioned ciliary body that causes contact between the iris and TM with resultant angle crowding 33 ), has a global prevalence of 0.6%. 3 , 34 , 35 Primary ACG occurs most frequently in women, Asians, people with hypermetropic (short) eyes and people with shallow anterior chambers. 34 Affected patients require urgent treatment (usually laser iridotomy) to reverse obstruction of the angle. 34

Several secondary types of ACG are seen. Neovascular glaucoma, new blood vessels that occlude the angle, may develop from central retinal vein occlusion or diabetic retinopathy and generally carries a poor visual prognosis. 1 , 36 Phacomorphic glaucoma involves angle-closure because of lens intumescence (advanced cataract), and cataract removal typically leads to good visual recovery. 37 Angle-closure may be caused by corneal endothelium abnormalities (eg, iridocorneal endothelium syndromes) 38 or large iris or ciliary body masses. 39 Several medications, including anticholinergics, may precipitate ACG in eyes with preexisting narrow angles. 1 , 40

Differentiating between OAG and ACG is usually done via gonioscopic examination with slit lamp viewing. 41 Gonioscopy has long been the gold standard for visualizing the anterior chamber angle (ACA); however, challenges, including lens-eye contact, lack of objective measurements, a steep learning curve, and inconsistent interpretations between physicians, exist. 41 , 42 Advanced ACA imaging techniques including swept-source optical coherence tomography (OCT), goniophotography systems, and deep learning algorithms have been developed to overcome the limitations of gonioscopy. 43

Examination

Approximately 50% of individuals in the resource-limited countries are unaware that they have glaucoma, underscoring the importance of patient awareness education in diagnosis and management. 3 , 44 , 45 The diagnosis of glaucoma involves risk assessment, measurement of visual acuity, IOP, and corneal thickness, OCT imaging of the retinal nerve fiber layer (RNFL) and ONH, and visual field testing. Because most patients with glaucoma are asymptomatic for years, it is recommended that those with risk factors (advanced age, family history, non-White race, high IOP, and steroid use) be referred to an eye care provider for a glaucoma assessment. 3 , 4 , 5

Intraocular pressure needs to be monitored regularly in patients at a high risk of developing glaucoma. It is commonly measured using rebound tonometry (iCare ic100; iCare) or the “gold standard” Goldmann applanation tonometry. The iCare tonometer measures IOP-dependent rebound velocity after brief corneal contact, whereas Goldmann applanation tonometry measures the force required to flatten a 3.06-mm diameter segment of the cornea. 46 Agreement in measurements is good between the 2 devices; however, the reliability of the iCare decreases at higher IOPs and with thicker central corneas. 47 , 48 , 49 Normal IOP ranges from 11 to 21 mm Hg 50 ; however, IOP should be evaluated with consideration of optic nerve defects and/or high central cornea values. 51 Up to 50% of glaucomatous eyes have normal IOP measurements, 3 , 52 which emphasizes the importance of performing additional diagnostic imaging when indicated.

Making the diagnosis of glaucoma, particularly at an early stage, can be difficult because there is no uniform standard for diagnosis. 3 Structural changes of early glaucoma can be seen with OCT imaging of the optic nerve and macula, and functional changes in advanced glaucoma can be detected with visual field testing. Normal appearances of the ONH, RNFL, and visual field are shown in Figure 2 A, C, and E, respectively. All glaucomas are defined by ONH degeneration with disc excavation ( Figure 2 B) and RNFL thinning ( Figure 2 D). 53 Optic nerve head damage is characterized by thinning of the neuroretinal rim, usually in the superior and inferior quadrants, although the remainder of the ONH may remain pink with a normal neuroretinal rim. 3 , 53 Glaucomatous damage leads to retinal ganglion cell apoptosis, which can be seen as thinning between the internal limiting membrane and ganglion cell layer on OCT. 53 As glaucoma progresses, ONH and RNFL abnormalities cause visual field defects ( Figure 2 F). Visual field defects are often not observed in the early stages of glaucoma because peripheral vision and Snellen visual acuity are preserved until RNFL damage reaches an advanced stage. 51

Comparison of optic nerve head (ONH), retinal nerve fiber layer (RNFL), and visual fields in normal and glaucomatous eyes. A, Normal ONH with round, elevated ONH and a small central cup. B, Glaucomatous ONH with excavation and thinning of neuroretinal rim. C, Optical coherence tomography (OCT) examination shows normal RNFL thickness. D, OCT examination shows RNFL thinning in glaucomatous eyes. E, A full field in both eyes is shown. F, Abnormal visual field results in glaucomatous eyes are shown. The right eye field shows a superior altitudinal defect, moderate inferior arcuate defects, and a nasal step. The left eye field shows a superior paracentral defect with nasal step that splits fixation, an early inferior arcuate scotoma, and nasal step.

A general correlation between OCT imaging and visual field examination can be observed; however, there is no widely accepted method for comparing the two, 54 and diagnosing glaucoma is ultimately up to the discretion of the physician. Once glaucoma has been diagnosed, its severity must be categorized—typically as mild, moderate, or severe. Because all glaucoma types present with structural damage, most classification systems grade severity on the basis of functional visual field abnormalities. Most recently (2015), the International Classification of Diseases, Tenth Revision, released a grading system that associates mild glaucoma with a general absence of visual field defects, moderate glaucoma with visual field abnormalities in 1 hemifield (but outside 5° of fixation), and severe glaucoma with abnormalities in both hemifields and visual field loss within 5° of fixation. 55

Management of Glaucoma

Medical therapy.

Guidelines from the American Academy of Ophthalmology Preferred Practice Pattern (2020) state that an initial IOP reduction of 20%-30% is a suitable goal to slow disease progression, even in eyes with normal tension glaucoma. 56 The IOP must be carefully monitored during each follow-up visit, and the IOP control goal should be lowered further if progression continues. 56

Intraocular pressure–lowering medications have been the first-line therapy for most patients with glaucoma for several decades ( Table 2 ). Pharmacotherapy for glaucoma has evolved significantly over the past several decades with the introduction of topical carbonic anhydrase inhibitors (CAIs), beta blockers, prostaglandin analogs, and alpha agonists. 57 These medications have greater effectiveness and more favorable safety profiles than the older topical (pilocarpine) and systemic (oral CAIs) treatments. 57 In accordance with the generally accepted pharmacotherapy principles, the desired IOP range should be achieved with the fewest medications and least adverse effects. 3 Because of their tendency to induce glaucoma, ocular and systemic corticosteroids should be administered with caution in at-risk patients. 29

Table 2

US Food and Drug Administration–Approved Medications Used for the Treatment of Glaucoma

Prostaglandin analogs (PGAs) are the most commonly used medications for the treatment of OAG and ocular hypertension. Prostaglandin analogs compensate for decreased TM outflow by increasing outflow through the uveoscleral pathway, 58 where aqueous humor moves through the ciliary muscle into the supraciliary and suprachoroidal spaces. 59 Prostaglandin analogs are administered once daily, are well tolerated, and have limited systemic adverse effects. 3 , 58 The main ocular adverse effects are eyelash growth, iris pigmentation, and uveitis. 56 Because most PGAs do not target the primary outflow pathway (TM), concerns have been raised about their long-term efficacy. 57 The recently approved latanoprostene bunod 0.024% may target the TM rather than the uveoscleral pathway, 57 , 60 and compared with timolol 0.5% over 3 months of follow-up, it has superior IOP-lowering efficacy and a comparable safety profile. 57 , 61 , 62 Prostaglandin analogs are a significant improvement over cholinergic agonists (such as pilocarpine), which induce miosis and increase conventional outflow by decreasing outflow resistance. 63 Pilocarpine, a mainstay of glaucoma treatment in the 1970s and 1980s, needed to be administered 4 times per day, a difficult regimen to maintain, which contributed to its being supplanted by beta blockers and PGAs. 3

Both CAIs and beta blockers lower the IOP by targeting the aqueous humor production in the ciliary body. After topical administration, CAIs penetrate the cornea and reach the ciliary body epithelium, where they reduce the production of bicarbonate ions. 64 The CAIs (dorzolamide 2% and brinzolamide 1%) are administered 2 or 3 times daily, 64 but they are generally less effective than PGAs and beta blockers, which limits their use as first-line therapy. Systemic CAIs (methazolamide and acetazolamide) are highly effective, which makes them useful in the treatment of ACG; however, their use is limited by their high incidence of adverse effects that cause 50% of patients to become intolerant after 1 month.

Beta adrenergic antagonists (beta blockers) block the sympathetic nerve endings in the ciliary body epithelium, which decreases the production of aqueous. 65 Beta blockers may be nonselective or cardioselective (β1 selective), the latter of which is well tolerated in patients with asthma and chronic obstructive pulmonary disease. 65 The advantages of beta blockers include their relatively low cost and once-daily administration. 3 , 5 Topically administered beta blockers enter the venous circulation but escape the first-pass metabolism in the liver, which predisposes the patient to pulmonary (bronchial constriction) and cardiac (arrythmias) disturbances. 5 , 66 Systemic absorption can be lessened by eyelid closure or gentle punctal occlusion for 2 minutes after topical administration. 3

Topical alpha-adrenergic agonists (brimonidine and iopidine) reduce the IOP by decreasing the aqueous humor production and increasing the outflow. 3 They are administered 2 or 3 times daily and are usually used as second-line agents in combination with other drugs. A retrospective study found that combination treatment (CAI+PGA) was more prevalent in everyday practice than alpha-2 agonists + PGA, suggesting that the administration of alpha-2 agonists may be accompanied by more adverse effects. 67

Rho kinase inhibitors are a recently introduced medication class that uses a combined mechanism of increasing the conventional outflow and decreasing the episcleral venous pressure. 68 Netarsudil 0.02%, a rho kinase inhibitor approved by the US Food and Drug Administration in 2017, has IOP-lowering efficacy comparable with that of timolol 0.5%, but with more frequent adverse effects. 59 , 69 , 70

Pharmacotherapy is an effective short-term treatment strategy; however, limitations to long-term use include cost, adverse effects, and failure to reach the target IOP. Nonadherence to the administration schedule is another significant issue because fewer than half of the patients with glaucoma regularly use antiglaucoma medications as prescribed after 1 year. 5 , 71

Laser Therapy

When pharmacotherapy fails to achieve the target IOP and prevent vision loss, laser and surgical procedures are indicated. Laser procedures effectively lower the IOP and minimize the long-term costs that are associated with the long-term use of multiple pressure-lowering medications. 5 A variety of laser procedures can be performed in glaucomatous eyes, with the procedure of choice depending on the etiology of the disease ( Table 3 ).

Table 3

Laser Procedures for the Treatment of Glaucoma

IOP, intraocular pressure; LPI, laser peripheral iridotomy; TM, trabecular meshwork.

Laser trabeculoplasty and ab-interno excimer trabeculostomy (Glautec AG) are both indicated for OAG that is refractory to pharmacotherapy. Laser trabeculoplasty—multiple spots of thermal laser applied directly to the TM—induces favorable structural changes that increase the aqueous humor outflow. 72 Argon laser trabeculoplasty, developed in 1979, uses a with a blue-green continuous-wave laser (488 and 514 nm) to disrupt the TM, whereas selective laser trabeculoplasty (SLT), developed in 1995, uses low energy, brief duration, large spots from a green, frequency-doubled laser to target melanin-containing cells and spare the TM tissue. 73 Selective laser trabeculoplasty has largely supplanted argon laser trabeculoplasty because of its favorable safety profile, comparable IOP-lowering efficacy, and ability for repeated treatment applications. 74 More recently introduced laser trabeculoplasty procedures include titanium-sapphire laser trabeculoplasty and pattern scanning trabeculoplasty. Limited short-term data suggest that both the procedures have efficacy and safety profiles similar to that of SLT. 74 Laser trabeculoplasty procedures are generally preferred over operations because they are less invasive and possess better safety profiles. 3 , 74 Ab-interno excimer trabeculostomy is a MIGS similar to laser trabeculoplasty that uses a 308-nm XeCl excimer laser to create microperforations in the TM and inner wall of the Schlemm canal. 75 Excimer trabeculostomy has a comparable safety profile and IOP-lowering efficacy similar to SLT over 2 years. 75

Patients with ACG require different laser procedures from those with OAG. A laser peripheral iridotomy creates a hole in the peripheral iris and is often performed to eliminate pupillary block, 76 whereas a laser peripheral iridoplasty uses low-power laser burns to relieve appositional angle closure (by shrinking the peripheral iris) in cases where laser peripheral iridotomy is ineffective. 77 When combined, both treatments have been shown to be safe and effective in lowering the IOP in eyes with acute primary ACG refractory to pharmacotherapy. 78 For eyes refractory to all other medical, surgical, and laser therapies, a series of cyclodestructive procedures that damage the ciliary body epithelium and decrease the IOP by reducing the aqueous humor secretion may be the final treatment option. 79 These procedures consist of endoscopic cytophotocoagulation (Endo Optiks), continuous-wave diode laser (IRIDEX Corp), or the newest alternative, MicroPulse transscleral laser therapy (IRIDEX Corp), which selectively targets the pigmented tissue of the ciliary body epithelium. 79 Cyclodestructive procedures are also useful for the secondary forms of glaucoma, such as uveitic, traumatic, or neovascular glaucoma; however, these procedures come have considerable risks and are particularly difficult to titrate. 79

Surgical Treatment

Operations are usually performed when medical and laser treatments have failed to achieve adequate IOP reduction. Surgical options consist of the traditional, bleb-based IOP-lowering operations (trabeculectomy and tube shunt implantation) and the newer, conjunctiva-sparing MIGSs ( Table 4 ). Bleb-based operations can effectively lower IOP; however, they may develop bleb-related complications and may have high reoperation rates. As a result, the current role of traditional procedures in the era of evolving MIGSs is unclear. Surgeons’ perspectives are changing 80 ; a recent practice preferences survey from the American Glaucoma Society (2017) found that trabeculectomy has fallen out of favor, with tube shunt implantation reported as the preferred incisional surgical treatment in 7 of 8 surgical centers. 81 When prospective MIGS trials are completed, the pendulum may swing in favor of MIGSs. 80

Table 4

Surgical Procedures for the Treatment of Glaucoma a

Trabecular Outflow Resistance

The juxtacanalicular tissue within the TM is the primary source of outflow resistance in eyes with POAG, with the inner wall of the Schlemm canal serving as an additional line of resistance. 82 , 83 , 84 To improve the aqueous outflow and lower the IOP, surgeons bypass the TM by directing the aqueous flow directly into the Schlemm canal or by rerouting the fluid from the anterior chamber into the subconjunctival space.

