Microfluidic bio-assays have emerged as the most privileged solutions and provide the basis for the realization of miniaturized bio-analytical systems and clinical diagnostic devices that are portable, user-friendly and cost-effective (Lab-on-a-chip). Two important steps that are implemented in a microfluidic bio-assay are: (a) the immobilization and/or patterning of target-specific bio-molecules on the surface of a microfluidic channel, for selectively capturing bio-targets like antigens or pathogens, followed by (b) sensitive detection of the bio-targets. In this thesis, we demonstrate microfluidic bio-assays based on novel methods for generating protein-patterns and on sensitive detection of the bio-targets. First, we introduce a simple and fast method for creating protein micropatterns both on a bare substrate and in-situ inside a microfluidic channel, in a matter of minutes, through electrostatic self-assembly of pre-functionalized magnetic beads. A lift-off patterned positively-charged aminosilane layer is used as the template for immobilizing the protein-coated negatively-charged beads. The number and arrangement of the beads can be well-controlled by altering the silane template design. Subsequently, we use patterned beads as assay substrates for performing on-chip bioassays. We demonstrate highly-sensitive full on-chip sandwich immunoassays for single and multi-analyte detection using beads as assay substrate. We successfully explored the possibility to lower the detection limit of immunoassays by concentrating the target antigens on a very small number of patterned beads. We also present the application of bead patterns as a platform for immuno-separation, culture and analysis of target (cancer) cells. Finally, we demonstrate a rapid on-chip immuno-histo-chemical assay on breast cancer tissues. We use luminescent lanthanide probes in place of conventional fluorescent probes, as labels for detection antibodies, for sensitive detection and quantification of biomarkers. Thanks to the time resolved microscopy and luminescent probes, the background noise due to the autofluorescence of the samples (i.e. tissue, cells) and microfluidic chips is successfully eliminated resulting in an improved signal-to-noise ratio when compared with the fluorescent microscopy results. Our assay results fully agree with the clinical analyses outcome, and this opens perspectives for a fully-integrated cancer detection platform for bedside diagnostics.