Aggressive periodontitis defined by recursive partitioning analysis of immunologic factors

Area of Science:

• Periodontology
• Immunology
• Computational Biology

Background:

• Periodontitis is classified into aggressive (AgP) and chronic (CP) forms, considered distinct clinical entities.
• Previous research suggests a non-linear model for periodontitis progression.

Purpose of the Study:

• To evaluate immunologic parameters as predictors for distinguishing AgP from CP.
• To extend findings on non-linear models of periodontitis progression.

Main Methods:

• Recursive Partitioning Analysis (RPA) was applied to three independent population samples.
• Seventeen immunologic parameters were analyzed, with control group means serving as cut-off points.
• Leave-one-out cross-validation (LOOCV) and Kullback-Leibler divergence (DKL) were used for model evaluation.

Main Results:

• Seven classification trees identified Interleukin (IL)-4, IL-1, and IL-2 as key predictors for AgP.
• Immunoglobulin (Ig)A, IgM, and CD4/CD8 ratio showed less predictive power for ruling out AgP.
• A model combining IL-4, IL-1, IL-2 was optimal for ruling out AgP, while a CD4/CD8 ratio model aided in ruling in AgP.

Conclusions:

• Immunologic parameters, when integrated into appropriate models, can effectively complement clinical examinations.
• These models show promise for improving the diagnosis of aggressive periodontitis.