Toric Contact Lens

Definition

Indications

• Corneal or lenticular astigmatism of ≥ 0.75 to 1.00 D that cannot be adequately corrected (or is poorly tolerated) with spectacles
• Anisometropia with astigmatism
• Irregular astigmatism (with RGP/scleral toric designs) - keratoconus, post-keratoplasty, post-refractive surgery ectasia
• Patients desiring cosmetic/occupational freedom from spectacles (sports, unilateral aphakia, high anisometropia)

Types of Astigmatism Corrected

1. With-the-rule astigmatism
2. Against-the-rule astigmatism
3. Oblique astigmatism

Types of Toric Lenses

• Made of hydrogel or silicone hydrogel
• Most commonly prescribed; comfortable, good initial fit

• Back-surface toric - corneal (front surface spherical, back surface toric) - used for high corneal astigmatism
• Front-surface toric - corrects residual lenticular astigmatism when back surface fits sphere cornea
• Bitoric - both surfaces toric - used when corneal astigmatism is very high (>2.50-3.00 D) so that the base curve can align with the cornea while the front surface corrects the residual refractive cylinder

The Core Design Problem: Stabilization

Stabilization Techniques

Fitting Methods

1. Diagnostic fitting - trial lens placed on the eye and assessed directly
2. Trial lens fitting - a set of trial toric lenses of varying power/axis used to refine the fit and over-refraction
3. Empirical fitting - lens ordered directly from refraction and keratometry data without a trial lens on the eye

Assessing Fit and Rotation

• Toric soft lenses carry laser-etched orientation/scribe marks, typically at the 6 o'clock position (or 3, 6, 9 o'clock)
• With the lens settled, the examiner notes whether the mark sits exactly at 6 o'clock, or has rotated to the right or left of it, and by how many degrees (usually estimated in 5-10° increments using a slit lamp graticule)
• Acceptable/stable rotation is usually ≤ 5-10°, provided it is consistent and repeatable

LARS Rule (Axis Compensation)

Advantages over Spectacles

• Better peripheral field, no minification/magnification, no prismatic aberration of spectacle lenses
• Better cosmesis, no fogging, unrestricted for sports/occupation
• More natural retinal image size, useful in high anisometropic astigmatism

Problems / Complications

• Lens rotation causing fluctuating or reduced vision
• Increased lens awareness/discomfort from ballast or truncation zones
• Reduced oxygen permeability in thicker ballast zones (hypoxia risk)
• Contact lens-associated giant papillary conjunctivitis (GPC), corneal neovascularization, deposits, and other standard soft-lens complications
• Higher cost and more complex/time-consuming fitting compared with spherical lenses

Contraindications

• Active ocular surface disease, severe dry eye, chronic blepharitis
• Irregular astigmatism unresponsive to soft toric correction (needs RGP/scleral toric)
• Poor patient compliance/hygiene, monocular patients (relative), recurrent lens-related infections

Artificial Intelligence (AI) in Ophthalmology

Introduction

Why AI is Important in Ophthalmology

• Large number of patients with diabetic retinopathy and glaucoma, but shortage of eye specialists, especially in rural areas
• Screening is repetitive and image-based - ideal task for a machine
• AI can work fast, is consistent, and can be used even by non-specialists (technicians, primary care staff) for screening
• Helps in early detection, reducing avoidable blindness

Main Applications

• This is the most advanced use of AI in ophthalmology
• Deep learning algorithms analyze fundus photographs and grade DR as none, mild, moderate, severe, or needing referral
• IDx-DR (now LumineticsCore) was the first AI system approved by the US FDA (2018) for fully autonomous DR screening, meaning it can give a result without a doctor checking the image
• EyeArt is another approved autonomous AI system, which also assists in detecting glaucoma and AMD

• AI helps detect optic disc changes, retinal nerve fibre layer loss on OCT, and visual field defects
• Useful for tracking disease progression over time by picking up subtle changes humans may miss
• Still more difficult than DR screening because glaucoma has no single fixed diagnostic criterion

• AI can detect drusen, geographic atrophy, and choroidal neovascularization on OCT and fundus images
• Helps predict which dry AMD patients are likely to progress to wet AMD

• AI algorithms (e.g., i-ROP) assist in grading plus disease in premature babies, useful where paediatric retina experts are scarce

• AI helps grade cataract severity from slit-lamp images and assists in IOL power calculation

• AI is used to detect keratoconus early from corneal topography, before it is clinically obvious

• Oculoplastics: analyzing eyelid and facial images
• Strabismus: photo-based angle estimation
• Surgical planning: refractive surgery outcome prediction, robotic-assisted and AI-guided surgical instruments
• Tele-ophthalmology: combining AI screening with remote reporting to reach rural populations

How It Works (Simple Explanation)

1. Thousands of labeled images (e.g., "DR present" / "DR absent") are fed to a deep learning model
2. The model (usually a Convolutional Neural Network, CNN) learns patterns associated with disease
3. When shown a new image, the trained model gives a probability or grading of disease
4. Output can be autonomous (final decision, no doctor needed) or assistive (flags cases for the doctor to review)

Advantages

• Fast and consistent results
• Can screen large populations, especially in areas with few ophthalmologists
• Reduces workload on specialists, allows them to focus on referred/complex cases
• Improves early detection and reduces preventable blindness
• Can detect very subtle changes that are hard for the human eye to catch

Limitations / Challenges

• Needs large, high-quality, diverse datasets to train - may not perform well on populations different from training data (generalizability problem)
• "Black box" problem - difficult to know exactly why the AI gave a certain result
• Ethical and legal issues - who is responsible if AI makes a wrong diagnosis
• Cannot fully replace clinical judgment, history-taking, and correlation with other systemic disease
• Cost of technology and need for good quality cameras/OCT machines
• Data privacy concerns

Conclusion