In the cancer management arena, the term “clinical benefit” has unfortunately achieved a problematic status.
Maurie Markman, MD
In the cancer management arena, the term “clinical benefit” has unfortunately achieved a problematic status. For example, a claim that a patient has experienced clinical benefit after receiving an experimental antineoplastic in a clinical trial may be based solely on the fact that the cancer has not progressed by some measurable but arbitrary metric for a limited period (eg, 2 months). This is highly misleading since all that may have been observed is the natural history of the cancer’s relatively slow progression in the patient.
In fact, this definition of clinical benefit also may require no evidence that the patient has symptomatically improved or even that the adverse effects experienced were tolerable—other than the statement the patient was not removed from the study due to toxicity. How can a term with such limited—if any—relevance to actual benefit be used to describe an individual patient’s clinical outcome?Another example of the questionable use of the concept is during the accelerated approval process in which drugs are conditionally approved based on a definition of clinical benefit that does not represent an objectively valid meaning of the term. A recently published commentary in the New England Journal of Medicine discussing the cost of new pharmaceuticals demonstrates the poorly considered use of this expression.1 The authors note that the FDA has given expedited approval for drugs despite a lack of known clinical benefit and then required additional trials for full/final approval. The commentators highlight the need for timely completion of these studies so as “to limit the period of uncertainty about true clinical effect.”1
However, this method of determining clinical benefit is also problematic. A phase III randomized trial is composed of a carefully selected group of individuals likely to be woefully inadequate in its inclusion of real-world patients, such as older people and those with common clinically relevant comorbidities including diabetes, heart disease, or morbid obesity. If such a trial achieves the statistical gold standard of P <.05, is that the definition of “true clinical effect”? Or, conversely, if the study fails to achieve this rather arbitrary level of statistical significance, does this mean patients who participated in the study experienced no clinical benefit?
To be clear, the FDA has legal authority to determine if an antineoplastic agent can be marketed in the United States for specific clinical indications and the agency can certainly elect to use whatever term it chooses to define the category of drugs that fulfill its criteria. But this regulatory approval process does not—and should not—necessarily equate with a medically appropriate or patient-centered statement regarding the meaning of clinical benefit.
Another example of a stunningly inappropriate use of the term “clinical benefit” is provided by a recently published commentary that lashes out against the immensely exciting and rapidly expanding arena of the paradigm-changing process known as precision cancer medicine.2 The author zeroes in on a single paper to suggest that a limited percentage of patients (6.4%) are able to receive a therapeutic strategy based on increasingly sophisticated molecular-based tumor profiling.2 The editorialist then declares that “only about 30% respond at all” to biomarker-driven therapies and concludes, “I estimate that precision oncology will benefit around 1.5% of patients with relapsed and refractory solid tumors.”2
While one might reasonably decide to simply ignore the stunningly arbitrary and biased nature of this individual’s comments, the concern here is that such rhetoric may reach the lay media, with the public and patients with cancer and their families reading about, or listening to, such unsubstantiated and scientifically questionable statements.Why is any of the preceding discussion important today? The answer is very simple: although the term clinical benefit is highly critically relevant, the determination of whether a therapeutic strategy reaches this status should not be based solely on whether a tumor has not grown by an arbitrary time metric (eg, 2 months) or by a drug regulatory agency that decides when and if to permit the marketing of an antineoplastic for 1 or multiple clinical indications. And it certainly should not be based on the arbitrary opinion of a single outspoken commentator.
Rather, clinical benefit should be defined by the individual patient. Such benefit entails far more than the mere shrinkage or relatively slow growth of a tumor mass or whether a trial’s P value reaches a statistical threshold. It is time to change how this term is employed in the realm of cancer management and in the routine discussion of clinical outcomes.