Region-specific statistical analysis of landmark configurations applied to rapid maxillary expansion data from cone-beam computerized tomography.

The aim of this study was to apply a region-specific method for the statistical analysis of landmark configurations to a rapid maxillary expansion data set obtained from cone-beam computerized tomography.A method of analyzing landmark configurations using persistent homology and dimensionality reduction is used to analyze region-specific treatment effects involving 2 types of maxillary expansion treatments, compared with a control group.The method identified maxillary width as the primary feature of variability in the data and showed that both treatment groups exhibited significantly greater maxillary expansion than the control group throughout the study period, with no significant differences found between the 2 treatments.The choice of distance used to compare landmark configurations and the focus on more specific regions of interest produce more detailed results. This approach can be used after a primary analysis with a larger number of landmarks has identified regions deserving further inquiry.