News >

Little Correlation Between Phase II, III Trials in Advanced Pancreatic Cancer

Published: Tuesday, Feb 07, 2017

Few phase II trials in advanced pancreatic cancer conformed to recommended design characteristics for pilot trials, a contributing factor to lack of success in the disease, Australian investigators reported.

A review of 148 phase II trials showed that 83% were single-arm studies, and only 11% of the studies evaluated new agents. Trials of biological agents rarely involved study populations enriched for a biomarker. Published data included little information on prognostic factors, and the studies exhibited wide variations in reported baseline factors.

Overall, 15% of the trials produced outcomes that led to phase III evaluations, Monica Tang, MBBS, of Prince of Wales Hospital and the University of Sydney in Australia, reported at the 2017 Gastrointestinal Cancers Symposium in San Francisco.

“These findings may explain the limited success of clinical trials in advanced and metastatic pancreatic cancer,” said Tang. “Overall survival has the strongest correlation between phase II and phase III trials.”

Over the past 25 years, survival in advanced pancreatic cancer has not improved appreciably, and fewer than 10% of patients remain alive at 5 years. A recent systematic review of phase III trials in pancreatic cancer showed the trials identified 3 agents or combinations with clinically meaningful activity.1 The analysis also showed that a 50% improvement in overall survival (OS) in phase II trials predicted clinically meaningful phase III results almost 80% of the time.

The National Cancer Institute (NCI) Trials Planning Meeting on Pancreatic Cancer Treatment established recommendations for conducting pilot trials in advanced pancreatic cancer.2 The recommendations addressed 3 aspects of trial design: patient selection (ECOG performance status 0-1, separation of patients with locally advanced and metastatic disease, uniform eligible criteria, and use of predictive biomarkers for enrichment); statistical design (survival as the preferred primary endpoint); and transition to phase III (consider input from multiple studies and initiate phase III evaluations only when pilot studies yield a robust signal).

Using the NCI planning meeting criteria for guidance, Tang and colleagues performed a systematic review of published phase II studies in advanced/metastatic pancreatic cancer from 1978 to 2015. Data analysis included 148 studies with a cumulative total of 7505 patients.

The results showed that 25 (16.9%) of the trials involved more than a single treatment arm, and randomized trials accounted for 12.2% (18) of the trials. Tang reported 37 (25%) trials evaluated biologic agents, but only 1 enriched for a biomarker; 7 others specified a biomarker as an exploratory or translational endpoint.

Collectively, the trials had limited reporting of prognostic factors. Age, sex, performance status, and disease type (locally advanced versus metastatic) were reported in 95% to 100% of the trials, and presence of liver metastases in 59%. Beyond those few, other prognostic markers were omitted in most reports, including CA19-9 (23%), patient-reported pain (4.1%), albumin (1.4%), bilirubin (0.7%), and lactate dehydrogenase (0%).

The published reports varied substantially with respect to baseline factors of the study population. For example, the proportion of patients with locally advanced disease ranged from 0% to 80%, and the proportion of patients with ECOG 0-1 performance stage ranged from 14.3% to 100%.

The primary endpoint was clearly identified in 68.9% of the trials, and objective response rate (ORR) was the most common endpoint (41.2%). Sample size and statistical assumptions were reported in 63.5% of the trials, although such reporting became more common in trials conducted after 2000. Tang said investigators reported the trials as successful in 55.4% of cases, although the target effect size was achieved in 26% of the studies. Subsequently, 14.9% of the trials transitioned to phase III.

A separate analysis of 27 agents evaluated in both phase II and phase III trials demonstrated inconsistency in the results. In general, phase III trials were less likely to yield positive results. Phase II and III trials had weak correlations for ORR and progression-free survival (PFS), and moderate correlations for OS.

Invited discussant Vincent Picozzi, MD, director of the pancreas cancer center at the Virginia Mason Medical Center in Seattle, pointed out that about 16% of phase II trials in oncology transition to phase III. The proportion is much lower for pancreatic cancer, on the order of 3%.

The bar needs to be set higher for transition from phase II to phase III evaluation in pancreatic cancer, Picozzi added. He suggested a 50% improvement in OS or 80% to 100% improvement in PFS in phase II for transition to phase III.

“Overall, this is a fairly grim report,” said Picozzi. “This abstract was selected not only because of its own merit but because it is reflective of a year in which there were a number of promising and highly touted trials in pancreatic cancer which unfortunately failed.”


  1. Rahib L, Fleshman JM, Matrisian LM, Berlin JD. Evaluation of pancreatic cancer clinical trials and benchmarks for clinically meaningful future trials: a systematic review. JAMA Oncol. 2016;2(9):1209-1216. doi: 10.1001/jamaoncol.2016.0585.
  2. Philip PA, Mooney M, Jaffe D, et al. Consensus report of the national cancer institute clinical trials planning meeting on pancreas cancer treatment. J Clin Oncol. 2009;27(33):5660-5669. doi: 10.1200/JCO.2009.21.9022.
  3. Tang M, Chen J, Goldstein D, et al. Correlation of phase 2 trials (Ph2t) results with outcomes of phase 3 trials (ph3t) of investigational agents (IA) in lovally advanced and metastatic pancreas cancer (LAMPC). Abstract presented at 2017 Gastrointestinal Cancers Symposium; January 19-21, 2017; San Francisco, CA. Abstract GI17.

View Conference Coverage
Online CME Activities
TitleExpiration DateCME Credits
Publication Bottom Border
Border Publication