Will My Child Do Better If She Enrolls on a Clinical Trial?

Abstract

The question of whether children with cancer who enroll in clinical trials have superior outcomes compared with those who do not participate has been pursued for more than four decades, with recent studies providing conflicting answers. Whether clinical trial participation influences outcome has important implications for how clinicians should present trial participation to patients and families. Methodological challenges limit generalizations about the impact of clinical trial participation on outcome when compared to non-participation. Oncologists should inform patients and families that clinical trials are the engine for future progress in identifying more effective therapies and that clinical trial participation is a reasonable option to consider for children with cancer. However, as noted by Truong et al, the rationale for trial enrollment should not include an expectation of better outcomes as compared with non-enrollment.

Condensed Abstract

Whether children with cancer who enroll in clinical trials have superior outcomes compared with those who do not participate has important implications for how clinicians should present trial participation to patients and families. While clinical trials are the engine for future progress in identifying more effective therapies, the rationale for trial enrollment should not include an expectation of better outcomes as compared with non-enrollment.

The analysis by Truong and colleagues using the Cancer in Young People in Canada database, which focuses on children with acute myeloid leukemia (AML), adds to the literature about the impact of clinical trial participation on outcome for children with cancer (1). The question of whether children with cancer who enroll in clinical trials have superior outcomes compared with those who do not participate has been pursued since at least the 1970s. A report from the United Kingdom in that era found that, “Inclusion in the [Medical Research Council] trial significantly improved the chances of survival of children with nephroblastoma” (2). A North American report found that, for children with acute lymphoblastic leukemia (ALL), “A striking difference in 4-year survival rate was observed between protocol and nonprotocol patients” (3). These and other early reports have had a major impact on the perceptions of pediatric oncologists regarding clinical trials participation, and have contributed to the belief that, “… the inclusion of the patient in clinical trials is widely recognised as the best ‘standard of care’ even for newly diagnosed patients” (4). These perceptions developed and persisted despite the paucity of high-quality evidence in support of the belief that cancer trial participation improves outcomes for patients who take part (5).

The impact of clinical trial participation on outcome has important implications for how clinicians should present trial participation to patients and families. While the Truong et al. report found no evidence of benefit to children with AML associated with clinical trial participation (1), a 2018 report using the same database found that, for children with ALL, enrollment onto a clinical trial was associated with significantly improved event-free—though not overall—survival (6). The authors of the ALL report noted that, “This information may be important to families and clinicians when deciding whether to enroll on a clinical trial at diagnosis.” Indeed, if participation in clinical trials truly leads to improved outcomes, then researchers are arguably obligated ethically to encourage participation in clinical trials on the grounds of the best interests of the child (5). Conversely, to the extent that informed decision making about clinical trial participation requires that either trial participation or receipt of standard of care be presented as reasonable options (7), conveying information about improved outcome for clinical trial participants risks implying to families that a decision not to enroll would be contrary to the child’s best interests.

Ascertaining whether trial participation improves outcomes for children with cancer presents methodological challenges. This difficulty occurs in part because there are multiple pathways—only some of which are causally related to trial participation—by which children enrolled in trials might appear to have better outcomes than children who are not enrolled. If superior outcomes for trial participants were consistently observed, three general hypotheses could explain the finding:

1. Patients enrolling on clinical trials differ from those not enrolling, and the differences have an impact on outcome.

2. Patients enrolling on clinical trials get better clinical care than patients not participating in clinical trials, independent of the specific interventions studied in the trial, and this has an impact on outcome.

3. The treatments provided in clinical trials are superior to those provided as standard therapy outside of clinical trials, and this has an impact on outcome.

Each of these hypotheses bears consideration when evaluating whether trial participation improves outcomes among children with cancer.

The first hypothesis for superior outcomes for clinical trial participants is selection bias—in other words, that patients who enroll in clinical trials are not representative of, and have a favorable baseline prognosis as compared with, the overall population of patients with the relevant diseases. This issue affects clinical trials for adults with cancer, as illustrated by a Southwest Oncology Group (SWOG) report that found that, for some cancer types, participation in clinical trials was associated with better survival, likely due to exclusion of patients with higher levels of comorbidities. Hence, in the SWOG analysis, differences in outcome could be explained by the different patient populations who entered clinical trials compared to those who did not (1). This explanation is less likely to be relevant in pediatric oncology, given the relatively high percentages of children with many cancer types who enroll in clinical trials. Consistent with this perspective, the children with AML in the analysis by Truong et al who enrolled in clinical trials were similar to—and may even have had worse prognoses than—those who did not enroll (1).