Traditional Incisional Operations

Trabeculectomy—the “gold standard” surgical glaucoma procedure for several decades—is the creation of a partial thickness scleral flap with excision of a segment of TM to create an alternate drainage route from the anterior chamber to the subconjunctival space. 85 , 86 Trabeculectomy can produce outstanding IOP control, particularly in eyes where an IOP near the low teens is targeted to slow glaucoma progression. 87 , 88 Trabeculectomy may be performed together with cataract extraction (CE) and/or administration of mitomycin C (MMC) on the surface of the sclera to prevent postoperative conjunctival fibrosis. 89 Trab-MMC alone, trab-MMC+CE, and trab-MMC in pseudophakic eyes were found to produce comparable IOP reductions and success rates after 5 years 90 ; however, other studies have found lower success rates with trab-MMC in pseudophakic eyes, probably because of postoperative inflammation after CE. 80 , 91

Tube shunt implantation, an alternative to trabeculectomy, has gained popularity in recent years. The implantation of tube shunts, often referred to as glaucoma drainage devices (GDDs), creates a permanent sclerostomy to drain the aqueous humor into the subconjunctival space. 92 The advantages of GDDs over trabeculectomy include decreased conjunctival scarring (by diverting aqueous drainage to the equatorial region of the eye and away from the limbus) and the formation of a permanent bleb (plate tube). 92 Most GDD designs are modeled after the early Molteno implant 93 and may be valved (promotes unidirectional flow) or nonvalved (passive-acting). 92 The Ahmed Baerveldt Comparison and Ahmed Versus Baerveldt studies compared the safety and efficacy of the valveless Baerveldt 350-mm 2 GDD (Johnson & Johnson) to that of the valved Ahmed-FP7 GDD (New World Medical Inc). Both devices were effective in reducing the IOP and the need for IOP-lowering medications, although a favorable IOP decrease, medication burden reduction, and safety profile (but with a higher incidence of hypotony) were seen with the valveless Baerveldt 350-mm 2 GDD at 5 years. 94 Recent advancements in valveless GDD operation include the development of the Ahmed ClearPath GDD (New World Medical Inc) and PAUL glaucoma implant (PGI; Advanced Ophthalmic Innovations). The Ahmed ClearPath GDD has several unique design features, such as a flexible, low-lying plate with anterior suture points to increase the ease of implantation, and a prethreaded 4-0 polypropylene ripcord to mitigate the risk of hypotony that has been reported in other GDD studies. 95 The PGI GDD has a smaller plate that occupies less space in the ACA and a relatively large endplate surface area through which the aqueous humor can be absorbed. 96 Early outcome data with the Ahmed ClearPath GDD and PGI found mean IOP reductions of 43% 97 and 51.6%, 96 at 6 months, respectively.

Both trabeculectomy and GDD implantation are effective treatment options for refractory glaucoma—eyes with poor results after both pharmacotherapy and laser. A 5-year comparison of trabeculectomy and tube shunt operation found that both techniques effectively lower the IOP (trabeculectomy: 49.5%; tube: 41.4%), with the tube group having a better safety profile. 97 In surgically naïve eyes with refractory glaucoma, the Primary Tube vs Trabeculectomy study found trabeculectomy to be superior, 98 whereas the Tube vs Trabeculectomy study reported similar outcomes in both groups at 5 years postoperatively in eyes that were not surgically naïve; however, eyes in the tube group had lower failure and reoperation rates. 17 , 97 Frequent complications within the early postoperative period included choroidal effusion (Tube, 14%; Trab, 13%) and shallow anterior chamber (Tube, 10%; Trab, 10%), and late postoperative complications included persistent corneal edema (Tube, 16%; Trab, 9%) and bleb encapsulation (Tube, 2%; Trab, 6%). 17 Many of the eyes needed postoperative interventions (Tube: 25%, Trab: 70%). 17 Craven et al 16 estimated that 25% of patients treated with trabeculectomy or a tube shunt needed additional interventions to address surgical failure.

Minimally Invasive Glaucoma Surgeries

The potential complications and surgical failures seen with traditional incisional operations speak to the need for better procedures for mild-to-moderate glaucoma that are minimally invasive yet durable. This has led to the introduction of MIGSs, which have revolutionized glaucoma care over the past decade. This group of novel procedures may sufficiently lower the IOP to delay or minimize the need for traditional incisional procedures, 82 and they are more suitable for patients with mild-to-moderate glaucoma. Minimally invasive glaucoma surgeries can be performed together with cataract operation, which makes them a valuable option for glaucomatous eyes with advanced cataracts (from aging, phacomorphic glaucoma, traumatic glaucoma, etc). Unlike the traditional filtration procedures, MIGSs are relatively simple to perform because they require surgical skills similar to those required for modern-day cataract surgery, 99 and they can be performed by cataract surgeons who are not glaucoma fellowship trained. Minimally invasive glaucoma surgeries have favorable safety profiles and are less invasive than traditional incisional operations. 100

One of the management challenges with performing MIGSs lies in whether to bypass or enhance the conventional aqueous outflow 101 because the currently available MIGS devices target 1 of the 3 pressure-lowering mechanisms: (1) the trabecular outflow pathway, referring to “angle-based” MIGSs that reroute the aqueous flow toward the Schlemm canal; (2) the subconjunctival space, referring to MIGSs that create a drainage pathway, diverting the aqueous humor to the subconjunctival space; (3) the suprachoroidal space, referring to MIGSs that increase the uveoscleral pathway outflow and divert the aqueous flow toward the suprachoroidal space. 100

MIGSs Targeting the Trabecular Outflow Pathway

Approximately 50%-75% of the outflow resistance lies within the TM and the inner wall of the Schlemm canal, whereas the remainder resides within the Schlemm canal and its distal collector channels. 102 , 103 , 104 , 105 This identifies the conventional outflow pathway as an attractive first target for the treatment of glaucoma. Angle-based MIGSs take advantage of the lower resistance within the Schlemm canal and divert the aqueous flow to the canal, thereby bypassing most of the outflow resistance. Despite this, however, a significant proportion of outflow resistance remains, thereby making these procedures unsuitable for patients with severe glaucoma who require significant IOP reduction. 80 Minimally invasive glaucoma surgeries that target the trabecular outflow pathway fall within the categories of trabecular bypass implant, ab-interno canaloplasty, ab-interno trabeculotomy (AIT), goniotomy, and the more recently introduced combined goniotomy/viscodilation and trabeculotomy/viscodilation procedures.

The iStent (Glaukos Corporation), the first trabecular bypass implant, has produced excellent results when implanted into glaucomatous eyes that are well-controlled on 1 IOP-lowering medication. 80 Additional IOP lowering is observed when placing more than 1 stent, which led to the development of the iStent inject and iStent inject W. 100 A study comparing the early outcomes of the iStent and iStent inject reported favorable IOP (iStent, 4.3%; iStent inject, 19.1%) and medication reduction results (iStent, 72.2%; iStent inject, 94.1%) in the iStent inject group at 12 months, with a similar safety profile observed in both the groups. 106 Ab-interno canaloplasty is typically performed with the iTrack microcatheter (Nova Eye Medical), and a retrospective comparison with ab-externo canaloplasty (iTrack with a 9-0 prolene tensioning suture) found comparable safety and efficacy. 107 Ab-interno trabeculotomy and goniotomy procedures bring the anterior chamber, Schlemm canal, and distal collector channels into direct communication through the disruption or partial excision of the TM. 108 The Trabectome (Neomedix), a long-standing AIT procedure, uses electrocauterization to ablate the TM and has been documented to safely and effectively reduce the IOP. 108 Recent advancements in excisional goniotomy include the Kahook Dual Blade (KDB; New World Medical) and KDB Glide (New World Medical) devices. Although limited data on KDB Glide exist within the literature, several studies of KDB have shown that it has a favorable safety profile and similar effectiveness to AIT procedures. 109 , 110

Angle-based MIGS procedures are easy to perform and have favorable safety profiles, but compared with traditional trabeculectomy, they have more limited abilities to lower IOP. 101 , 111 Distal outflow (collector channels and episcleral veins), which is often overlooked in the treatment of glaucoma, may play a pivotal role in IOP control and is unaffected by canalicular-based MIGS procedures. Studies with bovine and monkey eyes have found that collector channels may alter the pressure distribution within the Schlemm canal, suggesting that the aqueous outflow may depend on the location of these distal elements. 84 , 102 , 112 Resistance within the Schlemm canal and the collector channels appears to limit the outflow increase of trabecular bypass procedures to 13%-26% and IOP reduction to the mid-teens, but a greater pressure decrease is expected if a moderate dilation of the Schlemm canal and the collector channels is achieved. 84 , 113 , 114 Goniotomy and trabeculotomy may be performed concurrently with the implantation of an ophthalmic viscosurgical device (STREAMLINE Surgical Systems, New World Medical; OMNI360 Surgical Systems, Sight Sciences) to the Schlemm canal to reduce the distal outflow resistance and promote further IOP reduction. Interim analyses of the STREAMLINE and OMNI trials have shown effective, sustained IOP reductions and meaningful medication reductions at 6 and 12 months, respectively. 115 , 116

MIGSs Targeting the Subconjunctival Space

Minimally invasive glaucoma surgeries devices within this category work similarly to trabeculectomy by diverting the aqueous humor flow directly into the subconjunctival space. 100 The main disadvantage of this strategy is the potential for subconjunctival fibrosis, which for trabeculectomy may be prevented by the intraoperative application of MMC. 100 Subconjunctival MIGS devices, which are designed based on the Hagen-Poiseuille equation, include the ab-internally implanted XEN45 gel stent (Allergan) and the ab-externally implanted PRESERFLO microshunt (Santen). Both devices produce comparable safety profiles, IOP reductions, and overall surgical success at 2 years. 117 The analysis of both implantation approaches with an experimental microfluidic system found higher outflow resistance and less predictable bleb formation with ab-interno implantation. This may affect the long-term IOP control and could direct the development of future subconjunctival-based MIGS devices. 118

MIGSs Targeting the Suprachoroidal Space

The third category of MIGSs aims to increase the uveoscleral outflow. 100 The uveoscleral pathway is not limited by the pressure “floor” formed by episcleral venous pressure; thus, diverting the aqueous humor into the suprachoroidal space could have a greater lower IOP capacity. 119 Unfortunately, current studies have yet to realize such results. After the recall of CyPass (Alcon) in 2018 because of corneal endothelial cell loss from malpositioned devices, most suprachoroidal MIGSs are still under investigation. 119 A review of recent studies indicates favorable safety profiles and effective short-term IOP reductions to the mid-teens with the iStent SUPRA (Glaukos Corporation), STARflo (iSTAR Medical), and gold implant (SOLX, Inc). Longer follow-ups and more robust trial designs are still required for the US Food and Drug Administration approval of suprachoroidal MIGSs, 120 and long-term efficacy may be limited by fibroblast migration and proliferation. 121

The pathogenesis of glaucoma is multifactorial and incompletely understood, and diagnosis methods and management strategies are constantly being improved. Treatment outcomes, safety profiles, and recovery times have improved with the introduction of MIGSs. Future work should aim to develop MIGS devices with greater IOP-lowering capabilities than traditional incisional operations.

Potential Competing Interests

The authors report no competing interests.

Acknowledgments

The authors acknowledge Jason S. Calhoun, COA, for the ultrasound biomicroscopy image ( Figure 1 ).

A new treatment for glaucoma?

A Northwestern Medicine study in mice has identified new treatment targets for glaucoma, including preventing a severe pediatric form of glaucoma, as well as uncovering a possible new class of therapy for the most common form of glaucoma in adults. 

Using gene editing, the scientists in the study developed new models of glaucoma in mice that resembled primary congenital glaucoma. By injecting a new, long-lasting and non-toxic protein treatment (Hepta-ANGPT1) into mice, the scientists were able to replace the function of genes that, when mutated, cause glaucoma. With this injectable treatment, the scientists also successfully prevented glaucoma from ever forming in one model.

A new treatment for glaucoma?

• Feinberg School of Medicine
• Global Health

A Northwestern Medicine study in mice has identified new treatment targets for glaucoma, including preventing a severe pediatric form of glaucoma, as well as uncovering a possible new class of therapy for the most common form of glaucoma in adults. 

In people with high pressure glaucoma, fluid in the eye doesn’t properly drain and builds up pressure on the optic nerve, leading to vision loss. It affects 60 million people worldwide and is the most common cause of blindness in people over 60 years old. 

While there are a few treatments available for open angle glaucoma, the most common form of glaucoma in adults (eye drops, oral medication, laser treatments), there are no cures, and a severe form of glaucoma in children between birth and three years old known as primary congenital glaucoma can only be treated with surgery.  

“Although primary congenital glaucoma is much rarer than open angle glaucoma , it is devastating for children,” said corresponding author  Dr. Susan Quaggin , chief of nephrology and hypertension in the Department of Medicine at Northwestern University Feinberg School of Medicine. “New treatments and new classes of treatments are urgently needed to slow vision loss in both forms.

Using gene editing, the scientists in the study developed new models of glaucoma in mice that resembled primary congenital glaucoma. By injecting a new, long-lasting and non-toxic protein treatment (Hepta-ANGPT1) into mice, the scientists were able to replace the function of genes that, when mutated, cause glaucoma. With this injectable treatment, the scientists also successfully prevented glaucoma from ever forming in one model. This same therapy, when injected into the eyes of healthy adult mice, reduced pressure in the eyes, supporting it as a possible new class of therapy for the most common cause of glaucoma in adults (high intraocular pressure open angle glaucoma).

The study, “Cellular crosstalk regulates the aqueous humor outflow pathway and provides new targets for glaucoma therapies,”  was published Oct. 18 in the journal Nature Communications.

60 million people worldwide are affected by glaucoma

The next step is to develop the appropriate delivery system for the successful new protein treatment in patients and bring it to production, Quaggin said. 

Additionally, the scientists used bioinformatics and single cell RNA sequence data to understand and identify glaucoma pathways that can be explored in the future for additional therapeutic targets for the disease, such as ones that regulate communication with a specialized blood vessel in the eye (Schlemm’s canal) that is important for draining fluid and maintaining normal eye pressure. 

“Having a treatment that can promote remodeling and/or growth of a defective Schlemm’s canal to treat glaucoma would be fantastic,” Quaggin said. “These studies are the first step to that goal.

