# Concordance of Cephalometric Classifications of Divergence in Orthodontics

## TL;DR

This study compares five cephalometric methods for classifying facial divergence in orthodontics and finds significant differences in their classifications.

## Contribution

The study evaluates the concordance of five commonly used cephalometric methods for vertical skeletal classification in orthodontics.

## Key findings

- Divergence prevalence varied significantly between methods, with hypodivergent rates ranging from 16% to 30%.
- Agreement was substantial between SN/GoGn and MP/SN (κ = 0.656), but lowest when FAA was compared (κ ≈ 0.11-0.22).
- Overall agreement across all methods was fair (κ = 0.375), suggesting methods are not interchangeable.

## Abstract

Background

Vertical skeletal pattern is central to orthodontic diagnosis. We compared five cephalometric methods-Steiner (Sella-Nasion to mandibular plane, SN/GoGn), Downs (mandibular plane to Sella-Nasion, MP/SN), Tweed (Frankfort-mandibular plane angle, FMA), Arnett (palatal plane to mandibular plane, PP/MP), and Ricketts (facial axis angle, FAA)-for classifying facial divergence.

Methods

We retrospectively analyzed 300 lateral cephalograms (114 males, 186 females; mean age 26.4 ± 9.9 years) from skeletal Class I adults (A point-Nasion-B point angle, ANB, 0° < ANB < 4°). Each method classified subjects as hypo-, normo-, or hyperdivergent. Agreement was assessed using Cohen’s kappa (κ), and ordinal association using Kendall’s tau-b (τ_b).

Results

Divergence prevalence varied by method: hypodivergent 16% (48/300) with MP/SN to 30% (90/300) with SN/GoGn; hyperdivergent 15.3% (46/300) with PP/MP to 25% (75/300) with MP/SN. Agreement was substantial for SN/GoGn vs MP/SN (κ = 0.656, p < 0.001), moderate for several other pairs (κ ≈ 0.49-0.54), and lowest when FAA was compared (κ ≈ 0.11-0.22). Overall agreement was fair (κ = 0.375, p < 0.001).

Conclusions

Cephalometric methods are not interchangeable for vertical classification. Reporting at least two complementary indices is advisable; FAA should not be used alone. Standardized thresholds and multi-index confirmation may improve diagnostic consistency.