Enterprise Technology, Moffitt Cancer Center

Pete D'Addio, MS, CHCIO, CDH-E, CHISL

As someone who has been in technology for over 20 years, I have seen such an enormous groundswell of disruption and transformation in the last few years. The adjustment to succeed in such a rapidly moving ecology has been just as much of a culture change as compared to a technology change.

My responsibility is delivering foundational technology in an expanding cancer center care system. So, what is the best approach? Should we focus on the bigger strategy of AI opportunities in healthcare or determine areas of low-hanging fruit that can bring quick gains in using AI? It’s ultimately somewhere in the middle. A pragmatic approach is needed to deliver value. As we look at the landscape, we must foster a combination of creative AI solutions while working on partners to drive purpose-built solutions. There are considerations to account for the success of AI. AI training must be made available so physicians and staff can understand how to use AI responsibly and securely. This leads to building AI literacy, focusing on the ability of clinical and research staff to drive value with the use of AI.

There are so many good ideas and opportunities for AI in healthcare. So now, where to begin? Two of our main organizational focuses are data and imaging. My goal is to partner with our organization’s clinicians and researchers to understand the needs and strategy to deliver. With AI, ML and LLM, the ability to consume voluminous amounts of data and images to produce better and faster outcomes for clinicians and patients brings purpose-built use cases. This drives value for efficient patient care, as well as better research outcomes.

 “It's important to embrace ideas and prospects for AI, but ensure proper checks before moving forward”

Now, the focus of a technologist is to provide the right foundation to help deliver AI outcomes and useable models. Currently, my organization has increasing opportunities with imaging to build cancer and tumor ML models that produce better predictive results. Also, years and years of de-identified patient data are being fed into LLMs, providing results that could eventually lead to a cure for cancer. Now, my organization can mature these models with the potential to monetize them and provide entry points for data and model sharing. We have cultivated an infrastructure along with our scientific computing department to deliver the ability for our AI department to create those models to run on high-end HPC and GPU units.

As we look toward the future, it is important to build AI strategies now. AI will affect the workforce in the coming years and shape the way work is done. It’s important to embrace ideas and prospects for AI but ensure proper checks before moving forward. For me, AI is ‘trust but verify.'