The second hypothesis for superior outcome for clinical trial participants relates to the quality of clinical care received by patients participating in clinical trials as compared with the quality of care received by patients receiving standard therapy outside of a trial. A study reliably addressing this question would ideally evaluate only negative clinical trials, or limit the trial participants included in the analysis to those enrolled in the control arms of randomized trials, to avoid potential confounding related to the more effective treatment in the experimental arm. Recent single-institution reports do not demonstrate an impact of clinical trial participation on outcome (8, 9), suggesting that the proficiencies that are associated with treatment administered within the context of a clinical trial can be replicated outside of trials. In addition, regardless of whether or not individual children with cancer participate in trials, most are treated in specialized centers that actively conduct trials. As a result, differences between trial participants and non-participants due to treatment in different centers are likely reduced in pediatric oncology as compared with adult oncology or other areas of medicine. Furthermore, to the extent that factors associated with the conduct of clinical trials raise the quality bar at a center, all patients treated at that center—not just those treated within trials—can be expected to benefit (10–12). Although the limited data available suggest that quality of care may not meaningfully differ between children with cancer enrolled versus not enrolled on clinical trials, attention to and research into delivering quality care is important so that best practices can be identified and disseminated, as illustrated by quality improvement efforts of centers treating children with cystic fibrosis (13).

The third hypothesis that could explain superior outcomes among clinical trial participants relates to the potentially greater efficacy of the “experimental” therapies being investigated in clinical trials. The report of more favorable outcome for children with ALL enrolled on clinical trials in Canada from 2001 to 2012 occurred during a period of positive Children’s Oncology Group (COG) ALL clinical trials (6). Gratifyingly, a number of COG ALL phase 3 clinical trials in the past two decades have identified more effective therapies, including: the use of dexamethasone rather than prednisone (14, 15); the use of vincristine and escalating intravenous methotrexate without leucovorin rescue during the interim maintenance (IM) phases for standard risk ALL (16); the use of high-dose methotrexate for high-risk B-ALL (15); and the use of nelarabine for T-ALL (17). However, there is no guarantee that an ongoing COG clinical trial for which a child with cancer is eligible will also have a superior experimental arm; the observation is inherently grounded within a specific era of clinical trials and may not generalize to future trials.

In considering how to present trials to patients and families, it is important to remember that clinical trials are experiments. Despite strong rationales for experimental regimens, they occasionally prove inferior to standard therapy in terms of safety or efficacy. For example, there was reason to believe that replacing 6-mercaptopurine with 6-thioguanine would lead to improved outcome for children with ALL. Nevertheless, while event-free survival was superior for patients receiving thioguanine in a COG clinical trial (18), approximately one-quarter of patients developed liver toxicity (veno-occlusive disease or disproportionate thrombocytopenia), which required that treatment with thioguanine be discontinued and replaced with 6-mercaptopurine (18). The frequent and prolonged use of intermediate-dose methotrexate with low-dose leucovorin rescue was also a reasonable approach to test for the treatment of children with ALL (19). However, severe neurotoxicity was observed when this strategy was applied in POG-9405, requiring the clinical trial to be closed early (20). Another example is the use of clofarabine for children with very high-risk B-ALL (21). Despite preliminary data supporting the safety and activity of a clofarabine-based regimen (22, 23), its incorporation into a frontline ALL clinical trial resulted in unacceptable toxicity (21). These examples illustrate the need for caution about statements of more favorable outcome from clinical trial participation when approaching families of children with cancer.

In light of these considerations, what should clinical researchers tell patients and families about participation in trials? Most importantly, they should communicate that clinical trials are the engine by which progress is made over time for children with cancer, and that clinical trial participation may help identify superior treatments for future children. Beyond this, they should communicate that participation in clinical trials is a reasonable option to consider for many children with cancer. This statement is supported by more than 50 years of experience with multi-institutional pediatric oncology clinical trials in which standard of care treatment regimens have been compared to potentially superior regimens, with the outcome of the clinical trials defining the standard of care for the next generation of clinical trials as well as for subsequent patients treated outside of trials (24). The statement is also supported by the attention given to the development and implementation of these trials. As an example, clinical trials of the Children’s Oncology Group are developed by study teams with extensive experience in the diseases being studied, and then undergo independent review by a panel of experts coordinated by NCI. The trials must be approved by the NCI Central IRB prior to activation, and following activation they are monitored by the study team as well as by an independent Data and Safety Monitoring Committee. Nevertheless, clinical researchers need to convey to patients and families that this process for clinical trial development and monitoring cannot provide assurances that children participating in these trials will have superior outcomes compared to otherwise similar non-participating children. Similarly, the process cannot provide assurances that trial participants will not experience serious toxicities that would not occur with standard care. The process can, however, signify to patients and families that due diligence in evaluating for potential efficacy and for safety has been performed. In light of this robust development and review process, pediatric oncologists should communicate to patients and families that enrollment onto a Children’s Oncology Group clinical trial, or onto another clinical trial with comparably rigorous oversight, remains a reasonable treatment option.