Saddle-point p-values and confidence intervals based on log-rank tests when dependent subunits of clustered survival data are randomized by random allocation design

2021 ◽  
pp. 1-11
Author(s):  
Haidy A. Newer
Keyword(s):  
Saddle Point ◽  
Confidence Intervals ◽  
Survival Data ◽  
Random Allocation ◽  
Rank Tests ◽  
P Values ◽  
Clustered Survival Data ◽  
Log Rank Tests

Weighted log-rank tests for clustered censored data: Saddlepoint p-values and confidence intervals

2020 ◽  
Vol 29 (9) ◽  
pp. 2629-2636 ◽  
Author(s):  
Abd El-Raheem M Abd El-Raheem ◽  
Ehab F Abd-Elfattah
Keyword(s):  
Confidence Intervals ◽  
Clustered Data ◽  
Saddlepoint Approximation ◽  
Data Sets ◽  
Rank Tests ◽  
P Values ◽  
Exact Confidence Intervals ◽  
The Common ◽  
Asymptotic Normal ◽  
Log Rank Tests

Clustered data with censored failure times frequently arise in clinical trials and tumorigenicity studies. For such data, the common and extensively used class of two-sample tests is the weighted log-rank tests. In this article, a double saddlepoint approximation is used to calculate the p-values of the null permutation distribution of these tests. This technique is demonstrated using three real clustered data sets. Comprehensive simulation studies are conducted to appraise the efficiency of the saddlepoint approximation. This approximation is far superior to the asymptotic normal approximation. This precision allows us to determine almost exact confidence intervals for the treatment impact.

scholarly journals Nonparametric generalized fiducial inference for survival functions under censoring

Biometrika ◽  
2019 ◽  
Vol 106 (3) ◽  
pp. 501-518 ◽  
Author(s):  
Y Cui ◽  
J Hannig
Keyword(s):  
Confidence Intervals ◽  
Average Length ◽  
Asymptotic Methods ◽  
Survival Function ◽  
Right Censoring ◽  
Rank Tests ◽  
Fiducial Inference ◽  
Fiducial Distribution ◽  
Von Mises ◽  
Log Rank Tests

Summary Since the introduction of fiducial inference by Fisher in the 1930s, its application has been largely confined to relatively simple, parametric problems. In this paper, we present what might be the first time fiducial inference is systematically applied to estimation of a nonparametric survival function under right censoring. We find that the resulting fiducial distribution gives rise to surprisingly good statistical procedures applicable to both one-sample and two-sample problems. In particular, we use the fiducial distribution of a survival function to construct pointwise and curvewise confidence intervals for the survival function, and propose tests based on the curvewise confidence interval. We establish a functional Bernstein–von Mises theorem, and perform thorough simulation studies in scenarios with different levels of censoring. The proposed fiducial-based confidence intervals maintain coverage in situations where asymptotic methods often have substantial coverage problems. Furthermore, the average length of the proposed confidence intervals is often shorter than the length of confidence intervals for competing methods that maintain coverage. Finally, the proposed fiducial test is more powerful than various types of log-rank tests and sup log-rank tests in some scenarios. We illustrate the proposed fiducial test by comparing chemotherapy against chemotherapy combined with radiotherapy, using data from the treatment of locally unresectable gastric cancer.

Rank tests for clustered survival data when dependent subunits are randomized

2005 ◽  
Vol 25 (3) ◽  
pp. 361-373 ◽  
Author(s):  
Jong-Hyeon Jeong ◽  
Sin-Ho Jung
Keyword(s):  
Survival Data ◽  
Rank Tests ◽  
Clustered Survival Data

Six vs. Eight Cycles of Bi-Weekly CHOP-14 with or without Rituximab for Elderly Patients with Diffuse Large B-Cell Lymphoma (DLBCL): Results of the Completed RICOVER-60 Trial of the German High-Grade Non-Hodgkin Lymphoma Study Group (DSHNHL).

Blood ◽  
2006 ◽  
Vol 108 (11) ◽  
pp. 205-205 ◽  
Author(s):  
Michael Pfreundschuh ◽  
Marita Kloess ◽  
Samira Zeynalova ◽  
Eva Lengfelder ◽  
Astrid Franke ◽  
...  
Keyword(s):  
Multivariate Analysis ◽  
Elderly Patients ◽  
B Cell Lymphoma ◽  
Observation Time ◽  
Rank Tests ◽  
Bulky Disease ◽  
P Values ◽  
Intention To Treat ◽  
Non Hodgkin Lymphoma ◽  
Log Rank Tests

