Importance: Progress in the treatment of pancreatic adenocarcinoma has been minimal; it remains the only major cancer type with a 5-year survival rate of less than 10%.
Objective: To explore why a large proportion of advanced pancreatic cancer clinical trials executed over the past 25 years have had negative results and to identify benchmarks that could have predicted success.
Evidence review: Phase 3 studies of patients with advanced pancreatic cancer were identified by searching clinicaltrials.gov and the scientific literature.
Findings: Thirty-two phase 3 studies in 13 675 chemotherapy-naive patients resulted in 3 agents or combinations being considered clinically meaningful. Nineteen agents or combinations (70%) were tested in phase 2 trials preceding the phase 3 trial. In cases with paired phase 2 and 3 results, meeting the primary end point of the phase 2 trial predicted the outcome of the phase 3 trial 76% of the time but proceeded despite phase 2 negative results in 10 cases. We applied criteria for a clinically meaningful result identified by the American Society of Clinical Oncology (ASCO) Cancer Research Committee to these historical cases. Overall, progression-free and 1-year survival of experimental arms was compared with time period-controlled median values of control arms to normalize for the observed increase in response to gemcitabine over time.
Conclusions and relevance: Applying the benchmark of a 50% improvement in overall survival as the primary end point to phase 2 data, or secondary end points of a 90% increase in 1-year survival or an 80% to 100% increase in progression-free survival, showed the greatest ability to predict a clinically meaningful phase 3 trial. Had these criteria been applied to these trials over the past 25 years, more than 11 571 patients enrolled in phase 3 trials that did not meet the primary end point could theoretically have been diverted to earlier-stage trials in an attempt to more rapidly advance the field.