Empire Innovation Professor Nancy Guo joined the Binghamton University faculty in the fall 2024. Guo is experienced in leading foundation AI-based multidisciplinary research as PI of two NIH R01s and two NSF grants. She obtained more than $45.5 million in federal funding as PI/PD to develop technology and infrastructure to advance precision medicine. In addition to the software platform, Guo and her colleagues also developed a 7-gene lung cancer assay that accurately predicts the risk of tumour recurrence and metastasis while assessing the clinical benefits of chemotherapy. This invention has been validated in more than 1,600 patients, including a randomised phase 3 clinical trial. The FDA has recognised its significance by classifying it as a "Novel Technology" in its review process.
As the founding director of the Biomedical Informatics Resources Core of West Virginia Clinical & Translational Science Institute from 2009-17, she led statewide informatics initiatives and enhanced multi-state collaboration. She is fostering academic-industry partnerships for the clinical commercialisation of AI-based cancer treatment selection and drug development through her current NSF PFI-RP project.
Guo has not only been about software and assay development but also about identifying new drugs and new indications for existing drugs for the treatment of lung cancer and breast cancer. Through the patented AI software pipeline, Guo has made significant contributions to advancing cancer diagnostics and therapeutics.
Can you tell me about your role at Binghampton?
Guo: I'm a newly hired SUNY Empire innovation professor. Specifically, this is a track for artificial intelligence machine learning in healthcare. My expectation in this role is to establish a multidisciplinary, multi-institutional collaboration. Precision medicine is multidisciplinary. We need expertise from different fields. For example, we need people who understand AI, machine learning, and computer science, and we need people who understand hardware and biomedical engineering to make medical devices. We need clinical expertise involved from the beginning to understand their unmet needs. Inevitably, we also need expertise from biology and pharmacy if we are talking about medicine. We need experts to ensure the algorithms and the innovation we seek to develop make clinical and biological sense.
Why did you choose this path?
Guo: When I thought about my major in college at that time, I discussed it with my parents. My father was a university professor in engineering, specifically fluid dynamics, and my mom majored in chemistry. I was interested in understanding the human body and how it works systematically and precisely. Then, I took one class at university that taught us how to programme. Specifically, we learned how to use computer programming to compute a polling gene in DNA. I found that interesting. That's the only course where this was close to what I wanted to do.
At that time, few universities had a bioinformatics degree program. So, the best way to learn about these techniques is to learn computer science. I subsequently got into a project using computational methods to analyse EKG results. To analyse the heartbeat to detect heart attacks or abnormal heart events earlier.
How did you become involved with computational genomics?
So when I switched to computer science for my master's, that's about the time it started. The technology to measure genes in the genome was becoming more established. About 6000-7000 genes had been catalogued and measured at this time. I decided that I needed to learn computer science to get involved in this research.
Can you talk about your breakthroughs in cancer research?
Guo: We first developed this AI technology, and then applied it to analyse cancer patient genomes. We decided to focus on lung cancer as this type of cancer has the second highest incidence rates for both men and women and the top cancer-related mortality rate for both men and women. The current trend is that more and more non-smokers have lung cancer.
Through the research, we initially analysed the publicly available cancer patient data. Through this research, we could focus on 200 genes that we thought were important. Then, we collected samples from multiple US hospitals. From that, we reduced the number of genes to seven and developed a microfluidic chip. This medical device can predict early-stage patients who may develop tumours that will metastasise and may benefit from chemotherapy. We found that these seven genes can be used to accurately predict the risk of tumour occurrence and metastasis in the earliest stage of cancer patients. These patients would receive surgery because that's a current standard of care. Still, about 40% to even 60% of them will develop tumour recurrence, followed by metastasis, and these patients die within five years after surgery.
Before this research, there was just no way to tell whose tumour was more aggressive molecularly. We developed and published this method, and people came to me, oncologist colleagues, and they would ask, “I have a patient. Can I use your funding to decide whether he needs chemotherapy?” But this was research funding, and so there was no way they could use it.
In 2023, Guo received a $565,994 grant from the NSF’s Partnerships for Innovation Program to encourage the development of market-oriented prototypes and a multidisciplinary graduate training program in West Virginia and the Appalachian region. This grant helped to accelerate the market-oriented development of prototype products and the AI tools used to diagnose cancer patients using the gene assays developed by Guo and her team.
How do you commercialise this technology?
Guo: From the research side, we felt that we had solved the problem, but the challenge was to turn that into something that could be used in hospitals. At that time, I got a grant from the NIH National Institutes of Health that focused on turning this idea into a product. Through that program, I learned that you need to think about how to produce a commercial product so that clinicians and doctors can use it in a clinical setting. One aspect of this was a one-on-one coaching programme, and through that, I learned that you need to register for a company, get a small business grant, find a licensing partner, and go through all that process before you can get a new product in a clinic. Once we had gone through that process, in 2020, we submitted the technology to the FDA, which gave us the status of novel technology.
Through the process from research to commercialisation, I received tremendous help from the National Institutes of Health and National Science Foundation. They've been funding us and providing all kinds of programmes to train us in innovation and entrepreneurship, such as how to assess a current unmet clinical need.
What are the remaining challenges for your precision medicine research?
One ongoing area of research for us is how we translate research into products, and we're also looking at the next generation of technology. Everything we have done so far is based on 20 years of research, just from my group and all others, because we leverage many other public resources and data sets. So, before looking at the next generation of technology, especially AI, we must ensure that we stay at the forefront of biomedical research.
So that really has something to do with how we structure the institute we're building. We know what is needed to continue, and then try to fill the current gap. Filling the gap means we are branching out. For example, my expertise is in genomic analysis and computational genomics. However, some people are experienced with medical image analysis, so how do we work with them? How can we integrate genomic research with medical imaging.
There are many new emerging technologies, foundational AI models and large language models (LLM). We need to think about how to leverage emerging technologies. But we will not just stop at research. We can deliver products that can be implemented in the healthcare system. It is about improving existing technology, developing new technology and forming collaborations.
Nancy Guo is a SUNY Empire Innovation Professor at the School of Computing at Binghamton University's Thomas J. Watson College of Engineering and Applied Science.
