Related Experiment Video
Updated: Apr 21, 2026
14:27
Identification of Disease-related Spatial Covariance Patterns using Neuroimaging Data
Published on: June 26, 2013
15.2K
A PRIM approach to predictive-signature development for patient stratification
Gong Chen1, Hua Zhong, Anton Belousov
1Roche Pharmaceutical Research and Early Development, Pharmaceutical Sciences, Roche Innovation Center New York, Roche TCRC, Inc., 430 East 29th Street, New York, NY 10016, U.S.A.
Statistics in Medicine
|October 28, 2014
Summary
Identifying predictive signatures helps stratify patients for clinical trials. This study proposes a patient rule induction method to find these signatures, improving treatment response prediction in oncology.
Area of Science:
- Biostatistics
- Clinical Trial Design
- Translational Oncology
Background:
- Patient response heterogeneity significantly impacts clinical trial outcomes.
- Stratifying patients based on predictive signatures can reduce trial failures.
- Predictive signatures capture biological or demographic characteristics influencing treatment response.
Purpose of the Study:
- To propose a novel procedure for searching predictive signatures using patient rule induction.
- To optimize the search for predictive signatures by discussing objective functions and resampling schemes.
- To evaluate the proposed procedure's performance through simulations and real-world oncology data.
Main Methods:
- Patient Rule Induction Method (PRIM) for predictive signature discovery.
- Development and discussion of an objective function for signature search.
- Implementation of a resampling scheme to enhance search performance.
- Application to oncology datasets focusing on survival responses.
Main Results:
- The proposed procedure effectively identifies predictive signatures for patient stratification.
- Simulations characterized conditions under which the procedure performs optimally.
- Real-world data analysis demonstrated practical utility in oncology.
- Comparison with Adaptive Index Models highlighted respective advantages.
Conclusions:
- The developed patient rule induction method offers a robust approach for identifying predictive signatures.
- This methodology can enhance clinical trial efficiency by enabling precise patient stratification.
- The findings have significant implications for personalized medicine in oncology, particularly for survival outcomes.
