This article is more than 5 years old. Information may no longer be current.
Déjà vu all over again — or am I missing the take-home message?
Click Here to Manage Email Alerts
In my clinic recently, I saw a vigorously healthy 79-year-old woman newly diagnosed with diffuse large B-cell lymphoma.
She has virtually no significant past medical history, no major comorbidities, a left ventricular ejection fraction of 75%, and — pending the results of her staging evaluation — probably has advanced-stage disease.
I will most likely recommend chemotherapy with R-CHOP for six cycles at 21-day intervals. Based on the National Comprehensive Cancer Network International Prognostic Index (NCCN-IPI), the 5-year OS projection for this patient is 64% with this approach.
John Sweetenham
This is a common clinical scenario, and it’s now widely accepted that — although advanced age remains an adverse prognostic factor — older patients without comorbidities and with good performance status have excellent outcomes when treated with standard regimens.
On the face of it, this clinical vignette doesn’t lend itself to an editorial in HemOnc Today. However, this patient’s presentation happened to coincide with publication of a couple of articles related to diffuse large B-cell lymphoma (DLBCL) in the elderly, and a third intriguing paper about how we present and interpret data.
As the headline of this piece suggests, the subject of treating older patients with cancer is not new. It remains complex and controversial. The papers described below, in many respects, cover old ground and the findings are unsurprising. What is noteworthy — and maybe even slightly disturbing — is the nature of the “messaging” of some study findings.
‘A realistic snapshot’
The first of these papers describes a comparative effectiveness study — derived from data in the SEER–Medicare database — of more than 9,000 patients aged older than 66 years with DLBCL.
Hamlin and colleagues described their article as a “real-world” analysis of patterns of care and outcomes for older patients with DLBCL. In large part, the results were quite predictable.
They found that 23% of patients received no treatment (this figure rose to 33% among those aged older than 80 years).
Overall, researchers concluded OS was higher for patients who received rituximab (Rituxan; Genentech, Biogen Idec) plus chemotherapy and rituximab monotherapy compared with no treatment, and that suboptimal (ie, less than six cycles) of curative therapy was associated with a poor outcome.
There really are no new messages here and, as the authors acknowledge, the conclusions are limited by the retrospective nature of the study and the lack of granularity intrinsic to SEER data. Poor performance status, comorbidities, lack of financial resources, lack of caregivers and many other factors play into these treatment decisions, and the poor outcomes reported for many of these patients likely relate to these factors as much as the efficacy of treatment.
Despite that, however, these data do provide a realistic snapshot of the current treatment of older patients with this disease in the United States, and the many challenges to improving outcomes.
A different story
The second of these papers tells a very different story.
This study, conducted by the German High-Grade Non-Hodgkin’s Lymphoma Study Group, was designed to investigate the “pharmacokinetics, toxicity and efficacy of prolonged rituximab exposure in elderly patients with DLBCL.”
In their conclusions, Pfreundschuh and colleagues state: “Extended rituximab exposure … significantly improved outcome of elderly poor-prognosis patients without increasing toxicity” and “To our knowledge, results obtained … are the best reported for elderly patients with DLBCL to date.”
Given that I anticipate having to treat an older patient, these conclusions certainly made me sit up and take notice.
The report describes a phase 2 study of 189 patients with DLBCL who received an R-CHOP regimen given at 14-day intervals for six cycles. Rituximab scheduling was based on results of previous studies and on pharmacokinetic considerations to prolong rituximab exposure.
This study was based on the previous studies from this group and represents one in a sequence of carefully designed clinical studies in these patients. The survival curves for this study show remarkable OS rates for elderly patients with DLBCL, and comparisons with earlier studies seem to suggest an improvement in outcome with the new rituximab schedule.
Closer review of the data shows that this group defined “elderly” as age 61 years or older. The median age of the patient population was only 68 years, almost 90% of the patients had an ECOG performance status of 0, and almost 50% of the patients were low or low–intermediate risk according to the NCCN-IPI.
Further, a pre-phase treatment of vincristine and prednisone was given to all patients prior to receipt of protocol therapy. It isn’t clear whether patients who did not improve with this therapy were able to enter the study and were included in this analysis.
Overall, the patient population is clearly selected and very distinct from the “real-world” population of the first study. But potential selection bias in publication is hardly a new phenomenon, and the results of this study still remain provocative and relevant for this selected population. The authors should be congratulated for a fine study.
The problem I perceive with this paper is in the messaging. The description of the study population as “elderly” is a stretch when the lower age limit is 61 years. It may be that I am now finding this hitting a bit too close to home personally, but we need to achieve a universal definition of what constitutes “elderly” for study purposes.
At several places in the manuscript, the authors refer to “the standard six cycles of R-CHOP-14 plus rituximab.” This is concerning given the results of well-conducted randomized trials that show no difference in outcome for patients treated with R-CHOP for 14 days vs. the more widely used 21-day cycle. In fact, R-CHOP at 21 days for six cycles is the internationally accepted standard. The comments of the authors that this regimen “significantly improved all outcome parameters for elderly poor-prognosis patients” compared with their previous trial is speculative given that they are comparing to a historical control (in their conclusions, the authors do acknowledge this and cite an ongoing randomized trial).
Effects of spin
The last of the three papers was a thought-provoking study by Boutron and colleagues on the impact of “spin” in abstracts that report the results of randomized controlled trials in cancer.
I was surprised to learn there is a growing body of literature on the subject of spin in the reporting of results of cancer trials.
The paper describes a randomized trial in which clinicians were exposed to two versions of a meeting abstract that described a clinical trial with a non-significant outcome according to its primary endpoint.
One of the abstracts was an original, in which the language emphasized secondary outcomes or subgroup analyses — defined as having spin. The other was re-written to remove these nuances from the language.
The study showed clinicians were more likely to rate the experimental arm of the study as beneficial if the abstract contained spin. The authors went on to suggest a structured format for meeting abstracts that minimizes the chances for spin.
Although this study was restricted to consideration of meeting abstracts, it highlights the potential for the content of published manuscripts to send the wrong message, even when all data are presented accurately.
Conclusion
When considered together, these three papers demonstrate the choice of language and nuance by the authors — whether intentional or unintentional — can significantly influence the interpretation of data.
Although it’s probably just the impression of someone who has been in oncology for quite some time, I have the feeling there is less rigor in the peer review and editing processes with regard to these issues than was apparent a few years ago.
Do we need new guidelines around manuscript preparation to protect readers from the potential for spin? I don’t think so, but we should expect the peer review process to minimize this — especially at a time when manuscripts are sometimes prepared by ghost writers who work for industry sponsors of studies.
As oncologists and hematologists, we need to be especially vigilant regarding the messages contained in the scientific literature. We need to be aware of the subtle influences that the composition of manuscripts can have on our interpretation of results, and we must remember that the take-home message of clinical studies lies in the data.
What does that mean for me, as I prepare to treat the older patient I recently saw in my clinic? After careful consideration, I will recommend R-CHOP for 21 days.
References:
Boutron I. J Clin Oncol. 2014;32:4120-4126.
Hamlin PA. Oncologist. 2014;19:1249-1257.
Pfreundschuh M. J Clin Oncol. 2014;32:4127-4133.
For more information:
John Sweetenham, MD, is HemOnc Today’s Chief Medical Editor, Hematology. He also is senior director of clinical affairs and executive medical director at Huntsman Cancer Institute at the University of Utah. He can be reached at john.sweetenham@hci.utah.edu.
Disclosure: Sweetenham reports no relevant financial disclosures.