<- Back to digest

Journal of Pediatric Orthopaedics - 2026-10-19 - Journal Article

Knee Bone Age Estimation Using Capping Interval Index on MRI: Development and Validation of a Quantitative Method.

Lee W, Song KI, Yu SY, Cho YJ, Song MH, Cho TJ, Shin CH

retrospective cohortLOE IIIn = 695 (modeling n=417, validation n=278)N/A

Topics

pediatrics
PMID: 42895227DOI: 10.1097/BPO.0000000000003501View on PubMed ->

Key Takeaway

A two-parameter MRI-based capping interval index (CII) model estimates knee bone age with a mean error of only 0.1 ± 1.1 years and Spearman correlation of 0.960 in an independent validation cohort of 278 patients.

Summary Depth

Choose how much analysis to show on this article page.

Summary

This study developed and validated a quantitative knee MRI method for skeletal maturity estimation in patients ≤18 years without conditions affecting knee anatomy, using 26 coronal MRI parameters measured in 695 participants. Femoral and tibial capping interval indices (CII)—normalized distances from metaphyseal end to nearest ossified epiphysis—were the only parameters achieving R² >0.9 in both sexes and formed sex-specific exponential regression models. In the validation cohort, estimated knee bone age differed from chronological age by 0.1 ± 1.1 years (P=0.534) with rs=0.960.

Key Limitation

Chronological age substituted for true skeletal age as the reference standard, meaning the model's accuracy in patients with growth disturbances—the highest-stakes clinical use case—remains unvalidated.

Original Abstract

BACKGROUND

Knee bone age can differ from hand bone age and is more relevant to lower-extremity pediatric conditions. This study developed and validated a quantitative method for estimating skeletal maturity using knee MRI images.

METHODS

We identified 695 participants aged 18 years or younger without conditions affecting knee anatomy or maturation who underwent knee MRI. Participants were randomly allocated to the modeling (n=417) and validation (n=278) groups. Twenty-six maturation-related parameters were measured on coronal MRI, including the articular and peripheral cartilaginous thicknesses, physeal thicknesses, ossified epiphyseal width, and capping interval. The capping interval was defined as the distance from the medial or lateral metaphyseal end to the nearest ossified epiphysis. Each parameter was normalized to metaphyseal width and assessed by univariable linear regression against log-transformed chronological age. Top-ranked variables were entered into sex-specific multivariable models.

RESULTS

The sums of the medial and lateral capping intervals from the femur and tibia were the only parameters with R2>0.9 in both sexes and were defined as the femoral and tibial capping interval indices (CIIs), respectively. Final models were age=exp(-1.918[femoral CII]-1.797[tibial CII]+2.813) for males and age=exp(-2.056[femoral CII]-2.487[tibial CII]+2.738) for females. In the validation group, estimated knee bone age showed minimal difference from chronological age (0.1±1.1 y, P=0.534) and very strong correlation (rs=0.960).

CONCLUSIONS

Knee bone age estimated from the femoral and tibial CII showed close agreement with chronological age in the independent validation cohort. This study introduces a novel, simple quantitative method that enables estimation of bone age on knee MRI using a minimal number of parameters.

LEVEL OF EVIDENCE

Level III-diagnostic.