A new tool helps doctors choose the best treatment for breast cancer

Imagine doctors having a “crystal ball” that could show them in advance how well a specific breast cancer treatment will work for each individual patient. Researchers are working on something similar: an Artificial Intelligence program (a machine learning model) that can help make better decisions in treating a specific type of advanced (metastatic HR+/HER2-) breast cancer.

Why is this tool needed?

When breast cancer spreads, doctors often add drugs called CDK4/6 inhibitors to the standard hormonal treatment. The problem is that these drugs don’t work equally well for everyone. Some patients benefit greatly, while for others, the effects are minimal, but unpleasant side effects and additional costs can occur.

This new artificial intelligence program attempts to identify from the outset which patients will benefit most from the addition of these CDK4/6 inhibitors. This way, unnecessary aggressive treatments can be avoided for some patients, sparing them from side effects and expenses.

How does this “smart program” work?

Researchers have “trained” the program using information from many patients. They analyzed two types of data:

1. Clinical information: Details about the disease, such as how aggressive the tumor is, where the cancer has spread, and what previous treatments the patient has undergone.
2. Genetic (genomic) information: Specific changes in the genes of the cancerous tumor.

The program learned to combine this information to predict how long it will take until the disease might advance (this is called “progression-free survival”) if the CDK4/6 inhibitor is added to hormonal treatment.

What did the researchers find?

• They tested three variants of the program: one using only clinical information, one using only genetic information, and one combining both.
• The most performant was the program that combined both clinical and genetic information. It was able to divide patients into four risk groups:
  • Low risk: Patients in this group had the longest period without disease progression (on average 29 months).
  • High risk: Patients in this group had the shortest period without disease progression (on average 5.3 months).
  • There were also two intermediate risk groups.

This means the program can provide a much clearer picture of how likely a patient is to respond well to the combined treatment.

What’s next?

Although the results are promising, this study was conducted in a single hospital using historical data. The next step is to test this program on a larger number of patients from different hospitals to ensure it works equally well everywhere.

Why is this study important for patients?

This type of research is very important because it paves the way for more personalized treatment. Instead of offering the same treatment to everyone, doctors could use such tools to choose the best option for each individual patient, taking into account the unique characteristics of their disease. This could lead to more effective treatments, with fewer side effects and a better quality of life.