Whereas clinical trials of cancer drugs have methodological standards and conventional primary endpoints, these are not necessarily applicable to the clinical development of loco-regional treatments and new medical devices. The current challenge is to generate high-level clinical evidence for loco-regional treatments to define the benefits for patients. In this article, we argue that, to generate convincing evidence of clinical efficacy and safety, the collective coherence of the entire data package is often more important than the primary endpoint of one clinical trial. We also propose that, when a comprehensive clinical data package is not feasible, limited clinical data can be supplemented with other types of evidence. Emerging life science companies often define the "valley of death" after securing initial investment to translate an early medical device concept to a development stage that is attractive to funders. Unfortunately for this industry, there is a second "valley of death" if the focus and goal is only regulatory approval, to the neglect of clinical acceptance and reimbursement. For the emerging specialism of interventional oncology, it is critical to plan a clear line of sight for each new medical device to avoid the valleys of death and to demonstrate the clinical benefit. Increased international guidance to establish realistic yet convincing standards in this area may avoid attrition of potentially beneficial devices and therapeutic procedures in the valleys of death.