Which test for crossing survival curves? A user's guideline.
Crossing
Log-rank test
Non-proportional hazards
Oncology
Restricted-mean survival
Survival analysis
Time-to-event outcome
Journal
BMC medical research methodology
ISSN: 1471-2288
Titre abrégé: BMC Med Res Methodol
Pays: England
ID NLM: 100968545
Informations de publication
Date de publication:
30 01 2022
30 01 2022
Historique:
received:
24
06
2021
accepted:
18
01
2022
entrez:
31
1
2022
pubmed:
1
2
2022
medline:
22
3
2022
Statut:
epublish
Résumé
The exchange of knowledge between statisticians developing new methodology and clinicians, reviewers or authors applying them is fundamental. This is specifically true for clinical trials with time-to-event endpoints. Thereby, one of the most commonly arising questions is that of equal survival distributions in two-armed trial. The log-rank test is still the gold-standard to infer this question. However, in case of non-proportional hazards, its power can become poor and multiple extensions have been developed to overcome this issue. We aim to facilitate the choice of a test for the detection of survival differences in the case of crossing hazards. We restricted the review to the most recent two-armed clinical oncology trials with crossing survival curves. Each data set was reconstructed using a state-of-the-art reconstruction algorithm. To ensure reproduction quality, only publications with published number at risk at multiple time points, sufficient printing quality and a non-informative censoring pattern were included. This article depicts the p-values of the log-rank and Peto-Peto test as references and compares them with nine different tests developed for detection of survival differences in the presence of non-proportional or crossing hazards. We reviewed 1400 recent phase III clinical oncology trials and selected fifteen studies that met our eligibility criteria for data reconstruction. After including further three individual patient data sets, for nine out of eighteen studies significant differences in survival were found using the investigated tests. An important point that reviewers should pay attention to is that 28% of the studies with published survival curves did not report the number at risk. This makes reconstruction and plausibility checks almost impossible. The evaluation shows that inference methods constructed to detect differences in survival in presence of non-proportional hazards are beneficial and help to provide guidance in choosing a sensible alternative to the standard log-rank test.
Sections du résumé
BACKGROUND
The exchange of knowledge between statisticians developing new methodology and clinicians, reviewers or authors applying them is fundamental. This is specifically true for clinical trials with time-to-event endpoints. Thereby, one of the most commonly arising questions is that of equal survival distributions in two-armed trial. The log-rank test is still the gold-standard to infer this question. However, in case of non-proportional hazards, its power can become poor and multiple extensions have been developed to overcome this issue. We aim to facilitate the choice of a test for the detection of survival differences in the case of crossing hazards.
METHODS
We restricted the review to the most recent two-armed clinical oncology trials with crossing survival curves. Each data set was reconstructed using a state-of-the-art reconstruction algorithm. To ensure reproduction quality, only publications with published number at risk at multiple time points, sufficient printing quality and a non-informative censoring pattern were included. This article depicts the p-values of the log-rank and Peto-Peto test as references and compares them with nine different tests developed for detection of survival differences in the presence of non-proportional or crossing hazards.
RESULTS
We reviewed 1400 recent phase III clinical oncology trials and selected fifteen studies that met our eligibility criteria for data reconstruction. After including further three individual patient data sets, for nine out of eighteen studies significant differences in survival were found using the investigated tests. An important point that reviewers should pay attention to is that 28% of the studies with published survival curves did not report the number at risk. This makes reconstruction and plausibility checks almost impossible.
CONCLUSIONS
The evaluation shows that inference methods constructed to detect differences in survival in presence of non-proportional hazards are beneficial and help to provide guidance in choosing a sensible alternative to the standard log-rank test.
Identifiants
pubmed: 35094686
doi: 10.1186/s12874-022-01520-0
pii: 10.1186/s12874-022-01520-0
pmc: PMC8802494
doi:
Types de publication
Journal Article
Research Support, Non-U.S. Gov't
Review
Langues
eng
Sous-ensembles de citation
IM
Pagination
34Informations de copyright
© 2022. The Author(s).
Références
N Engl J Med. 2015 Nov 5;373(19):1803-13
pubmed: 26406148
BMC Med Res Methodol. 2012 Feb 01;12:9
pubmed: 22297116
Biometrics. 2000 Dec;56(4):971-83
pubmed: 11129494
Crit Rev Oncol Hematol. 2021 Jun;162:103350
pubmed: 33989767
J Clin Oncol. 2018 Dec 1;36(34):3361-3369
pubmed: 30153096
BMC Cancer. 2019 Dec 27;19(1):1253
pubmed: 31881856
Ann Hematol. 2020 Jul;99(7):1551-1560
pubmed: 32504186
Psychol Methods. 2013 Dec;18(4):572-82
pubmed: 24079924
Ann Oncol. 2019 Jun 1;30(6):970-976
pubmed: 31050707
N Engl J Med. 2019 Jul 25;381(4):317-327
pubmed: 31157963
Biometrics. 2010 Mar;66(1):30-8
pubmed: 19397582
Ann Surg Oncol. 2018 Aug;25(8):2441-2448
pubmed: 29948420
Ann Oncol. 2019 Dec 1;30(12):1959-1968
pubmed: 31562758
N Engl J Med. 2016 Nov 10;375(19):1856-1867
pubmed: 27718784
Biometrics. 2020 Dec;76(4):1157-1166
pubmed: 32061098
Stat Methods Med Res. 2020 Oct;29(10):2830-2850
pubmed: 32183613
Biometrics. 2016 Mar;72(1):39-45
pubmed: 26393315
Cancer Sci. 2020 Feb;111(2):513-527
pubmed: 31789476
Leukemia. 2020 Jan;34(1):296-300
pubmed: 31427720
BMC Med Res Methodol. 2016 Feb 11;16:16
pubmed: 26869168
Int J Clin Oncol. 2019 Mar;24(3):274-287
pubmed: 30515674
Lancet Oncol. 2020 Mar;21(3):345-357
pubmed: 32035020
Ann Oncol. 2018 Sep 1;29(9):1995-2002
pubmed: 30084934
J Clin Oncol. 2016 May 20;34(15):1813-9
pubmed: 26884584
JAMA Cardiol. 2017 Nov 1;2(11):1179-1180
pubmed: 28877311
Stata J. 2017 Apr;17(2):405-421
pubmed: 29445320
J Thorac Oncol. 2018 Nov;13(11):1743-1749
pubmed: 30017831
Leukemia. 2019 Feb;33(2):379-389
pubmed: 30555165
PLoS One. 2015 Jan 23;10(1):e0116774
pubmed: 25615624
Biometrics. 2018 Jun;74(2):694-702
pubmed: 28901017
Pharm Stat. 2020 Jul;19(4):399-409
pubmed: 31916378
Trials. 2019 Mar 18;20(1):172
pubmed: 30885277
N Engl J Med. 2017 Mar 16;376(11):1015-1026
pubmed: 28212060
J Clin Oncol. 2020 Feb 20;38(6):593-601
pubmed: 31829912
J Clin Oncol. 2014 Aug 1;32(22):2380-5
pubmed: 24982461
Trials. 2020 Apr 6;21(1):315
pubmed: 32252820
Biometrics. 2021 Oct 5;:
pubmed: 34608996