“Our hope is that this study leads to the first targeted therapy that effectively promotes (aqueous humor) fluid outflow from the front of an eye, reversing the underlying biologic defect in patients with glaucoma.”

Other Northwestern co-authors are  Ben Thompson  (first),  Dr. Jing Jin ,  Pan Liu  and medical student Raj Purohit. This study builds on major teamwork and an ongoing collaboration with University of Madison-Wisconsin co-authors Terri Young and Stuart Thomson.

Funding for the study was provided by the National Eye Institute (grant numbers R01 EY025799, P30 EY016665 and R01 EY014685), the National Heart, Lung and Blood Institute (grant number R01 HL124120), the National Institutes of Health Office of the Director (grant number 1S10OD025120), the National Institute of Diabetes and Digestive and Kidney Diseases (grant number P30 DK114857), the National Cancer Institute (grant number CCSG P30 CA060553), and Research to Prevent Blindness.

Editor’s Picks

Construction crews work to protect Northwestern’s South Campus shoreline

New biomaterial regrows damaged cartilage in joints, chronicling chicago, one column at a time, related stories.

HIV prevention medication users fear being stigmatized as ‘promiscuous’

Neurological symptoms in long covid patients persist up to three years, cancerous uterine tumors more aggressive in black patients than white patients.

Thank you for visiting nature.com. You are using a browser version with limited support for CSS. To obtain the best experience, we recommend you use a more up to date browser (or turn off compatibility mode in Internet Explorer). In the meantime, to ensure continued support, we are displaying the site without styles and JavaScript.

• View all journals
• Explore content
• About the journal
• Publish with us
• Sign up for alerts
• Open access
• Published: 23 August 2024

The use of artificial neural networks in studying the progression of glaucoma

• Filip Târcoveanu 1 ,
• Florin Leon 2 ,
• Cătălin Lisa 3 ,
• Silvia Curteanu 3 ,
• Andreea Feraru 4 ,
• Kashif Ali 5 &
• Nicoleta Anton   ORCID: orcid.org/0000-0002-4987-5049 1  

Scientific Reports volume  14 , Article number:  19597 ( 2024 ) Cite this article

1 Altmetric

Metrics details

• Applied mathematics
• Eye diseases
• Optic nerve diseases

In ophthalmology, artificial intelligence methods show great promise due to their potential to enhance clinical observations with predictive capabilities and support physicians in diagnosing and treating patients. This paper focuses on modelling glaucoma evolution because it requires early diagnosis, individualized treatment, and lifelong monitoring. Glaucoma is a chronic, progressive, irreversible, multifactorial optic neuropathy that primarily affects elderly individuals. It is important to emphasize that the processed data are taken from medical records, unlike other studies in the literature that rely on image acquisition and processing. Although more challenging to handle, this approach has the advantage of including a wide range of parameters in large numbers, which can highlight their potential influence. Artificial neural networks are used to study glaucoma progression, designed through successive trials for near-optimal configurations using the NeuroSolutions and PyTorch frameworks. Furthermore, different problems are formulated to demonstrate the influence of various structural and functional parameters on the study of glaucoma progression. Optimal neural networks were obtained using a program written in Python using the PyTorch deep learning framework. For various tasks, very small errors in training and validation, under 5%, were obtained. It has been demonstrated that very good results can be achieved, making them credible and useful for medical practice.

Similar content being viewed by others

The impact of artificial intelligence in the diagnosis and management of glaucoma

An objective structural and functional reference standard in glaucoma

Deep learning in glaucoma with optical coherence tomography: a review

Introduction.

Artificial intelligence (AI) is currently one of the fastest-evolving fields, notable for its diverse techniques and applications areas. Modern medicine requires the acquisition, analysis, and processing of large amounts of data to address complex clinical tasks. In this context, AI serves as a valuable tool for clinicians, which aids in therapeutic decision-making and provides predictions that extend beyond available data.

Artificial neural networks (ANN) are modelling and classification tools with the ability to learn different types of relationships from available data. They can approximate any continuous nonlinear function, and are effective for modelling complex nonlinear systems. ANNs draw inspiration from biological neurons and the structure of the nervous system. A biological neuron comprises three main components: the cell body, dendrites, and axon, with synapses facilitating connections. These elements have artificial counterparts in the basic units of neural networks. An artificial neuron receives inputs, possibly from other neurons, each associated with a connection weight. It calculates a weighted sum of these inputs and then applies an activation function to compute the output, which can be further transmitted to other neurons.

Glaucoma is a chronic, progressive, irreversible, multifactorial optic neuropathy that mainly affects elderly people, requiring early diagnosis, individualized treatment, and lifelong monitoring.

There are many approaches in the literature that report the study of glaucoma evolution using neural networks, original articles, and numerous reviews. Mainly, there are two categories of approaches based on the data’s origin: from images or medical records , with the former being more numerous given the ease and accuracy of image acquisition. There are specific AI algorithms for segmenting and automatically enhancing ocular images from optical coherence tomography (OCT) and processing fundus images.

On the other hand, using data from medical records is more laborious because it requires comprehensive datasets, but, by processing these data with AI algorithms, more information can be obtained, useful for the ophthalmologist, as the dataset can include numerous analyses, possibly influencing glaucoma.

The most common use of artificial intelligence techniques in ophthalmology has been in the early diagnosis of glaucoma when there are diagnostic doubts. Neural networks, in particular, have played a significant role in determining the need for early antiglaucoma therapy to prevent disease progression.

The first studies on using ANNs in interpreting incipient perimetric lesions in glaucoma belong to Anton et al. 1 . The authors concluded that neural networks can differentiate incipient lesions caused by glaucoma from those caused by other diseases with 97% accuracy.

In 2005, Bowd et al. 2 used two algorithms (RVM—Relevance Vector Machine and SVM—Support Vector Machine) for classifying healthy eyes and eyes affected by glaucoma, using information based on the retinal nerve fiber layer (RNFL) and thickness measurements obtained through scanning laser polarimetry (SLP).

Previously, the same author 3 had used different artificial intelligence tools to determine the evolutionary changes in the visual field in glaucoma patients and predict the glaucoma stage. Similar concerns were held by Simon et al. 4 , Hernández et al. 5 , Grewal et al. 6 , Parsaei et al. 7 , who used various artificial intelligence tools (neural networks with different learning algorithms) to determine possible visual field progression in glaucoma patients.

Zhu et al. 8 developed a neural network using a Bayesian-type function to establish the relationship between structure and visual function in glaucoma; the results showed that the network could improve the prediction of visual function. Based on the analysis of the specialized literature, they showed that diabetes is an individual risk factor for the development of open-angle glaucoma.

An interesting study on the utility of ANNs in ophthalmology is found in a doctoral thesis at UMF Iași 9 . For the first time, artificial neural networks are used to demonstrate the existence of a relationship between glaucoma and diabetes, as well as in predicting the progression of diabetic ocular changes (diabetic retinopathy) in patients with glaucoma and diabetes. The constructed neural models demonstrated the possibility of using them in predicting mean deviation (MD) deterioration, with the best results obtained using Jordan Elman-type neural networks 10 .

In a recent review, Devalla et al. 11 discuss the role of AI in glaucoma diagnosis and prognosis. They show that while early studies relied on simple ANNs to detect glaucoma using perimetric data (visual field), modern deep learning (DL) systems have successfully exploited high-resolution images from fundus photography and OCT. The use of AI algorithms can help, by collecting data from multiple tests, to detect anomalies, and perform relevant screenings. This reduces the workload for clinicians, minimizes diagnostic errors, and improves the quality of care for glaucoma patients. This will lead to early detection of glaucoma and promote research and development of new drugs in treatment. Also, pre-perimetric changes can be identified, likely leading to advances in research and clinical practice.

Bizios et al. 12 conducted a study comparing artificial neural networks and support vector machines for glaucoma diagnosis, measuring the thickness of the retinal nerve fiber layer obtained with OCT. Both algorithms provided satisfactory results.

Yousefi et al. 13 developed a DL-based algorithm that detected visual field (VF) progression earlier than conventional strategies. Although lowering intraocular pressure (IOP) has proven to be therapeutically effective in delaying glaucoma progression, some authors have shown that disease progression is still inevitable, suggesting that optimal treatment regimens for different forms of glaucoma have not yet been achieved.

Guangzhou et al. 14 published a study addressing the task of optic disc classification in glaucoma. A total of 163 eyes from 105 glaucoma patients were utilized, divided into training and testing datasets, to establish classification criteria to aid clinical assistance in glaucoma management. Ocular images were obtained using optical coherence tomography techniques and retinal perfusion was investigated using laser flowgraphy. A total of 91 parameters were extracted from each eye, and general patient information was included. Classification methods tested in the study included neural networks, naive Bayes classifiers, support vector machines, and gradient boosted decision trees. The results of the performance comparison of the three classification methods showed that neural networks achieved the best classification with an accuracy of 87.8% using only nine ocular parameters.

The study of glaucoma progression, through simulation with artificial intelligence tools, consisted of structural , functional , or mixed approaches .

Functionally, several studies have been conducted aiming to compare the performance of various machine learning classifiers with global indices from STATPAC and human experts 15 , 16 , 17 .

Visual fields are used as input data, and the classifiers’ ability to predict the development of abnormal fields in subsequent examinations of hypertensive eyes (OHT) with initially normal visual fields is evaluated, as well as to identify and quantify areas of progression in standard automated perimetry fields 18 .

In addition, some studies compare the performance of machine learning algorithms with traditional algorithms in detecting glaucomatous visual field defects. The studies employ various techniques such as multi-layer perceptrons (MLP), support vector machines (SVM), Gaussian classifiers (MoG), generalized Gaussian classifiers (MGG), variational Bayesian independent component analysis models (vB-ICA-mm), and deep convolutional neural networks (CNN) 19 , 20 . Results indicate that machine learning algorithms can improve STATPAC global indices and identify glaucomatous changes earlier in standard visual fields, sometimes up to 3.92 years earlier than traditional methods.

It is concluded that machine learning algorithms have the potential to reduce testing time by decreasing the number of visual field location measurements. Moreover, the results show that machine learning algorithms have higher diagnostic performance compared to traditional algorithms, and the proposed glaucoma progression index based on CNN and machine learning outperforms global, regional, and point indices in detecting longitudinal visual field progression in glaucoma.

On the other hand, various studies have focused on detecting structural progression in glaucoma by analysing OCT and fundus images with AI and DL algorithms.

The first study mentioned was conducted by Christopher et al. 21 , which used PCA to identify new glaucoma-related structural features based on SS-OCT images. This study obtained significantly higher accuracy for glaucoma detection with an AUC of 0.95.

The second study, conducted by Medeiros et al. 22 , used a deep learning convolutional neural network to predict average RNFL thickness based on fundus photographs, achieving good performance in discriminating glaucomatous eyes from healthy eyes.

Devalla et al. 11 reviewed the role of AI in glaucoma diagnosis and prognosis, discussing the advantages and challenges of using AI systems in clinics and predicting potential areas of progress in the future in this field.

Another study, conducted by Ran et al. 23 , discussed the potential clinical impact of DL models and identified areas for future research.

In 2022, Hood et al. 24 described an automated method for detecting glaucoma without the need for a clinician’s judgment, using a simple anatomical artifact model to help distinguish artifacts from actual glaucomatous changes in OCT probability maps. In the same year, Li et al. 25 developed a deep learning system for the automatic detection of angle closure in AS-OCT images. The results showed that the deep learning system performed better compared to the one based on quantitative features.

Chaurasia et al. 26 conducted, in 2022, a meta-analysis to determine the overall performance of AI in glaucoma diagnosis and identify potential factors affecting their implementation. The results showed that AI algorithms performed well in detecting glaucoma.

The specialized literature contains various hybrid studies (functional and structural approaches) that have used artificial intelligence for diagnosing and managing glaucoma. The studies included various combinations of techniques, such as ANN, optical coherence tomography (OCT), standard automated perimetry (SAP), short-wavelength automated perimetry (SWAP), and confocal scanning laser ophthalmoscopy (CSLO).

As early as 1996, Brigatti et al. 27 used a two-layer back-propagation network to assign each eye an estimated probability of having glaucoma and correctly identified 88% of all eyes with 90% sensitivity and 84% specificity.

Another study conducted by Mardin et al. 28 used six different combinations of morphological data and visual field data to diagnose the disease, applying machine learning classifiers. The combination of Heidelberg Retina Tomograph (HRT), standard Octopus output (HRT/PERI1), and standard deviations (HRT/PERI2) achieved the highest sensitivity (95%) and specificity (91%).

Other studies 16 , 29 , 30 have investigated the use of structural and functional measurements, such as OCT and standard automated perimetry SAP, using machine learning classifiers to identify eyes with glaucoma with high accuracy. These studies have shown that the optimized combination of data can improve diagnostic accuracy. In addition, some studies have investigated the use of an ANN-based approach to differentiate the diagnosis of open-angle glaucoma from glaucoma suspicion without a visual field test 18 and found that this approach can be a cost-effective screening tool. Overall, these studies have shown that AI can be a powerful tool for diagnosing and managing open-angle glaucoma.

Modern AI algorithms are specifically adapted to extract significant features from complex and high-dimensional data for screening, diagnosis, management, and follow-up of glaucoma based on the interpretation of functional and/or structural parameters 12 , 31 .

Numerous other studies in specialized literature demonstrate the successful use of artificial intelligence tools in Ophthalmology, such as evaluating visual fields, optic nerves, retinal nerve fiber layers, thus providing better accuracy in identifying glaucoma progression and retinal changes in diabetes. The main advantages of using these techniques in medical diagnosis are: the ability to process large amounts of data, low probability of overlooking relevant information, and reduced time for diagnosing.

Singh et al. 32 , 33 , 34 , 35 , 36 have important contributions in identification of symptoms for glaucoma and glaucoma affected retinal images, using Machine Learning or Deep Learning Techniques. They show that important steps in applying machine learning techniques for automatic detection of glaucoma are retinal image acquisition, image preprocessing, feature extraction, classification of symptoms of glaucoma affected image, and evaluating the performance 33 . In order to overcome some difficulties related to conventional diagnostic methods utilized by ophthalmologists, Singh et al. 36 propose a computer-assisted diagnosis systems by using machine learning approaches to classify retinal pictures as “healthy” or “infected”. Different AI methodologies are explored for glaucoma detection: Differential Evolution based multi objective feature selection technique 32 , Particle Swarm Optimization 34 , 36 , Artificial Bee Colony, Binary Cuckoo Search, Random Forest, Support Vector Machine, used individually or in ensembles 36 . The efficiency of these methos has been proved by the obtained results—accuracy with value between 0.91 and 0.99.

