The aim of this thesis is to optimize and obtain fundamental information on i) the photodetection (PD) of early stage cancers in the tracheo-bronchial tree by spectroscopic imaging of the tissue autofluorescence, ii) photodynamic therapy (PDT) applied in dermatology based on the use of protoporphyrin IX (PPIX) precursors. These two areas of photomedecine have been the subject of investigation at the air and soil pollution laboratory (LPAS) for the last 10 years. PD and PDT are areas of research that evolve in parallel, both having similar objectives: i.e. improving the patients' chances of survival by the early detection and treatment of cancer. Three main studies are presented, each in relation to a particular type of chromophore : Endogenous fluorophores generate the autofluorescence of biological tissue; some of these fluorophores can be synthesized endogenously in excess of their normal concentration following the administration of a biochemical precursor; finally, exogenous luminophores are artificially synthesized and administered to the organism where they can be used for detection. The introduction of this thesis presents several aspects of cancer as well as the theoretical background necessary for this study. At first we address time-resolved autofluorescence spectroscopy of bronchial tissue and the mechanisms at the origin of the contrasts existing between normal tissue and early stage tumours. Several hypotheses explain the reduction in intensity and spectral changes of the autofluorescence observed between early epidermal carcinoma and normal bronchial mucosa : 1) the reduction in the concentration of fluorophores in the malignant tissue, 2) the quenching of the fluorescence due to a change in the physico-chemical environment of the tumour, 3) an architectural effect such as a thickening of the malignant epithelium due to an abnormal cellular proliferation and 4) an increase in the concentration of light absorbers such as haemoglobin. In the course of our measurements no statistically significant difference in the fluorescence lifetime was observed between carcinoma in situ, moderate dysplasia and normal tissue, which leads us to conclude that hypotheses 1, 3 and/or 4 may well be at the origin of the differences in fluorescence intensity. The fluorescence lifetime decays are consistent, identifiable by a triple-exponential decay in the green and red part of the spectrum. This seems to indicate that one or two dominant fluorophores are involved which possibly, given the measured lifetimes, could be collagen and/or elastine. The second part of this thesis concerns the modulation of the production of PPIX (an endogenously induced fluorophore) in the skin, following the administration of 5-aminolevulinic acid or one of its esterified derivatives. These investigations have been carried out on the epidermis of human volunteers as well as on the epidermal equivalent (Epidex™). The final goal of these studies was on the one hand to increase the production of PPIX well below the skin surface to improve certain dermatological treatments such as the PDT of nodular basal cell carcinomas (BCC), and, on the other hand, the ability to use the results obtained with the Epidex™ model to produce PPIX specifically in hair follicles, which permits the development of a new method of hair removal by PDT. The tests using the Epidex™ model demonstrated that it is possible, by adding Desferal® (an iron chelator) or (L+) ascorbic acid iron (II) salt, to increase respectively decrease the production of PPIX as well as to modulate its elimination. These results open new perspectives such as the specific use of zinc-protoporphyrin (produced by skin cells following the dermatological use of Desferal®), and the identification of a specific precursor of PPIX for hair follicles. This precursor is an excellent "candidate" in the field of research for the development of a new method of hair removal by PDT. In the third part of this thesis we describe a preliminary study of optical sensors for the intravascular partial pressure of oxygen via the measurement of phosphorescence lifetime quenching. The phosphorescence lifetime of these species is directly related to the partial pressure of the oxygen dissolved (pO2). These oxygen sensors allow us, for instance, to measure the reduction of the pO2 during PDT, without phototoxicity. Palladium and ruthenium complexes were among the most promising compounds, and certain types of nanoparticules were used as vectors. Another interesting lead consists in the covalent modification of these "luminophores-molecules": in this context, the experiments conducted in collaboration with the "Laboratoire de photonique et interfaces" of the EPFL, demonstrate that we can drastically reduce the speed of leakage outside the vascularisation of these molecules and/or improve their retention by the nanoparticules.