Small-sample inference for incomplete longitudinal data with truncation and censoring in tumor xenograft models

Area of Science:

• Oncology
• Biostatistics
• Pharmacology

Background:

• Xenograft models are crucial for evaluating anticancer drug efficacy in preclinical cancer research.
• Tumor volume measurements in xenograft studies often result in incomplete data due to animal mortality, tumor regrowth, or specific censoring criteria.
• Small sample sizes in these models limit the applicability of traditional statistical inference methods.

Purpose of the Study:

• To develop and present novel statistical methods for analyzing incomplete tumor volume data in cancer xenograft models.
• To provide robust parametric test procedures that account for informative censoring and small sample sizes.
• To compare treatment effects accurately in preclinical cancer studies with complex data structures.

Main Methods:

• Proposed two parametric test procedures utilizing the Expectation-Maximization (EM) algorithm.
• Employed Bayesian methods to handle informative censoring and incomplete repeated measurements.
• Applied these methods to analyze a real xenograft study involving temozolomide and irinotecan.

Main Results:

• The developed methods effectively address challenges posed by missing data in xenograft studies.
• Parametric procedures provide reliable comparisons of treatment effects even with informative censoring.
• Demonstrated the utility of the methods in a practical preclinical cancer drug study.

Conclusions:

• The proposed EM algorithm and Bayesian methods offer a statistically sound approach for analyzing xenograft data with missing observations.
• These methods enhance the reliability of preclinical cancer drug efficacy assessment.
• Accurate analysis of xenograft studies is vital for advancing cancer drug development.