Getting to know Chief Data Officer Caroline Chung, M.D.
In October 2021, radiation oncologist Caroline Chung, M.D., became MD Anderson’s first-ever chief data officer.
Her charge? To shape MD Anderson’s data strategy and lead its implementation from an operational and cultural perspective.
Chung joined MD Anderson in 2016 from Princess Margaret Cancer Centre in Toronto, Canada, to spearhead the safe and novel uses of MRI to further advance personalized treatment in Radiation Oncology. Under her leadership, the effort has grown into a strategic initiative around Advanced Imaging and her passion for maximizing the use of high-quality imaging data to improve patient outcomes brought her to co-chair the Tumor Measurement Initiative, MD Anderson’s platform to support standardized, automated, quantitative imaging-based tumor measurement. Most recently, Chung has led the effort to establish an institutional Data Governance and Provenance Office to ensure strong stewardship of our data while maximizing the effective use and accelerating discoveries from our data to serve our mission of ending cancer.
Here, she answers nine questions about her new role, data and artificial intelligence (AI), and transforming personalized medicine for our cancer patients.
What brought you to MD Anderson?
The institution’s people, collaborative environment and tangible passion and dedication toward providing cancer patients with world-class care through innovation, research, education and prevention attracted me to MD Anderson and continue to inspire me every day. It is wonderful to work at an institution that has such strong alignment with my personal goals of delivering patient-centered, data-driven care and achieving the best outcomes possible – all of which we accomplish through data.
What are you most excited about in your new role as MD Anderson’s first chief data officer?
I am most excited about partnering with our patients and everyone across MD Anderson to shape our data-centric approach to ending cancer. We want to responsibly and effectively use the massive amounts of data generated each day to determine the best possible treatments and care paths for each patient, based on their diagnosis and personal goals. I am excited to work with everyone across MD Anderson to build a culture that recognizes the value and importance of ensuring high quality data to serve our mission to end cancer.
What role does AI play in medicine?
Asingle cancer patient generates over 2 GB of data each year in imaging and electronic medical records data, so the sheer volume of information can be overwhelming. With more advances and new applications of technology — such as telemedicine, deeper characterization of tumors and -omics research — we’re expecting to see an even larger increase in data.
AI has the potential to help us gain the greatest value and use out of the exponential growth of data and paint a clearer picture of each person’s cancer journey, deliver more personalized care, increase safety of care, and improve patient-centered outcomes.
What is MD Anderson doing to pioneer AI-enabled solutions in medicine?
We need to look at the data in context in order for AI to reach its full potential. At MD Anderson, we are shaping our approach to data around a set of principles that start with the core view of keeping our data in context. I recently highlighted these messages in a perspective piece in Cancer Research. Essentially, metadata is the contextual data about the data. It provides information about the who, what, where, how and even why the data was collected that give context to the data and, thereby, informs the definition, quality and lineage of the data. This is important because it ensures that we can trust the data and use it appropriately to ensure confidence in our data-driven decisions while serving as good stewards of the data.
What does this mean for cancer patients?
By looking at the data in context, we consider the patient as a whole person and record the situation within the moment the data was captured. We can interpret the information that considers your holistic journey over time up until the present moment and provide the contextual assessment that enables personalized medicine when we have the ability to look at your:
By building a robust flow of data (the content) and metadata (the descriptors that provide the context around the content) and developing predictive models, AI has great potential to help us make profound leaps and bounds toward truly personalized, patient-centric care.
How do a data-driven approach and AI help us discover new treatments and improve diagnoses?
A data-driven approach aims to both ask and answer questions with an open mind based on objective investigation of the data.
In order to do so, we need to put an emphasis on gathering detailed, high quality and consistent data. Having multidisciplinary teams including your clinical care team but also other researchers and data scientists from across the organization come together around the data allows us to consider new questions and insights that may not have been obvious before. It is a true testament to the power of team data science and data-driven approach.
How will AI improve cancer treatment advances?
The human body is an intricate connection of systems (e.g., nervous system, gastrointestinal system, endocrine system). AI is enabling us to model much more complex systems and the complicated interaction between systems with greater efficiency. By generating a digital twin, AI could enable us to model different systems and even a patient’s entire body with all the different integrated systems to help inform and predict treatment responses. Using digital twin technology to accelerate our understanding of cancer could lead to the discovery of more effective treatments and guide personalized cancer treatments.
For example, MD Anderson’s Tumor Measurement Initiative focuses on improving the quality of our imaging measurements. Enhanced by technology such as AI, it has the potential to improve our ability to confidently determine tumor response to treatment. This will improve our ability to assess how effective a potential new drug or treatment will be against certain cancers. It may also improve our ability to make more timely adjustments to treatments in clinical care. Utilizing the data and AI, we may see new patterns of treatment response that may lead to future discoveries in cancer treatment. For example, the contextual data in combination with AI may reveal that tumors with specific genetic mutations behave a certain way or combining cancer treatments with other associated medications may lead to better responses.
AI can also accelerate tasks that would take humans years to complete. One example of a task that has been accelerated with the use of machine learning is the ability to model the shapes of large molecules and how they fold and interact with each other. Speeding up this process with AI will help scientists speed up the development of new drugs or biologicals personally tailored for each patient.
What are you passionate about outside of work?
My fascination with imaging and data is stimulated by my love of art, food and learning new things. I love to paint, cook and read on a vast range of topics. I love sharing the products of my experiences and experiments (never follow recipes) with those around me. It fascinates me how the brain takes, interprets and gets affected by all our senses – sight, sound, touch, smell and taste.
If you weren’t a doctor, what would you be?
Being in Texas and collaborating with NASA, I may say space cowboy, which I was for Halloween in 2020 when kids came by for socially distanced treats.
At the end of the day, I love what I do and feel blessed to be constantly inspired by my patients — by their strength, their struggles, their hopes and resilience. They are a grounding motivation for all my efforts in clinical care, research, and in my new leadership role.