The LINUX trial investigated the effectiveness of using artificial intelligence to guide precision medicine for patients with advanced HR+/HER2– breast cancer who failed initial treatments. By utilizing digital pathology and deep learning to analyze tumor images, researchers categorized patients into four distinct molecular subtypes (SNF1–4) without the high cost of traditional genomic sequencing. Each group received a specific targeted therapy tailored to their subtype’s unique biological characteristics or a standard treatment chosen by their doctor. The study found that this AI-assisted subtyping significantly improved response rates and survival outcomes, particularly for the SNF2 and SNF4 groups. While some experimental regimens were more successful than others, the overall results demonstrate that personalized treatment strategies can outperform traditional "one-size-fits-all" chemotherapy. These findings suggest a scalable way to integrate machine learning into clinical oncology to enhance patient care and therapeutic efficacy.

References:

• Fan L, Zhang W J, Li H P, et al. Precision treatment with artificial intelligence assisted subtyping enhances therapeutic efficacy in HR+/HER2− breast cancer: The LINUXtrial[J]. Cancer Cell, 2025.