The objective of the paper involves creating a database with medical records of glaucoma patients and formulating problems to determine relevant input and output parameters related to glaucoma progression. This includes selecting or developing suitable AI methodologies (artificial neural networks) and software, and comparing the effectiveness of different methods. Additionally, the results are compared with existing literature to highlight the benefits of simulations for glaucoma patients and ophthalmologists' decisions.

An important contribution of the research is the use of data from medical records, containing multiple information from investigations and visits, compared to most studies in the literature that address this topic and that use data from acquired images. The use of data from medical files allows the consideration of various information with possible influence in glaucoma. ANNs are developed using the NeuroSolutions and PyTorch frameworks. The best results are obtained in the second case, when the errors in the testing phase are very small, which gives credibility to the models and high chances that they can provide reliable and useful information to the specialist doctor. Therefore, the main original contribution of this work lie in the applied methodology, which includes extensive and diverse data extracted from the medical records of glaucoma patients, the formulation of problems to evaluate glaucoma progression, and the implementation of various neural network models using PyTorch .

The work is organized in the following sections: " Introduction ", experimental part that includes " Datasets " and " Modelling methodology ", " Results ", " Discussion ", " Conclusions " and References.

Most simulation approaches in the literature for studying glaucoma use data acquired from images. These have several advantages related to the fact that they are numerous and accurate since the images are obtained with high-performance devices, and any number of points desired by the user can be read from an image.

Much more difficult and, therefore, less used, is the collection of data from medical records, each entry representing a consultation and results of a whole set of tests. Time and accuracy are the major difficulties in building the dataset. However, once such a task is formulated and addressed, the advantages become evident through the information that can be obtained about the factors influencing the disease, its progression, and even the most suitable treatment.

In this paper, glaucoma, a chronic, progressive, irreversible, and multifactorial optic neuropathy, is analysed using data from medical records.

To create the dataset, medical records of glaucoma patients examined at Countess of Chester Hospital in the UK were used, and the study was retrospective, following patients who had at least three consecutive check-ups between 2018 and 2021 and underwent all necessary investigations to be included in the study. Precise inclusion and exclusion criteria 37 were established, and two distinct datasets were created.

Inclusion criteria for this study were: open-angle glaucoma or intraocular hypertension, with or without associated diabetes. Only patients who had the following investigations recorded at each of the three visits were selected: visual field assessment, macula and optic nerve OCT, intraocular pressure measurement on both eyes, visual acuity measurement, pachymetry, HbA1c measurement, and baseline IOP documented at the time of glaucoma diagnosis.

Criteria for the diagnosis of primary open-angle glaucoma (POAG) considered were: age over 40 years, IOP greater than 21 mm Hg without treatment at the time of diagnosis, open-angle in the anterior chamber upon gonioscopy, typical optic nerve damage in glaucoma (C/D ratio > 0.5), abnormal visual field (by Humphrey Field Analyzerperimetry), and retinal nerve fiber layer thickness impairment measured by Heidelberg OCT. Patients with ocular hypertension (OHT) were also included, with diagnostic criteria such as: age over 40 years, IOP greater than 21 mm Hg without treatment at the time of diagnosis, open-angle in the anterior chamber upon gonioscopy. These patients may have a normal visual field, normal optic nerve appearance (C/D ratio < 0.5), and normal retinal nerve layer thickness measured by OCT.

Our approach, based on modeling with artificial neural networks, starts with the following considerations regarding to normal values ​​and deviations from them, for the quantities considered in the evaluation of glaucoma progression:

Mean Deviation (MD), with normal range greater than − 2 dB, measures the overall depression of the visual field. A value less than − 2 dB suggests abnormality, with more negative values indicating more severe loss.

Pattern Standard Deviation (PSD), with normal range close to 0 dB, indicates the irregularity of the visual field loss. Higher values suggest more localized defects; normal values are usually near 0, with higher values indicating glaucomatous damage.

Visual Field Index (VFI), with normal range of 100%, represents the percentage of normal visual function. Lower percentages indicate more significant visual field loss; normal values are near 100%, with lower values indicating more severe glaucoma.

Retinal Nerve Fiber Layer (RNFL), with normal range of 80–100 µm, is a critical indicator of glaucomatous damage. Values below the normal range suggest thinning due to glaucoma.

Intraocular Pressure (IOP), with normal range of 10–21 mmHg, is a significant risk factor for glaucoma. Values above 21 mmHg may indicate glaucoma or risk of developing it, although some patients with normal IOP can also have glaucoma (normal-tension glaucoma).

The medical investigations represented the 47 attributes, from which the inputs and outputs were selected for the modeling action, corresponding to the formulation of various problems through which the progression of glaucoma could be evaluated. Fifty patients were included in the database, each with three check-ups, resulting in 150 visits with 300 entries (for both eyes). These constitute the records or instances of the database. From this information, two datasets were formed: considering the visits of a patient as distinct records (dataset 1) and attributing these visits to the patient who performed them (dataset 2).

Compared to the modeling on dataset 1, where the values corresponding to each visit were considered as separate records (as if they belonged to different patients), in dataset 2 these values were attributed to the same patient, becoming additional inputs. The database thus configured includes 100 records (instances).

Various tasks were formulated in the sense of choosing different sets of input and output parameters because in this way there is the possibility of obtaining multiple and varied information. From this point of view, factors that may contribute to the progression of glaucoma were selected, such as: patient age, glaucoma age, intraocular pressure, central corneal thickness, phakic or pseudophakic status, as well as factors monitoring the severity of the disease: visual field values (VFI), mean deviation (MD), retinal nerve fiber layer thickness (RNFL). A binary encoding (0 and 1, Yes and No) was used to represent glaucoma progression. Inputs and outputs were different for different datasets, the distinct approaches aiming to obtain additional information about the studied conditions. Thus, modelling the dependence between functional and structural glaucomatous damage and common glaucoma risk factors was carried out in various ways, the results obtained providing information about the severity of the disease and recommended treatment. Alternatives for testing different influencing factors in glaucoma represent a study method of this condition, with possibilities to extract different information.

Modelling methodology

The tools used to monitor the evolution of glaucoma were artificial neural networks, designed using two methods: successive attempts with the NeuroSolutions program and an original implementation in PyTorch using the Python programming language.

Different types of neural models were tested, for which parameters that might influence the modelling results were varied: network topology, learning method, data partitioning into training and validation sets, transfer functions, the number of training epochs. The first model design method was based on trial and error , the error being evaluated at each variant, i.e., combination of parameters.

The first attempt to predict glaucoma progression using neural networks was based on the NeuroSolutions program, a specialized software product developed by Neurodimension . It allows the design, training, and validation of ANN models, as well as making predictions with them. Featuring a graphical user interface, the program can be easily operated by non-specialists who should nevertheless possess basic knowledge ANNs. It should be noted that the essential elements ensuring success in modelling are the quality of the dataset and the formulation of the task. Specifically, the data must be accurate, sufficiently large, and uniformly distributed across the investigated domain. In terms of task formulation, the inputs and outputs must be correctly chosen, considering the interdependencies and functional relationships between parameters.

A second methodology applied in this article is based on original programs written in Python using the PyTorch deep learning framework. PyTorch has implementations of several gradient-descent optimization algorithms, such as Adam (Adaptive Moment Estimation), RMSprop (Root Mean Square Propagation), or SGD (Stochastic Gradient Descent). In our case, the Adam algorithm was used. The parameters of the algorithm, such as the learning rate and regularization factor, should also be set. PyTorch has already implemented several loss such functions, but the choice here depends on the nature of the task. For example, for a classification task, the cross-entropy function can be used, while for a regression task, the mean square error can be used.

The training and prediction sections can be repeated a number of times to statistically estimate the performance of a combination of parameters. Both data partitioning into training and testing sets and optimization algorithm results are stochastic, so a statistical evaluation is beneficial.

In the case of both methods, different problems were formulated, aiming to track glaucoma progression through structural, functional, or combined (hybrid) parameters. By considering different input–output combinations, the attempts were made to highlight factors with significant influence on disease progression, these factors coming from data included in the patient files. In this way, through the predictions of neural models (credible predictions in the case of models with small testing errors), the ophthalmologist can have additional information and suggestions regarding the treatment to be followed.

Institutional review board

The study was approved by the Ethics Commission of the University Hospital Sf. Spiridon Iasi, approval no. 38/24.04.2023 and Countess of Chester Hospital, Liverpool Rd, Chester approved no. 28.04.2023, in compliance with ethical and deontological rules for medical and research practice. The study was conducted in accordance with the Helsinki Declaration.

Informed consent

Informed consent was obtained from all subjects involved in the study.

Neural networks models determined using NeuroSolutions

Different tasks were formulated and solved on the two datasets, denoted 1 and 2, distinguished by differentiating the parameters corresponding to the three visits (dataset 2).

The following case studies were performed using dataset 1, which neglects the third visits.

Case study 1: modelling the dependence between functional and structural glaucomatous damage and common glaucoma risk factors. The factors that can contribute to glaucoma progression (patient age, glaucoma age, IOP, Base IOP, IOL, CCT) were selected as inputs and the variables that monitor the severity of the disease (VFI, PSD, RNFL) were selected as outputs. The lower the VFI, the higher the PSD, and the lower the RNFL values, the more advanced the glaucoma.

The testing plan included various neural network methodology options, evaluated by the mean square error ( MSE ) and the correlation coefficient ( r ). Specifically, we used neural networks with one or two layers and a variable number of intermediate neurons, different percentages for splitting data between training and testing sets (85–15%, 80–20%, 75–25%, 70–30%), training algorithms (backpropagation with momentum and Levenberg–Marquardt), and different numbers of training epochs (500, 2000, 10000, 20000).

An MLP(6:24:8:3) network (multilayer perceptron with 6 inputs and 3 outputs, and 2 hidden layers with 24 and 8 neurons) recorded very small training errors: an MSE ranging from 0.00012 (minimum) to 0.0163 (maximum), and an r ranging from 0.9493 to 0.99958. The average relative errors in training and testing phases for the three outputs were: 0.34% and 5.44% for VFI, 0.18% and 9.85% for PSD, and 1.96% and 14.79% for RNFL.

The results obtained are considered satisfactory in the context of formulating a complex task where the simultaneous prediction of three real-value outputs was sought: VFI, PSD, RNFL. But to increase the degree of confidence, other methods that improved the generalization capacity were further tested.

Case study 2: modelling the dependence between IOP, CCT, and the degree of visual field loss. In this study, a simpler model was sought with a single output, namely VFI. This parameter reflects the integrity of the visual field and extrapolates the degree of visual field loss in glaucoma. Three inputs were considered: baseline IOP, current visit IOP, and corneal thickness (pachymetry), as these three characteristics help to form an idea about the desired pressure control. The working parameters were: 75% of data for training and 25% for testing, the use of the Levenberg–Marquardt algorithm for training, the hyperbolic tangent ( tanh ) as the activation function, and different types of network architectures: multilayer perceptron, generalized feed-forward network, and Jordan-Elman network. The choice of the working characteristics for these models was made after comparing the results (i.e., errors) of successive trials.

The best ANN model was MLP(3:24:8:1) with an average percentage error of 0.20% in the training phase and 3.23% in the testing phase; these results can be appreciated as very good.

Figures  1 and 2 show the training and testing phases, respectively comparisons between VFI obtained from clinical observations and predictions of the best neural model.

VFI—comparisons between clinical observations and predictions of the MLP(3:24:8:1) network in the training phase.

VFI—comparisons between clinical observations and predictions of the MLP(3:24:8:1) network in the testing phase.

In Fig.  1 , corresponding to the training phase, the placement of the points on the line proves that the response of the neural network coincides with the VFI values obtained from clinical observations. Figure  2 reflects the fact that VFI obtained experimentally (from clinical observations, blue rectangles) has values close to those provided by the neural network (red rectangles). This being the testing phase, the good generalization capacity of the neural model is demonstrated, thus the high probability of predicting the evolution of glaucoma.

Case study 3: modelling the dependency between IOP, CCT, and the degree of optic nerve damage. Similar to the previous case study, the same inputs were maintained, namely baseline IOP, current visit IOP, and corneal thickness, and the output considered was the thickness of the retinal nerve fiber layers (RNFL), which reflects the degree of structural damage in glaucoma based on the same three inputs.

The best model, MLP(3:24:8:1), was obtained with MSE  = 0.0000845, r  = 0.9997, average training error = 0.11%, and average testing error = 5.44%, which also represent satisfactory results.

In this case, where RNFL was considered as the only output, the average testing error was 5.44%, compared to the modelling in case study 1, where RNFL was one of the three outputs, and the testing error was 14.79%. This improvement comes from changing the task formulation of mapping the inputs to outputs, which led to more accurate predictions.

Case study 4, inputs: patient age, patient sex, glaucoma age, diabetes age, HbA1c, B-IOP, IOP, IOL, CCT, output: VFI. The best model was MLP(9:36:9:1), obtained with MSE  = 0.000001, r  = 0.999984, and E p  = 0.05%.

Case study 5, inputs: patient age, patient sex, glaucoma age, diabetes age, HbA1c, B-IOP, IOP, IOL, CCT, output: MD. The best model was MLP(9:36:9:1), obtained with MSE  = 0.000004, r  = 0.999933, and E p  = 0.05%.

Case study 6, inputs: patient age, patient sex, glaucoma age, diabetes age, HbA1c, B-IOP, IOP, IOL, CCT, output: PSD. The best model was MLP(9:36:9:1), obtained with MSE  = 0.000004, r  = 0.999633, and E p  = 2.088632%.

Case study 7, inputs: patient age, patient sex, glaucoma age, diabetes age, HbA1c, B-IOP, IOP, IOL, CCT, output: RNFL. The best model was MLP(9:36:9:1), obtained with MSE  = 0, r  = 1, and E p  = 0.00265%. Figures  3 and 4 show graphical results for this case, specifically the training and testing phases for the best model. A correlation coefficient of 1 demonstrates that the model has well learned the relationships and patterns in the data (Fig.  3 ). In the testing phase, the r coefficient value is also close to 1, which gives credibility to the determined model (Fig.  4 ). The figures include the coefficient of determination R 2 as a performance metric, which is the square of the correlation coefficient r .

Experimental RNFL the predictions provided by an MLP(9:36:9:1) in the training phase.

