Cancer immunotherapy has a prediction problem…

Source: NCI – Killer T cells,

We have long known some patients achieve durable responses to checkpoint blockade while others derive little or no benefit. The biological signals distinguishing these groups remain frustratingly scattered across genomic and biochemical data, immune profiles, and clinical outcomes.

AI tools may offer some promise in connecting the dots by processing datasets way too large for manual analysis.

At the recent SITC meeting, three presentations demonstrated different approaches to this challenge with varying degrees of success. What emerged wasn’t so much a story about whether AI works in oncology research, as opposed to which AI approaches are adding value beyond conventional methods.

Together, they illustrate both the potential possibilities, as well as the persistent pitfalls of AI in translational oncology…

To continue reading our latest highlights on oncology new product development including commentary and analysis BSB subscribers can log-in or you can click to access the content.

This content is restricted to subscribers