Digital Twins Predict Treatment Response in Triple-Negative Breast Cancer

Triple-negative breast cancer (TNBC) is an aggressive, fast-growing subtype that accounts for about 15% of all breast cancer cases. Its tumor cells lack estrogen receptors, progesterone receptors, and the HER2 protein, limiting the use of hormone therapies and HER2-targeted drugs. With these treatment limitations in mind, researchers have developed a computer model based on symbolic artificial intelligence to help identify additional options when standard therapies fail.

Developed by a University College London team in collaboration with the University of Cambridge, the model uses an individual patient’s detailed genetic fingerprint to recommend potential treatment options for TNBC. It relies on symbolic AI, which uses logical reasoning and allows clinicians to see how recommendations are generated, providing transparency that “black box” artificial intelligence models lack.

To evaluate the approach, the researchers created computer-generated tumors, or digital twins, based on tumors from 31 patients in the PARTNER clinical trial. The trial’s full name is Platinum and PolyAdenosine 5'-Diphosphoribose Polymerisation Inhibitor for Research and Treatment of Neo-Adjuvant Early and Relapse-Risk Breast Cancer. Using these digital twins, the team simulated three scenarios: chemotherapy alone, chemotherapy combined with olaparib, and no treatment.

The researchers then compared the simulations with clinical trial outcomes and found that the model correctly predicted whether treatment would work in 74% of cases. Among 13 patients who did not respond fully, the digital twins identified at least one additional targeted treatment for 12 patients that might have improved outcomes. For 10 of these patients, the model predicted that adding the treatment could produce a sufficient response that no invasive cancer would likely remain in the breast or lymph nodes afterward.

The inclusion of olaparib reflects a biological vulnerability found in some TNBC tumors. Homologous recombination deficiency affects roughly 15% to 20% or more of cases and can make tumors vulnerable to drugs that block poly-ADP ribose polymerase, or PARP, an enzyme involved in repairing single-strand DNA breaks. Olaparib is a PARP inhibitor and was used in the PARTNER trial, led by the University of Cambridge and Addenbrooke’s Hospital, Cambridge, between 2016 and 2024. The findings were published in Cell Reports Medicine on October 26, 2026.

Related Links
University College London