BACKGROUND: The pharmaceutical industry's productivity has been declining over the last two decades and high attrition rates and reduced regulatory approvals are being seen. The development of oncology drugs is particularly challenging with low rates of approval for novel treatments when compared with other therapeutic areas. Reliably establishing the potential of novel treatment and the corresponding optimal dosage is a key component to ensure efficient overall development. A growing interest lies in terminating developments of poor treatments quickly while enabling accelerated development for highly promising interventions. METHODS: One approach to reliably establish the optimal dosage and the potential of a novel treatment and thereby improve efficiency in the drug development pathway is the use of novel statistical designs that make efficient use of the data collected. RESULTS: In this paper, we discuss different (seamless) strategies for early oncology development and illustrate their strengths and weaknesses through real trial examples. We provide some directions for good practices in early oncology development, discuss frequently seen missed opportunities for improved efficiency and some future opportunities that have yet to fully develop their potential in early oncology treatment development. DISCUSSION: Modern methods for dose-finding have the potential to shorten and improve dose-finding and only small changes to current approaches are required to realise this potential.