Will a drug allow my patient to live a longer, better life? That question lies at the heart of drug trial outcomes.
But sometimes those core goals get buried in the design and interpretation of drug trials.
The recent CodeBreaK 200 trial represents a prime example of that.
CodeBreaK is a phase 3 randomized controlled trial that evaluated the new targeted drug sotorasib against docetaxel chemotherapy in patients with KRAS G12C mutation positive non–small cell lung cancer (NSCLC) who have already progressed on both immunotherapy and chemotherapy — essentially a last-line therapy trial. Sotorasib received accelerated approval for this indication based on objective response rate from a single-arm trial in 2021, but this phase 3 trial provides the first comparative progression-free and overall survival data.
Here are the results from CodeBreaK 200: the median overall survival was 10.6 months for the new drug (sotorasib) vs 11.3 months for docetaxel, less than 1 year in either arm. The hazard ratio (HR) is slightly over 1, with the 95% CI showing, in the worst-case scenario, a 33% increase in risk for deaths with sotorasib (HR, 1.01; 95% CI, 0.77-1.33).
Progression-free survival (PFS) results showed that compared with docetaxel, sotorasib improved median PFS by 1 month (5.6 vs 4.5 months; HR, 0.66).
For a new drug that could potentially worsen overall survival compared with chemotherapy, the researchers should at minimum hold off expressing enthusiasm for the new drug until more data can clarify its position in the treatment algorithm. The conclusion cannot be that the drug should now get full approval, which would be the likely next step after such enthusiasm about the results; I'd argue that its earlier accelerated approval should in fact be revoked for now. I definitely would not conclude that it is superior to docetaxel nor that sotorasib beats docetaxel "hands down."
But that's exactly what the many experts have done. This surprises me because sotorasib clearly failed to benefit patients. And even those who advocate using PFS as a surrogate endpoint generally agree that when survival is less than 1 year, overall survival takes precedence.
Some of these experts may be confusing excitement surrounding a scientific discovery with the need to celebrate trial results, even when they're not impressive. Yes, being able to drug KRAS is a huge scientific discovery. It's also true that the CodeBreaK 200 results are underwhelming. Both statements can simultaneously be true. We don't need to buy into hype.
The trial authors, however, appear to be playing into the hype — touting the "hands down" superiority of a new drug that only improved median PFS by 1 month compared with chemotherapy. The overall survival findings, if anything, favored docetaxel. We should seriously question our common sense in oncology if we genuinely believe these results are cause for celebration.
However, the more surprising fact about this trial was that it was initially appropriately powered for overall survival. The publication mentions that "per regulatory feedback," the protocol was "amended to decrease" the sample size and "only power the trial for PFS," to "limit the overall number of patients treated with docetaxel."
This change to the analysis midway through the trial defies our code of ethics when conducting clinical trials. First, as discussed, in last-line therapy when median overall survival is already less than 1 year, PFS alone does not imply any clinical benefit, so why would the regulators recommend changing the protocol in a way that shifted the primary endpoint from overall survival to PFS? I understand why the industry would want to get away with PFS but for the regulators to suggest this change defies common sense.
The argument that overall survival takes forever to measure doesn't hold up here and neither does the postprotocol therapy argument because this is patients' last line of treatment.
The trial does mention one reason for this change: limiting the overall number of patients in the control arm. But why? Are we already convinced that sotorasib is better than is docetaxel? If so, why even do the trial? We do randomized trials because there is clinical equipoise. Being randomized to control arm has saved many lives in the history of medicine and oncology.
Take the iLLUMINATE trial in patients with high cardiovascular risk, where adding the drug torcetrapib to statin increased the risk for mortality by 58% and the risk for cardiovascular events by 25%. In oncology, adding rilotumumab, a targeted drug, to chemotherapy among MET-positive patients with advanced gastroesophageal cancer shortened median survival by 2 months, which was statistically significant.
And then there's the ongoing, heated debate in the oncology community surrounding the role of surrogate endpoints such as PFS. The proponents of PFS argue that although it is not as clinically relevant as overall survival, this surrogate endpoint gives earlier results and avoids confounding by subsequent therapies. Following this logic, PFS would have minimal to no utility in this population: patients with a poor prognosis (median survival less than 1 year) and with no available later-line therapies.
But I and others have repeatedly shown in our research that improving PFS does not necessarily translate to improving survival or quality of life. Because PFS has no correlation with survival in many cases, the word "survival" in PFS is misleading, and we have argued that PFS should be rephrased as progression-free interval instead.
When the median overall survival in the control arm seems longer, and the HR is over 1, as in the CodeBreaK trial, the principle of primum non nocere dictates that we don't accept this new therapy as standard of care until we have better, clearer, cleaner results that establish the superiority of the new drug. Yes, we can simultaneously be happy about the research progress that has allowed us to have a drug targeted against KRAS mutation but also not accept sotorasib as the standard of care.
When you consider the price of sotorasib, with an average wholesale price per month of more than $22,000 USD, prematurely celebrating this drug becomes all the more unacceptable. But even if the price were more reasonable, this should still not be considered a practice-changing trial.
Perhaps in the future, we will build on this experience and have a better drug in a better-defined population where the efficacy is unequivocal, but that is not the case with CodeBreaK 200.
This trial was named CodeBreaK. We indeed broke several codes" our codes of ethics, statistical principles, and common sense in its design and interpretation.
Acknowledgment: I dedicate this column to my 2023 Winter GLPH 488 class of Queen's University BHSc program where I taught how to critically appraise a clinical trial using CodeBreaK 200 as an example and promised to turn our critical appraisal discussion into a commentary.
Bishal Gyawali, MD, PhD, is an associate professor in the Departments of Oncology and Public Health Sciences and a scientist in the Division of Cancer Care and Epidemiology at Queen's University in Kingston, Ontario, Canada, and is also affiliated faculty at the Program on Regulation, Therapeutics, and Law in the Department of Medicine at Brigham and Women's Hospital in Boston. His clinical and research interests revolve around cancer policy, global oncology, evidence-based oncology, financial toxicities of cancer treatment, clinical trial methods, and supportive care. He tweets at @oncology_bg.
Lead image: Bishal Gyawali, MD, PhD
Medscape Oncology © 2023
Cite this: The CodeBreaK Trial: We Broke the Code in More Ways Than One - Medscape - Apr 26, 2023.