In recent years, bone fracture detection and classification has been a widely discussed topic and many researchers have proposed different methods to tackle this problem. Despite this, a universal approach able to classify all the fractures in the human body has not yet been defined. We aim to analyze and evaluate a selection of papers, chosen according to their representative approach, where the authors applied different deep learning techniques to classify bone fractures, in order to select the strengths of each of them and try to delineate a generalized strategy. Each study is summarized and evaluated using a radar graph with six values: area under the curve (AUC), test accuracy, sensitivity, specificity, dataset size and labelling reliability. Plus, we defined the key points which should be taken into account when trying to accomplish this purpose and we compared each study with our baseline. In recent years, deep learning and, in particular, the convolution neural network (CNN), has achieved results comparable to those of humans in bone fracture classification. Adopting a correct generalization, we are reasonably sure that a computer-aided diagnosis (CAD) system, correctly designed to assist doctors, would save a considerable amount of time and would limit the number of wrong diagnoses.

The aim of this study was to evaluate how imaging parameters at clinical dental CBCT affect the accuracy in quantifying trabecular bone structures, contrast-to-noise ratio (CNR) and radiation dose. 15 radius samples were examined using CBCT (Accuitomo FPD; J. Morita Mfg., Kyoto, Japan). Nine imaging protocols were used, differing in current, voltage, rotation degree, voxel size, imaging area and rotation time. Radiation doses were measured using a kerma area product-meter. After segmentation, six bone structure parameters and CNRs were quantified. Micro-CT (μCT) images with an isotropic resolution of 20 μm were used as a gold standard. Structure parameters obtained by CBCT were strongly correlated to those by μCT, with correlation coefficients >0.90 for all studied parameters. Bone volume and trabecular thickness were not affected by changes in imaging parameters. Increased tube current from 5 to 8 mA, decreased isotropic voxel size from 125 to 80 μm and decreased rotation angle from 360° to 180° affected correlations for trabecular termini negatively. Decreasing rotation degree also weakened correlations for trabecular separation and trabecular number at 80 μm voxel size. Changes in the rotation degree and tube current affected CNR significantly. The radiation dose varied between 269 and 1153 mGy cm(2). Trabecular bone structure can be accurately quantified by clinical dental CBCT in vitro, and the obtained structure parameters are strongly related to those obtained by μCT. A fair CNR and strong correlations can be obtained with a low radiation dose, indicating the possibility for monitoring trabecular bone structure also in vivo.