Nowadays, to diagnose a lung cancer, a bronchoscopy is performed and a biological sample is extracted and analyzed by the anatomical pathology department of the hospital. Currently, there are no commercially available techniques allowing a real-time, in vivo, label-free diagnosis of lung cancer. The PREDICTION project aims to develop a biosensing tool gathering all the attributes mentioned above by combining optics, biochemistry and mechanics. The role of my research is to focus on the mechanics and to develop an instrumented catheter, acting as a shield of the biosensor. The choice of the material and the design were made based on the optical properties (visible under fluoroscopy) and the mechanical characteristics (trade-off between rigidity and compliance). In order to provide a stable measurement, the distal extremity of the instrumented catheter was shaped in the form of a conical needle. A window was patterned on the side of the instrumented catheter to expose the biosensor to the targeted tissue. The instrumented catheter was designed to be able to embed one biosensor and one control fibre. Its measurement integrity has been validated through in vitro and ex vivo experiments. In order to improve navigation outside the scope of the working channel of the bronchoscope, i.e. add one degree of freedom to the catheter, Shape Memory Polymers were investigated. Two prototypes were designed. The first prototype combines a soft pneumatic actuator with a shape memory polymer strip acting as a stiffness tuner. The Shape Memory Polymer structure proved to be efficient to fix the shape of the soft pneumatic actuator and also to increase the force it can provide. The second prototype combines a catheter with a Shape Memory Polymer strip. The experimental results proved the ability of the Shape Memory Polymer strip to develop a force high enough to bend a catheter with an adequate bending angle for in vivo lung navigation. To conclude, the work produced during this PhD resulted in the development of an instrumented catheter allowing real time, ex vivo, label-free diagnosis of lung cancer. Further work should be done on the instrumented catheter dimensions and sterilization to apply these results to in vivo diagnosis. / Doctorat en Sciences de l'ingénieur et technologie / info:eu-repo/semantics/nonPublished