Titles and Affiliations
Chair, Department of Medical Physics
Enid A. Haupt Endowed Chair in Medical Physics
Research area
Developing new mathematical tools to interpret and understand large sets of data in order to gain a deeper understanding of cancer.
Impact
Mathematical approaches can be used to gain insight into how complex, interacting systems drive cancer, how cancer affects a patient, and how treatments affect cancer. As part of the Mathematical Oncology Initiative, Dr. Deasy has assembled a team of mathematicians, biologists, oncologists, and other scientists to develop mathematical models and tools that can be used to interpret many kinds of data. These tools can help us gain a deeper understanding of the overall picture of cancer, including areas such as disease evolution, treatment response, identifying subtypes, patient risk of toxicity, and more. Their work will contribute to the advancement of precision medicine for cancer.
Progress Thus Far
Dr. Deasy and his team have advanced two complementary mathematical tools and applied them across breast and other cancers. The first tool, NetFlow, arranges patient samples into a relational map that reveals subtle patient subgroups invisible in any single type of data, uncovering subgroups at higher or lower risk. Applied to triple-negative breast cancer (TNBC), it revealed a particularly aggressive subgroup whose tumors are largely not recognized by the immune system. The second tool reads patterns directly from pathology slides and MRI scans; when applied to MRIs taken during chemotherapy, it revealed early signals of treatment response that traditional models missed. The team also found that where immune cells are located within a tumor, not just their overall number, strongly predicts survival in TNBC.
The team reached several other milestones this year: a map of more than 13,000 brain metastasis cases that can help doctors plan radiation treatment more precisely, a new tool for estimating a patient’s risk of developing a second breast cancer after radiation, and confirmation that certain pregnancy-related proteins drive worse lung cancer outcomes specifically in women. Together, this work is moving the Initiative from simply describing patterns in cancer data toward tools doctors can use to identify subtypes, read scans, monitor treatment, and personalize care.
What’s next
In the coming year, the team will pursue two main goals. First, they aim to train a computer model that can recognize their newly identified high-risk breast cancer subtype simply by looking at a tumor sample, without needing separate genetic testing. Second, they are working to build a tool that can reliably distinguish between tissue damage caused by radiation and tumor recurrence after brain radiosurgery, a persistent diagnostic problem after which clinical decisions diverge sharply. Early results are promising, and the tool will next undergo multi-institutional validation and prospective testing. The team will also extend NetFlow to additional breast cancer cohorts and treatment response settings, and will pursue further research into why certain pregnancy-related proteins seem to worsen lung cancer in women.
Biography
Dr. Joseph O. Deasy is Chair of the Department of Medical Physics, and holder of the Enid A. Haupt Endowed Chair in Medical Physics, at Memorial Sloan Kettering Cancer Center, New York.
Dr. Deasy is an attending physicist at Memorial Sloan Kettering Cancer Center (MSK). He received his PhD in Physics from the University of Kentucky in 1992. Thereafter he completed a NIH-funded post-doctoral fellowship at the University of Wisconsin-Madison, with mentors Rock Mackie and Jack Fowler. Before arriving at MSK in 2010, Dr. Deasy spent 11 years in the Department of Radiation Oncology, Washington University in St. Louis, first in the physics division under the direction of James Purdy, and later as the first Director of the Division of Bioinformatics and Outcomes Research. Dr. Deasy is the co-author of about 140 peer-reviewed publications and has been the principal investigator of several NIH grants. Dr. Deasy’s current interests are in applying mathematical modeling and machine learning to the analysis of imaging, genomic, and treatment datasets in order to understand the relationship between treatment, patient, and disease characteristics and the probability of disease progression and treatment response.
