X-ray CT has always been a technique of choice to image bone at different scales. The development of X-ray micro CT which has been strongly motivated by the application to bone microarchitecture imaging now plays a major role in pre-clinical animal imaging and in microscopic imaging studies. After a brief review of progresses in X-ray CT, the presentation will focus on 3D synchrotron nano-CT and its ability to provide quantitative attenuation or phase maps at a given energy with a high signal-to-noise ratio. We will discuss the notion of quantitative imaging and highlight the different parameters involved in image acquisition and their impact on image quality. We will present recent applications of synchrotron nano CT to image bone tissue at the cellular level with an isotropic spatial resolution. The osteocyte system, which is increasingly considered as being fundamental in bone tissue remodeling, aging and reparation, can be imaged thanks to synchrotron nano-CT. The development of dedicated image analysis algorithms allows to go beyond imaging and extract quantitative parameters on cell lacunae. We will present new data on the lacuno-canalicular network in a series of 27 human femoral samples. This analysis opens further perspective to estimate relevant bone mechanobiology parameters to model the poroelastic behavior of bone. Then we will illustrate some developments in CT imaging related to the promises of artificial intelligence. In this context, deep learning methods are increasingly used for both improving image quality and image analysis. We will present applications of convolutional neural networks to X-ray dose reduction in dental CT, super resolution in HR-pQCT imaging, and image reconstruction in X-Ray spectral CT which is an emerging quantitative imaging technique at the patient scale. Finally we will show the use of CNNs to segment the large data volumes obtained from 3D synchrotron nano-CT and applications to the analysis of osteocytes and chondrocytes in the knee joint.