Adjusting Simon's optimal two-stage design for heterogeneous populations based on stratification or using historical...

Area of Science:

• Oncology
• Biostatistics
• Clinical Trial Design

Background:

• Cancer patient populations are often heterogeneous, impacting clinical trial outcomes.
• Covariates like clinical, demographical, and biological factors significantly influence patient prognosis.
• Classical clinical trial designs may exhibit inflated Type I and Type II errors in heterogeneous settings.

Purpose of the Study:

• To develop and evaluate modifications of Simon's optimal two-stage design for heterogeneous cancer patient populations.
• To enhance the operating characteristics of clinical trial designs when patient response probabilities vary.
• To provide more accurate statistical power and error control in cancer studies.

Main Methods:

• Proposed two modifications to Simon's optimal two-stage design.
• Method 1: Stratification into finite subgroups with distinct response probabilities.
• Method 2: Employing logistic regression with historical controls to estimate patient-specific response probabilities.

Main Results:

• The proposed modifications aim to correct for heterogeneity in cancer trial populations.
• Simulation examples demonstrate the performance of both modified designs.
• The methods are designed to provide more reliable Type I and Type II error rates.

Conclusions:

• Modified Simon's two-stage designs offer improved accuracy for heterogeneous cancer patient populations.
• Accounting for patient covariates through subgrouping or regression modeling enhances clinical trial reliability.
• These approaches are crucial for robust cancer clinical trial design and interpretation.