# Predictive Models for Stylohyoid Complex Elongation: A Multivariate Statistical Analysis with Evidence-Based Diagnostic Criteria

## TL;DR

This study uses statistical models to determine diagnostic criteria for stylohyoid complex elongation, finding a 32.5 mm cutoff more accurate than traditional values.

## Contribution

The study introduces evidence-based diagnostic criteria and predictive models for stylohyoid elongation using advanced statistical methods.

## Key findings

- A 32.5 mm cutoff for styloid process length showed high diagnostic accuracy (AUC=0.87).
- Age and female gender were significant predictors of elongation.
- Four distinct phenotypic groups were identified through clustering.

## Abstract

Stylohyoid complex elongation represents significant anatomical variations with clinical implications, yet comprehensive morphometric analyses using advanced statistical modeling remain limited in establishing evidence-based diagnostic criteria.

This cross-sectional study analyzed 100 digital panoramic radiographs from a Colombian population. Advanced statistical methods included multivariate regression analysis, receiver operating characteristic (ROC) curve analysis, cluster analysis, and factor analysis. Morphometric measurements were validated using intraclass correlation coefficients and Bland-Altman analysis.

Mean styloid process length was 36.79±10.15 mm. A 97% prevalence of elongation >25mm was observed. Multivariate logistic regression identified age (β=0.31, p<0.001) and female gender (β=4.23, p=0.030) as independent predictors. ROC analysis revealed optimal diagnostic cutoff at 32.5 mm with excellent performance (AUC=0.87, sensitivity=89.2%, specificity=78.6%). Factor analysis identified three principal components explaining 78.4% of total variance. K-means clustering revealed four distinct phenotypic groups.

This study establishes evidence-based diagnostic criteria for stylohyoid complex evaluation through advanced statistical modeling. The 32.5 mm cutoff demonstrates superior performance compared to traditional values, while predictive models provide reliable risk assessment capabilities for precision medicine applications.

Key words:Eagle syndrome, styloid process, panoramic radiography, physiological calcification, temporal bone, predictive models.