Matching methods in precision oncology: An introduction and illustrative example.

Area of Science:

• Oncology
• Genomics
• Biostatistics

Background:

• Randomized controlled trials (RCTs) are rare in precision oncology.
• Matching methods offer an alternative for evaluating precision oncology's effectiveness.
• The British Columbia Personalized OncoGenomics (POG) program uses whole-genome and transcriptome analysis (WGTA) for advanced cancer care.

Purpose of the Study:

• To introduce and illustrate matching methods for evaluating precision oncology in the absence of RCTs.
• To compare propensity score matching (PSM) and genetic matching for cohort creation.
• To assess the impact of WGTA-informed treatment on patient survival.

Main Methods:

• A cohort of 230 POG patients (2014-2015) and matched POG-naive controls were analyzed.
• 1:1 propensity score matching (PSM) and genetic matching were used to create comparable cohorts.
• Survival differences were explored between POG and POG-naive patients, stratified by WGTA-informed treatment changes.

Main Results:

• Genetic matching demonstrated superior covariate balancing compared to PSM.
• Overall survival did not significantly differ between POG and POG-naive patients (p > 0.05).
• Patients receiving WGTA-informed treatment changes showed significantly reduced hazard of death (HR: 0.33-0.41) compared to POG-naive controls.

Conclusions:

• The clinical effectiveness of precision oncology hinges on the rate of genomics-informed treatment modifications.
• This study provides a framework for evaluating precision oncology when RCT data are unavailable.
• Matching methods can support reliable effect estimation in precision oncology research.