Pragmatic Approaches to Cancer Management in the Absence of Randomized Trial Data

Maurie Markman, MD

When available, data generated from well-designed, phase 3 randomized trials are appropriately considered the highest level of evidence for optimal management for patients with cancer. Unfortunately, it is increasingly recognized that such studies often fail to reasonably represent the actual population of patients seen in routine community practice. Concerns include limited numbers of patients older than 65 or 70 years and individuals with common comorbid medical conditions who participate in these studies.^1,2

The magnitude of this deficiency was emphasized by a study reported several years ago, which examined the Surveillance, Epidemiology, and End Results–Medicare database for the noninvestigative administration of bevacizumab (Avastin) to patients with metastatic cancers of the lung, breast, or colon who were aged at least 65 years.³ The clinical trials that led to regulatory approval of this agent in the management of these advanced cancers specifically excluded individuals with certain contraindicated cardiac, vascular, and bleeding conditions. Remarkably, in the reported analysis, approximately one-third of the individuals who received this agent were also found to have Medicare billing claims for 1 or more of the so-called contraindicated conditions. These data raise a serious question regarding the relevance of actual clinical practice to trial data employed for regulatory antineoplastic drug approval The first 2 words of this commentary are a second issue here: “When available.”

Randomized trials are generally quite complex and expensive to conduct, require large numbers of patients and investigators, and take considerable time (usually measured in years) to complete. As a result, it is reasonable to suggest randomized trials realistically define the care in only a limited number of specific cancer settings. It is also relevant to note that by the time results of many randomized trials are available, new management paradigms may have been introduced into the clinic based on phase 2 or even robust phase 1 study results that make the study findings only modestly or even minimally relevant.

Further, a large percentage of randomized cancer trials conducted today focus on establishing a role for new therapeutic strategies rather than answering perhaps less exciting but equally compelling clinical questions, such as the optimal sequencing of available antineoplastic drug regimens for patients with metastatic disease. Another example is the absence of studies directly comparing checkpoint inhibitors approved in each clinical setting to determine whether there are relevant differences in efficacy or toxicity.

As a result, clinical guidelines or pathways developed as decision support strategies or as insurance-mandated payment mechanisms are frequently based as much on opinion as on objective, statistically significant study results. Although this statement does not negate the potential use of such efforts, it is important to note the prominent role of expert opinion vs randomized trial data.

Finally, it is essential to emphasize the following in this discussion: The absence of evidence-based phase 3 trial data in support of a particular approach to cancer management is not equivalent to evidence of absence of clinical value.

For example, fallopian tube cancer and primary cancer of the peritoneum have been and continue to be appropriately managed (eg, surgical, antineoplastic drug therapy) in an identical manner to epithelial ovarian cancer (EOC). Further, these 2 malignant entities are routinely included in clinical trials examining novel approaches in the management of EOC. However, in the absence of phase 3, randomized trials conducted in these conditions that have independently examined outcomes, is it fair to conclude there is equivalent activity of bevacizumab or of PARP inhibitors compared with the administration of these agents in EOC?

The issue highlighted here is one of many where members of the scientific investigative communities have carefully considered the available evidence and concluded the assumptions are both scientifically valid and of clinical value in disease management. In fact, the alternative to making such rational decisions would be to simply declare there are no meaningful data upon which clinicians can base management decisions.

One might describe the process noted above as being a pragmatic strategy to define optimal approaches to treatment in the absence of existing phase 3 trial data or where it is unlikely such information will even exist.

The relevance of this critical process to multiple questions in cancer management cannot be overemphasized. As previously noted, research subjects in phase 3 trials often poorly represent the population of individuals seen in clinical practice. Of course, oncologists always desire to offer appropriate patients the potential benefits associated with the delivery of agents revealed to improve outcome in the research setting, regardless of whether the clinical characteristics of the study population completely match that of the patient they are considering for noninvestigative drug delivery.

But to proceed in the unfortunate absence of adequate safety data (eg, bevacizumab (Avastin) in the presence of a patient history of contraindicated comorbidity) clinicians may need to do their own “trial.” For example, to observe for excessive toxicity (eg, bone marrow suppression in a patient previously treated with external beam radiation) the new drug may initially be administered at a lower dose or on a less frequent schedule than recommended in the product label, with subsequent dose escalation assuming acceptable toxicity.

In addition, clinicians may elect to employ strategies to mitigate potential toxicity that have not been documented through the conduct of phase 3 randomized trials, but that do not require regulatory approval for their use and, importantly, have no realistic risk of causing harm. For example, a recent study prospectively followed patients receiving oxaliplatin as adjuvant therapy for stage III colon cancer and noted that individuals who had higher levels of exercise after treatment experienced a reduced degree of peripheral neuropathy.⁴ What is wrong with strongly recommending an exercise regimen for such patients in the absence of phase 3 trial data that documented its value?

There is no inherent conflict in this discussion of pragmatic approaches to disease management to complement the value of data generated from phase 3 randomized trials. Where available, discussions of disease management should always begin with a critical acknowledgment of relevant data from such studies, but the specific plan for an individual patient needs to also focus on the unique clinical features of the individual and that patient’s input into the decision-making process.

Maurie Markman, MD, is president of Medicine & Science at Cancer Treatment Centers of America, a part of City of Hope.

References

1. Murthy VH, Krumholz HM, Gross CP. Participation in cancer clinical trials: race-, sex-, and age-based disparities. JAMA. 2004;291(22):2720-6. doi:10.1001/jama.291.22.2720
2. Abi Jaoude J, Kouzy R, Mainwaring W, et al. Performance status restrictions in phase III cancer clinical trials. J Natl Compr Canc Netw. 2020;18(10):1322-1326. doi:10.6004/jnccn.2020.7578
3. Hershman DL, Wright JD, Lim E, Buono DL, Tsai WY, Neugut AI. Contraindicated use of bevacizumab and toxicity in elderly patients with cancer. J Clin Oncol. 2013;31(28):3592-3599. doi:10.1200/JCO.2012.48.4857
4. Lee S, Ma C, Shi Q, et al. Potential mediators of oxaliplatin-induced peripheral neuropathy from adjuvant therapy in stage III colon cancer: findings from CALBG (Alliance)/SWOG 80702. J Clin Oncol. 2022;41(5):1079-1091. doi:10.1200/JCO.22.01637