Mandibular growth prediction: mean growth increments versus mathematical models

Area of Science:

• Orthodontics
• Pediatric Growth and Development
• Biostatistics

Background:

• Accurate prediction of craniofacial growth is crucial in orthodontics.
• Previous methods often relied on linear extrapolation or mean growth increments.
• Understanding population growth curves is essential for individual patient assessment.

Purpose of the Study:

• To compare the accuracy of two methods for predicting cephalometric distance sella-gnathion at age 15.
• To evaluate predictions based on mean annual velocities versus a polynomial growth curve model.
• To identify biases in different growth prediction approaches.

Main Methods:

• Utilized a sample of 223 children with previous cephalometric measurements at ages 11, 12, and/or 13.
• Estimated the sella-gnathion distance at age 15 using two predictive models.
• Calculated root mean square error (RMSE) to assess prediction accuracy for both methods.

Main Results:

• Root mean square errors decreased with age for both sexes, indicating improved prediction accuracy.
• RMSE values were comparable between the mean increment and polynomial methods at age 13.
• Predictions based on mean increments showed significant bias, over- or underestimating growth for children deviating from the average.
• Polynomial model predictions were conditional on individual child size, thus unbiased.

Conclusions:

• Polynomial growth models provide more accurate and unbiased predictions of craniofacial growth compared to mean increment methods.
• The bias in mean increment predictions stems from an inability to account for changing variance in growth.
• Conditional growth models are superior for personalized orthodontic treatment planning.