Researchers at Dana-Farber Cancer Institute have developed a deep learning model that can identify previously unnoticed features in clear cell renal carcinoma (ccRCC) from pathology slides. The model can identify tumor microheterogeneity and immune properties, such as immune infiltration, that may help predict the response to immunotherapy in patients with this rare form of kidney cancer. The findings suggest that the pathology slides of ccRCC tumors contain important biological information that could aid in understanding the biology of this cancer and potentially be used for other types of cancer as well.
Renal cell carcinoma is one of the most common cancers worldwide, and ccRCC accounts for the majority of metastatic cases. While some tumors have responded to immune checkpoint inhibitors, there is currently no diagnostic measure to determine whether a specific tumor will respond.
The AI tool was initially trained to assess a tumor’s nuclear grade, which measures how far tumor cells deviate from normal cells. It successfully assessed nuclear grade and was also able to identify differences in grade across the tumor sample. Inspired by these findings, the researchers refined the model to quantify tumor microheterogeneity and immune properties across the slide.
The analysis of the pathology slides revealed that some tumors were homogenous, while others showed a range of different nuclear grades and immune infiltration patterns. The researchers then turned to data from a clinical trial and found that tumor microheterogeneity and immune infiltration were associated with improved overall survival in patients taking immune checkpoint inhibitors.
The AI tool has the potential to provide valuable insights into tumor biology and guide personalized cancer care. The researchers are now testing the tool in a clinical trial involving a combination immunotherapy as a first-line treatment for ccRCC.
This study demonstrates the growing convergence of AI and cancer biology, presenting an opportunity to measure key features of tumors and their immune microenvironment. With scalable AI tools, more information can potentially be extracted from pathology slides, leading to improved understanding and care for patients with cancer.
Sources:
– Cell Reports Medicine
– Dana-Farber Cancer Institute