Experimental RNFL and the predictions provided by an MLP(9:36:9:1) in the testing phase.

For cases studies 4–7, the presented results are very good, and the obtained models are relatively simple.

Dataset 2 includes information related to the progression of glaucoma in 50 patients aged between 50 and 94 years, 52% of whom are male, with a glaucoma history ranging from 0 to 28 years. In this case, the modelling was carried out under the following conditions. The input data considered were: VFI, MD, PSD, and RNFL, measured at each of the three evaluation visits for both eyes, and indexed as 1, 2, and 3. The output variable, glaucoma progression, was encoded as (1, 0) for “yes”(there is progression in glaucoma) and (0, 1) for “no” (there is no progression in glaucoma). Multilayer perceptron neural models were constructed with 12 inputs, 2 outputs, and one or two hidden layers of neurons. The number of neurons used in the hidden layers varied between 6 and 30. The tanh activation function and the backpropagation with momentum algorithm were used.

The best model, MLP(12:24:6:2), had the following performance in the training phase:   MSE  = 0.000223, r  = 0.999782, and E p  = 0.192070. For the testing phase, several results are given in Table 1 . The probability of a correct answer for the prediction of glaucoma progression was 93.3%, which is a satisfactory accuracy.

Neural networks models determined using PyTorch

PyTorch is an open-source deep learning framework developed by Facebook’s AI Research lab. It provides a flexible and efficient platform for building and training neural networks. It supports computations on both CPU (central processing unit/computer’s processor) and GPU (graphics processing unit), which enables faster processing of large datasets. It has an intuitive design and a robust library of pre-built functions and models, which make it popular among researchers and practitioners in machine learning. One of its key features is the dynamic computation graph, i.e., a network where nodes represent mathematical operations and edges represent the data flow between these operations. This graph structure is essential for defining and optimizing complex computations in neural network models. Automatic differentiation is another fundamental feature of PyTorch . The “autograd” system automatically tracks all operations on tensors (multi-dimensional arrays) and enable the automatic computation of gradients during gradient-descent optimization such as backpropagation and its more advanced versions. By applying the chain rule to the computational graph, PyTorch efficiently calculates the derivatives needed to update model parameters. This capability simplifies the training process, making it more straightforward to implement and optimize deep learning models.

The first task to be solved uses dataset 2, the dataset where three visits were considered for a patient. In this case, there are 12 inputs: VFI 1 , VFI 2 , VFI 3 , MD 1 , MD 2 , MD 3 , PSD 1 , PSD 2 , PSD 3 , RNFL 1 , RNFL 2 , RNFL 3 , and the output of the model is represented by glaucoma progression, encoded by YES or NO. The configured dataset includes 100 records (instances).

The proposed architectures have two fully-connected layers with different number of neurons in the two hidden layers, using the tanh activation function. Since this a binary classification problem expressed using one-hot encoding, a softmax layer is used before the output. Softmax is a function that converts an array of real-valued numbers into a probability distribution.

The training algorithm is Adam (Adaptive Moment Estimation), a well-known optimization algorithm based on the principles of gradient descent. It computes individual adaptive learning rates for each network parameter using estimates of first and second moments of gradients. The specified learning rate is 0.001, which defines the size of steps taken to update the model parameters during training. A smaller learning rate, such as 0.001, implies smaller steps, often leading to more stable convergence. Weight decay is set to 10 −5 , serving as a regularization term during training. It helps prevent overfitting by adding a penalty to the model weights.

In order to determine the best architecture, fivefold cross-validation (CV) was implemented, meaning that in one iteration 80% of the data is used for training and 20% for testing. This methodology helps assess the generalization capabilities of the neural models. We also included a statistical approach of evaluating the architectures, running the CV process multiple times (in this case, 200 times). This decreases the influence of particular data splits in a single CV operation. The number of training epochs ranged between 50 and 100. It was observed that a larger number of epochs led to overfitting.

The best result was provided by a feedforward network, MLP(12:33:11:2), with two intermediate layers containing 33 and 11 neurons, respectively. The average error for the training phase was 0.09%, and 0.19% for testing, which represents a very good result.

After this process of identifying the best configuration, we used the process of splitting the data into 80% for training and 20% for testing, and selecting the model with the lowest testing error for prediction out of 20 repeated trials.

Another task to be solved with this method was modelling the evolution of glaucoma reflected by the values of the third visit. In this case, glaucoma progression was tracked by predicting the value of a specific parameter at the third visit of the patient included in the dataset. These parameters were, in turn, VFI 3 , MD 3 , PSD 3 , RNFL 3 , and IOP 3 . Considering 16 input sizes, 5 tasks were formulated and solved, corresponding to the mentioned outputs (Fig.  5 ).

Modelling glaucoma progression by predicting the results of the third visit.

The training methodology is similar, but in this case we are dealing with regression problems, therefore the architectures involve a fully-connected layer with 1 output instead of a softmax layer.

Table 2 displays the results of the runs, showing the best determined network along with the corresponding errors observed during the training and testing phases.

Next, some detailed results are presented, either in graphical format or in tables. For example, for MD 3 , Figs.  6 and 7 compare the experimental data (from the dataset) with those provided by the determined neural model, both for the training phase and for the testing phase, in an Microsoft Excel trendline format.

MD 3 experimental and obtained through neural modelling in the training phase.

MD 3 experimental and obtained through neural modelling in the testing phase.

In other words, comparing the experimental data with those provided by the best neural model involves evaluating the distance of the points from the straight line. The closer the points are to the straight line (or even on the line), it means that the values of the two datasets (experimental/investigations and prediction) are closer. The graphs also show the coefficient of determination R 2 whose value is close to 1.

The results presented in Table 2 (row 2) and Figs.  6 and 7 show that MD at the third visit can be approximated with an accuracy of 98.75% based on the values recorded at the previous visits.

For the PSD parameter (task 3 in Table 2 ), Figs.  8 and 9 show comparatively some of the training and testing data, this time through bar graphs.

PSD 3 experimental and obtained through neural modelling in the training phase.

PSD 3 experimental and obtained through neural modelling in the testing phase.

An accuracy of 99.25% for estimating PSD at the third visit represents a reliable prediction for this parameter.

For task 4, with the output being RNFL at the third visit, the comparison of the values from the dataset with those provided by the model (for a sample of data) is presented in table format (Table 3 ), which allows visualization and comparison of values for each pair of data.

As in the previous tasks, the RNFL prediction at the third visit can be made, based on the determined model, with very high accuracy, specifically 98.33%, which means a faithful estimation of the structural progression.

For the last task (noted 5 in Table 2 ), with the IOP output at the third visit, the experimental vs. NN comparisons are given in graphical form in Figs.  10 and 11 .

IOP 3 experimental and obtained through neural modelling in the training phase.

IOP 3 experimental and obtained through neural modelling in the testing phase.

Another proposed task consisted in developing a neural network model to check if, during the three visits, it was necessary to add more drops to control glaucoma or if surgical manoeuvres were needed to control pressure. Thus, 11 inputs were used: base IOP, IOP 1 , IOP 2 , IOP 3 , VFI 1 , VFI 2 , VFI 3 , RNFL 1 , RNFL 2 , RNFL 3 , and CCT, with a single output, TRATMOD (YES or NO): “YES” for the situation where it was necessary to add more drops to control glaucoma or surgical manoeuvres were needed to control pressure, and “NO” for cases that were stable.

The PyTorch program determined an MLP(11:25:2) model with an average training error of 0.40% and an average testing error of 0.70%, making only one wrong classification out of 100; thus, 99 instances were correctly classified.

For the first modelling methodology, namely determining neural networks through trial and error, different tasks were formulated, each type of modelling potentially contributing to obtain different information regarding disease progression and necessary treatment.

In these tasks, various types of neural models were tested, both structurally (one or two intermediate layers with different numbers of neurons) and parametrically (different transfer functions, training algorithms). Also, different variants (percentages) of data splitting into training and testing sets were tried, with performance evaluation in all cases based on errors (percentage, mean squared) or the correlation coefficient. Thus, the best model determination was done through trial and error.

For the tasks of dataset 1, a satisfactory generalization capacity was obtained, especially for the last four case studies containing a single output size, thus a simpler formulation.

Compared to the studies encountered in the literature addressing glaucoma through simulation using artificial intelligence tools, it is noted that these approaches use data extracted from OCT images 13 , 14 , 38 , 39 with relatively few parameters.

Considering the modelling results obtained on dataset 2 (which is, in fact, the complete dataset), namely one wrong answer out of 15 possible (95.33% correct answers), these are close to what other researchers have reported or even better. Shon et al. 40 predicted glaucoma progression based on visual field (VF) test information using neural networks and, although they had a much larger dataset than the one included in this study, they achieved 86.40% accuracy.

Shuldiner et al. 41 tested several artificial intelligence tools: support vector machine, neural networks, random forest, logistic regression, naïve Bayes, and hybrid models for predicting glaucoma progression, obtaining an accuracy of 72%. They included in the input data visual field parameters, MD (mean deviation), and PSD (pattern standard deviation). Logistic regression applied by these researchers to the available data highlighted the fact that the advanced age of patients and the high value of PSD can influence the rapid progression of glaucoma.

In recent studies, Anton et al. 42 , 43 demonstrated, using neural networks, that changes in visual field parameters, MD and PSD, are related to the influence of sleep apnoea syndrome on glaucoma.

Glaucoma causes visual disability, so it is crucial to find ways to predict its progression, enabling timely intervention to reduce its effects. Most existing studies in the literature rely on artificial intelligence tools to predict glaucoma progression based on spatial features embedded in fundus images 2 , 3 , 7 , 25 , 38 , 41 , 42 , 43 , 44 , 45 , 46 , 47 , 48 , 49 , 50 , 51 . All these studies demonstrate the ability of AI techniques to identify glaucoma and its progression from fundus images, suggesting that they can assist Ophthalmologists quickly and accurately. It can also be argued that artificial intelligence strategies can detect glaucoma progression earlier than conventional methods 48 , 51 . Most of these existing studies in the specialized literature, which are using fundus images report an accuracy of over 90% for predicting glaucoma progression. Compared to these, it is worth mentioning that the models developed in this paper use data collected from patients’ medical records.

Regarding the second modelling methodology, it included the implementation of a neural network design method, specifically a PyTorch program, which provides very good functional neural network models, i.e., network parameters. It is essential to note that very small errors were obtained during the testing phase, giving credibility to the models used in predicting glaucoma.

The program is flexible, allowing for the selection of the best values for the parameters; it works with: optimizer type, learning rate, weight decay, number of attempts, number of epochs, and data splitting into training and testing sets.

It was found that a larger network can provide better results; the same configuration, MLP(16:42:14:1), yielded very small errors in training and testing for all parameters for which predictions were made (VFI, MD, PSD, RNFL, IOP). The average testing errors were below 5%. A comparison can be made with previously reported results 37 obtained with the dedicated Weka machine learning software. In both cases, the same dataset (dataset 2) was used, and the same task was solved, answering the question of whether glaucoma had progressed, considering the same 12 inputs. The best result obtained with Weka was a MLP(12:8:1), which provided 92 correct answers out of 100 possible, so a relative error of 8%, compared to the network determined in PyTorch , with 99 correct classifications out of a total of 100 instances.

The PyTorch ANN models provided very good results for predicting the visual field parameters (VFI, MD, and PSD) based on previous visit investigations. In the literature, similar research 16 presents the possibility of making predictions for 24–2 visual field defects up to 5.5 years into the future. It is a complex approach, with a large number of parameters, based on deep learning neural networks that uses points from visual field images. Indeed, relatively small prediction errors are obtained, but, in our approach, the predictions for the visual field parameter values are smaller than those reported by Wen and colleagues 16 . Also, our methodology assumes the use of patient medical record data by a simple method.

When using AI models to analyze eye-related glaucoma illness datasets, several potential disadvantages and repercussions must be considered. The diversity of medical records means that more parameters may need to be considered for studying glaucoma progression than those offered by images alone. This paper uses such a diverse database as a challenge to find solutions that are accurate and methodologically sound. While it can be demonstrated that credible and useful results can be obtained with appropriate tools, several risks remain.

One significant risk is the potential for false diagnoses, including false-negative results, which can lead to missed diagnoses and delayed treatment. Such errors could have serious repercussions, including blindness if the glaucoma progresses untreated. If clinical decisions are based directly on AI system interpretations, the responsibility of the ophthalmologist becomes critical, necessitating the development of medico-legal protocols that involve both the clinician and the AI methods.

Additionally, there are limitations to using AI in clinical practice, such as patient mistrust in remote screening and doctors' dependency on technology. Despite these limitations, AI systems combined with ocular imaging in telemedicine could offer a cost-effective long-term solution to enhance screening efficiency and patient monitoring. This underscores the importance of integrating AI carefully and responsibly into medical practice to maximize benefits and minimize risks.

Conclusions

Glaucoma, a cause of irreversible blindness, represents a public health issue (global prevalence of 3.5%, affecting individuals aged between 40 and 80 years old and an estimation of 112 million cases worldwide by 2040). Early diagnosis and accurate monitoring of glaucoma progression are essential for controlling this disease and providing personalized treatment. In recent years, AI tools have been increasingly used for diagnosis, monitoring, and prognosis of glaucoma.

The initial trial and error method is simple but time-consuming. Moreover, it does not guarantee the determination of the best model. However, good results have been obtained, depending on the task formulated and the parameters considered.

Very good neural networks were obtained using the PyTorch deep learning framework. For various tasks, very small errors in training and testing, under 5%, were obtained. The networks can model both binary outputs, indicating whether glaucoma progresses, and real-valued outputs for variables that provide information on glaucoma progression (VFI, MD, PSD, RNFL, IOL). Predictions made by the PyTorch neural networks achieved over 98% accuracy for both functional (visual field parameters) and structural (retinal nerve fiber layer thickness) progression of glaucoma. This approach proved superior for developing efficient neural models to estimate the course of glaucomatous disease in clinical practice.

As future research directions, we propose testing the already developed software tools on additional datasets, as well as exploring other AI tools that are likely to be effective in simulation studies of glaucoma (for instance neuro-fuzzy technique).

Data availability

The datasets used and analyzed during the current study, as well as the source code of the software, are available from the corresponding author on reasonable request.

Antón, A., Jordano, J. & Maquet, J. D. Sistema experto de diagnóstico de glaucoma. Archivos de la Sociedad Española de Oftalmologia 69 (1), 23–28 (1995).

