Cross-sectional data accurately model longitudinal growth in the craniofacial skeleton

Area of Science:

• Craniofacial biology
• Growth modeling
• Biostatistics

Background:

• Longitudinal sampling is optimal for biological growth studies but faces practical limitations like cost and patient health.
• Cross-sectional samples offer a more feasible alternative, but their utility for estimating growth milestones remains uncertain.
• The Craniofacial Growth Consortium Study (CGCS) provides a valuable dataset for investigating this discrepancy.

Purpose of the Study:

• To assess the accuracy of growth parameter estimation using cross-sectional data compared to longitudinal data.
• To determine the minimum sample size required for reliable growth milestone prediction from cross-sectional data.
• To validate a method for growth modeling when repeated radiologic imaging is not feasible.

Main Methods:

• Utilized existing longitudinal data from the Craniofacial Growth Consortium Study (CGCS).
• Extracted cross-sectional samples of varying sizes (5 to full sample) from the CGCS.
• Analyzed linear trait measurements, performing sex-specific predictions of growth parameters.
• Calculated mean absolute differences between cross-sectional estimates and longitudinal growth rates.

Main Results:

• Cross-sectional sample means approximated longitudinal growth rates from the full CGCS sample.
• Mean absolute differences were below 1 mm with cross-sectional sample sizes exceeding approximately 200 individuals.
• Accurate estimation of growth parameters and milestones is achievable using cross-sectional data.

Conclusions:

• Cross-sectional data can reliably estimate population-level growth parameters and milestones, comparable to longitudinal data.
• This methodology enhances the utility of cross-sectional datasets for growth modeling.
• The findings are applicable to various growth types and situations where repeated imaging is impractical, such as cone-beam CT scans.