Subgroup Testing in the Change-Plane Cox Model

Area of Science:

• Biostatistics
• Epidemiology
• Oncology

Background:

• Survival outcomes are critical in biomedical and epidemiological research.
• Treatment effects can differ across patient subgroups, influenced by covariates like tumor mutational burden.
• Change-plane Cox models identify subgroups with differential treatment effects in survival data.

Purpose of the Study:

• To introduce a novel likelihood ratio test for change-plane Cox models.
• To enhance the power of detecting treatment effect modifications in survival analysis, particularly in small samples.
• To provide a more robust statistical method for subgroup analysis in clinical studies.

Main Methods:

• Development of a new test statistic based on the likelihood ratio test.
• Establishment of asymptotic distributions for the test statistic under null and local alternative hypotheses.
• Extensive simulation studies to evaluate finite sample performance.
• Application to real-world nonsmall cell lung cancer data.

Main Results:

• The proposed likelihood ratio test demonstrates enhanced power compared to existing score test methods.
• Asymptotic properties of the test statistic are theoretically established.
• Simulation studies confirm the method's effectiveness and reliability in various scenarios.
• The test successfully identified relevant patterns in nonsmall cell lung cancer survival data.

Conclusions:

• The novel likelihood ratio test offers a powerful and practical approach for change-plane Cox models.
• This method improves the identification of subgroups with varied treatment effects in survival data.
• The approach has significant utility in analyzing complex clinical trial and epidemiological data, including cancer research.