by H. Jack West MD
When the press release for KEYNOTE-091 came out in mid-January, it was only logical to compare the outcomes for patients with those of the 2021 IMpower010 trial — both trials were looking at adjuvant immunotherapy following surgery for early-stage non–small cell lung cancer (NSCLC).
The KEYNOTE-091 trial announced an improvement in disease-free survival for patients with resected stage IB-IIIA NSCLC after receiving a year of adjuvant pembrolizumab.
The IMpower010 trial reported similarly promising results: A year of adjuvant atezolizumab led to significant improvement in disease-free survival in the same patient population. This finding prompted the recent FDA approval of adjuvant atezolizumab in this setting.
While we await the formal publication of KEYNOTE-091, one distinction in the design of these two studies is the positioning of adjuvant chemotherapy — an established standard of care for patients with resected stage IB-IIIA NSCLC that is associated with a modest but consistent survival benefit. Specifically, the IMpower010 trial randomly assigned patients only after they had received adjuvant chemotherapy, while KEYNOTE-091 did not require it, an approach that may be framed as reflecting “real-world practice.”
KEYNOTE-091 is not the only recent clinical trial to forgo adjuvant cisplatin-based chemotherapy in its design. Take the randomized trial of gefitinib vs adjuvant chemotherapy in EGFR mutation–positive Chinese patients with resected stage IB-IIIA NSCLC. The trial omitted adjuvant chemotherapy in the investigational arm, even though chemotherapy has been shown to confer a significant improvement in overall survival — a benefit we have not yet seen from adjuvant EGFR inhibitors. The same logic applies to the 2020 ADAURA trial of adjuvant Osimertinib.
This shift away from adjuvant chemotherapy in clinical trials reflects real-world practice. Studies show considerable geographic variation in the use of adjuvant chemotherapy across the United States, with carboplatin-based doublets often selected instead of cisplatin-based regimens.
However, these practices do not reflect evidence-based guidelines. The clinical practice guidelines from the National Comprehensive Cancer Network (NCCN) and the European Society for Medical Oncology have unequivocally directed eligible patients with resected stage IB-IIIA NSCLC to first pursue cisplatin-based chemotherapy on the basis of a decade’s worth of trials demonstrating a consistent benefit. And comparator arms in studies like the Intergroup-run ECOG 1505 trial use cisplatin-based doublet chemotherapy for four cycles, not best supportive care.
I think the growing tendency to omit adjuvant cisplatin-based chemotherapy in clinical trials is troubling. The danger is that patients will miss out on adjuvant chemotherapy in this setting because their oncologists have assumed from clinical trials that an alternative can effectively replace chemotherapy. A central characteristic of clinical trials is that they define optimal practice, as much in their design and control arms as in the novel practices introduced. My concern is that we are justifying and exonerating suboptimal practice as normal when we run clinical trials that don’t include adjuvant chemotherapy in these settings.
Those who defend this shift away from adjuvant chemotherapy cite the fact that many patients in the real world are not fit enough to pursue the rigors of adjuvant cisplatin-based chemotherapy. They also argue that clinical trial populations do not reflect the variability in performance status, comorbidities, and diversity of patients seen in the clinic, and thus having clinical data that support the decision to deviate from guidelines is important.
I am not saying that patients and their physicians cannot deviate from standard-of-care treatments when it’s an informed decision. It is critical to decide on the goals of care with patients and discuss the sometimes-marginal benefits associated with guideline-recommended treatments, and these discussions should be appropriately individualized.
But individual patient characteristics don’t explain the tremendous variability in the use of adjuvant chemo across health systems and geographical regions. In other words, the range of patients and their tumor biology is far more constant than oncologists’ commitment to optimal, evidence-based care. It seems as though many patients aren’t getting adjuvant chemotherapy simply because they or their oncologist just don’t feel like it.
And my point is that we should not excuse suboptimal care by rebranding it “real-world practice” and ratifying it in the control arms of landmark clinical trials.
There is definitely value in running trials specifically for patients with marginal performance status, poor renal function, and other real-world characteristics that render these patients ineligible for many of our clinical trials. However, their inclusion in trials that define optimal standards of care shouldn’t redefine and lower our standards. Personally, I’d have far more validating visits with my primary care physician if we just relaxed our body mass index and cholesterol goals to accommodate real-world findings in the US, but that wouldn’t be good for our health. If we don’t aim higher, we’re apt to be complacent or even aim lower.
In a previous column, I highlighted another deviation from recommended practice: low levels of molecular marker testing for patients with advanced NSCLC across community-based practice in the US. Fewer than half of these patients are getting next-generation sequencing, even though it is clearly recommended by the NCCN.
Yet, experts aren’t arguing that we should stop trying to increase molecular marker testing rates, or that we should stop sequencing tumors in trials simply because comprehensive testing doesn’t represent real-world practice.
That would be inane. It is also not in the corporate interest of pharmaceutical companies trying to identify every eligible patient for their targeted therapies.
Likewise, it should also be considered inane to lower our standards for early lung cancer patients receiving adjuvant therapies — yet, in this instance we are willing to rebrand suboptimal care as “real-world practice” for the sake of broadening clinical trial eligibility and finishing trials faster.
Doing this not only reduces the quality of clinical trials, but it also lowers our expectations of what should be standard and optimal care for our patients.