Regularized phase retrieval algorithms for X-ray phase tomography of 3D bone cell culture analysis

International audienceRegenerative medicine is receiving increased interest. In the field of bone tissue engineering, a recurrent problem is to select efficient association of scaffolds and bone cells to maximize bone formation. Sensitive imaging modalities are required to quantify both cells and mineralized tissue. Tothis aim, X-ray phase contrast imaging is an attractive modality since it offers three orders of magnitude higher sensitivity than conventional radiology. In-line phase contrast imaging consists in letting a spatially coherent beam propagate after passing through a sample [1]. Projections areacquired at several angles, for several sample-to-detector distances. The phase can be extracted from projections using a so-called phase retrieval algorithm. A tomographic reconstruction algorithm is then applied to these phase projections to obtain a 3D-map of the refractive index in the object.In this work, we compare several phase retrieval methods based on different priors [3, 4, 5] in porous bone scaffolds seeded with bone forming cells. These algorithms not only enable to discriminate bone cells, newly formed bone and scaffolds, but quantify the volume and density of the calcified fraction, as well as the volume of extra-cellular matrix generated by the bone cells, by thresholding the refractive index maps. Further, the use of heterogeneous object algorithms enabled quantification of the refractive index of the soft tissue compartment, as well as segmentation of osteoblasts trapped in the pre-bone matrix, which was not possible using standard μCT