The promise of effective targeted therapy for lung cancer requires rigorous identification of potential targets combined with intensive discovery and development efforts aimed at developing effective "drugs" for these targets. We now recognize that getting the right drug to the right target in the right patient is more complicated than one could have imagined a decade ago. As knowledge of targets and development of agents have proliferated and advanced, so too have data demonstrating the biologic heterogeneity of tumors. The finding that lung cancers are genetically diverse and can exhibit several pathways of resistance in response to targeted agents makes the prospect for curative therapy more daunting. It is becoming increasingly clear that single-agent treatment will be the exception rather than the rule. This information raises important new questions about the development and assessment of novel agents in lung cancer treatment: (1) How do we identify the most important drug targets for tumor initiation and maintenance? (2) What is the best way to assess drug candidates that may only be relevant in a small fraction of patients? (3) What models do we use to predict clinical response and identify effective combinations? And (4) how do we bring combination regimens to the clinic, particularly when the agents are not yet approved individually and may be under development from different companies? The Fifth Cambridge Conference on Novel Agents in the Treatment of Lung Cancer was held in Cambridge, Massachusetts, on October 1-2, 2007, to discuss these questions by reviewing recent progress in the field and advancing recommendations for research and patient care. New information, conclusions, and recommendations considered significant for the field by the program faculty are summarized here and presented at greater length in the individual articles and accompanying discussions that comprise the full conference proceedings. A CME activity based on this summary is also available at www.informedicalcme.com/cme.