Strict cancer trial eligibility criteria ensure patient safety, but ‘pendulum has swung too far’
Clinical trials are the cornerstone of oncology research.
Despite their essential role in evidence-based medicine, enrollment barriers are more formidable than ever.
Limited study availability at some academic and community centers, as well as lack of provider recommendation, largely explain why fewer than one in 20 patients with cancer participate in clinical trials.
However, an estimated 17% of individuals with cancer cannot participate in trials even if they wanted to because they fail to meet eligibility criteria, according to a report released in April by American Cancer Society Cancer Action Network, the society’s advocacy affiliate.
In some cases — such as with trials designed to assess treatments that target specific genetic mutations or immune markers in the tumor microenvironment — patient populations must be limited to genetically defined subsets. In other cases, broader characteristics — such as age, HIV status or prior cancer history — further narrow the eligible patient pool.
Eligibility criteria are intended to define appropriate patient populations so investigators can easily interpret results, accurately assess the efficacy of an investigational therapy and also ensure patient safety. However, because protocols for new trials frequently mirror those used in prior studies, some experts suggest investigators too often rely on a template to determine who can — and who cannot — participate.
“We have seen eligibility criteria drift in the territory that relies more on precedence than scientific rationale,” Mark Fleury, PhD, MS, principal of policy development and emerging science at American Cancer Society Cancer Action Network, told HemOnc Today. “The ‘cut-and-paste’ of previous trial protocols into new trials without stopping, stepping back and questioning on a trial-by-trial basis which criteria are important is where the issue lies.”
HemOnc Today spoke with investigators, clinicians and policy advocates about the benefits and drawbacks of strict eligibility criteria for oncology trials, whether their use limits the research community’s ability to accurately project a treatment’s efficacy and safety in a real-world population, the increased calls from the clinical and advocacy communities to relax eligibility criteria for specific patient groups, and the likelihood that trial designs will be altered to expand the eligible patient pool.
First, do no harm
Participants in oncology clinical trials must meet appropriate diagnosis and disease stage criteria prior to enrollment. After that, other factors — including age and health status — come into play.
“The primary goal is patient safety,” Fleury said. “[We want] to ensure we do not endanger patients on a clinical trial. We want the criteria to be designed so that we do not cause harm to patients.”
A constant struggle for balance exists in the research community to make trial results interpretable and generalizable, according to Matthew Galsky, MD, professor of medicine at Icahn School of Medicine at Mount Sinai and director of the genitourinary medical oncology program at Tisch Cancer Institute.
Thus, patients with prior cancer history often are excluded from oncology trials because researchers want to ensure all outcomes they observe relate to the malignancy that the therapy under investigation is designed to treat.
Eligibility criteria also are designed to minimize risk for trial participants.
“There are differences among patients not only with regard to the potential efficacy of treatments, but with regard to their potential safety, as well,” Galsky said in an interview. “For example, drugs that are metabolized by the kidney may be more dangerous for patients with poor kidney or liver function.”
This has shifted the emphasis away from generalizability and more toward interpretability and safety.
“Perhaps [it has moved] too far in that direction,” Galsky said. “We really should be somewhere in between.”
Clinical trial evidence largely determines which therapies receive regulatory approval, creating new treatment options for patients with life-threatening diseases. The patient population enrolled on a trial is one of the most important variables to ensure an investigational therapy has a fair opportunity to demonstrate its effectiveness compared with other available treatments, Edward S. Kim, MD, chair of the department of solid tumor oncology at Levine Cancer Institute at Atrium Health and a HemOnc Today Editorial Board Member, said in an interview.
“However, the running joke has now become that a patient has to be a triathlete to enroll on a clinical trial,” Kim told HemOnc Today. “That is very ironic, considering these usually are patients who have an incurable cancer and have been treated with more than one prior standard therapy.”
Although strict eligibility criteria help ensure the safety of trial participants, there is a downstream drawback.
Patients deemed ineligible for a trial still may be prescribed a therapy once it is approved, even though evidence is limited — or in some cases nonexistent — about how the treatment will work in this real-world setting.
Mitchell and colleagues retrospectively analyzed data from patients with metastatic renal cell carcinoma treated between January 2007 and May 2011 at oncology practices within Duke University Health System, as well as the ACORN network of 16 community oncology practices.
The cohort included 438 patients who received first-line treatment with the mTOR inhibitor temsirolimus (Torisel, Pfizer), or the tyrosine kinase inhibitors sunitinib (Sutent, Pfizer), sorafenib (Nexavar, Bayer) or pazopanib (Votrient, Novartis).