Abstract Interval reduction from 3 (CHOP-21) to 2 weeks (CHOP-14; Pfreundschuh et al., Blood, 2004) and the addition of rituximab to CHOP-21 (R-CHOP-21; Coiffier et al., NEJM, 2002) improved outcome in elderly patients with DLBCL to a similar extent compared to CHOP-21. In the RICOVER-60 trial, elderly patients (61–80 years) were randomized to receive 6 or 8 cycles of CHOP-14 with or without rituximab given on days 1, 15, 29, 43, 57, 71, 85, and 99. Radiotherapy was planned to sites of initial bulk and/or extranodal involvement. Between 07/2000 and 06/2005, 1222 patients with CD20+ DLBCL and informed consent were recruited and evaluable (median age 68 years; IPI=1: 30%, IPI=2: 28%; IPI=3: 26%; IPI=4,5: 16%). The primary endpoint was event-free survival (EFS) with events defined as additional therapy, failure to achieve complete remission, progressive disease, relapse, or death. As by intention to treat, the 3-year EFS rate was 47% after 6×CHOP-14 (n=307), 53% after 8×CHOP-14 (n=305), 66% after 6×R-CHOP-14 (n=306), and 63% after 8×R-CHOP-14 (n=304). Regarding time-to-event data, we found a relevant interaction term (RR 1.4, p=0.068) between treatment contrasts and performd according to the protocol single-arm comparisons using 6×CHOP-14 without rituximab as the reference arm instead of a 2×2 factorial analysis. The p-values for the univariate log-rank tests of the improvement in EFS over 6×CHOP-14 were p=0.036 for 8×CHOP-14, and p<0.001 for 6×R-CHOP-14 and 8×R-CHOP-14, respectively. After a median observation time of 34.5 months, the estimated 3-year overall survival (OS) rates were 68% for 6×CHOP-14, 66% for 8×CHOP-14, 78% for 6×R-CHOP-14, and 72% for 8×R-CHOP-14. The p-values for the univariate log-rank tests of the OS improvement over 6×CHOP-14 were p=0.836 for 8×CHOP-14, p=0.018 for 6×R-CHOP-14 and p=0.260 for 8×R-CHOP-14. In a multivariate analysis using 6×CHOP-14 without rituximab as the reference and adjusting for the stratification variables (elevated LDH, advanced stage III&IV, ECOG performance state >1, bulky disease, >1 extranodal site, and age >70), both rituximab arms had a significantly improved EFS (6×R-CHOP-14: relative risk [RR]=0.51, p<0.001; 8×R-CHOP-14: RR 0.54, p<0.001; 8×CHOP14: RR 0.76, p=0.017). However, in the multivariate analysis for OS adjusting for the stratification variables, only the OS after 6×R-CHOP-14 (RR 0.63; p=0.003), but not after 8×R-CHOP-14 (RR=0.78; p=0.102) was significantly better than the OS after 6×CHOP-14. In summary, only 6 cycles of R-CHOP-14 significantly improved both EFS and OS over 6×CHOP-14, while 8×R-CHOP-14 did so only with respect to EFS. The reasons for the less favorable OS after 8×R-CHOP-14 compared to 6×R-CHOP-14 will be analyzed and discussed. The results after 6 cycles of R-CHOP-14 in this largest randomized trial of DLBLC performed to date are the best reported for elderly patients with CD20+ DLBCL. 6×R-CHOP-14 should be considered as reference standard in future trials for elderly patients with DLBCL.

Combining stratified and unstratified log-rank tests in paired survival data

2010 ◽  
Vol 29 (16) ◽  
pp. 1735-1745 ◽  
Author(s):  
David Oakes ◽  
Changyong Feng
Keyword(s):  
Survival Data ◽  
Rank Tests ◽  
Log Rank Tests

scholarly journals Outcomes of Patients with Multiple Myeloma Harboring Gain/Amplification 1q in the Era of Modern Therapy

Blood ◽  
2020 ◽  
Vol 136 (Supplement 1) ◽  
pp. 45-46
Author(s):  
Xiao Hu ◽  
Cherng-Horng Wu ◽  
Janet Cowan ◽  
Raymond L. Comenzo ◽  
Cindy Varga
Keyword(s):  
Multiple Myeloma ◽  
Survival Data ◽  
Research Funding ◽  
The United States ◽  
Cell Transplant ◽  
Newly Diagnosed ◽  
Rank Tests ◽  
Treatment Regimens ◽  
The Impact ◽  
Log Rank Tests

