Tumor digital twins uncover therapeutic strategies in triple-negative breast cancer.
TL;DR
A TNBC digital twin framework, built from prior knowledge and formulated as a mechanistic qualitative network in which nodes represent genes and proteins and edges denote the interactions between them, offering an interpretable alternative to black-box predictive models, enabling transparent decision-making in precision oncology.
Abstract
Triple-negative breast cancer (TNBC) is an aggressive malignancy with limited therapeutic options and a high risk of relapse and mortality. Consequently, there is a critical need for personalized treatment strategies that reflect the unique biology, clonal architecture, and evolutionary dynamics of individual tumors. Here, we present a TNBC digital twin framework, built from prior knowledge and formulated as a mechanistic qualitative network in which nodes represent genes and proteins and edges denote the interactions between them. Parameterized with patient-specific molecular data, the model identifies potential resistance mechanisms and proposes tailored treatment plans to overcome them. We validate the model using clinical data from PARTNER, a large, randomized controlled trial evaluating platinum-based chemotherapy with or without PARP inhibitors. The model accurately predicts treatment responses and identifies mutational and transcriptional biomarkers associated with clinical outcome. Our approach offers an interpretable alternative to black-box predictive models, enabling transparent decision-making in precision oncology.