Related Experiment Videos
Analysis of multiple tumor data from a rodent carcinogenicity experiment
D M Finkelstein1, D A Schoenfeld
1Department of Biostatistics, Harvard School of Public Health, Massachusetts General Hospital, Boston.
Biometrics
|March 1, 1989
Summary
This study introduces a new statistical test for rodent carcinogenicity studies. The test counts tumor-bearing sites to accurately assess substance safety without being affected by mortality rates.
Area of Science:
- Toxicology and Carcinogenesis
- Biostatistics
- Preclinical Safety Assessment
Background:
- Rodent tumorigenicity studies are crucial for evaluating substance safety for human exposure.
- Current statistical methods for analyzing tumor development across multiple body sites can lead to an inflated false positive rate due to test multiplicity.
- There is a need for robust statistical methods to simultaneously test for treatment effects across multiple tumor sites.
Purpose of the Study:
- To propose a novel statistical test for analyzing rodent carcinogenicity data.
- To address the issue of inflated false positive rates in multi-site tumor analysis.
- To develop a test that is independent of tumor lethality and mortality differences.
Main Methods:
- A new statistical test based on the count of tumor-bearing sites is proposed.
- The performance of the proposed test is evaluated through simulations.
- Comparisons are made against existing methods, including Bonferroni adjustments.
Main Results:
- The proposed test demonstrates reliable performance in simulations.
- The test is shown to be appropriate regardless of tumor lethality or mortality.
- The test is illustrated using data from the large ED01 experiment.
Conclusions:
- The proposed site-count test offers a valid approach for analyzing rodent carcinogenicity data.
- This method helps mitigate the problem of multiple testing in carcinogenicity studies.
- The test provides a reliable tool for substance safety evaluation in preclinical research.