Google Scholar  

Bowd, C. et al. Relevance vector machine and support vector machine classifier analysis of scanning laser polarimetry retinal nerve fiber layer measurements. Invest. Ophthalmol. Vis. Sci. 46 , 1322–1329 (2005).

Article   PubMed   Google Scholar  

Bowd, C. et al. Comparing neural networks and linear discriminant functions for glaucoma detection using confocal scanning laser ophthalmoscopy of the optic disc. Invest. Ophthalmol. Vis. Sci. 43 (11), 3444–3454 (2002).

PubMed   Google Scholar  

Simon, M., Alonso, L. & Alfonso, A. A hybrid visual field classifier to support early glaucoma diagnosis. Inteligencia Artificial Revista Iberoamericana de Inteligencia Artificial 9 (26), 9–17 (2005).

Hernández, G. E., Santos-García, G. & Inés, F. B. Identification of glaucoma stages with artificial neural networks using retinal nerve fibre layer analysis and visual field parameters. Innov. Hybrid Intell. Syst. 44 , 418–424 (2007).

Grewal, D., Jain, R., Grewal, S. & Rihani, V. Artificial neural network-based glaucoma diagnosis using retinal nerve fiber layer analysis. Eur. J. Ophthalmol. 18 (6), 915–921 (2018).

Article   Google Scholar  

Parsaei, H., Moradi, P. & Parsaei, R. Development and verification of artificial neural network classifiers for eye diseases diagnosis. in Proceedings of the 14th ICBME . 398–402 (2008).

Zhou, M. et al. Diabetes mellitus as a risk factor for open-angle glaucoma: A systematic review and meta-analysis. PLoS One 9 (8), e102972 (2014).

Article   ADS   PubMed   PubMed Central   Google Scholar  

Anton Apreutesei, N. The influence of glaucoma on the ocular changes of diabetes. PhD thesis UMF Iași, (2015).

Anton Apreutesei, N. et al. Predictions of ocular changes caused by diabetes in glaucoma patients. Comput. Methods Progr. Biomed. 154 , 183–190 (2018).

Devalla, S. et al. Glaucoma management in the era of artificial intelligence. Br. J. Ophthalmol. 104 (3), 301–311 (2019).

Bizios, D., Heijl, A., Hougaard, J. L. & Bengtsson, B. Machine learning classifiers for glaucoma diagnosis based on classification of retinal nerve fibre layer thickness parameters measured by Stratus OCT. Acta Ophthalmol. 88 (1), 44–52 (2010).

Yousefi, S., Kiwaki, T. & Zheng, Y. Detection of longitudinal visual field progression in glaucoma using machine learning. Am. J. Ophthalmol. 193 , 71–79 (2018).

Guangzhou, A. et al. Comparison of machine-learning classification models for glaucoma management. J. Health Ing. 2018 , 1–8 (2018).

Bizios, D., Heijl, A. & Bengtsson, B. Trained artificial neural network for glaucoma diagnosis using visual field data: A comparison with conventional algorithms. J. Glaucoma 16 , 20–28 (2007).

Wen, J. C. et al. Forecasting future Humphrey visual fields using deep learning. PLoS One 14 , e0214875 (2019).

Article   CAS   PubMed   PubMed Central   Google Scholar  

Sample, P. A. et al. Using machine learning classifiers to identify glaucomatous change earlier in standard visual fields. Invest. Ophthalmol. Vis. Sci. 43 , 2660–2665 (2002).

Sample, P. A. et al. Unsupervised machine learning with independent component analysis to identify areas of progression in glaucomatous visual fields. Invest. Ophtalmol. Vis. Sci. 46 , 3684–3692 (2005).

Goldbaum, M. et al. Comparing machine learning classifiers for diagnosing glaucoma from standard automated perimetry. Invest. Ophthalmol. Vis. Sci. 43 (1), 162–169 (2002).

Chan, K. et al. Comparison of machine learning and traditional classifiers in glaucoma diagnosis. IEEE Trans. Biomed. Eng. 49 , 963–974 (2002).

Christopher, M. et al. Retinal nerve fiber layer features identified by unsupervised machine learning on optical coherence tomography scans predict glaucoma progression. Invest. Ophthalmol. Vis. Sci. 59 (7), 2748–2756 (2018).

Article   PubMed   PubMed Central   Google Scholar  

Medeiros, F., Jammal, A. A. & Thompson, A. C. From machine to machine: An OCT-trained deep learning algorithm for objective quantification of glaucomatous damage in fundus photographs. Ophthalmology 126 (4), 513–521 (2019).

Ran, A. R. et al. Deep learning in glaucoma with optical coherence tomography: A review. Eye 35 (1), 188–201 (2021).

Hood, D. C. et al. Detecting glaucoma with only OCT: Implications for the clinic, research, screening, and AI development. Prog. Retin. Eye Res. 90 (9), 101052 (2022).

Article   CAS   PubMed   Google Scholar  

Li, F. et al. A deep-learning system predicts glaucoma incidence and progression using retinal photographs. J. Clin. Invest. 132 (11), e157968 (2022).

Chaurasia, A. K., Greatbatch, C. J. & Hewitt, A. W. Diagnostic accuracy of artificial intelligence in glaucoma screening and clinical practice. Glaucoma 31 (5), 285–299 (2022).

Brigatti, L., Hoffman, D. & Caprioli, J. Neural networks to identify glaucoma with structural and functional measurements. Am. J. Ophthalmol. 121 (5), 511–521 (1996).

Mardin, C. Y., Peters, A., Horn, F., Jünemann, A. G. & Lausen, B. Improving glaucoma diagnosis by the combination of perimetry and HRT measurements. J. Glaucoma 15 (4), 299–305 (2006).

Yousefi, S. et al. Monitoring glaucomatous functional loss using an artificial intelligence-enabled dashboard. Ophthalmology 127 (9), 1170–1178 (2020).

Nair, M., Tagare, S., Venkatesh, R. & Odayappan, A. Artificial intelligence in glaucoma. Indian J. Ophthalmol. 70 (5), 1868–1869 (2022).

Ittoop, S. M., Jaccard, N., Lanouette, G. & Kahook, M. Y. The role of artificial intelligence in the diagnosis and management of glaucoma. J. Glaucoma 31 (3), 137–146 (2022).

Singh, L. K. & Garg, H. Detection of glaucoma in retinal images based on multiobjective approach. Int. J. Appl. Evolut. Comput. 11 (2), 13 (2020).

CAS   Google Scholar  

Singh, L. K., Garg, H. & Pooja, A. Automated glaucoma type identification using machine learning or deep learning techniques. Adv. Mach. Intell. Interact. Med. Image Anal. 241–263 (2019).

Singh, L. K., Munish, K. & Shankar, T. A novel hybrid robust architecture for automatic screening of glaucoma using fundus photos, built on feature selection and machine learning-nature driven computing. Expert Syst. 39 , e13069 (2022).

Singh, L. K., Munish, K., Shankar, T. & Singh, R. Deep-learning based system for effective and automatic blood vessel segmentation from Retinal fundus images. Multimed. Tools Appl. 83 , 1–45 (2023).

Singh, L. K., Munish, K., Shankar, T. & Singh, R. Nature-inspired computing and machine learning based classification approach for glaucoma in retinal fundus images. Multimed. Tools Appl. 82 , 1–49 (2023).

Article   CAS   Google Scholar  

Târcoveanu, F. E. et al. Classification algorithms used in predicting glaucoma progression. Healthcare 10 , 1831 (2022).

Elze, T. et al. Patterns of functional vision loss in glaucoma determined with archetypal analysis. J. R. Soc. Interface 12 (10), 1098 (2015).

Raja, C. & Gangatharan, N. Appropriate sub-band selection in wavelet packet decomposition for automated glaucoma diagnosis. Int. J. Autom. Comput. https://doi.org/10.1007/s11633-014-0858-6 (2015).

Shon, K., Sung, K. & Shin, J. Can artificial intelligence predict glaucomatous visual field progression? A spatial-ordinal convolutional neural network model. Am. J. Ophthalmol. 233 , 124–134 (2022).

Shuldiner, S. et al. Predicting eyes at risk for rapid glaucoma progression based on an initial visual field test using machine learning. PLoS One 16 (4), e0249856 (2021).

Anton, N. et al. Use of artificial neural networks to predict the progression of glaucoma in patients with sleep apnea. Appl. Sci. 12 (12), 6061 (2022).

Anton, N. et al. Assessing changes in diabetic retinopathy caused by diabetes mellitus and glaucoma using support vector machines in combination with differential evolution algorithm. Appl. Sci. 11 (9), 3944 (2021).

Wang, S. Y., Tseng, B. & Hernandez-Boussard, T. Deep learning approaches for predicting glaucoma progression using electronic health records and natural language processing. Ophthalmol. Sci. 2 (2), 100127 (2020).

Thompson, A., Jammal, A. & Medeiros, F. A review of deep learning for screening, diagnosis, and detection of glaucoma progression. Trans. Vis. Sci. Tech. 9 (2), 42–42 (2020).

Hemelings, R. et al. Accurate prediction of glaucoma from colour fundus images with a convolutional neural network that relies on active and transfer learning. Acta Ophthalmol. 98 (1), e94–e100 (2020).

Mirzania, D., Thompson, A. C. & Muir, K. W. Applications of deep learning in detection of glaucoma: A systematic review. Eur. J. Ophthalmol. 31 (4), 1618–1642 (2020).

Zafar, A. et al. A Comprehensive convolutional neural network survey to detect glaucoma disease. Mob. Inf. Syst. Hindawi 2022 , 3971516 (2022).

Mursch-Edlmayr, A. S. et al. Artificial intelligence algorithms to diagnose glaucoma and detect glaucoma progression: Translation to clinical practice. Trans. Vis. Sci. Tech. 9 (2), 55 (2020).

Belgacem, R., Malek, I. T., Trabelsi, H. & Jabri, I. A supervised machine learning algorithm SKVMs used for both classification and screening of glaucoma disease. N. Front. Ophthalmol. 4 (4), 1–27 (2018).

Li, Z. et al. Efficacy of a deep learning system for detecting glaucomatous optic neuropathy based on color fundus photographs. Ophthalmology 125 (8), 1199–1206 (2018).

Download references

Acknowledgements

This work was supported by Exploratory Research Project PN-III-P4-ID-PCE-2020-0551, no. 91/2021, financed by UEFISCDI Romania. Funding: Research Project PN-III-P4-ID-PCE-2020-0551, no. 91/2021.

Research Project PN-III-P4-ID-PCE-2020-0551, no. 91/2021.

Author information

Authors and affiliations.

Ophthalmology Department, Faculty of Medicine, University of Medicine and Pharmacy “Gr. T. Popa” Iasi, University Street No 16, 700115, Iasi, Romania

Filip Târcoveanu & Nicoleta Anton

Faculty of Automatic Control and Computer Engineering, “Gheorghe Asachi” Technical University of Iasi, 27 Mangeron Street, 700050, Iasi, Romania

Florin Leon

Department of Chemical Engineering, Faculty of Chemical Engineering and Environmental Protection “Cristofor Simionescu”, “Gheorghe Asachi” Technical University of Iasi, 73 Mangeron Street, 700050, Iasi, Romania

Cătălin Lisa & Silvia Curteanu

Faculty of Economic Science, “Vasile Alecsandri” University of Bacau, Calea Marasesti 156, 600115, Bacau, Romania

Andreea Feraru

Countess of Chester Hospital, Liverpool Rd, Chester, CH21UL, UK

You can also search for this author in PubMed   Google Scholar

Contributions

Conceptualization, N.A., F.T. and S.C.; methodology, N.A., F.T., S.C, F.L; software, formal analysis C.L, A.F; resources, data curation, validation F.T, N.A, F.L, S.C, K.A; investigation, F.T, K.A, N.A and F.L, visualization, supervision N.A, S.C; writing—original draft, writing—review and editing, N.A., S.C; project administration, funding acquisition S.C, F.L.

Corresponding authors

Correspondence to Silvia Curteanu or Nicoleta Anton .

Ethics declarations

Competing interests.

The authors declare no competing interests.

Additional information

Publisher's note.

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Rights and permissions

Open Access This article is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License, which permits any non-commercial use, sharing, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if you modified the licensed material. You do not have permission under this licence to share adapted material derived from this article or parts of it. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by-nc-nd/4.0/ .

Reprints and permissions

About this article

Cite this article.

Târcoveanu, F., Leon, F., Lisa, C. et al. The use of artificial neural networks in studying the progression of glaucoma. Sci Rep 14 , 19597 (2024). https://doi.org/10.1038/s41598-024-70748-1

Download citation

Received : 10 May 2024

Accepted : 20 August 2024

Published : 23 August 2024

DOI : https://doi.org/10.1038/s41598-024-70748-1

Share this article

Anyone you share the following link with will be able to read this content:

Sorry, a shareable link is not currently available for this article.

Provided by the Springer Nature SharedIt content-sharing initiative

• Artificial neural networks
• Simulation studies
• Modelling and optimization

By submitting a comment you agree to abide by our Terms and Community Guidelines . If you find something abusive or that does not comply with our terms or guidelines please flag it as inappropriate.

Quick links

• Explore articles by subject
• Guide to authors
• Editorial policies

Sign up for the Nature Briefing: AI and Robotics newsletter — what matters in AI and robotics research, free to your inbox weekly.

• Alzheimer's disease & dementia
• Arthritis & Rheumatism
• Attention deficit disorders
• Autism spectrum disorders
• Biomedical technology
• Diseases, Conditions, Syndromes
• Endocrinology & Metabolism
• Gastroenterology
• Gerontology & Geriatrics
• Health informatics
• Inflammatory disorders
• Medical economics
• Medical research
• Medications
• Neuroscience
• Obstetrics & gynaecology
• Oncology & Cancer
• Ophthalmology
• Overweight & Obesity
• Parkinson's & Movement disorders
• Psychology & Psychiatry
• Radiology & Imaging
• Sleep disorders
• Sports medicine & Kinesiology
• Vaccination
• Breast cancer
• Cardiovascular disease
• Chronic obstructive pulmonary disease
• Colon cancer
• Coronary artery disease
• Heart attack
• Heart disease
• High blood pressure
• Kidney disease
• Lung cancer
• Multiple sclerosis
• Myocardial infarction
• Ovarian cancer
• Post traumatic stress disorder
• Rheumatoid arthritis
• Schizophrenia
• Skin cancer
• Type 2 diabetes
• Full List »

share this!