Background: Multiple myeloma (MM) is the second most common hematological malignancy in the United States. Fluorescence in-situ hybridization (FISH) is a popular tool to detect cytogenetic alterations which in turn, can contribute to the risk stratification of patients with MM. Gain or amplification of CKS1B gene at chromosome 1q21 region (gain 1q) is detected in 35-40% of newly diagnosed MM cases, and has been reported to be associated with inferior prognostic outcomes. It also frequently occurs with the deletion of CDKN2C at chromosome 1p32.3 (del 1p). There is a distinct lack of data on patients harboring this cytogenetic alteration in the era of novel agents. We sought to look at outcomes of this patient population at a single institution over the last 5 years. Methods: This retrospective study included all patients with MM as defined by the International Myeloma Working Group (IMWG) with available FISH studies identifying gain 1q between 01/01/2015 to 04/30/2020 at Tufts Medical Center. The study was approved by the Institutional Review Board. Baseline demography, disease characteristics, and treatment history were extracted from the electronic medical records. With death as primary event, overall survival (OS) was defined as the survival time from the discovery of gain 1q to death. Progression free survival (PFS) was defined as the time from discovery of gain 1q to first progression/relapse or death, whichever occurred first. Kaplan-Meier method was used to estimate survival data. Differences in survival between two groups were analyzed by log-rank tests. Multivariable cox regression adjusting for baseline characteristics and significant concurrent cytogenetic alterations were performed to explore the impact of treatment regimens on survival. Results: Of the forty-nine subjects included in this study, the age range was 39 to 85 years; 31 patients (63.3%) were over the age of 65 years, and 28 (57.1%) were male. Twenty-eight (57.1%) subjects with gain 1q were newly diagnosed while the remaining 21 (42.9%) were identified at relapse. Gain 1q was present in more than 20% of clonal cells in 73.5% of subjects and 29.6% had del 1p as well. Patients with gain 1q were more likely to have deletion 13q (65.3%) and hyperdiploidy (61.2%). Regarding treatment, 75.7% of patients received bortezomib, 70.3% received lenalidomide, 38.9% underwent autologous stem cell transplant (ASCT) and 64.9% received daratumumab. At the time of analysis, 41 patients were still alive. For the entire cohort, the estimated median OS was not reached (NR) (95% confidence interval [CI], 24.1-NR), and the estimated median PFS was 15.27 months (95% CI, 4.77-NR). In log-rank tests, presence of extra medullary disease was associated with shorter PFS (4.8 vs 24.1 months, P=0.003), while IGH abnormalities including complex IGH rearrangements or losses were associated with longer PFS (NR vs 8.2 months, P=0.046). Lenalidomide-based treatment was associated with prolonged OS (NR vs 17.2 months, P=0.048). Bortezomib-based therapy and upfront ASCT were associated with improved PFS (15.3 vs 4.7 months, P=0.036; NR vs 4.8 months, P =0.019 respectively). Further multivariate analyses adjusting for age, number of CKS1B copies, International Staging System stage, baseline creatinine, clone size, del 1p, lactate dehydrogenase, extra medullary disease, and IGH abnormalities revealed that administration of daratumumab after the discovery of gain 1q was associated with superior OS (Hazard Ratio [HR]=0.023x10^-2, 95% CI [0.002x10^-4, 0.299], P =0.022) compared with those not receiving this agent; both the use of bortezomib (HR=0.210, 95% CI [0.064, 0.687], P =0.010) and daratumumab (HR=0.126, 95% CI [0.015, 1.036], P =0.054) were associated with prolonged PFS. The use of lenalidomide or upfront ASCT lost prognostic benefit after adjusting for additional variables in multivariate models. Conclusions: The outcomes of MM patients with gain 1q were evaluated according to clinical characteristics, concurrent chromosomal alterations and treatment regimens. In our small cohort, daratumumab and daratumumab-bortezomib combination regimens were found to have a favorable impact on survival. Future prospective clinical trials with larger sample sizes are warranted to confirm the results and further improve the outcomes of MM patients with this cytogenetic alteration. Disclosures Comenzo: Amgen: Consultancy; Takeda: Consultancy, Research Funding; Sanofi: Consultancy; Unum: Consultancy; Caleum: Consultancy; Prothena: Consultancy, Research Funding; Janssen: Consultancy, Research Funding; Karyopharm: Consultancy, Research Funding.

A Class of Weighted Log-Rank Tests for Survival Data When the Event is Rare

2000 ◽  
Vol 95 (449) ◽  
pp. 249-258 ◽  
Author(s):  
Steven Buyske ◽  
Richard Fagerstrom ◽  
Zhiliang Ying
Keyword(s):  
Survival Data ◽  
Rank Tests ◽  
Log Rank Tests
Sign in / Sign up

Export Citation Format

Share Document