Researchers compared these patients with those enrolled on phase 3 clinical trials that led to FDA approval of targeted therapies for metastatic renal cell carcinoma.
Patients treated in the real-world setting with one of the three TKIs appeared more likely than those enrolled in the phase 3 trials to have poor-risk disease based on Memorial Sloan Kettering Cancer Center criteria (7.4% vs. 2.9%; P < .001) and less likely to have favorable-risk disease (30.1% vs. 43.8%; P < .001). Patients in the real-world setting were nearly 20 times more likely to have ECOG performance status greater than 1 (11.1% vs. 0.6%; P < .001).
The researchers determined 39% of patients in their analysis would have been excluded from the phase 3 clinical trials that led to the regulatory approvals for the drugs they received.
“There are a lot of standard reasons why patients would be excluded from a clinical trial but, if a person will be getting the drugs in a clinical setting, it is important to know if they can expect to see the same benefits or if they might experience more dangerous side effects,” Mitchell said in a press release. “Physicians can’t know for sure because the trial data do not directly apply.”
Unless clinical trials become more inclusive, physicians must view trial data with a more critical eye, Mitchell said. Also, FDA approval must not represent the final step in the quest to measure a therapy’s effectiveness. Instead, postmarketing data collection and analysis should intensify.
“There is a lot of talk about the need to continue studying these drugs even after they have crossed that approval finish line and are being used,” Mitchell said. “We should do a better job of making sure we’re still seeing beneficial results and not unintended side effects.”
‘Problem with our system’
A study by Mikkael A. Sekeres, MD, MS, director of the leukemia program in the department of hematology and medical oncology at Cleveland Clinic Taussig Cancer Institute, and colleagues further support the notion that strict clinical trial eligibility criteria exclude patients who are most likely to benefit from cancer treatment.
One member of the research team reviewed several studies and observed similarities in the eligibility criteria for many cancer trials. A formal study ensued to determine whether the eligibility criteria were sensitive to known toxicities of the drugs under investigation for treatment of hematologic malignancies.
The investigators identified 97 studies of therapies for leukemia, lymphoma, multiple myeloma or myelodysplastic syndrome.
Results showed the known toxicities of the drugs evaluated in the trials did not correlate with the renal or cardio-hepatic eligibility criteria of the trials. Eligibility criteria were not sensitive to the known toxicities of those therapies, nor did they correlate with the adverse events observed on the trials, Sekeres said.
“It would be legitimate for a trial assessing a drug with known liver toxicities to have strict criteria for patients with liver conditions,” Sekeres told HemOnc Today. “Yet, for the same drug, we found there also were strict eligibility criteria for patients with kidney conditions, heart conditions and neurological conditions, even though the drug had not shown in early phase or preclinical work to have those end-organ effects.
“We have a problem with our system,” Sekeres added. “We can no longer allow ‘cut-and-paste’ eligibility criteria from one trial to another. The pendulum has swung too far.”
This trend is particularly problematic given the already-limited participation in oncology trials.
An estimated 70% of Americans express an inclination or strong willingness to participate in clinical trials; however, less than 5% of U.S. adults with cancer actually do so.
One in four cancer trials terminates early, and the most common reason for early closure is failure to complete enrollment, according to a study Galsky and colleagues presented at Genitourinary Cancers Symposium in 2014.
Galsky and colleagues used ClinicalTrials.gov to identify 7,776 interventional phase 2 or phase 3 adult cancer trials registered between 2005 and 2011. Of these, 25% (95% CI, 19-31) terminated prematurely.
The most common reason for termination was poor accrual (38.7%), followed by sponsor cancellation (10.6%), availability of results at interim analysis (9.9%), toxicity or adverse events (8.2%), and lack of funding (5.6%).
“Relaxing eligibility criteria could help to address this problem,” Galsky told HemOnc Today. “Professional organizations now have this problem on their radar. [They] have appropriately called for eligibility criteria that are pointless and not associated with a high likelihood of compromising clinical trial safety or efficacy to be eliminated. This is definitely needed. Because trial enrollment is such a problem, we do not want to put up any unnecessary barriers.”
Education and awareness
A groundswell of anecdotal support in the clinical and advocacy communities to eliminate or relax some eligibility criteria has transformed within the past couple years into coordinated, evidenced-based endeavors.
ASCO and Friends of Cancer Res-earch — a nonprofit think tank and advocacy organization — examined five specific eligibility criteria to determine whether they could be modified to broaden trial participation. These criteria included minimum age, HIV infection, brain metastases, organ dysfunction, and prior cancer history or concurrent malignancy.
