Optical Coherence Microscopy (OCM) is a three-dimensional imaging technique that provides cross-sectional views of the subsurface microstructure of biological tissue, with a high axial and lateral resolution. In OCM, a low time coherence light source is split into two beams, propagating along the reference and object arms. Light backscattered by variation of the index of refraction in the sample is then recombined with the strong reference field. The high intensity of the reference beam provides a coherent amplification of the weak sample field, resulting in a high sensitivity. By detecting in the Fourier domain, only a two-dimensional scan is required, as the whole depth structure is extracted from a single spectrum acquisition. For volumetric imaging, this parallel acquisition confers a tremendous speed advantage over classical techniques, such as confocal fuorescence microscopy. In addition, the contrast in OCM relies on fundamental properties of the object, namely scattering and absorption. This intrinsic contrast offers the advantage of not requiring any staining process of the sample. Above all, relying on the sensitive phase information, motions on a nanometer scale can be revealed in a non-contact and non-invasive manner. However, depending on the application, this asset may become a drawback as the signal lacks specificity, in comparison to exogenous labels. The aim of this thesis work was to evaluate and implement contrast enhancement mechanisms in OCM for 3D cell imaging. In brief, an improvement of the sensitivity to weakly scattering objects, such as single cells combined with the molecular specificity of nanolabels were investigated. The developed technique was then employed for the imaging of dynamic cellular processes. First, we realized a dark-field illumination scheme for OCM (dfOCM). This design efficiently filters scattered light and suppresses specular reflections arising from glass interfaces, almost compulsory in cell preparation. The sensitivity to backscattered light was drastically enhanced in comparison to the classical approach. After a careful analysis of the imaging performances, tomograms of living cells were acquired, where even subcellular structures could be identified. Second, we implemented photothermal optical lock-in OCM (poli-OCM), which uses the photothermal contrast to exclusively localize gold nanoparticles within scattering medium. We provided a theoretical analysis of the signal-to-noise ratio and explained thoroughly the lock-in detection effect. Significant parameters such as resolution and depth of field were measured experimentally and we applied the combined dfOCM/poli-OCM technique for the 3D imaging of gold nanoparticles in living cells. Finally, dfOCM has been successfully applied to the monitoring of dynamic cellular processes, opening new horizons for further investigations.