How many new treatments are “breakthroughs”? Evaluation of innovations in cancer

Abstract

6066 Background: Research efforts tested in randomized controlled trials (RCT) have played a vital role in the development of new treatments for cancer. The research conducted by the National Cancer Institute (NCI) cooperative groups (CG) is universally considered of vital importance to this progress. However, how often new interventions tested in the NCI-CG-RCTs are highly successful or “breakthrough” is not known. Methods: All completed phase 3 RCTs (563 trials, ∼149,000 patients) conducted by 6 NCI-CGs was reviewed. Published and unpublished data were used. Breakthroughs were defined in two ways: semi-quantitatively (score 1–6): the intervention was so successful according to the original researchers that it should be adopted immediately as a new therapeutic standard (=6), and quantitatively: the intervention effect resulted in a log hazard ratio of ≤ -0.5 in terms of survival or event-free survival. Results: According to the original investigators, 14% (n=81) of experimental testing resulted in breakthrough. 7% of interventions dramatically improved survival in terms of log hazard ratio of ≤ -0.5, while 13% led to a similar improvement in event-free survival. By disease, quantitatively (23%) and semi-quantitatively (12%) the highest number of breakthroughs were for leukemias. Induction and curative therapies jointly resulted in 62% and 73% of breakthroughs by quantitative and semi-quantitative assessment, respectively. 15% of childhood vs. 5% of adult cancer trials resulted in dramatic survival improvements. However, assessing semi-quantitatively, there were 18% and 14% breakthroughs in childhood and adult cancers, respectively. Conclusions: 14% of the experimental treatments that progressed to testing in RCTs resulted in discovery of breakthrough interventions. This is the first comprehensive assessment of success of innovative therapies in cancer and in clinical medicine. No systematic research was found in the literature permitting to contrast our findings with the estimate of the proportion of breakthrough interventions from non-RCT, phase I & II trials. Researchers, physicians, policy makers and the public should understand what it takes to produce a successful, breakthrough intervention in cancer. No significant financial relationships to disclose.