Working groups — which included clinical investigators, FDA medical reviewers, drug and biotech manufacturers, biostatisticians, pharmacologists and patient advocates — reviewed evidence from each patient population and consulted with members of the research community.
“[We] took a step back and tried to figure out how we got to this place of overly strict eligibility criteria,” Kim told HemOnc Today.
A joint research statement — published along with four supporting manuscripts last fall in Journal of Clinical Oncology — emphasized the need for trial enrollment criteria to “strive for inclusiveness,” and that any rationale for exclusion “should be clearly articulated and reflect expected toxicities associated with the therapy under investigation.”
The statement outlined the following consensus recommendations:
- Pediatric-specific cohorts should be included in initial dose-finding trials based on strong scientific rationale for benefit;
- Later-phase trials for diseases that span pediatric and adult populations should include patients aged older than 12 years;
- Individuals with HIV who are healthy and demonstrate low risk for AIDS-related outcomes should be eligible unless there is a specific reason for exclusion;
- Patients with treated or clinically stable brain metastases should be excluded from trials only if there are compelling reasons;
- Renal function criteria should enable liberal creatinine clearance unless the investigational agent involves renal excretion; and
- Individuals with prior or concurrent malignancies should be allowed to participate in trials, particularly if there is very low risk that the other malignancy will interfere with efficacy or safety endpoints.
Aside from minimum age requirements for trial enrollment, HIV positivity is the most frustrating exclusion criteria in clinical trial protocols, Kim said.
“This means that if [basketball legend] Magic Johnson — one of the most well-known celebrities living with HIV — is diagnosed with cancer, he would be excluded from the majority of clinical trials,” Kim said. “This is ridiculous. We know the data show people with HIV live just as long as anyone in the general population, so why are we still hanging on to these antiquated eligibility criteria?”
Patients with brain metastases traditionally have been excluded from clinical trials because previous data suggested their prognosis is worse.
“We are testing drugs that have great penetration into the cerebral spinal fluid across the blood-brain barrier, predictively in preclinical experiments,” Kim said. “When the clinical trial starts, we exclude the patients who could benefit the most — those with disease in the brain. Why would we exclude those patients in a clinical trial when preclinical profiles show it is where the drug has an advantage?”
A movement already is underway to broaden eligibility criteria for patients with organ dysfunction, Sekeres said.
“At Cleveland Clinic, we have codified our clinical trial template instructions to ... only limit eligibility criteria based upon organ dysfunction if a drug being tested directly affects that organ,” Sekeres said.
The ASCO-Friends of Cancer Research recommendations already have affected major change, according to Joseph M. Unger, PhD, MS, assistant member of the cancer prevention program in the division of public health sciences and affiliate assistant professor in the department of health services research at University of Washington School of Public Health.
SWOG — of which Unger is a member — now is re-examining whether the common exclusion criteria those organizations highlighted can be relaxed or removed from its trial protocols, Unger said.
“I assume that all of the major network groups are doing the same, as these recommendations are having a wide impact on trial designs,” Unger said.
Awareness and education are essential to further these efforts, Galsky said.
“ASCO has done a good job ensuring that this topic is front and center,” he said. “The disconnect between the FDA and our industry colleagues is the gap that now needs to be bridged. We all want to relax eligibility criteria to make trials more efficient, but there is still this background concern about safety. For certain eligibility criteria, the FDA still tends to err on the side of caution, and it will require more education and experience for the tide to turn.”
Nonessential eligibility criteria
Although awareness must increase before certain trial restrictions are relaxed or eliminated, others can be abandoned with no negative consequences, according to the experts with whom HemOnc Today spoke.
These stipulations — such as missing documentation or paperwork — are known as nonessential eligibility criteria and have no bearing on a treatment’s efficacy or safety profile.
“This is the issue we are squarely focused on — eligibility criteria that have been created but do not serve a safety or scientific purpose,” Fleury said. “These are the ‘cut-and-paste’ criteria that nobody took the time to look at. It is concerning when we have a willing patient who would otherwise enroll on a trial and contribute data safely but is kept off the trial because its design was not well thought out.”
Sekeres and colleagues compared outcomes of patients eligible for SWOG Leukemia Committee phase 2 or phase 3 trials with outcomes of those who were ineligible. Their analysis included 13 studies conducted since 2005. Researchers categorized reasons for trial ineligibility, as well as baseline characteristics of patients — such as ECOG performance status — serious adverse events, complete remission status and OS.
