Individual participant data from digital sources informed and improved precision in the evaluation of predictive biomarkers in Bayesian network meta-analysis

TL;DR

This study develops a Bayesian network meta-regression model that integrates aggregate and individual participant data from digital sources to enhance the evaluation of predictive biomarkers in cancer treatments, especially when IPD is unavailable.

Contribution

The paper introduces a novel meta-analytic approach combining aggregate data and digital-derived IPD to improve predictive biomarker assessment in oncology.

Findings

IPD integration reduced uncertainty in subgroup treatment effects by up to 49%.

The model showed no differential effect of taxane in hormone receptor subgroups.

EGFR inhibitors were more effective in KRAS wild type colorectal cancer patients.

Abstract

Objective: We aimed to develop a meta-analytic model for evaluation of predictive biomarkers and targeted therapies, utilising data from digital sources when individual participant data (IPD) from randomised controlled trials (RCTs) are unavailable. Methods: A Bayesian network meta-regression model, combining aggregate data (AD) from RCTs and IPD, was developed for modelling time-to-event data to evaluate predictive biomarkers. IPD were sourced from electronic health records, using target trial emulation approach, or digitised Kaplan-Meier curves. The model is illustrated using two examples; breast cancer with a hormone receptor biomarker, and metastatic colorectal cancer with the Kirsten Rat Sarcoma (KRAS) biomarker. Results: The model developed allowed for estimation of treatment effects in two subgroups of patients defined by their biomarker status. Effectiveness of taxane did not differ in hormone receptor positive and negative breast cancer patients. Epidermal growth factor receptor (EGFR) inhibitors were more effective than chemotherapy in KRAS wild type colorectal cancer patients but not in patients with KRAS mutant status. Use of IPD reduced uncertainty of the sub-group specific treatment effect estimates by up to 49%. Conclusion: Utilisation of IPD allowed for more detailed evaluation of predictive biomarkers and cancer therapies and improved precision of the estimates compared to use of AD alone.