"Cancer of unknown primary (site)" (CUP) represents a frequent and often difficult task for diagnostic histopathology. In the initial CUP situation, immunohistochemistry, which is relatively cost-efficient and non-burdensome for the patient, contributes substantially to the targeted search for and identification of the primary tumor. It may be performed in a two-step algorithm (keratins as the first step, group and selective markers as the second step). However, in true CUP syndrome--a heterogeneous disease whose etiology and pathogenesis is still poorly understood--the primary tumor is rarely identified, and the immunohistochemical marker profile is often anomalous. Nevertheless, immunohistochemistry remains the most important special method, aiming at the phenotypical classification and particularly the identification of certain favorable therapy-responsive subsets. Novel methods for subtyping of CUP based on the analysis of gene expression profiles need to be further evaluated before their possible diagnostic application. Future basic and clinical research is required to unravel the still enigmatic tumor biology of the CUP syndrome and to improve the hitherto very unfavorable prognosis by developing new efficient therapy regimens.