Maurie Markman, MD
Although oncology researchers seek to describe the significance of clinical trial outcomes with scientific rigor, the terminology used to communicate those findings in the peer-reviewed literature sometimes falls short in conveying the impact that a given therapy may have on patients in a real-world scenario. The quest for more precise language is an ongoing struggle, both for oncology specialists and for generalists seeking to translate clinical data into meaningful information.
Two recent letters to the editor in peer-reviewed oncology publications provide poignant examples of this dilemma. In the first letter,1
the authors question the use of the description “increasing survival in pancreatic cancer” for the title of an article2
about clinical trial findings, suggesting that the “title strikes us as inappropriately rosy, given the modest benefits and substantial toxic effects observed.”1
The second letter3
expresses concern about the description in an article abstract4 about a combination therapy program as being “well tolerated” considering the degree of objectively reported toxic events.
The intent of noting these specific letters to the editor is not to suggest anything of concern in the original peer-reviewed manuscripts. In fact, in both cases the authors of the primary report have provided rational justifications for their use of these descriptive terms.5,6
Rather, the focus of this commentary is to emphasize the difficulty of objectively describing clinical trial outcomes when specific words or terms realistically may have quite subjective meanings.
Meaning of “Clinical Benefit” Unclear
For example, consider the term, clinical benefit rate. This description, which is increasingly utilized in the clinical trial domain, commonly combines the experience of patients whose cancers have demonstrated objective biological activity, defined as a complete or partial response/remission usually using Response Evaluation Criteria in Solid Tumors (RECIST),7
plus those whose cancers have not progressed (again, usually employing RECIST) for a protocol-specified period of time that is often a duration of no more than several months.8
The fundamental problem here, of course, is that there is absolutely nothing inherent in the failure of a cancer to grow at a particular rate or to demonstrate a limited or even a major degree of shrinkage that indicates that a patient has actually attained any level of clinical benefit.
That is, does mere shrinkage of one or more cancerous masses indicate that a patient with cancer-related pain has experienced improvement in that pain, or that an individual with cancer-related bowel obstruction has experienced relief of that obstruction or an improvement in appetite, nutritional status, or quality of life? Further, what if the decrease in the size of tumor masses is associated with substantial chemotherapy-induced peripheral neuropathy or mucositis? In this setting, can such a response truly be classified as achieving clinical benefit?
Similarly, a decline in the size of one or more mass lesions can reasonably be considered to represent a biological effect of an antineoplastic strategy. However, such a conclusion would not necessarily be appropriate for the category of response known as stable disease. That a cancer has not progressed (by standard RECIST) in 2, or 3, or 4 months may only represent the inherent natural history of a given malignancy in a specific patient. With or without therapy, the cancer may exhibit the same or a similar time frame for one or more masses to demonstrate a measurable increase in size. Therefore, why assume that the simple absence of progression by RECIST for a limited period of time actually represents clinical benefit? And, as previously noted, the use of this terminology is even more problematic if the so-called clinical benefit is associated with serious symptomatic adverse events.
“Objective Response” Can Vary in Impact
Even the more neutral term of objective response, as in, “The patient achieved a partial or a complete response of measurable disease,” requires further discussion. In the evaluation of the relative degree of biological activity of a specific antineoplastic program in a clinical trial, and in particular a direct comparison between “regimen A” versus “regimen B” in a well-designed phase III randomized study, the utility of defining an objective response is clear. Even if this definition may appear to be considerably more than a little arbitrary (eg, RECIST), the study investigators will at least be employing the same criteria to declare a response within the specific study, and researchers will be using similar vocabulary when discussing outcomes across different trials. Of course, this assumes the criteria are rigorously applied within a given study and by the community of clinical researchers.