49 Researchers at the University of California, San Francisco (UCSF) have developed a new method to predict how different types of breast cancer may respond to treatment, combining patient data from the I-SPY2 breast cancer trial with rapid testing on laboratory-grown mini-tumours. The advance could help accelerate more personalised breast cancer treatment, including for triple-negative breast cancer, which is particularly aggressive and difficult to treat. In the study, published on 6th August in Cell Reports Medicine, researchers found that the organoids could mimic the treatment response of a patient’s tumour. “The breast tumour organoids modelled how corresponding patient tumours responded to therapies and identified candidate combination therapies for cancers that don’t respond to standard treatment,” said senior author Jennifer M. Rosenbluth, MD, a medical oncologist and the Sulochana Pradhan, MD, Endowed Professor in Breast Cancer at UCSF. “These findings support organoid modelling as a bridge between clinical biomarkers and precision treatment strategies in breast cancer.” The researchers created the breast cancer organoids by placing patient-derived tumour cells into a gel designed to support the biology of the original tumour. Over several weeks, the cells clustered into tiny spheres that reflected the structure and biological characteristics of the source tumour. Each cluster contained up to thousands of cells, making them too small to see clearly without a microscope, while the largest appeared as translucent clusters within the gel. The team then evaluated this biobank of organoids derived from early-stage invasive breast cancers to investigate the mechanisms behind treatment resistance. Response-predictive subtypes, molecular subtypes developed during the I-SPY2 trial, provided the researchers with a framework for anticipating how tumours might respond to current therapies, including immunotherapy, PARP inhibitors, platinum chemotherapy drugs, and dual HER2-targeted therapies. Using clinical data from patients enrolled in I-SPY2, the researchers combined response-predictive subtypes with biomarker data to predict how the organoids would respond to different treatments. Because many of the organoids were derived from Triple-Negative Breast Cancer (TNBC) tumours, the researchers used these models to validate a prediction model for response to veliparib plus platinum chemotherapy (VP), a combination often used to treat TNBC despite the cancer’s tendency to resist treatment. To identify alternative strategies for overcoming this resistance, the researchers selected tumour organoid TORG40, which showed the highest predicted and subsequently validated resistance to VP. The team screened 386 small-molecule inhibitors against this organoid, including ABT-263, a drug that helps eliminate damaged cells. They then compared ABT-263 alone with ABT-263 combined with the chemotherapy drug cisplatin. The combination enhanced activity against resistant tumour cells and produced a particularly potent effect in TORG40. The screen also identified other promising candidates, including a class of drugs known as HSP90 inhibitors. The researchers were able to link these findings to a subset of I-SPY patients who had responded more favourably to this type of treatment. “The breast cancer organoids were found to express important cancer biomarkers, many of which can be targeted with drugs,” said Tam Binh V. Bui, MSc, the study’s first author and a PhD candidate at UCSF. “These organoids allowed us to study the effects of drugs directly in human tissue and prioritise the most promising therapies for this subtype.” The study did not assess how organoid-guided treatment decisions would perform over time. The organoids also could not capture the full complexity of a living organ or recreate the environment inside the body, including blood vessels, immune cells, and other processes that influence the signals tumour cells receive. Even so, the researchers hope physicians may one day be able to use patient-derived organoids to develop more personalised approaches to cancer care. “By combining computational analyses of large molecular and clinical datasets with organoid model systems, this proof-of-principle study demonstrated the utility of matching I-SPY2 resistance biomarkers and signatures to residual disease tumour organoid cultures,” Rosenbluth said. “Our findings highlight the value of a reverse translational approach that integrates patient-level clinical trial data and testing in organoid models to inform drug discovery and future personalised treatment strategies for patients.” Breast cancer remains one of the most common cancers worldwide, with ongoing efforts focused on improving outcomes for aggressive subtypes such as triple-negative disease by better matching therapies to individual tumour biology. The I-SPY2 trial has already generated extensive molecular and clinical data that continue to inform these precision approaches. Further research will be needed to determine whether and how these laboratory models could eventually guide real-time clinical decisions. For now, the findings provide a foundation for integrating organoid testing with established biomarkers as researchers work towards more personalised breast cancer treatment.