Common cancers may arise from several different mutations, and each causative mutation may require different treatment approaches. There are also several mechanisms by which malignancies may become resistant to therapy, and each mechanism will also require a different therapeutic strategy. Hence, the paradigm of devising therapies based on tumor type is suboptimal. Each common malignancy may now be regarded as a collection of morphologically similar but molecularly distinct orphan diseases, each requiring unique approaches. Current strategies that employ randomized clinical trials (RCTs) in unselected patients carry a high risk of misleading results. Available data suggest that it is reasonable to grant marketing approval for new anticancer agents based solely on high single-agent response rates in small phase I-II studies involving molecularly-defined patient groups where benefit from other therapies is unlikely. This could markedly speed patient access to important therapies while reducing health care costs by slashing drug development costs. Feasible post-approval surveillance procedures could provide ongoing monitoring of drug safety. While assessment of drug combinations would be more complex due to variable contributions of each component, new strategies have been proposed. In addition to savings from more efficient clinical trials methods, it is essential that we also markedly reduce costs of complying with clinical research regulations. Compliance is too cumbersome and expensive, and current regulatory inflexibility markedly slows progress while escalating health care costs. This requires urgent attention. Regulatory approaches intended to enhance safety may instead potentially cost far more life-years than they save by delaying approval of effective therapies.