Multivariate data analysis for the detection of surface defects in the dental enamel during orthodontic treatment
by
José Italo Cortez
31 poniente 1304, Puebla, Puebla, México
Coauthors: Cortez José Italo, González Flores Marcos, Caldera Miguel Javier, Perea González Gloria Patricia, Vega Galina Javier, Liliana Cortez

In this work, the acquisition and data processing of the rugosity of the surface dental enamel is shown. The purpose of this is the detection of physical defects (de-mineralized zones) in the dental enamel through the comparison of the mean values of the samples generated in orthodontic treatment. The objective of the work is the development of software which allows the comparison of mean values of two samples to synthesize and acquire the covariance matrix from a data matrix of each sample and later obtain an estimation of both matrices. From this matrix its inverse is obtained and finally the Hottelling statistics T2 and F are calculated for multivariate data, to demonstrate if there are differences in the mean vectors. The experiment carried out with non-invasive technology developed in the Laboratory of “Non-Invasive Methods for Dental Diagnosis” allowed corroboration that the orthodontic treatment generates physical changes in the dental enamel, since it was proven that the null hypothesis in which it is assumed that the mean vectors are equal is rejected, the experiment zone of the dental organ was the middle third, in which the samples without treatment and application of toothpaste, without treatment and application of etching acid, without treatment and application of adhesive, with toothpaste and etching acid, with toothpaste and adhesive, and finally application of etching acid with adhesive were compared, with the objective of proving the hypothesis that if the mean vectors are equal or not. The results obtained allowed the hypothesis that the mean vectors are equal to be rejected, which is made evident in the different stages of the orthodontic treatment. The results obtained are satisfactory and work is continued in this technique for future analysis and different zones of the tooth.

Key Words: rugosity of the dental enamel, multivariate data analysis, covariance, Hottelling and F statistics, inverse matrix, data matrix.

Date received: June 22, 2006