Clinical Pharmacology and Pharmacy, Department of Public Health, University of Southern Denmark, Odense, Denmark.
Department of Clinical Biochemistry and Pharmacology, Odense University Hospital, Odense, Denmark.
Basic Clin Pharmacol Toxicol. 2019 Mar;124(3):240-244. doi: 10.1111/bcpt.13168. Epub 2018 Dec 12.
Overall survival (OS) is the undisputed gold standard efficacy end-point in cancer drug trials. It is with growing concern that we observe how progression-free survival (PFS) gains ground as surrogate end-point in its place. PFS has appeal because it is resource-efficient, but it has severe shortcomings. Our concern is that uncritical use of PFS will harm the evidence-based evaluation of cancer drugs when considering them for standard use in publicly financed health care systems. PFS is only valid as a surrogate end-point for OS if it correlates strongly with OS and if the cancer drug being investigated has the same effect on PFS and OS such that effects on one predict effects on the other. The latter might be less obvious than the former but is no less critical. Research indicates that in a majority of cases, correlation between surrogate end-points and OS is of medium strength or lower. PFS is therefore unreliable as a surrogate for OS. We do not find it justified to use PFS as surrogate for OS without first having assessed its validity. Stakeholders who take part in evaluating cancer drugs considered for standard use in a health care system must be particularly vigilant about this issue to minimize the risk of introducing cancer drugs that have an unacceptable cost-risk-benefit profile.
总生存期（OS）是癌症药物试验中无可争议的黄金标准疗效终点。我们越来越关注到无进展生存期（PFS）作为替代终点的地位正在逐渐上升。PFS 具有吸引力，因为它节省资源，但它也有严重的缺陷。我们担心，在考虑将癌症药物用于公共资助的医疗保健系统的标准治疗时，如果不加批判地使用 PFS，将会损害基于证据的癌症药物评估。只有当 PFS 与 OS 高度相关，并且正在研究的癌症药物对 PFS 和 OS 具有相同的影响，即对一个的影响可以预测对另一个的影响时，PFS 才可以作为 OS 的替代终点。后者可能不如前者明显，但同样至关重要。研究表明，在大多数情况下，替代终点与 OS 之间的相关性为中等强度或更低。因此，PFS 作为 OS 的替代物是不可靠的。如果不首先评估其有效性，我们认为将 PFS 用作 OS 的替代物是不合理的。参与评估拟在医疗保健系统中作为标准治疗使用的癌症药物的利益相关者必须特别关注这个问题，以最大程度地降低引入具有不可接受的成本-效益风险特征的癌症药物的风险。
