Revolutionary Study Reveals AI’s Ability to Predict Prognosis in Triple-Negative Breast Cancer

Researchers from Karolinska Institutet in Sweden studied how well different AI models can predict the prognosis of triple-negative breast cancer by analyzing immune cells within tumors. Their findings appear in the journal eClinicalMedicine and mark a significant advancement in using AI for cancer care.

Tumor-infiltrating lymphocytes (TILs) are immune cells that help fight cancer. Their presence in tumors indicates that the immune system is targeting cancer cells. TILs can predict how patients with triple-negative breast cancer respond to treatment and how the disease may progress. However, pathologists often report varying results when assessing these immune cells. AI can help standardize this evaluation, but demonstrating its effectiveness in healthcare has been challenging.

The researchers tested ten AI models to see how accurately they could analyze TILs in tissue samples from triple-negative breast cancer. The results showed varied performance among the models. However, eight of the ten models had good prognostic abilities, showing promise in predicting patient health outcomes.

Balazs Acs, a researcher involved in the study, noted, “Even models trained on fewer samples showed good prognostic ability, suggesting that tumor-infiltrating lymphocytes are a strong biomarker.”

The study emphasized the need for larger datasets to compare AI tools effectively and ensure their reliability for healthcare use. Acs stated that independent studies are essential to replicate real clinical conditions. This testing will help verify the effectiveness of AI tools in clinical settings.

For further reference, the study can be found in eClinicalMedicine: Vidal, J. M., et al. (2024). The analytical and clinical validity of AI algorithms to score TILs in TNBC: can we use different machine learning models interchangeably? DOI: 10.1016/j.eclinm.2024.102928.