Enhancing Clinical Trials in Ontario

---

By Robert A. Phillips, Janet Dancey, Karen Arts and Kay Friel

For the past fifty years, we have promised the public that support of basic science will reveal the fundamental defects in disease processes and that this understanding will stimulate remarkable new therapies that exploit this understanding. Despite a plethora of discoveries and suggestions of new targets, the explosion of new therapies has not arrived. In fact, the number of new molecular drugs approved by the FDA has decreased in recent years.

Last year, Dr. Richard Pazdur, Director of the FDA Office of Oncology Drug Products, reported on the very active area of oncology clinical trials . This office was created in 2005 to facilitate testing of the rapidly developing oncology drug pipeline. However, as he described, the office has received since its inception over 4,000 investigational new drug (IND) applications and presumably many of these lead to clinical trials. Despite this activity since 2005, the FDA has only approved 53 new agents and only 18 of these were new molecular entities – about 4 uniquely new agents per year.

In looking at this disappointing record, analysts have pointed fingers at all aspects of the drug development process from preclinical candidate identification and selection, to clinical trials, to regulatory review and approval. This article will focus only on the clinical trials methodology and on how Ontario is introducing new processes to improve clinical trials.

Clinical Trials
The clinical trials process is clearly the most expensive component of the drug development pipeline. Expense is partly due to the high rate of trial failure and in part due to the escalating costs of conducting studies designed to lead to regulatory approval. Although industry conducts its trials to be completed quickly, the overall operation of clinical trials is inefficient as drugs often fail in late trials resulting in delayed analysis of potentially important new therapies, and perhaps even more important, wasting precious dollars that could have been used to investigate more promising agents.

Academic-led cooperative groups also conduct trials testing new therapies. Compared to industry, these academic led studies, often funded through public dollars, lack sufficient resources to ensure the timely execution, but nonetheless conduct a significant proportion of clinical therapeutics research. Why do we get so few new drugs when, at the present time, there are more than 6,000 industry or academic-sponsored clinical trials being conducted in the United States and Canada?

For any new therapy to be approved, it must be tested in cancer patients and these tests usually occur in three separates phases. In a Phase I study clinicians test the new agent on a small group of patients to determine a safe dosage range and identify side effects. In a Phase II study, a larger group of patients with a specific cancer receive the new agent to determine possible effectiveness against that cancer and to further evaluate its safety. In a Phase III study, large groups of patients receive either the new therapy or the conventional therapy to determine the comparative effectiveness, monitor side effects, and look for safety problems.

Focusing just on matters related to clinical trials, there are three major issues that dramatically reduce their effectiveness.

First, the initiation of trials is very complex, involving many redundant processes that delay implementation without adding value. Second, a high proportion of trials, particularly those conducted by academics, fail to accrue enough patients to allow making a definitive conclusion about the efficacy of the new therapy being tested; this problem affects mainly the large Phase III trials. Third, even for Phase III trials that get started and accrue adequately, the majority fail to confirm the promising results obtained in the Phase II trial that indicated the possible effectiveness of the new therapy.

Unnecessary redundancy in implementation of clinical trials
All clinical trials investigators recognize the enormous administrative burden entailed in starting clinical trials. Dilts and his colleagues have quantified the steps and evaluated the need for the various decision points through their extensive analysis of trials conducted by the Eastern Cooperative Oncology Group (ECOG) , an academic adult cancer clinical trial organization funded by the National Cancer Institute’s Cancer Therapy Evaluation Program (CTEP). They found that activation of a trial required 481 distinct processes and required 783 days (median, range: 285 – 1542 days) for Phase III trials, counting from the date of executive approval excluding activities required at individual sites. Activation time was equal or greater than the actual time to complete the trial in many instances. Delays resulting from review and approval at individual sites may also be significant.

For multicentre trials, many of the processes are repeated at numerous sites, with multiple groups examining protocols, many lawyers examining contracts, and many ethics boards reviewing the trials. There is little evidence that these redundant reviews and analyses add any value to the trials.

Primarily these processes simply increase the complexity of the review process. They increase the cost of activating a trial and delay the onset of trials, very often resulting in trials not being completed or being abandoned because the question being explored in the trial is no longer relevant or because more interesting trials have been started which need the same group of patients.

