09/01/2026
Two Australian-led studies published in The Lancet Oncology have found that AI models are as effective as human pathologists at counting tumor-infiltrating lymphocytes (TILs) — immune cells in breast tissue whose levels predict cancer outcomes. Higher TIL counts indicate a stronger immune response and are associated with better outcomes in several breast cancer types. The first study, analyzing data from more than 5,600 breast cancer patients, found that while AI and pathologist scores weren't identical, both provided equivalent prognostic information. The second study showed AI could go further by mapping the spatial organization of immune cells within tumors and identifying "hotspots" — information that provided additional prognostic value beyond simple TIL counting. "AI may allow us to extract information that the human eye cannot readily quantify, such as the spatial organization of immune cells within a tumor," said Professor Sherene Loi of Peter MacCallum Cancer Center, who led both studies. Researchers say AI-based TIL assessment should be more widely deployed, particularly in regions where pathologist access is limited. Both studies were published in The Lancet Oncology.
📷 Getty Images