Mesoscopic three-dimensional hard X-ray imaging of central and peripheral nervous system

Abstract

Micro computed tomography (μCT), either by means of hard X rays from synchrotronradiation (SR) facilities, or from advanced laboratory sources, has been proven as a powerful method for the nondestructive three-dimensional visualization of biological specimens with isotropic micro- and even nanometer resolution. The established absorption-contrast modality of μCT has been sometimes associated with the need for contrast agents, whereas the more advanced phase-contrast modality has yielded superior results for biological specimens without staining. For around three decades, phase-contrast μCT has been considered between a hundred and a thousand times better than absorption-contrast μCT, based on the ratio of the imaginary and the real part of the complex refractive index that could be determined using the two modalities at desired photon energies. The results of the present study elucidate that for formalin-fixed, paraffin-embedded nervous tissues, conventional μCT delivers a much better contrast than originally expected. Related measurements were performed at a SR facility using monochromatic X rays. The photon energies were not equal for absorption- and phase-contrast measurements, but selected to obtain optimized contrast within a reasonable period of time. The choice of the photon energy, which is much smaller for absorption-contrast μCT, explains that the contrast difference between phase- and absorption-contrast μCT, indicated by the contrast-to-noise ratio of anatomical regions in the respective datasets being about two times better for phase μCT, is much smaller than reported in literature. It should be highlighted that μCT in absorption- and phase contrast are complementary methods and a combination might give additional quantitative insights into the three-dimensional images. For example, one can register the data and build a joint histogram from the common volume to segment anatomical features indistinguishable using just one imaging modality. The main relevance of such results lies in the opportunity to employ laboratory-based μCT, which are much better accessible and cost-effective than μCT at SR facilities. This approach was benchmarked on peripheral nerve reconstruction. The threedimensional visualization of regenerating nerves inside collagen scaffolds was feasible and included the automatic extraction of anatomical features to quantify the regeneration. Indeed, the characteristic parameters, revealed from the conventional μCT data, were significantly different between regenerating and control nerves. The approach including specimen preparation, data acquisition, and analysis has been useful for the investigations of the anatomical alterations in medial temporal epilepsy and the time-critical diagnosis of vasculitis prior to the standard histology.