Evaluation of the Relationship Between the Age of the Patient and the Mandibular Trabecular Bone Structure Through Dental Cone Beam Tomographic Images by Means of a Convolutional Neural Network

Abstract

 Osteoporosis is a systemic condition that affects bone mineral density in individuals, affecting mainly women and is diagnosed by DXA examination. The research used 137 cone beam tomography scans, from which 1389 mandibles samples from female patients and 633 from male patients were selected for CNN training, which is composed of 3 dense layers with 100 neurons each and Relu activation function, update of weights by the Adam algorithm, using MaxPooling in each convolution; the dense layer uses 100 neurons, with the activation function. The neural network training accuracy was 98% for males and 94% for females, with area under the ROC curve (AUC) equal to 0.94 and 0.81 for the respective genders. The accuracy obtained in validating the CNN was 98% for males and 89% for females. In the supervised test, using 5 cuts from 10 exams of the test set of each sex; approximately 100% accuracy was obtained for both sexes. Thus, it is concluded that the proposed model is capable of classifying the samples in the proposed age groups and proved to be robust and solid in the tests.