A team of scientists at Technion – Israel Institute of Technology has developed an advanced artificial intelligence system that may reshape how doctors determine whether breast cancer patients should undergo chemotherapy.
The newly developed model analyzes standard pathology samples to estimate both the likelihood of cancer returning and the potential benefit a patient may gain from chemotherapy. The findings, published in The Lancet Oncology and presented at a leading European oncology conference, represent a significant milestone in cancer research.
This innovation could have immediate implications, particularly in regions like Israel, where approximately 5,000 new breast cancer cases are diagnosed annually. By offering a faster and more accessible alternative to existing methods, the technology may help reduce unnecessary treatments while ensuring patients who truly need chemotherapy receive it.
Moving beyond costly genomic testing
Currently, oncologists often rely on genomic testing tools such as Oncotype DX to guide treatment decisions after surgery. While effective, these tests can be expensive, time-consuming, and not always widely available. Additionally, results may take weeks to process, delaying critical care decisions.
The AI model developed at Technion aims to overcome these limitations by using data already collected during routine diagnostic procedures. Instead of examining genetic material, the system evaluates high-resolution images of tumor tissue.
Using deep learning techniques, the model detects subtle patterns within cancer cells and their surrounding environment—details that are often too complex for the human eye to interpret accurately. These patterns are then translated into a predictive score that reflects both recurrence risk and the likely effectiveness of chemotherapy.
Backed by large-scale clinical data
The research was led by Gil Shamai, alongside Ron Kimmel and Dvir Aran, in collaboration with international institutions including Dana-Farber Cancer Institute, Mount Sinai Medical Center, University of Chicago Medical Center, and IPATIMUP Medical Center.
To validate the model, researchers used data from the TAILORx trial, one of the largest breast cancer clinical trials ever conducted, involving more than 10,000 participants. This allowed the team to assess not just the risk of recurrence, but whether chemotherapy actually improved patient outcomes.
The system was further tested on thousands of additional cases across multiple countries, including leading hospitals in Israel such as Sheba Medical Center. Results showed consistent accuracy across different healthcare environments and imaging conditions.
A new direction for cancer care
Researchers say the model represents a fundamental shift in how treatment decisions could be made. By focusing directly on tissue analysis rather than genetic testing, doctors may soon have access to quicker, more cost-effective tools for guiding therapy.
If widely adopted, this technology could reduce overtreatment, lower healthcare costs, and improve patient outcomes globally. While further implementation and regulatory approval are still required, the study highlights the growing role of artificial intelligence in precision medicine.
