Metallic nanoparticles have unique optical properties that can be exploited for molecular imaging in tissue. Image contrast depends on the nature of the particles, properties of the target tissue, and the imaging system. Maximizing image contrast for a particular application requires an understanding of the interplay of these factors. We demonstrate an approach that integrates the use of reflectance spectroscopy and imaging of particles in water and various tissue phantoms to evaluate the expected image contrast. We illustrate the application of this methodology for gold and silver nanospheres targeted against a biomarker expressed in epithelial tissue; predictions of contrast properties using diffuse reflectance spectroscopy were compared with widefield and high-resolution images of labeled tissue phantoms. The results show that the predicted image contrast based on spectroscopy agrees well with widefield and high-resolution imaging, and illustrate that gold and silver nanospheres at subnanomolar concentration are sufficient to produce contrast in both imaging modes. However, the effective contrast achieved with a particular type of nanoparticle can differ dramatically depending on the imaging modality. The ability to predict and optimize image contrast properties is a crucial step in the effective use of these nanomaterials for biomedical imaging applications.