Launch of AI model to track cancer tumors at the cellular level

In a post on the website of NVIDIA Technology Corporation (USA) on February 4, the AI model that BioTuring recently announced can help scientists clarify how and why certain relationships between cells will trigger cancer development.

BioTuring has announced an AI model that can generate detailed images of cancerous tumors at cellular resolution. These images will provide data on the size and shape of cells, which genes are activated, and importantly, the relative spatial location of millions of different cells in a tissue sample.

Compared to computational biology, the new AI model has advantages in information processing time, providing high-resolution, real-time insights into the interactions between cancer cells and the immune system.

“There are about 30 trillion cells in the human body, and a large tumor will have several million cells,” said Son Pham, CEO of BioTuring. “Imagine if you were analyzing high-resolution satellite imagery to understand how a city works, our model, from a biological perspective, would show you every house, what’s in it, who’s talking and what they’re saying,” he said, explaining that BioTuring’s AI model will help answer some of the complex challenges in clinical cancer research — by showing how cells interact.

BioTuring, a member of the NVIDIA Inception program for startups, used NVIDIA equipment and technology to create a high-resolution map of cells. By understanding how cancer cells grow and spread in the human body, scientists hope the AI model can improve screening methods for early cancer detection.

Additionally, researchers could use the model’s cellular detail to better understand tumor heterogeneity, or how cancer tumors in the same patient have cells that are significantly different from one another. The visual granularity of the AI model could also aid in cancer drug development.