Expansion Pattern of Miniscrew Assisted Rapid Maxillary Expansion (MARPE) After Midpalatal Suture Separation: Clustering-Based Classification.

objectivesThis study aimed to establish a clustering-based classification model for skeletal expansion outcomes after miniscrew-assisted rapid maxillary expansion (MARPE) and evaluate its clinical implications. MATERIAL AND

methodsThe study samples comprised 61 patients with transverse discrepancy who underwent maxillary skeletal expansion using MSE I. Cone-beam computed tomography images before (T0) and after (T1) expansion were analysed with nine measurements assessing transverse changes in the maxilla, molars and appliance. Gaussian mixture model clustering, random forest models and decision trees identified expansion patterns and key classifiers, establishing a classification system. Two- and one-way analyses of variance compared intergroup characteristics.

resultsThe clustering-based classification model identified four distinct clusters of skeletal expansion patterns, establishing two key classifiers: expansion efficiency (good/poor) and parallelism (parallel/V-shaped). Poor responders had an older age and showed significantly lower efficiency, more V-shaped tendencies, greater activation loss and longer intervals until interincisal diastema than good responders (p < 0.001). V-shaped responders exhibited less efficiency, reduced posterior expansion and increased miniscrew slippering (p < 0.001). Twenty days until interincisal diastema can be used to monitor expansion efficiency. Age is a useful predictor of efficiency, except during the ambiguous period of 16.81-20.14 years (17.91-19.88 for males, 16.95-18.94 for females) and for outliers.

conclusionThis study introduces a concrete classification system for MARPE outcomes, emphasising the role of expansion efficiency and parallelism in optimising clinical results. These four identified phenotypes provide a data-driven validation of the existing concepts of expansion patterns and their implications for treatment.