March 10, 2023

This article has been reviewed according to Science X's editorial process and policies . Editors have highlighted the following attributes while ensuring the content's credibility:

fact-checked

peer-reviewed publication

trusted source

Researchers discover therapeutic target to aid in glaucoma treatment

by Indiana University School of Medicine

Indiana University School of Medicine researchers have identified a new therapeutic target that could lead to more effective treatment of glaucoma.

Glaucoma is a neurodegenerative disease that causes vision loss and blindness due to a damaged optic nerve. More than 200,000 people are affected by glaucoma in the United States each year. Unfortunately, there is currently no treatment. In a newly published paper in Communications Biology , researchers found neurons use mitochondria for a steady source of energy, and restoring mitochondrial homeostasis in the diseased neurons can protect the optic nerve cells from being damaged.

"Age-related neurodegenerative disease, which includes glaucoma, Parkinson's disease, and amyotrophic lateral sclerosis (ALS), is the biggest global health problem," said Arupratan Das, Ph.D., assistant professor of ophthalmology and principal investigator of the study.

"The fundamental mechanisms that we discovered can be used to protect neurons in glaucoma and be tested for the other diseases. We have identified a critical step of complex mitochondrial homeostasis process, which rejuvenates the dying neuron, similar to giving a lifeline to a dying person."

The research team, led by Michelle Surma and Kavitha Anbarasu from the Department of Ophthalmology, used induced pluripotent stem cells (iPSCs) from patients with and without glaucoma as well as clustered regularly interspaced short palindromic repeats (CRISPR) engineered human embryonic stem cells with glaucoma mutation.

Using stem cell differentiated retinal ganglion cells (hRGCs) of the optic nerve, electron microscopy and metabolic analysis, researchers identified glaucomatous retinal ganglion cells suffer mitochondrial deficiency with more metabolic burden on each mitochondrion. This leads to mitochondrial damage and degeneration. Mitochondria are the tube like structures in cells which produce adenosine triphosphate, cell's energy source.

However, the process could be reversed by enhancing mitochondrial biogenesis by a pharmacological agent. The team showed retinal ganglion cells are highly efficient in degrading bad mitochondria, but at the same time producing more to maintain homeostasis.

"Finding that retinal ganglion cells with glaucoma produce more adenosine triphosphate even with less mitochondria was astonishing," Das said. "However, when triggered to produce more mitochondria, the adenosine triphosphate production load was distributed among more mitochondrion which restored the organelle physiology. It is similar to a situation where a heavy stone is carried by fewer people versus a greater number of people—each person will have less pain and injury, just like each mitochondrion will have less difficulty and damage."

In the future, Das would like to test if these mechanisms protect the optic nerve in animal models under injury before testing in humans to hopefully lead to new clinical interventions.

Explore further

Feedback to editors

Study identifies metabolic switch essential for generation of memory T cells and anti-tumor immunity

21 hours ago

Multiple sclerosis appears to protect against Alzheimer's disease

Aug 23, 2024

Good sleep habits important for overweight adults, study suggests

Mediterranean diet supplement can affect epigenetics associated with healthy aging

New method for quantifying boredom in the body during temporary stress

Cancer researchers develop new method that uses internal clock inside tumor cells to optimize therapies

Strength training activates cellular waste disposal, interdisciplinary research reveals

Being a 'weekend warrior' could be as good for brain health as exercising throughout the week

Simple blood test for Alzheimer's disease could change how the disease is detected and diagnosed

Chlamydia can settle in the intestine, organoid experiments reveal

Related stories.

Novel gene therapy shows promise for treating multiple eye diseases

Jan 30, 2023

New study shows people with glaucoma are at significant risk of Alzheimer's disease

Sep 30, 2022

Research advance may lead to new treatments for glaucoma

Mar 22, 2016

Researchers stimulate neuroprotection and optic nerve regeneration in mouse models of glaucoma

Dec 20, 2022

High-tech imaging offers new way to detect signs of early glaucoma

Aug 2, 2022

Scientists discover gene therapy provides neuroprotection to prevent glaucoma vision loss

Jul 22, 2021

Recommended for you

GnRH neurons in the mouse olfactory bulb shown to translate socially relevant odors into male reproductive behavior

Study suggests even mild concussions can have lifelong brain impacts

Low-dose carbon monoxide may explain the paradoxical reduced risk of Parkinson's disease among smokers

Scientists uncover new mechanism of 'forgetting' in brain neurons that could inform Parkinson's treatment

Spike mutations that help SARS-CoV-2 infect the brain discovered

Let us know if there is a problem with our content.

Use this form if you have come across a typo, inaccuracy or would like to send an edit request for the content on this page. For general inquiries, please use our contact form . For general feedback, use the public comments section below (please adhere to guidelines ).

Please select the most appropriate category to facilitate processing of your request

Thank you for taking time to provide your feedback to the editors.

Your feedback is important to us. However, we do not guarantee individual replies due to the high volume of messages.

E-mail the story

Your email address is used only to let the recipient know who sent the email. Neither your address nor the recipient's address will be used for any other purpose. The information you enter will appear in your e-mail message and is not retained by Medical Xpress in any form.

Newsletter sign up

Get weekly and/or daily updates delivered to your inbox. You can unsubscribe at any time and we'll never share your details to third parties.

More information Privacy policy

Donate and enjoy an ad-free experience

We keep our content available to everyone. Consider supporting Science X's mission by getting a premium account.

E-mail newsletter

More From Forbes

A promising new approach for glaucoma.

• Share to Facebook
• Share to Twitter
• Share to Linkedin

Eye looking at camera

This story is part of a series on the current progression in Regenerative Medicine. This piece is part of a series dedicated to the eye and improvements in restoring vision.

In 1999, I defined regenerative medicine as the collection of interventions that restore tissues and organs damaged by disease, injured by trauma, or worn by time to normal function. I include a full spectrum of chemical, gene, and protein-based medicines, cell-based therapies, and biomechanical interventions that achieve that goal.

Glaucoma is a debilitating eye ailment that can cause irreversible harm to the optic nerve and eventually result in loss of vision. The condition is often asymptomatic in its early stages, making it challenging to diagnose and treat . It occurs when the fluid pressure inside the eye increases, damaging the optic nerve and causing vision loss. Fortunately, recent medical breakthroughs in RGC (retinal ganglion cell) replacement therapy offer hope for those suffering from this condition. This revolutionary treatment involves transplanting healthy RGCs to replace the damaged ones and restore visual function. Doing so aims to prevent further damage to the optic nerve and potentially reverse the damage already done. With this new treatment, there is hope for individuals with glaucoma.

What are Retinal Ganglion Cells?

Retinal Ganglion Cells (RGCs) are specialized neurons that play a critical role in the visual system by transmitting information from the retina to the brain via the optic nerve. These highly complex cells comprise more than a dozen molecularly, functionally, and topographically unique subtypes. Each subtype has distinct morphology, connectivity, and response properties, allowing it to perform different visual system functions.

RGCs are the final output neurons of the retina. They are responsible for encoding visual information into electrical signals sent to the brain. These signals are then processed by higher visual centers in the brain, which enables us to see and perceive the world around us.

Despite their complexity, RGCs have been a promising target for therapeutic intervention as they do not regenerate naturally once they deteriorate. Any damage to these cells can result in permanent vision loss. Scientists are actively working to develop new therapies that can help protect and restore RGCs in individuals with retinal degenerative diseases such as glaucoma and age-related macular degeneration.

Best High-Yield Savings Accounts Of 2024

Best 5% interest savings accounts of 2024, stem cells & organoids for glaucoma.

Stem cell-based transplantation has emerged as a viable option for replacing lost or damaged retinal ganglion cells (RGCs). As stem cells can differentiate into various cell types, including RGCs, they present a potential source for RGC replacement therapy. However, the limited availability of RGCs in stem cell-derived cultures makes the integration of RGCs into the host retina a complex task. Organoids offer a promising alternative to traditional branch cell-based transplantation methods.

Organoids are self-organizing three-dimensional structures that closely resemble the complexity and organization of the retina. They are generated by culturing stem cells in a controlled environment that mimics the developmental process of the retina. Recent advances in organoid protocols have made it possible to create RGCs from both human and mouse stem cells, overcoming the limitations of traditional stem cell-based transplantation methods.

Methods of Retinal Cell Replacement

Cell delivery can be either suprachoroidal, intravitreal or subretinal.

Subretinal space and vitreous cavity transplantation are two potential methods for RGC replacement. Subretinal space transplantation involves transplanting cells into the subretinal space, which is located between the retinal pigment epithelium and the photoreceptor layer. The transplanted cells can integrate into the host retina and potentially restore vision.

Vitreous cavity transplantation, on the other hand, involves transplanting cells into the vitreous cavity of the eye, which is the gelatinous substance located behind the lens. The transplanted cells can secrete neurotrophic factors that promote the survival of the remaining RGCs and stimulate the regeneration of damaged ones.

Clinical Trials Exploring Cell Therapies

Numerous studies are currently investigating the possibility of using RGC replacement therapy as a treatment for glaucoma. One such study was conducted on rats with glaucoma, and it showed significant improvement in visual function and RGC survival when RGC-like cells were transplanted into them. These RGC-like cells were differentiated from human-induced pluripotent stem cells, which can develop into any cell in the human body.

The study demonstrated that human-induced pluripotent stem cells could be a potential source for RGC transplantation in glaucoma patients. However, further research is necessary to determine the safety and effectiveness of this therapy before it can be implemented in clinical practice.

The Challenges of Making Retinal Ganglion Cells a Treatment

Still, despite the promise of studies, there are some challenges. A review done by a team in Spain assessed these challenges and more. One of the significant challenges is understanding the origin of RGCs and how to replicate their natural development in laboratory settings.

Another area for improvement is the difficulty in scaling up the production of RGCs and ensuring they can be produced in large quantities. Additionally, once produced, there are challenges in integrating and ensuring the survival of transplanted RGCs in the host tissue.

Regrowing RGC axons is also a significant challenge, as it requires identifying the factors that promote axon growth and determining how to apply them effectively. Finally, achieving functional RGC replacement is a significant hurdle, as it involves creating RGCs that can function as well as, or better than, the RGCs they are replacing.

This illustrative graph displays three commonly used types/formats of donor cells in retinal ... [+] ganglion cell (RGC) replacement studies.

However, recent advances in stem cell technologies and organoid protocols discussed before offer new solutions to these limitations. For example, organoid protocols enable the generation of RGCs that more closely resemble those found in the human retina, thus improving the accuracy and efficacy of transplantation. Additionally, stem cell technologies have been used to create RGCs capable of integrating into the host tissue and regenerating axons.

RGC replacement therapy shows potential as a new therapeutic approach for treating glaucoma. Although there are challenges to overcome regarding scaling up RGC production and achieving reliable and functional integration, recent advancements in stem cell technologies and organoid protocols offer promising solutions. Clinical trials investigating the potential of RGC replacement therapy have reported positive outcomes, providing hope for restoring vision in glaucoma patients.

To learn more about the eye, read more stories at www.williamhaseltine.com

• Editorial Standards
• Reprints & Permissions

Glaucoma is a common ocular condition whereby damage is caused to the optic nerve connecting the eye to the brain.

What is Open-Angle Glaucoma?

Glaucoma pathophysiology, glaucoma causes, feature articles, glaucoma diagnosis, glaucoma management, glaucoma classifications, glaucoma symptoms, types of glaucoma, what is normal-tension glaucoma, what is traumatic glaucoma, optic nerve cupping in glaucoma, latest glaucoma news and research.

McGill discovery offers hope for targeted stem cell treatments

A new technique developed by McGill researchers for mechanically manipulating stem cells could lead to new stem cell treatments, which have yet to fulfill their therapeutic potential.

How states pioneered the medical marijuana movement in America

From the war on drugs to the current wave of legalization efforts, public opinion in the United States has shifted on few issues like it has with marijuana.

Case Western Reserve University receives $1.5M grant for oral drug trial to prevent blindness in RP patients

There's only one U.S. Food and Drug Administration-approved therapy for an inherited retinal disease, and dozens of retinitis pigmentosa (RP) genes for which no therapy is available.

Screen time significantly associated with myopia in children, research shows

The epidemiological evidence on the association between screen time exposure and myopia in children and adolescents.

Newly discovered neuron type guides retinal blood vessel formation

Scientists have known for years that a lattice of blood vessels nourishes cells in the retina that allow us to see – but it's been a mystery how the intricate structure is created.

Soft ERG multi-electrode system developed for enhanced eye disease monitoring

Eye diseases are becoming more prevalent worldwide, partly because of the aging population, but also because of our greatly increased screen time compared to previous generations.

Innovative "smart" contact lens may revolutionize glaucoma detection and monitoring

Most people with early-stage glaucoma don't know they have it, even though early treatment is key to reducing vision loss.

Does your ADHD medication increase your risk of glaucoma?

The risks of glaucoma among users of ADHD medications.

Google Gemini and Bard pass the ophthalmology board examination

The performance of two AI chatbots in the ophthalmology board examination.

Ultra-processed foods linked to higher glaucoma risk, study warns

Study finds a significant relationship between the highest intake of ultra-processed foods, especially sweets, and an increased risk of developing glaucoma among Spanish university graduates, highlighting the importance of dietary quality for ocular health.

As AI eye exams prove their worth, lessons for future tech emerge

Christian Espinoza, director of a Southern California drug-treatment provider, recently began employing a powerful new assistant: an artificial intelligence algorithm that can perform eye exams with pictures taken by a retinal camera. It makes quick diagnoses, without a doctor present.

AI system can equal or surpass ophthalmologists in diagnosis, treatment of glaucoma and retina disorders

A large language model (LLM) artificial intelligence (AI) system can match, or in some cases outperform, human ophthalmologists in the diagnosis and treatment of patients with glaucoma and retina disease, according to research from New York Eye and Ear Infirmary of Mount Sinai (NYEE).

Novel genetic locations associated with primary open-angle glaucoma identified

An international genetic study using multiancestry biobanks has identified novel genetic locations associated with primary open-angle glaucoma (POAG), the most common type of glaucoma and the leading cause of irreversible blindness globally.

Study uncovers key genes and cell types that may affect pathogenesis of primary open-angle glaucoma

Although primary open-angle glaucoma (POAG) is the leading cause of blindness in people over the age of 55, there remains no cure for the disease and its biological mechanisms are not well understood.

Breakthrough $20 million OCT project aims to revolutionize eye health screening

In the United States, more than one-fourth of adults over age 40 have an eye disease, including glaucoma, cataracts or age-related macular degeneration, or a chronic health condition that affects the eyes, such as diabetic retinopathy.

