Biology first: Key insights on AI, RWD, and the future of oncology from AACR 2026

Neil Bence, PhD: This is a really exciting time in drug discovery and development. We have access to more data and more modalities than ever before, and that’s translating into a deep understanding of disease. This is one of the most complicated team sports on the planet. We have to consider the complexities of the disease and move beyond a simple single-oncogene hypothesis to account for the tumor's full composition.

All roads lead back to a deep understanding of that biology. As we invent new medicines, you’re changing the disease. Cancer adapts, it evolves, it evades, and then you have to think about what is next. This is where what Tempus has done to help open the doors for access to patient data is part of the collective change and the evolution in our industry that we need.

Neil Bence, PhD: To me, it’s anything that allows us to see the critical finding where we say, 'Oh, it was sitting right underneath our nose.’ That’s an exploitable, targetable node of biology. A couple of examples:

• Profiling beyond the basics: We are starting to see evidence of profiling tumors beyond just DNA, RNA, and protein to include post-translational modifications, giving us additional layers of insight.
• Leveraging transcriptomics: We know that looking at transcripts and gene expression is the best way to stratify disease and patient subgroups.
• Embracing spatial technologies: The other thing I’m really excited about is the spatial side of things. One of the things I worry about is that we do all this tumor profiling, find a really good target, take it into the clinic, and see underwhelming responses. This brings up the topic of heterogeneity. What if it’s a good target for only a certain subset of the cells within that tumor? If we can look at spatial and single-cell transcriptomics to start piecing together what’s really going on in the tumor, I think that’s going to be important.

Neil Bence, PhD: It’s absolutely critical. This is a part of the iterative cycle of drug discovery and development. What can we control?

• Access to data: First and foremost, we can work to gain access to data from patients on-study and from the general standard-of-care population.
• The right model systems: The second thing is to make sure we have the right model systems. If we don’t have the right biological models, how do we know if we’re on the right track when we make that big investment to go into the clinic?

If we have access to real-world patient data and patient-derived models that represent that disease state, we can piece together a continuous narrative. One of the most important things about how we’re leveraging patient-derived organoids is that, at the end of the day, you’re getting as close to the patients that you’re trying to treat. This helps us ensure we’re modulating those nodes of biology and seeing the appropriate responses in preclinical models to give us confidence. That laser focus on a clear path to clinical proof of concept is absolutely critical.

Neil Bence, PhD: From my perspective, AI is very real and tangible. I’m seeing evidence of it helping in virtually every aspect of this complicated team sport.

• Simplifying operations: Everything from simplifying communication and documentation on the regulatory side.
• Deepening biological understanding: Understanding the disease state, patient stratification, and identifying unique subpopulations—seeing the forest through the trees in all that patient data.
• *Accelerating molecular invention: On the molecular invention side, in terms of inventing new drugs, you’re seeing significant advancements.

It’s not a replacement for people. I believe what you often hear is true: the people who learn how to use AI and apply it will replace those who don’t. Every single person on my team is tasked with incorporating it into their work. They can take advantage of even simple things like what LLMs allow you to do. It can allow the average biologist to masquerade for a little while as a clinician, understanding the standard of care landscape and market forces. So when they enter into that conversation with their clinical or commercial colleagues, they’re coming with a starting position that everyone can lean into and refine.