Project 2: Statistical Methods for Biomarkers and Patient Reported Outcomes in Cancer Trials PROJECT SUMMARY The primary goal of this project is to develop and evaluate new statistical methodology for the identification of patient reported outcomes (PROs) and biomarkers in cancer research. We will develop several novel statis- tical methods for establishing go/no-go decisions and for the evaluation of PROs (such as quality of life) and biomarkers in cancer clinical trials. We further propose methods in meta-analysis for assessing and identifying important biomarkers. The four proposed aims present a unified approach for the identification of biomarkers and PROs useful in cancer research.
The aims are closely tied together, having the common goal of trying to identify important biomarkers and PROs, such as quality of life, which may be highly associated with biomarkers in cancer clinical trials.
The first aim develops new statistical methods for assessing the probability of success (POS) in cancer clinical trials. For example, these new methods will use biomarkers and covariate information in determining the benefit of proceeding from a phase II to a phase III clinical trial.
Aims 2 and 3 focus on the identification of both PROs and biomarkers when trying to associate a time-to-event endpoint to a biomarker or PRO.
In Aim 2, we develop new statistical methods for assessing model fit in the joint modeling of longitudinal and survival data, with the goal of identifying the biomarkers and PROs that are most highly associated with survival outcomes. This development will have a significant clinical impact on how patients are monitored and treated, with an ultimate goal of curing cancer.
In Aim 3, we broaden our research scope by proposing a flexible class of new statistical models. This class allows for the inclusion of multi-dimensional longitudinal measures with missing data and survival outcomes of different types and different observed patterns, such as disease progression and overall survival, which are known as semi-competing risks survival times. The models devel- oped in this aim will allow us to develop new statistical methods for more accurately assessing and identifying a set of important biomarkers and/or PROs in this more general semi-competing risks setting.
Aim 4 tackles important statistical problems in network meta-analysis (NMA), in which the goal is to compare and assess treatments using meta-regression models that may not have been formally compared head to head in random- ized trials. Again, the goal of these methods is to use information on PROs and/or biomarkers as covariates in a NMA model to compare treatments and assess the relevance of biomarkers in cancer clinical trials. These four aims are closely linked with the common goal of assessing and identifying important biomarkers and/or PROs in cancer clinical trials, which can shed light on the treatment and improve our understanding of the disease. Moreover, these four aims are clinically significant and have the potential for considerable clinical impact.

Public Health Relevance

Project 2: Statistical Methods for Biomarkers and Patient Reported Outcomes in Cancer Trials PROJECT NARRATIVE The primary goal of this project is to develop and evaluate new statistical methodology for the identification of patient reported outcomes (PROs), such as quality of life, in cancer clinical trials. We will also develop several novel statistical methods for establishing formal methods of decision making in cancer clinical trials. These goals address extremely important issues in cancer research because they will allow us to assess simultaneously the efficacy of a treatment, the quality of life, and the trade-offs between the two. In addition, these methods will help us to assess the costs, benefits, and risks associated with conducting large comparative clinical trials using the efficacy and safety results of much smaller trials.

Agency
National Institute of Health (NIH)
Institute
National Cancer Institute (NCI)
Type
Research Program Projects (P01)
Project #
2P01CA142538-06
Application #
8794727
Study Section
Special Emphasis Panel (ZCA1-RPRB-B (O1))
Project Start
2010-04-01
Project End
2020-03-31
Budget Start
2015-05-15
Budget End
2016-03-31
Support Year
6
Fiscal Year
2015
Total Cost
$232,039
Indirect Cost
$63,421
Name
University of North Carolina Chapel Hill
Department
Type
DUNS #
608195277
City
Chapel Hill
State
NC
Country
United States
Zip Code
27599
Jung, Sin-Ho; Lee, Ho Yun; Chow, Shein-Chung (2018) Statistical Methods for Conditional Survival Analysis. J Biopharm Stat 28:927-938
Jiang, Yuchao; Wang, Rujin; Urrutia, Eugene et al. (2018) CODEX2: full-spectrum copy number variation detection by high-throughput DNA sequencing. Genome Biol 19:202
Kim, Soyoung; Zeng, Donglin; Cai, Jianwen (2018) Analysis of multiple survival events in generalized case-cohort designs. Biometrics :
Chung, Yunro; Ivanova, Anastasia; Hudgens, Michael G et al. (2018) Partial likelihood estimation of isotonic proportional hazards models. Biometrika 105:133-148
Liang, Shuhan; Lu, Wenbin; Song, Rui (2018) Deep advantage learning for optimal dynamic treatment regime. Stat Theory Relat Fields 2:80-88
Chen, Xiaolin; Cai, Jianwen (2018) Reweighted estimators for additive hazard model with censoring indicators missing at random. Lifetime Data Anal 24:224-249
Yang, Yuchen; Huh, Ruth; Culpepper, Houston W et al. (2018) SAFE-clustering: Single-cell Aggregated (From Ensemble) Clustering for Single-cell RNA-seq Data. Bioinformatics :
Nasution, Marlina D; Wang, Xiaofei (2018) Statistical issues and advances in cancer precision medicine research. J Biopharm Stat 28:215-216
Wu, Yuan; Chambers, Christina D; Xu, Ronghui (2018) Semiparametric sieve maximum likelihood estimation under cure model with partly interval censored and left truncated data for application to spontaneous abortion. Lifetime Data Anal :
Ibrahim, Joseph G; Kim, Sungduk; Chen, Ming-Hui et al. (2018) Bayesian multivariate skew meta-regression models for individual patient data. Stat Methods Med Res :962280218801147

Showing the most recent 10 out of 549 publications