A total of 2,361 patients enrolled in the 13 studies, but 247 were deemed ineligible. Sekeres and colleagues included 169 of the ineligible patients — all of whom were treated on the SWOG studies, but found to have been ineligible after the fact — in their analysis; the majority (73%) of the 78 ineligible trial participants who were not included in the analysis did not have the appropriate disease of interest.
More than half (60%) of the 169 trial-ineligible — but treated on study — individuals analyzed had missing documentation.
Multivariable analysis revealed no significant difference in OS between treated eligible and ineligible patients. Complete remission rates differed significantly for only two of the 13 analyzed trials — one favoring eligible patients, the other ineligible patients. Both eligible and ineligible patients treated on SWOG studies had similar rates of adverse events.
The results suggest nonessential eligibility criteria — particularly those related to missing documentation — can be relaxed without affecting outcomes, Sekeres said. The findings also highlight one way in which poor enrollment is unnecessarily exacerbated by trial design, he added.
Focus on patients
Although support appears to be growing for less stringent trial eligibility criteria, the restrictions — aside from those deemed nonessential — may not disappear imminently, Unger said.
“The next steps are to examine the common exclusion criteria identified by ASCO and Friends of Cancer Research and determine if they can be safely removed from clinical trials without affecting safety,” Unger told HemOnc Today. “To implement change, it is necessary to conduct studies that examine the relationship between common exclusion criteria and trial outcomes. Sometimes these studies can take several years to conduct, but this is truly important.
“Although this will be an ongoing process that will unfold during the next 5 to 10 years, it has already gotten off to a good start with the ASCO recommendations,” he added. “It is vitally important for us to do what we can to help patients with cancer. It all boils down to being about the patients, as well as about the research.”
The first key barrier is to eliminate ‘cut-and-paste’ eligibility criteria, Sekeres said.
“The second barrier to overcome is to stop penalizing pharmaceutical companies for enrolling real-world patients,” Sekeres said. “We cannot allow real-world patients who have comorbidities that lead to side effects to be inappropriately tagged to a study drug and have that jeopardize that drug’s potential approval.
“The only way this will happen is if FDA mandates that real-world data be collected prior to drug approval,” he added. “I am passionate about this because it has gone too far.”
Ultimately, more intense scrutiny on trial eligibility criteria is necessary to ensure restrictions are practical and clinically relevant for the real world, not the “super-athlete clinical trial world,” Kim said.
“Eligibility criteria should help in the evaluation of a drug and not be an appendage that needs to be removed,” he said. – by Jennifer Southall
American Cancer Society Cancer Action Network. 2018. Barriers to Patient Enrollment in Therapeutic Clinical Trials for Cancer: A Landscape Report. Available at: www.acscan.org/policy-resources/clinical-trial-barriers. Accessed on July 12, 2018.
Kim ES, et al. J Clin Oncol. 2017;doi:10.1200/JCO.2017.73.7916.
Mitchell AP, et al. J Oncol Pract. 2015;doi:10.1200/JOP.2015.004929.
Pestine E, et al. Ann Oncol. 2018;doi:10.1093/annonc/mdy138.
Statler A, et al. Leukemia. 2016;doi:10.1038/leu.2016.374.
Statler A, et al. Blood. 2018;doi:10.1182/blood-2018-01-826693.
Stensland KD, et al. J Natl Cancer Inst. 2014;doi:10.1093/jnci/dju229.
Unger JM, et al. Am Soc Clin Oncol Educ Book. 2016;35:185-198.
For more information:
Mark Fleury, PhD, MS, can be reached at American Cancer Society Cancer Action Network Inc., 555 11th St. NW, Suite 300, Washington, D.C. 20004; email: firstname.lastname@example.org.
Matthew Galsky, MD, can be reached at Mount Sinai Health System, 1470 Madison Ave., New York, NY 10029; email: email@example.com.
Edward S. Kim, MD, can be reached at Levine Cancer Institute, 1021 Morehead Medical Drive, Charlotte, NC 28204; email: firstname.lastname@example.org.
Mikkael A. Sekeres, MD, MS, can be reached at Cleveland Clinic Taussig Cancer Institute, 2010 E. 90th St., Cleveland, OH 44195; email: email@example.com.
Joseph M. Unger, PhD, MS, can be reached at Fred Hutchinson Cancer Research Center, 1100 Fairview Ave. North, Seattle, WA 98109; email: firstname.lastname@example.org.
Disclosures: Fleury, Galsky, Kim, Sekeres and Unger report no relevant financial disclosures.