Résumé : Background. The development of non-invasive imaging techniques has been stimulated by the shortcomings of histopathology. Currently the only valid diagnostic technique in dermatology is skin biopsy which remains a painful, invasive intervention for the patient. Moreover, this approach is not always convenient for monitoring and follow-up of a skin disease. Optical imaging technologies could solve these shortcomings as they are fast, precise, repeatable and painless. There are four established non-invasive skin imaging techniques used in daily practice: dermoscopy, high-frequency ultrasound, reflectance confocal microscopy (RCM) and conventional optical coherence tomography (C-OCT). In imaging there is a trade-off between resolution and penetration depth. The former permits the visualization of cells, if the resolution is at least 3 µm. The latter enables the recognition of patterns and structures in deeper layers of the skin if the penetration depth is deeper than 150 µm. New non-invasive techniques using infrared light sources have been developed recently. The technique used in this work is a high-definition optical coherence tomography (HD-OCT).Objectives. The overall aims of this thesis were the feasibility of HD-OCT to visualize in/ex vivo, in real time and in 3-D the cellular and structural morphology of the skin, secondly the assessment of the capability of this technology to measure in vivo and real time the cutaneous optical properties, and finally the determination of the contribution of this technique to the non-invasive near-infrared imaging technologies. Five specific objectives have been established: i) could cells be observed in their 3-D microenvironment in normal and diseased skin, ii) could we describe morphologic features of cells and structures in normal and diseased skin (m_HD-OCT), iii) could these morphologic features be quantified by optical property analysis (o_HD-OCT), iv) was it possible to perform accurate thickness measurements in normal and diseased skin, and finally v) what was the diagnostic potential of this technique?Methodology. HD-OCT uses a combination of parallel time-domain interferometry, high power tungsten lamp (with Gaussian filter, very low lateral coherence and ultra-high bandwidth (1300 nm +/- 100 nm)), and last but not least, full field illumination with real time focus tracking. A constant homogeneous resolution of 3 µm resolution in all three dimensions is obtained up to a depth of 570 µm. Hence, the system is capable of capturing real time full 3-D images. Moreover, the in vivo assessment of optical properties of the skin is only applicable to OCT when operating in focus-tracking mode, which is the case for HD-OCT. The means to obtain answers to the five specific questions were the comparison of en face HD-OCT images with RCM and HD-OCT cross-sectional images with histopathology and C-OCT. Results. At least 160 line pares were observed by imaging a high resolution phantom with HD-OCT. This suggested a 3 µm lateral resolution. The presence of cells such as keratinocytes, melanocytes, inflammatory cells, fibroblasts and melanophages in their 3-D cutaneous microenvironment in vivo as well as ex vivo has been demonstrated . A qualitative description of structures and patterns in normal and diseased skin could be performed by HD-OCT. Clear structural changes of the epidermis, dermo-epidermal junction, papillary dermis and reticular dermis related to intrinsic skin ageing could be observed. Lobulated structures, surrounded by stretched stromal fibers and arborizing vessels, could be demonstrated in nodular basal cell carcinoma (BCC). The o_HD-OCT of normal and diseased skin could be assessed in vivo. This approach permitted the quantitative assessment of the OCT signal attenuation profiles of normal healthy skin, actinic keratosis (AK) and squamous cell carcinoma (SCC). Differences in signal attenuation profiles could be demonstrated between these three groups. These differences were also observed between BCC subtypes. The slope of the exponential attenuation of the signal in the upper part of the epidermis was very high in benign nevi. The more malignant the lesion the lower the slope. Thickness measurements of epidermis and papillary dermis could be performed by m_HD-OCT, based on a cross-sectional images and their corresponding en face image. More accurate measurements of epidermal and papillary dermal thickness could be performed based on the optical analysis of a skin volume by o_HD-OCT. The diagnostic potential of HD-OCT in comparison with dermoscopy, RCM and C-OCT could be assessed regarding i) melanoma, ii) BCC differentiation from BCC imitators and BCC sub-differentiation and iii) SCC differentiation from AK. A much higher diagnostic potential could be demonstrated for o_HD-OCT in comparison with m_HD-OCT concerning melanoma detection. The diagnostic potential of HD-OCT to discriminate BCC from clinical BCC imitators was moderate. However, HD-OCT seemed to have high potential in sub-differentiation of BCC subtypes: i) it seemed to be the best technique to include and exclude a superficial BCC, ii) the technique appeared to be the best approach to exclude nodular BCC, and iii) HD-OCT looked to be the best technique to include an infiltrative BCC. Finally, HD-OCT has proven to be a powerful method to discriminate AK from SCC.Conclusions. HD-OCT is able to capture real time 3-D imaging with a sufficiently high optical resolution and penetration depth to allow the visualization of cells in and ex vivo in their micro-architectural context. At the same time, HD-OCT permits the recognition of patterns and structures in a sufficiently large volume of skin (1.5 mm³). HD-OCT closes therefore the gap between RCM with a high resolution but low penetration depth and C-OCT with a low resolution but high penetration depth. Moreover, HD-OCT permits, in contrast to RCM and C-OCT, the real time in vivo analysis of optical properties of the skin. HD-OCT seems to be a promising tool for early diagnosis of melanoma, BCC sub-differentiation and differentiation between SCC and AK.Future perspectives. Multicenter validation studies are needed to determine the diagnostic performance of this promising new technology, especially in other clinical settings combining both morphological and optical property analysis. This combined analysis could be a valuable method not only for diagnosis, monitoring and therapeutic guidance of dermatologic diseases but it could also be helpful in the management of non-dermatologic conditions such as diabetic micro-angiopathy, infantile cystinosis or even osteoporosis.