Does visual impairment influence mental disorders?

Whether visual impairment (VI) affects mental disorders in the older population.

Genomics meets retinal imaging: A powerful fusion can predict future health

The retina is said to provide a window into a person's systemic health. In a new study published January 24th in Science Translational Medicine, physician-researchers from Mass Eye and Ear, a member of Mass General Brigham, and the Broad Institute of MIT and Harvard combined retinal imaging, genetics and big data to estimate how likely a person is to develop eye and systemic diseases in the future.

Preparing to hang up the car keys as we age

Lewis Morgenstern has made up his mind. When he turns 65 in four years, he's going to sign an advance directive for driving.

Study identifies genetic variants that influence response to preterm birth treatment

A UC San Francisco-led study has for the first time identified genetic variants that predict whether patients will respond to treatment for preterm birth, a condition that affects one in 10 infants born in the United States.

Innovative diabetes detection method with DiaNet v2, utilizing retinal imaging technology

Innovative diabetes detection method with DiaNet v2, utilizing retinal imaging technology.

• Trending Stories
• Latest Interviews
• Top Health Articles

Global and Local Efforts to Take Action Against Hepatitis

Lindsey Hiebert and James Amugsi

In this interview, we explore global and local efforts to combat viral hepatitis with Lindsey Hiebert, Deputy Director of the Coalition for Global Hepatitis Elimination (CGHE), and James Amugsi, a Mandela Washington Fellow and Physician Assistant at Sandema Hospital in Ghana. Together, they provide valuable insights into the challenges, successes, and the importance of partnerships in the fight against hepatitis.

Addressing Important Cardiac Biology Questions with Shotgun Top-Down Proteomics

In this interview conducted at Pittcon 2024, we spoke to Professor John Yates about capturing cardiomyocyte cell-to-cell heterogeneity via shotgun top-down proteomics.

A Discussion with Hologic’s Tim Simpson on the Future of Cervical Cancer Screening

Tim Simpson

Hologic’s Tim Simpson Discusses the Future of Cervical Cancer Screening.

Newsletters you may be interested in

Your AI Powered Scientific Assistant

Hi, I'm Azthena, you can trust me to find commercial scientific answers from News-Medical.net.

A few things you need to know before we start. Please read and accept to continue.

• Use of “Azthena” is subject to the terms and conditions of use as set out by OpenAI .
• Content provided on any AZoNetwork sites are subject to the site Terms & Conditions and Privacy Policy .
• Large Language Models can make mistakes. Consider checking important information.

Great. Ask your question.

Azthena may occasionally provide inaccurate responses. Read the full terms .

While we only use edited and approved content for Azthena answers, it may on occasions provide incorrect responses. Please confirm any data provided with the related suppliers or authors. We do not provide medical advice, if you search for medical information you must always consult a medical professional before acting on any information provided.

Your questions, but not your email details will be shared with OpenAI and retained for 30 days in accordance with their privacy principles.

Please do not ask questions that use sensitive or confidential information.

Read the full Terms & Conditions .

Provide Feedback

• Allergy & Immunology
• Anesthesiology
• Critical Care
• Dermatology
• Diabetes & Endocrinology
• Emergency Medicine
• Family Medicine
• Gastroenterology
• General Surgery
• Hematology - Oncology
• Hospital Medicine
• Infectious Diseases
• Internal Medicine
• Multispecialty
• Ob/Gyn & Women's Health
• Ophthalmology
• Orthopedics
• Pathology & Lab Medicine
• Plastic Surgery
• Public Health
• Pulmonary Medicine
• Rheumatology
• Transplantation
• Today on Medscape
• Business of Medicine
• Medical Lifestyle
• Science & Technology
• Medical Students
• Pharmacists

Low-Dose Steroids Reduce Glaucoma Risk in Kids

Edited by Manasi Talwadekar

August 21, 2024

Use of low-dose glucocorticoids after cataract surgery reduces the risk for secondary glaucoma in children with longer axial lengths over 5 years compared with high-dose glucocorticoids.

METHODOLOGY:

• To examine the relationship between the postoperative dose of glucocorticoids and the development of secondary glaucoma following pediatric cataract surgery, researchers retrospectively reviewed the records of children who underwent cataract surgery until 2016 and prospectively followed those treated from 2017 to 2021.
• The study included 267 consecutive children (388 eyes; 120 girls) before the age of 10 years from two specialized centers in Denmark.
• Participants received treatment with either low-dose (n = 173) or high-dose (n = 95) glucocorticoids after surgery, with a median follow-up period of 40.5 months for the low-dose group and 89 months for the high-dose group.
• The primary outcome was the incidence of glaucoma or ocular hypertension lasting over 3 months after surgery; the secondary outcome was the incidence of opacification of the visual axis.
• Overall, 21 children (34 eyes) were diagnosed with glaucoma or ocular hypertension, with both eyes affected in 13 children.
• Use of low-dose glucocorticoids after surgery significantly reduced the risk for secondary glaucoma or ocular hypertension in children with axial lengths of ≥ 18 mm vs high-dose treatment ( P = .0008). This effect was not found in those with axial lengths < 18 mm.
• The risk for glaucoma or ocular hypertension was estimated to be 10% in the high-dose group and 1% in the low-dose group in the first 2 years after surgery.
• Opacity of the visual axis was reported in 57 children (69 eyes). The incidences were similar with high and low doses of corticosteroids.

IN PRACTICE:

"We report a lower risk of glaucoma 5 years after cataract surgery in children with axial lengths ≥ 18 mm receiving low-dose versus high-dose postoperative local and topical glucocorticoid treatment. This effect was not present in children with shorter axial lengths," the authors wrote.

The study was led by Diana Chabané Schmidt from the Copenhagen University Hospital, Rigshospitalet in Copenhagen, Denmark. It was published online on August 12, 2024, in Acta Ophthalmologica .

LIMITATIONS:

The findings were limited by noninterventional design, selection bias, and the presence of unknown confounders. The wide age range could have affected the risk for secondary glaucoma.

DISCLOSURES:

The Danish Eye Research Foundation, Helsefonden, Fight for Sight, Synoptik Foundation, Fabrikant Einar Willumsens Mindelegat, Beckett-Fonden, Th. Maigaards efterfølger fru Lily Benthine Lunds fond af 1.6.1978, Kong Christian den Tiendes Fond, and Helene and Viggo Bruuns Foundation supported the study. The authors did not declare any conflicts of interest.

This article was created using several editorial tools, including AI, as part of the process. Human editors reviewed this content before publication.

Send comments and news tips to [email protected] .

TOP PICKS FOR YOU

• Perspective
• Drugs & Diseases
• Global Coverage
• Additional Resources
• What's Medical About Marijuana?
• Why Doctors Are Disenchanted With Medicare
• Steroid Implants Best for Improving Vision in Uveitis ME
• Diseases & Conditions Unilateral Glaucoma
• Diseases & Conditions Neovascular Glaucoma
• Diseases & Conditions Uveitic Glaucoma
• Diseases & Conditions Juvenile Glaucoma
• Hyphema Glaucoma
• Unilateral Glaucoma
• Uveitic Glaucoma
• Phacomorphic Glaucoma
• Malignant Glaucoma
• Juvenile Glaucoma
• Ocular Trauma: 8 Potentially Devastating Eye Injuries
• FDA Approves Implant for Glaucoma
• A Better Way to Assess Glaucoma?
• Stiff Arteries Can Predict Risk for Glaucoma

Blood Test May Predict Whether Patients With Glaucoma Will Continue To Lose Vision Following Treatment, Research Suggests

July 17, 2024

Statistical Modelling of Glaucoma Studies

• More on this topic

Professor Ngai Hang Chan, Head and Chair Professor from CityU’s Department of Biostatistics, and his team is working on developing an innovative “Interpretable Artificial Intelligence (IAI)” framework with a holistic approach to IAI medical diagnostics, with glaucoma as an exemplary example.  

The research team has established long-term collaborative efforts and successes in developing world-leading technology in glaucoma study.  One key to developing a set of effective diagnostic tools is the domain knowledge from experts not only in machine learning and artificial intelligence algorithms, but also in Ophthalmology and Visual Sciences.  

Although AI-based medical systems have been rapidly developing, they are tricky for clinicians to understand clearly how a model can obtain its prediction/classification results, which undermines the trust that doctors and patients can put. The team will leverage their strong expertise in statistical modelling and genetic analysis to construct novel interpretable and robust classifiers under the IAI framework for screening and detecting diseases, say glaucoma, and for identification of their progression.   

By carrying out multi-site clinical trials at City University of Hong Kong, The Chinese University of Hong Kong, University of California at San Diego, and Duke University, the team will be able to collect large volumes of relevant medical data across different geographical locations and demographic profiles to achieve efficient models and results.

It should be emphasised that the proposed IAI framework is the first attempt of leveraging the complementary modelling advantages of both statistics and computer science to solve an exemplary problem in medical study in which both early detection and rates of disease progression based on medical imaging are two fundamental, yet challenging, problems that concern clinicians, researchers and patients alike.  

The cross-disciplinary collaborative effort of their plan is meaningful and strategically relevant as statistics focuses on capturing key signals, modelling structures, and identifying significant variables upon which machine learning tools can discover generalisable predictive patterns that cannot be readily expressed via statistical models alone .

  Successful completion of this project would serve as an important milestone to apply IAI technology in advancing our understanding of many related diseases, such as Alzheimer’s disease.  Their research results will be fundamental and crucial in delivering health risk prediction models, providing informed health guidance, and improving care delivery systems.  In the long term, the proposed research will relieve the imminent growing burden arising from the ageing population, which increasingly challenges the healthcare system and the economy, not only in Hong Kong but also in the Greater Bay Area and beyond.

Professor Ngai Hang Chan and his team is working on developing an innovative “Interpretable Artificial Intelligence (IAI)” framework with a holistic approach to IAI medical diagnostics, with glaucoma as an exemplary example. 

Find out more about  the Department of Biostatistics  at CityU.

Enter a Search Term

National glaucoma research.

We accelerate groundbreaking research into the root causes of leading eye diseases and support pioneering prevention strategies and treatments to end glaucoma.

Our 360° Research Approach

New knowledge about what causes glaucoma, new ways to predict progression and treat glaucoma, controlling eye pressure in new ways, imaging and exploring the eye-brain connection, protecting and regenerating the optic nerve.

BrightFocus takes a 360 ° approach to fund innovative scientific research worldwide to defeat glaucoma, exploring the full range of scientific paths toward better treatments and ultimately a cure. 

Glaucoma is a group of eye diseases united under one name. Ultimately, glaucoma threatens sight by damaging the optic nerve, at the back of the eye which carries light signals from the eye to the brain. However, our knowledge of how and when glaucoma damages nerve cells remains imprecise.

Currently approved treatments for glaucoma primarily focus on eye pressure. Numerous therapies exist to lower eye pressure effectively; however, the bulk of them (eyedrops and surgeries) require skill and consistency to achieve results.

Easier methods are needed, as well as new therapies to address other underlying causes of glaucoma besides intraocular pressure (IOP).

Elevated eye pressure, or intraocular pressure (IOP), is present in most forms of glaucoma. This can happen when the fluid that constantly bathes the front of the eye, called aqueous humor, gets clogged and backed up.

Eye changes associated with glaucoma contribute to tiny blind spots, known as “visual field defects,” which, if they worsen, might advance to vision loss and blindness. The chance of that, and the speed at which it happens, vary greatly from person to person.

Early diagnosis is key, and much progress has been made in imaging the eye to detect the tiniest changes that may precede glaucoma.

Unlike most cells in the body, which repair themselves, the nerve cells providing our vision do not regrow once damaged.

BrightFocus is supporting research into ways of protecting cells threatened by advancing glaucoma as well as regenerating those cells after vision loss.

Research We Fund

BrightFocus drives innovative research worldwide on Alzheimer’s, macular degeneration, and glaucoma. Search our grant awards to learn more.

Breakthrough Spotlight

Breakthrough News

Researchers have found a way to block a “master switch” in the eye and the brain that controls inflammation, introducing a possible target for combatting glaucoma and other neurodegenerative diseases.

Insights and Breakthroughs

Well-designed research pays off. With further research, each of these discoveries may contribute to the development of new treatments and preventions.

Investigating Eye Pressure to Prevent Glaucoma-Related Vision Loss

Experimental Drug Could Improve Vision in Glaucoma

New Study Reveals Health Disparities in Glaucoma Care

FDA Approves First-of-its-Kind Glaucoma Treatment

Breakthrough Discovery Reveals New Pathway to Lowering Eye Pressure

Blocking a “Master Switch” in the Eye’s Immune Cells to Treat Glaucoma

New Genetic Targets for Glaucoma Treatment Identified Through Artificial Intelligence

Can Marijuana Help Glaucoma?

With Positive Initial Results, Innovative Glaucoma Therapy Moves Ahead

Could Vision Lost to Glaucoma Be Restored?

Beyond Eye Pressure: A Potential New Path for Treating Glaucoma

Researcher Q&A: Growing Cells for Glaucoma

Donate to national glaucoma research.

Your gift can help lead to treatments and a cure to end glaucoma. Fund the latest, promising research and help provide valuable information to families living with this disease.

I would like to donate

Learn more about glaucoma.

Glaucoma is a group of eye diseases that can damage the optic nerve and result in vision loss and blindness. In this section, you can find out more about glaucoma and how you can manage care for yourself or a loved one.

• Glaucoma Overview

Connect with @BrightFocus

Let’s stay in touch. BrightFocus is a top source of accurate, helpful information on the latest scientific research and care for Alzheimer’s, glaucoma, and macular degeneration. 

The Cure is in Sight.

Glaucoma research.

Our multidisciplinary consortium of researchers is working together to speed the cure for glaucoma and restore vision.

Catalyst for a Cure principal investigators report on their latest challenges and progress.

Glaucoma Research Foundation provides seed money for creative pilot research projects that hold promise.

Glaucoma Events

The Annual Gala is part of Glaucoma 360, a series of events uniting research, industry, and philanthropy with one mission — to find a cure for glaucoma.

The Patient Summit conference helps patients understand and live with glaucoma by highlighting advances in treatment options and and reviewing practical information.

New Horizons Forum unites key clinical, industry, financial, and FDA leaders in a unique exchange on research innovation and advances in glaucoma treatment.

For Eyecare Professionals

To help you manage the vast amount of glaucoma information online, we provide an organized list of helpful links. You’ll find guidance for support groups , free booklets , finding a doctor , and more.