Pitt professor wins NSF CAREER Award for AI-driven immunotherapy research

The University of Pittsburgh's Natasa Miskov-Zivanov, assistant professor of electrical and computer engineering, has received a prestigious Faculty Early Career Development (CAREER) Award of $581,503 from the National Science Foundation (NSF) for her project titled "Artificial Intelligence-Driven Framework for Efficient and Explainable Immunotherapy Design." Through her novel approach and the development of an automated system that leverages AI and knowledge graphs to design more effective lymphocytes, she hopes to transform the design of life-saving immunotherapies.

Chimeric antigen receptor (CAR) T cell-based therapies have revolutionized the treatment of blood cancers such as leukemia and lymphoma, demonstrating the power of synthetic signaling receptors for immunotherapy. For these therapies, patient T cells are harvested, engineered with a CAR, and then reintroduced into the patient. However, CAR T cells have been less successful at treating solid tumors. The recognition and infiltration of solid tumors requires new CAR T cell designs.

While researchers continue to explore the most potent configurations, the combinatorial complexity of potential therapeutic lymphocyte designs, including CAR T cells, is vast. Miskov-Zivanov will create a system that will survey the scientific literature and databases to efficiently retrieve and integrate existing expert knowledge with experimental data and recommend more effective CAR T cells and tumor infiltrating lymphocytes (TILs).

Miskov-Zivanov, a computer engineer whose post-doctoral research has been in computational and systems biology, hopes that the automated framework she is developing can reliably accelerate this process. "As a computer engineer," she said, "I am driven to automate complex design processes. I explore how tasks traditionally performed manually by biologists can be streamlined and executed automatically using computational approaches."

In 2023, Miskov-Zivanov received an NSF EAGER Award to create a tool that uses Natural Language Processing (NLP) to extract information from scientific literature and, with the integration of experimental data, help synthetic biologists engineer new CAR T cells.

Building off that two-year project, she will develop a new system that uses both traditional NLP approaches and more recent large language models (LLM) together with neural networks to read, analyze, and interpret research papers and experimental data and conduct comprehensive in silico experiments on a wide range of cell designs. 

By developing and testing new prompting methods, Miskov-Zivanov hopes that "instead of a researcher having to go through tens of thousands of papers, many that are irrelevant, the system can extract meaningful data and insights."

Educating future engineers

The extracted information will be presented as knowledge graphs (KGs), which she will further refine and analyze using graph neural networks (GNNs) to predict the most effective therapeutic cell designs. This past year at Pitt, Miskov-Zivanov also introduced a new graduate-level course focused on KGs and the methods for their construction and application. She sees great promise in integrating information and structured knowledge within KGs with data-driven predictions enabled by GNNs, and she aims to uncover novel and meaningful connections and relationships that will significantly advance the engineering of new therapies. Equipping the next generation of engineers with these tools is essential for addressing critical, life-saving research challenges. 

Miskov-Zivanov hopes to build "reliable methodology to engineer and test thousands of designs for immunotherapeutic cells such as TILs and CARs with diverse and potent receptor systems." Her work seeks to advance immunotherapy while developing new algorithmic processes to identify and present trustworthy, predictive scientific data and research.

I am so grateful and honored to have received this CAREER Award. My interest in immunotherapy began twelve years ago when I read about a young girl whose leukemia was cured by this kind of treatment. Since then, I've been inspired to help advance this research. As a computer engineer, I'm especially motivated by opportunities to apply computing to make a meaningful impact in other fields."

Natasa Miskov-Zivanov, assistant professor of electrical and computer engineering, University of Pittsburgh

"Our department of electrical and computer engineering is so proud of Professor Natasa Miskov-Zivanov and her latest achievement, the NSF CAREER Award," said Alan George, Department Chair, R&H Mickle Endowed Chair, and professor of electrical and computer engineering. "She leads the MeLoDy (Mechanisms and Logic of Dynamics) Laboratory, where her research spans a broad range of cutting-edge topics by leveraging her expertise in digital circuits, synthetic biology, artificial intelligence, and dynamic systems. Natasa's new project promises to significantly advance the field of immunotherapy design. She is an innovator in both the laboratory and the classroom, and I am so excited about her future as a rising star in the field."