Slow or inadequate accrual to trials
Several companies have analyzed accrual to trials and found that often 35 per cent of sites that open trials never recruit any patients and another 15 per cent only recruit a single patient. In general 90 per cent of the patients come from 10 per cent of the sites across academic and industry-led trials. The limited patient accrual from individual sites often drives industry to sign up considerably more sites and conduct international trials to ensure adequate and rapid accrual, driving up development costs. In an analysis of CTEP sponsored trials, Dilts and his group found that in cooperative groups, 15 per cent of sites did not even achieve 10 per cent of their accrual target. Perhaps not surprisingly, they found that trials that opened rapidly accrued better than trials that opened slowly. They found that 50 per cent of trials failed to reach 25 per cent of minimum accrual required; overall only 64 per cent of trials reached the minimum desired accrual at the time of closure.

Clearly, for the centres that fail to recruit more than one or two patients and for trials that fail to meet accrual objectives, the time, effort and expense to open the trial are wasted. However, for patients who agree to risk their health and survival to be recruited to these ‘incomplete’ trials there is arguably moral responsibility to complete the trial so that the efforts and commitment of these patients are not wasted.

From this data it is clear that decreasing start-up times could help to dramatically improve recruitment to trials as trial questions would remain relevant throughout the duration of the proposed study. In addition, centres need to monitor closely the trials ongoing in their centre to keep the number of trials that compete for the same group of patients to a minimum.

Phase III trials often fail to confirm results obtained in Phase II
As mentioned above, the usual process followed for assessing all new therapies is to initially test a small group of patients in a Phase II study to determine whether or not the new therapy has potential efficacy in treating a particular cancer. Because Phase II trials are relatively small, often uncontrolled and accrue select patient populations recruited from a limited number of sites, one would expect that Phase III studies with much larger numbers of patients would fail to confirm a Phase II study. However, a couple of recent studies have indicated that the number of trials failing to confirm Phase II results is unexpectedly high.

Zia et al. conducted a thorough literature review of all chemotherapy trials for solid tumours published in English journals between July 1998 and June 2003. They searched for Phase III studies and then searched to determine if they were based on a published Phase II study. Among 181 Phase III studies found, 43 used the identical treatment procedure of one or more preceding Phase II studies. Surprisingly, only 12 Phase III studies (28%) confirmed the positive result of the Phase II study.

Djulbegovic and colleagues have recently studied all the Phase III trials conducted by the National Cancer Institute since 1955; they used both published and unpublished data in their analysis. They considered a trial positive if the new treatment or process was significantly better than standard treatment. Thirty percent of trials had a significant result; of the significant trials, 80 per cent favoured the experimental therapy and 20 per cent the standard therapy. Thus, only 24 per cent of trials confirmed benefit of the experimental therapy. While Djulbegovic et al. did not attempt to examine how many of these trials were based on prior Phase II trials, it is likely that most of the trials using new agents were based on Phase II results. Their data are highly consistent with the low success found in the study by Zia et al.

A success rate of 25 per cent is sufficiently low that individuals designing trials need to look carefully at the reasons for this poor level of success. It seems likely that either the Phase II trials had insufficient power to allow reasonable prediction or that sponsors of Phase III trials are not being sufficiently rigorous in their analysis of Phase II results.

New Ontario Approaches to Improving Clinical Trial Efficiency and Effectiveness
Through leadership from the Ontario Institute for Cancer Research (OICR) and its predecessor organization, the Ontario Cancer Research Network (OCRN), several initiatives have been launched to resolve some of the issues outlined above and to enhance clinical trials activity in Ontario. The important principle behind all of the initiatives is to encourage institutions to work together, reducing redundancy, improving efficiency and taking advantage of the unique capabilities of each institution.

