In an interview, Dr Gyawali outlined when he feels it is best to use a noninferiority study approach in cancer trials, and when it may be more appropriate to run superiority trials.

Cancer Therapy Advisor (CTA): You found that the lack of justification for running a noninferiority design was associated with industry funding for a study, but essentially, justified or not, you found there was no association with running this type of trial and OS. What is the influence of the findings on patients, then?

Dr Gyawali: We were concerned if noninferior trials, on the aggregate, were actually harmful. It was reassuring to find no detrimental effect on OS. However, this is a pooled effect and each trial individually can be harmful. For example, the [hazard ratio] was over 1 in 41% of the trials in our study. Also, there is the possibility of publication bias against the truly harmful noninferiority studies. Furthermore, the lack of effect on OS can also be interpreted as the drugs tested via this approach do not improve OS. So patients should be aware if there are any other benefits that justify this design.

The association of lack of justification and industry funding is not surprising because academic centers, when they run non-inferiority trials with public funds, usually do so when there is a compelling reason —  whereas for industry, noninferiority designs have become a way to sneak in to get regulatory approval.

Related Articles

CTA: It seems there are really only 2 situations for which noninferiority trials should be used: when you are dealing with a reduction/change in dose or dose schedule, or when you are looking at a formulation switch (ie, from intravenous to oral administration). Improvement in QoL might be a worthwhile third justification for use, too. Should noninferiority trials be used in any other circumstances? How about in the case of oncology biosimilars?

Dr Gyawali: Noninferiority designs are justified when there are compensatory benefits: the 3 circumstances that you’ve so well articulated. The fourth circumstance would be when the newer intervention is cheaper than the standard of care. It could be due to lower dose/frequency of the same drug or a different drug costing less than the standard. Biosimilars would fall into this category and may be reasonably tested in a noninferiority design, but “equivalence design trials” are the most commonly used approach for testing biosimilars, as they need to be proven equivalent to the reference molecule.

CTA: Should we stop measuring/looking at progression-free survival (PFS) in cancer altogether? What would prevent pharma from choosing end points that are easily (and consistently) met in noninferiority trials?

Dr Gyawali: Although there is some debate about the use of PFS in cancer drug trials, there should be no debate that PFS or any surrogates should not be used in noninferiority trials. Noninferiority [trials], by definition, test whether a new drug is worse than the standard, so why would any patient want to take a new drug that didn’t worsen PFS? What if the PFS was noninferior, but survival [was] significantly worse? Thus, OS should always be tested in any noninferiority trial of new cancer drugs. For this to happen, regulatory agencies and ethics [institutional review boards] should take a proactive step.

CTA: Can you think of any examples in which a drug that was deemed “noninferior” got preferred formulary placement/coverage status over a drug that was used as the control?

Dr Gyawali: I haven’t checked with other formularies, but lenvatinib for HCC [hepatocellular carcinoma] got category 1 recommendation by The National Comprehensive Cancer Network® (NCCN®). It was tested against sorafenib (and sorafenib also continues to receive category 1). Lenvatinib got approval a few months ago. If we check next year, I am confident we will see more patients with HCC getting lenvatinib than sorafenib in the first line.3


  1. FDA
  2. Gyawali B, Tessema FA, Jung EH, Kesselheim AS. Assessing the justification, funding, success, and survival outcomes of randomized noninferiority trials of cancer drugs: a systematic review and pooled analysis. JAMA Netw Open. 2019;2(8):e199570.
  3. Gyawali B, Kesselheim AS. US Food and Drug Administration approval of new drugs based on noninferiority trials in oncology: a dangerous precedent? JAMA Oncol. 2019;5(5):607-608.