Data Fuels Cancer Research and Treatment
EDITOR’S NOTE: In honor of National Women Physician Day on Feb. 3, Moffitt Cancer Center’s Women in Oncology group is highlighting the ways women physicians and researchers are contributing to the prevention and cure of cancer.
In a world bursting with data, its influence often goes unnoticed. Data plays an unassuming but integral role in shaping our lives. It fuels our next show suggestion from our favorite streaming service. It autocompletes our next text message. It gives us the fastest route to our destination. The study of data helps us better understand the world and prepare for the future.
At Moffitt Cancer Center, data flows in from all angles. Electronic health records. Biospecimens. DNA. Clinical trials. The sources are endless. Collecting and combining individual bits of data helps our scientists tell broader stories. These stories ultimately guide how we treat patients, assess outcomes and develop new therapies.
Central to these efforts are a handful of innovative women scientists who are harnessing the power of data and shaping the future of cancer care.
Leading the Charge
Dana Rollison, PhD, is a narrator of the stories data tells. As chief data officer and associate center director for data science, she is fueling the engine of cancer research at Moffitt with her expertise and vision. Rollison is a multifaceted leader, known for her passion as a foodie, her love for photography and her talent as a ballroom dancer. But at her core, Rollison is a dedicated data scientist.
The journey that led Rollison into the world of cancer research began with her mother’s experience with Hodgkin lymphoma. In the 1970s, Hodgkin lymphoma became one of the first cancers to be successfully treated with radiation therapy, and her mother was among the early beneficiaries of this lifesaving treatment. Sadly, her mother’s battle with cancer didn’t end there. She was later diagnosed with breast cancer and a spinal cord tumor. She ultimately passed away at the age of 49.
Rollison’s childhood and early adulthood were marked by her mother’s courageous fight. “She’s really the inspiration behind what I do,” Rollison said.
Rollison’s interest in medicine has been lifelong, but her focus was never on clinical care. Instead, she was drawn to understanding the factors driving disease risk. This passion led her to cancer research, and she joined Moffitt in 2004.
Data has the potential to save lives, Rollison says. As an example, she points to the increasing incidence of colorectal cancer in younger people and the decreasing mortality rates from cervical cancer. These aren’t mere statistics. They represent vital insights derived from meticulous data collection in cancer registries.
Nearly two decades ago, Moffitt began investing in data science, which rapidly evolved into the Total Cancer Care research protocol. Total Cancer Care uses data and biospecimens collected from patients over time to improve personalized treatment and prevention efforts. However, in the beginning, the data was not collected in a centralized source, making it challenging to harness its full potential. Patient-reported data, tumor information, clinical records from the Cancer Registry and electronic health records were spread across different platforms.
In 2010, Moffitt made a significant move by investing in a centralized data warehouse and establishing an access team to serve as a front door to this valuable resource. As data volumes expanded, Rollison’s role in shaping the data landscape also grew.
Under her leadership, the data warehouse underwent a profound transformation. With the introduction of more complex data types, such as textual information in electronic health records and imaging data, the challenge of downstream analysis became evident. Rollison embarked on a multiyear journey to harness the power of cloud computing.
Recognizing the need for data scientists at the faculty level and understanding the importance of multimodal algorithms and the integration of various data sources, Moffitt created the Division of Quantitative Science, a collaborative effort uniting Biostatistics and Bioinformatics with Integrated Mathematical Oncology, under Rollison’s leadership. Moffitt was also one of the first cancer centers to have a Machine Learning Program dedicated to cancer research.
“The algorithm is only as good as the data you feed into it,” Rollison explained. She stresses the importance of data quality in the development and implementation of artificial intelligence algorithms. Trust is key, and physicians need to have confidence in the algorithms incorporated into clinical care.
From DNA to Drug Discovery
For Aleksandra Karolak, PhD, the stories that can be told through data start with DNA. Genetic sequences fill the dry-erase whiteboard hanging in her office, right next to a rainbow of triathlon medals. Handheld models of molecules decorate her desk. All things science and research get Karolak excited, but what makes her perk up most is drug discovery.
Aleksandra Karolak, PhD, focuses her research on how to use artificial intelligence and other computational tools such as molecular simulations to speed up mutational analysis and drug discovery.
