Blood-based biomarkers are valuable diagnostic tools for the management of lung cancer patients. They support not only differential diagnosis and histological subtyping, but are also applied for estimation of prognosis, stratification for specific therapies, monitoring of therapy response, surveillance monitoring and early detection of residual or progressive disease. Early diagnosis of lung cancer in high risk populations (screening) is a promising future indication but poses high medical and economic challenges to marker performance. The five mostly used classical 'tumor markers' show characteristic profiles of sensitivity and specificity for non-small cell lung cancer (NSCLC) like cytokeratin 19-fragments (CYFRA 21-1), carcino-embryonic antigen (CEA) and squamous cancer cell antigen (SCCA) as well as for small cell lung cancer (SCLC) like progastrin-releasing peptide (ProGRP) and neuron-specific enolase (NSE). Combined use and pattern recognition approaches enable highly accurate diagnosis, subtyping and therapy monitoring. For the interpretation of serial measurements on an individual level, marker-specific algorithms have to be developed. So-called companion diagnostics identify druggable molecular changes in signaling pathways of tumor tissue that can be addressed by targeted therapies. New highly sensitive technologies enable the convenient and serial molecular characterization on circulating tumor DNA (ctDNA) in the blood, too. This approach is helpful when biopsies are not available and to overcome tumor molecular heterogeneity and plasticity. As only a portion of patients have such druggable molecular changes, future strategies will imply the combined use of classical and new ctDNA-based biomarkers to optimize the management of lung cancer patients during the course of disease.