Immunotherapy: How is AI Revolutionizing Pathology

Stacey Smith, Pharma Tech Outlook | Monday, March 15, 2021

Specialists are focusing on using AI to reliably differentiate between PD-L1 expression on specific immune cells and tumor cells. This data will be used to help classify patients who can benefit from immune checkpoint blockade

FREMONT, CA: Immunotherapy has resulted in long-term responses for a growing number of patients, including those who had a negative prognosis at first. Those that respond to immunotherapy are, unfortunately, in the minority. As a result, a big priority in the field is exploring novel strategies to classify and stratify patients that are more likely to react to particular immunotherapeutic approaches.

Imaging a patient's cancer and the immune system's responses may offer crucial knowledge that can further influence immunotherapeutic care choices during treatment. Advances of imaging, such as real-time, noninvasive visualization of specific cells, or the application of Artificial Intelligence (AI) to pathology, have seen a lot of press recently in the immunotherapy world.

Applying Artificial Intelligence to Transform Pathology

Although alternate imaging modalities have various advantages, definitive diagnoses for certain diseases, including cancer, include tissue sample review. Humans are limited in their ability to evaluate and understand pathology images, despite the fact that they provide a wealth of detail. Research has shown that AI can outperform human pathologists. Experts are incorporating data annotated from millions of experiments by hundreds of pathologists into the deep-learning model to integrate more patterns to strengthen this technique. Real-world problems, such as variations in slide quality, arise with such large datasets. To address these issues, researchers are developing specific deep-learning models.

PD-L1 (Programmed Death-Ligand 1), also known as cluster of differentiation 274 or B7 homolog 1, is a protein, which in humans is encoded by the CD274 gene has been a significant biomarker for specific checkpoint inhibitor-based therapeutic treatments. Still, comprehensive pathological platforms to reliably quantify this biomarker are lacking. Specialists are focusing on using AI to reliably differentiate between PD-L1 expression on specific immune cells and tumor cells. This data will be used to help classify patients who can benefit from immune checkpoint blockade.