A better statistical method of predicting postsurgery soft tissue response in Class II patients

Area of Science:

• Orthodontics
• Oral and Maxillofacial Surgery
• Biostatistics

Background:

• Accurate prediction of postsurgical soft tissue response is crucial for Class II malocclusion treatment.
• Conventional statistical methods may not fully capture the complex interplay of variables influencing soft tissue changes.

Purpose of the Study:

• To introduce and evaluate a superior statistical methodology for predicting postsurgical soft tissue outcomes.
• To compare the predictive accuracy of partial least squares (PLS) against ordinary least squares (OLS) in Class II patients.

Main Methods:

• Eighty patients with severe Class II malocclusions undergoing surgical correction were analyzed.
• Two multivariate prediction equation methods, ordinary least squares (OLS) and partial least squares (PLS), were applied.
• Leave-one-out cross-validation was employed to assess predictive performance, calculating bias and mean absolute prediction error.

Main Results:

• The partial least squares (PLS) method demonstrated significantly higher prediction accuracy.
• PLS provided a more precise estimation of soft tissue profile changes compared to OLS.

Conclusions:

• The multivariate PLS method is a more satisfactory approach for predicting soft tissue profile changes post-surgery in Class II malocclusion correction.
• This finding suggests PLS can enhance the predictability and outcomes of surgical orthodontic treatments.