Inappropriate use of progression-free survival in cancer drug approvalsBMJ 2020; 368 doi: https://doi.org/10.1136/bmj.m770 (Published 10 March 2020) Cite this as: BMJ 2020;368:m770
All rapid responses
Trial entry criteria are more problematic than end points
Progression-free survival (PFS) is not a valid surrogate for drug efficacy but many cancer drugs, which eventually show overall survival benefit, do quite often show early PFS benefit in the initial stages of a clinical trial. Furthermore, effective salvage therapy can artificially minimise the observed benefit of a drug in a trial particularly in good prognosis cancers such as breast or prostate cancer.
The critical flaw in design of trials with a PFS end point is the not the end point itself but the stringent inclusion and exclusion criteria of trial protocols. . The phase 3 cancer clinical trials do not reflect the real world population at all. Trials often test drugs in a highly selected group of patients who are much fitter and younger with less morbidity. The drugs which are licensed based on a narrow set of population are then promoted as a standard treatment to a much wider population. .
Licensing authorities should insist on the trial population to be representative of the real world population in future.
Licensing authorities should insist on real world phase 4 data collection as a requisite for continuation of license beyond an initial period of few years for the ongoing trials.
1 Naci H, Davis C, Savović J, et al. Design characteristics, risk of bias, and reporting of randomised controlled trials supporting approvals of cancer drugs by European Medicines Agency, 2014-16: cross sectional analysis. BMJ 2019;366:l5221. doi:10.1136/bmj.l5221
2 Inappropriate use of progression-free survival in cancer drug approvals | The BMJ. https://www.bmj.com/content/368/bmj.m770 (accessed 11 Mar 2020).
3 Tannock IF, de Wit R, Berry WR, et al. Docetaxel plus prednisone or mitoxantrone plus prednisone for advanced prostate cancer. N Engl J Med 2004;351:1502–12. doi:10.1056/NEJMoa040720
4 Templeton AJ, Vera-Badillo FE, Wang L, et al. Translating clinical trials to clinical practice: outcomes of men with metastatic castration resistant prostate cancer treated with docetaxel and prednisone in and out of clinical trials. Ann Oncol 2013;24:2972–7. doi:10.1093/annonc/mdt397
Competing interests: Past and Ongoing participation in pharmaceutical company sponsored clinical trials. Participation in Pharmaceutical company advisory boards and lectures. Pharmaceutical company sponsorship of conference attendances.
I would like to add my thoughts to the inappropriate use of progression-free survival in cancer drug approvals and note, in my opinion, the inappropriate use of Kaplan-Meier curves as a whole.
In a Kaplan-Meier curve, when the disease progresses or a death occurs, the probability of survival decreases, yet if someone leaves the study (cannot be followed up, excluded for other reasons such as dies due to an unrelated cause etc) the even is censored and does not decrease the probability of survival. You then have Kaplan-Meier curves which end with the number at risk of 0, meaning there is nobody left to progress or die in the study population, yet one cannot even obtain a median survival figure, because the curve did not cross the 50% line. The further the line remains above zero at the end, in my opinion, the more meaningless the data. Yet it "appears" to the untrained observer that a curve above or to the right of the comparator curve somehow implies an increase in survival,. And one wonders how many "untrained" observers there are who look at these curves, and how many people really understand what the data in front of them is telling them or not telling them.
Competing interests: No competing interests