Oops, you're using an old version of your browser so some of the features on this page may not be displaying properly.

MINIMAL Requirements: Google Chrome 24+Mozilla Firefox 20+Internet Explorer 11Opera 15–18Apple Safari 7SeaMonkey 2.15-2.23

Poster Display session 3

3612 - Validation of progression-free survival (PFS) as surrogate endpoint in randomised trials of immune checkpoint inhibitors in advanced solid cancers


30 Sep 2019


Poster Display session 3



Tumour Site


Peey Sei Kok


Annals of Oncology (2019) 30 (suppl_5): v475-v532. 10.1093/annonc/mdz253


P.S. Kok1, W. Yoon2, I. Marschner1, S.J. Lord1, M. Friedlander3, C.K. Lee4

Author affiliations

  • 1 Clinical Trials Centre, National Health and Medical Research Council Clinical Trials Centre, 2050 - Camperdown/AU
  • 2 Medical Oncology, Wollongong Hospital, Wollongong/AU
  • 3 Medical Oncology, Prince of Wales Clinical School UNSW and Prince of Wales Hospital, Sydney/AU
  • 4 Medical Oncology, St George Hospital Cancer Care Centre, 2217 - Kogarah/AU


Login to get immediate access to this content.

If you do not have an ESMO account, please create one for free.

Abstract 3612


Combination of ICIs and chemotherapy (ICI-C) has become a more favourable design in recent trials. We have previously reported 6-month PFS rate (PFS6) is a good surrogate for 12-month overall survival rate (OS12) within treatment (ICI) arms, but had limited data to examine surrogate endpoints for treatment comparisons. We extend our work to validate prior findings by updating with data from more recent randomized controlled trials (RCTs), especially studies testing combination ICI-C.


We performed a literature search to identify eligible Phase (Ph) 2 and 3 RCTs of ICI. We excluded studies with sample size of < 100 or control arms without active treatment. Efficacy data including objective response rate (ORR), PFS, OS, PFS6 and OS12 were extracted. We examined the associations of relative treatment comparisons between ORR risk ratio (RR) with hazard ratios (HRs) for PFS and OS. We also studied the correlations between ORR, PFS6 and OS12 within ICI arm. A correlation coefficient (r) of 0 indicates lack of association whereas 1 implies perfect surrogacy.


Forty-six RCTs (5 Ph2; 41 Ph3) with 51 treatment comparisons were identified. Associations between ORR RR with PFS HR, ORR RR with OS HR, and PFS HR with OS HR were r = 0.67, r = 0.42 and r = 0.52 respectively. For the 16 RCTs of ICI-C, these associations were r = 0.79, r = 0.58 and r = 0.64 respectively. For the 35 RCTs of ICI monotherapy, these associations were r = 0.71, r = 0.40 and r = 0.54. Within the ICI arms, associations between ORR and PFS6, ORR and OS12, and PFS6 and OS12 were r = 0.67, r = 0.54 and r = 0.79. Within ICI-C trials, these associations were r = 0.60, r = 0.56 and r = 0.60 respectively. Within ICI monotherapy trials, these associations were r = 0.51, r = 0.46 and r = 0.80 respectively.


In treatment comparison, the correlation between PFS and OS was modest, but stronger in ICI-C trials. Within ICI arms, PFS6 had stronger correlation with OS12 than ORR. PFS is recommended as a surrogate endpoint for ICI trials over ORR.

Clinical trial identification

Editorial acknowledgement

Legal entity responsible for the study

The authors.


Has not received any funding.


M. Friedlander: Honoraria (self), Advisory / Consultancy, Research grant / Funding (institution), Lecturer fees: AstraZeneca; Honoraria (self), Advisory / Consultancy: Lilly; Honoraria (self), Advisory / Consultancy: MSD; Honoraria (self), Advisory / Consultancy: Takeda. All other authors have declared no conflicts of interest.

This site uses cookies. Some of these cookies are essential, while others help us improve your experience by providing insights into how the site is being used.

For more detailed information on the cookies we use, please check our Privacy Policy.

Customise settings
  • Necessary cookies enable core functionality. The website cannot function properly without these cookies, and you can only disable them by changing your browser preferences.