Karolak originally completed her postdoctoral research as part of Moffitt’s Integrated Mathematical Oncology Department from 2015 to 2018. When she rejoined Moffitt for the Machine Learning Program in 2021, Karolak brought with her a background in computational chemistry, mathematical modeling and machine learning. This background gave her a unique perspective, enabling her to follow the data from DNA organization to protein mutations to the need for new drugs to target different clinical pathways. Her lab now focuses on how to use artificial intelligence (AI) and other computational tools such as molecular simulations to speed up mutational analysis and drug discovery.
The key to success, she says, is being able to test the potential for new molecules, experiment, make changes and continue testing. “It’s not just about designing new compounds,” she said. “It’s about testing them and making sure they’re effective.”
Of course, the power of AI is limited by the amount and quality of data that is available. To gain insights from small sets of data, Karolak works with other researchers throughout Moffitt as well as collaborators from other cancer centers to expand the data. This typically means identifying related data sets that could be combined, including data that is related but presented in a different format such as image data vs. molecular data.
She also shares insights with physicians to show how useful machine learning can be in analyzing patient data, providing earlier diagnoses and identifying drug targets.
“We have a great advantage that we work in a hospital,” Karolak said. “We have research facilities where ideas can be communicated and tested immediately. So in this way, we’re trying to add AI as a tool that can help automate multiple levels of drug discovery. And I believe we have a chance of succeeding in drug discovery and optimization and addressing the resistance problems properly.”
Patterns Drive Predictions
Most people think of statistics and informatics as crunching numbers or looking at mathematical formulas, but Moffitt’s Biostatistics and Bioinformatics team is developing methods and tools that directly impact the way researchers investigate cancer. The team’s expertise includes bioinformatics, biostatistics and statistical genomics. One of its members is Xiaoqing Yu, PhD.
For Yu, an avid Lego architect outside of work, embarking on any project requires a strong vision of the big picture before laying down the first building block. After all, the outcomes of a cancer research study are only as good as the first step in designing the study.
Xiaoqing Yu, PhD, helps personalize treatment approaches in clinical research by constantly evaluating the data and interpreting the results in real time to inform investigators if changes need to be made.
At Moffitt, the Biostatistics and Bioinformatics team is involved in designing clinical trials from initiation to evaluation. The biostatistics side helps set up the trial design and determines the number of patients needed to have statistically relevant results. They also help decide the dosing strategy of therapies. The bioinformatics side, like Yu’s lab, helps personalize treatment approaches in clinical research. Through analysis of genome sequencing or microarray gene expression data from patients, bioinformaticians can identify mutations or genetic variants that could impact response to therapy.
The team is constantly evaluating the data and interpreting the results in real time to inform investigators if any changes need to be made. They also help answer some of the most complex questions, like how to improve immunotherapies, one of the fastest-growing areas of cancer research.
In her daily work, Yu finds patterns in data as she collaborates with clinicians and other data scientists. She does this by developing bioinformatics tools for analyzing and actualizing data. One of her current projects includes a focus on bladder cancer, where she analyzes the raw data from patients before and after treatment to identify somatic mutations unique to these patients. She then predicts whether a mutation can be identified as a unique antigen and later be used as a target for immunotherapy.
“It’s all about finding the patterns — the right biomarkers — and determining if they can move forward for further validation from the experiment side,” Yu explained, noting that these discoveries can improve personalized medicine.
Like Karolak, Yu also uses machine learning and artificial intelligence for her work on single-cell data analysis.
“Analyzing data can be very time consuming,” Yu said. “We collect a huge amount of single-cell data. Can we build a model from this already analyzed data to make predictions? We’re designing deep-learning AI models to do that.”
Looking Forward to the Stories Yet to Be Told
The potential applications of data are boundless. In 2023, Rollison’s team launched the Moffitt Cancer Analytics Platform, known as MCAP. The platform brings together a wide range of high-quality data sources and analytics tools that will improve clinical and operational decision-making, cancer research and care delivery. Moving forward, Rollison expects Moffitt to continue building on the data that’s available from the clinical side and adding tools to empower researchers.
“Moffitt is ahead of the curve in how we’re thinking about data and how we’re building our data strategy to be enterprisewide,” she said. “We’re really focused on building synergies between clinical and research initiatives, and we’re investing in data infrastructure that can benefit the most people. There are very few institutions that have a data strategy at the highest levels of the organization the way we do.”
