AI-driven predictive biomarker discovery with contrastive learning to improve clinical trial outcomes.

Journal: Cancer cell
Published Date:

Abstract

Modern clinical trials can capture tens of thousands of clinicogenomic measurements per individual. Discovering predictive biomarkers, as opposed to prognostic markers, remains challenging. To address this, we present a neural network framework based on contrastive learning-the Predictive Biomarker Modeling Framework (PBMF)-that explores potential predictive biomarkers in an automated, systematic, and unbiased manner. Applied retrospectively to real clinicogenomic datasets, particularly for immuno-oncology (IO) trials, our algorithm identifies biomarkers of IO-treated individuals who survive longer than those treated with other therapies. We demonstrate how our framework retrospectively contributes to a phase 3 clinical trial by uncovering a predictive, interpretable biomarker based solely on early study data. Patients identified with this predictive biomarker show a 15% improvement in survival risk compared to those in the original trial. The PBMF offers a general-purpose, rapid, and robust approach to inform biomarker strategy, providing actionable outcomes for clinical decision-making.

Authors

  • Gustavo Arango-Argoty
    Oncology Data Science, Oncology R&D, AstraZeneca, Waltham, MA, USA. gustavo.arango@astrazeneca.com.
  • Damian E Bikiel
    Oncology Data Science, Oncology R&D, AstraZeneca, Waltham, MA, USA.
  • Gerald J Sun
    Oncology Data Science, Oncology R&D, AstraZeneca, Waltham, MA, USA.
  • Elly Kipkogei
    Oncology Data Science, Oncology R&D, AstraZeneca, Waltham, MA, USA.
  • Kaitlin M Smith
    Oncology Data Science, Oncology R&D, AstraZeneca, Waltham, MA, USA.
  • Sebastian Carrasco Pro
    Life Sciences, Tempus AI, Boston, MA, USA.
  • Elizabeth Y Choe
    Oncology Data Science, Oncology R&D, AstraZeneca, Waltham, MA, USA.
  • Etai Jacob
    Oncology Data Science, Oncology R&D, AstraZeneca, Waltham, MA, USA.