Ontario Cancer Research Ethics Board
Many groups recognized that a centralized ethics review process had the potential to improve ethics review and reduce unnecessary redundancies in initiating multicentre clinical trials. Following meetings with Ontario leaders of the ethics community in 2002-2003, a decision was made to create a coordinated review process under the auspices of OCRN. In December 2003, the Ontario Cancer Research Ethics Board (OCREB) was launched with three institutions formally adopting OCREB as their board of record for multicentre trials. OCREB has grown rapidly and is now used by 19 centres in Ontario, including all of the major cancer centres. By creating a 23 member panel dedicated to oncology trials, OCREB has reduced conflicts of interest, ensured a high quality of review and reduced review times to 55 days, most of which is time taken by investigators to respond to queries from OCREB. In general, OCREB is considered a successful model for centralized ethics review. The details about OCREB can be found on the OICR web site (www.oicr.on.ca/ocreb/index.htm) and in a recent publication.

Testing the Effectiveness of a Coordinated Process to Initiate Trials
Based on the OCREB success, it is logical that further collaboration during other aspects of initiation of trials could also bring additional benefits and reduce start-up times. With this goal in mind, OICR has recently piloted a single point of contact program where sponsors of trials, public or private, can work with an office at OICR to facilitate trial contracts, budgets and ethics submissions across multiple Ontario institutions. In tests of five trial start-ups over the past 18 months, this office has worked with sponsors to identify interested sites, develop a common budget and facilitate a common contract for all sites, and prepare an ethics submission for OCREB. The initial tests of this process have proven the feasibility of a centralized process with a trials being initiated within 3 months.

Creating High Impact Clinical Trials
One likely cause of the infrequent success of Phase III trials replicating Phase II results is that the Phase II trials did not collect sufficient data to allow a reasonable prediction of success. Given the development of sensitive, high-throughput technologies, and the fact that most new therapies are based on interventions targeting well-known cellular pathways, it should be possible to collect enough data in clinical trials to know whether and, more importantly, why a therapy works. Clinical trials that collect clinical data as well as informative biological and pharmacologic data should allow better decisions in both Phase II and Phase III, but especially in Phase II.

In this context OICR created a program in High Impact Clinical Trials. This initiative, led by Dr. Janet Dancey, is bringing together clinical and basic scientists to design and implement sophisticated laboratory and imaging studies that will greatly inform the results of clinical trials in Ontario. It is anticipated that when this process is implemented in late 2009 and early 2010, the data collected together with the clinical results on the efficacy of the new treatment will more accurately inform the best next steps in the development of the therapy.

Conclusions
The current clinical trial process is broken. Initiation of clinical trials is inefficient, involving unproductive, redundant processes. These inefficiencies lead to trials that are often not completed or which lead to results that fail to confirm previously promising preliminary data. The path toward improvement is clear, but will require close cooperation of cancer centres, ethics boards, regulatory agencies, sponsors and governments. Ontario has begun the process in tackling some specific issues such as ethics review and start-up of trials, but much more remains to be done. If we continue our current inefficient processes, we will not be able to take advantage of all the new therapeutic targets being identified by research, and we will not be able to develop new drugs in a cost-effective manner.

References
1.     CancerNetwork.com: http://www.cancernetwork.com/practice-management/display/article/10165/1263819
2.     Dilts, D.M; Sandler, A.; Cheng, S.; Crites, J.; Ferranti, L.; Wu, A.; Gray, R.; MacDonald, J.; Marinucci, D.; Comis, R. Clin Cancer Res, 2008. 14(11): 3427-33.
3.     Dilts, D.M. http://www.cmrhc.org/presentations/doc_download/34-are-academic-medical-center-clinical-trials-going-the-way-of-oldsmobile.html
4.     Zia, M.I.; Siu, L.L.; Pond, G.R.; Chen, E.X. J Clin Oncol, 2005. 23(28): 6982-91.
5.     Djulbegovic, B.;Kumar, A.;Soares, H.P.;Hozo, I.; Bepler, G.; Clarke, M.; Bennett, C.L. Arch Intern Med, 2008. 168(6): 632-42.
6.     Chaddah, M.R. Curr Oncol. 2008 January; 15(1): 49–52

Robert A. Phillips, Deputy Director, Ontario Institute for Cancer Research; Janet Dancey, Program Director, High Impact Clinical Trials, Ontario Institute for Cancer Research; Karen Arts, Director, Business Development – Clinical Trials, Ontario Institute for Cancer Research; Kay Friel, Director, Clinical Trials Programs, Ontario Institute for Cancer Research