Related Experiment Video
Updated: Aug 8, 2026
11:14
Creating Anatomically Accurate and Reproducible Intracranial Xenografts of Human Brain Tumors
Published on: September 24, 2014
Repeated-measures models with constrained parameters for incomplete data in tumour xenograft experiments
Ming Tan1, Hong-Bin Fang, Guo-Liang Tian
1Division of Biostatistics, University of Maryland Greenebaum Cancer Center, 22 South Greene Street, Baltimore, MD 21201, USA. mtan@umm.edu
Statistics in Medicine
|November 4, 2004
Summary
This study introduces a new statistical method for analyzing cancer xenograft experiments, addressing missing data from mouse deaths or significant tumor changes. The method improves the estimation of drug efficacy, crucial for advancing cancer treatments.
Area of Science:
- Pharmacology
- Biostatistics
- Oncology
Background:
- Xenograft models are vital in cancer drug development for assessing compound efficacy.
- Tumor volume measurements in xenografts are key outcome variables.
- Incomplete data due to mouse death, sacrifice, or drastic tumor changes complicates analysis.
Purpose of the Study:
- To develop a statistical method for analyzing xenograft data with informative censoring.
- To accurately estimate dose-response relationships in cancer drug development.
- To address challenges posed by intrinsic tumor growth and incomplete measurements.
Main Methods:
- A maximum likelihood method utilizing the expectation/conditional maximization (ECM) algorithm.
- Accounting for informative censoring and model parameter constraints.
- Application to a real xenograft study of temozolomide and irinotecan.
Main Results:
- The proposed ECM-based method provides robust estimation of dose-response relationships.
- The analysis successfully handled missing data and informative censoring in xenograft experiments.
- Accurate efficacy assessment of anti-tumor agents is facilitated.
Conclusions:
- The developed statistical approach enhances the reliability of xenograft model data analysis.
- This method is crucial for advancing promising cancer drug candidates to clinical trials.
- Improved statistical methodologies are essential for modern oncology research.
