Ocular Centiles

Early risk detection through age-specific growth curves: the chance to intervene before myopia begins

Just as height-weight centiles in pediatrics are used to monitor a child's growth, ocular biometry data can be tracked using age-specific centile curves. This approach makes it possible to identify at-risk children even before myopia has clinically appeared.

The Centile Concept and Clinical Significance

A child's axial length being at the 75th centile means their eye is longer than 75% of children the same age. Even if myopia hasn't started yet, this indicates that axial elongation is above the expected trajectory and myopia risk is elevated.

Chen et al. (2016) demonstrated that axial length above the 75th centile predicts future high myopia development with high sensitivity. Sanz Diez et al. (2019) proposed standardizing growth curves for monitoring refractive development.

The centile system answers whether a single measurement is "normal or abnormal." But the real power lies in serial measurements: a child staying within their centile band indicates stability, while crossing between bands signals accelerating or decelerating growth.

Chen et al. 2016; Tideman et al. 2018; Sanz Diez et al. 2019; Flitcroft 2014

Axial Length (AL) Centiles

Axial length is the most direct and reliable biometric indicator of myopia progression. While refraction values can vary with cycloplegic status and measurement method, axial length is objective and reproducible.

Clinical threshold values:

AL < 22 mm (adult): high likelihood of hyperopia
AL 22–24.5 mm: normal / mild myopia range
AL ≥ 24.5 mm (high for age in children): myopia risk
AL ≥ 26 mm: high myopia risk, retinal examination recommended
AL ≥ 28 mm: pathological myopia risk, regular fundus monitoring required

AL Growth Rate

Annual axial elongation rate is the most sensitive indicator of treatment efficacy. Normal emmetropic growth rate between ages 3-8 is ~0.1 mm/year, while untreated myopic children can reach 0.2–0.4 mm/year. A growth rate > 0.3 mm/year is a warning signal requiring aggressive treatment modification.

MiyoLab automatically calculates AL growth rate for each patient, compares it against age norms, and generates alerts through the clinical decision support system when abnormal acceleration is detected.

Tideman et al. 2016 (Rotterdam); He et al. 2015; Mutti et al. 2007 (OLSM)

AL/CR Ratio: Size-Independent Refractive Indicator

The axial length to corneal radius (AL/CR) ratio is a powerful biometric index that predicts refractive status independent of eye size. It is more reliable than AL alone for comparing children of different ages and sizes.

According to a 2025 meta-analysis of 9 studies covering 24,030 children (Clin Exp Optom), age-specific AL/CR thresholds achieve AUC 0.93 sensitivity in myopia screening, outperforming AL or SER alone.

Clinical threshold values:

AL/CR < 2.9: emmetropic / hyperopic
AL/CR 2.9–3.0: pre-myopic monitoring zone
AL/CR > 3.0: myopia highly likely present
AL/CR > 3.2: strong predictor of high myopia

MiyoLab automatically calculates the AL/CR ratio when AL and corneal measurements are entered, compares it against age-expected ranges, and alerts when anomalies are detected.

2025 meta-analysis (Clin Exp Optom, 9 studies, n=24,030); Mutti et al. 2007; Tideman et al. 2016

Refraction Centiles and Emmetropization Monitoring

Most children are born hyperopic (+2 to +3 D). During normal emmetropization, this hyperopia gradually decreases to between 0 and +0.75 D by school age. Disruption of this process heralds myopia development.

Children with less than +0.75 D of hyperopic "buffer" at age 6-7 are in the highest risk group for myopia development (Zadnik et al. 1999). MiyoLab tracks this emmetropization trajectory within the BHVI model and provides early warning.

When refraction centiles are evaluated together with axial length centiles, clinical decision power increases. A child rising in both refraction and AL centiles is a very strong indicator of myopia risk.

Zadnik et al. 1999; Flitcroft 2014; Sankaridurg et al. 2015

View All References →