Department of Otolaryngology researchers in the Rocco Lab along with pathology experts at The Ohio State University Comprehensive Cancer Center – Arthur G. James Cancer Hospital and Richard J. Solove Research Institute (OSUCCC – James) are collaborating on head and neck cancer research by integrating artificial intelligence (AI) with groundbreaking tumor modeling techniques. This work, led by Bhavna Kumar, MS, program director; Edmund Mroz, PhD, research scientist in the Department of Otolaryngology – Head and Neck Surgery; and Anil Parwani, MD, PhD, chair and clinical professor of Pathology, is poised to revolutionize cancer diagnosis and treatment while paving the way for breakthroughs in other cancer types.
Leveraging AI for biomarker detection
A key innovation of the research team involves using AI to analyze biomarkers, including estrogen receptor alpha (ERα), which has been linked to improved survival rates of head and neck cancers associated with the human papillomavirus (HPV).
“The observation that estrogen receptor alpha is expressed in head and neck tumors was very novel and interesting,” Kumar, the project coordinator, says. “What’s even more intriguing is that patients with estrogen receptor alpha–positive tumors seem to have better outcomes and survival rates.”
The team analyzed more than 300 tumor samples using an AI algorithm initially designed for breast cancer. While the algorithm successfully identified ERα expression, the researchers are refining it to enhance accuracy for head and neck tumors.
The team employs a rigorous validation process, which combines AI-generated results with manual results found by head and neck pathologists Abberly Lott Limbach, MD, at The Ohio State University Wexner Medical Center and William Faquin, MD, PhD, at Mass General Brigham, to ensure accuracy. They also are exploring AI’s potential to detect other key biomarkers, such as p16, an HPV indicator, with the goals of understanding the spatial relationships among tumor cells and integrating AI-assisted diagnostics into clinical practice.
“AI could transform how we diagnose and treat head and neck cancers,” Dr. Mroz says, emphasizing the technology’s potential to streamline pathology workflows and enable more precise decision-making.
Kumar adds, “The incorporation of digital AI in pathology is still in its infancy, but the potential is tremendous. By training AI algorithms on our growing database of digitally scanned tumor samples, we hope to accelerate analysis and uncover new insights to guide clinical decision-making.”
Advancing tumor research with living cell models
In parallel, researchers are exploring innovative methods to study tumors at the cellular level. Dr. Kumar also leads a project cultivating living tumor cells from fresh patient samples. Unlike fixed-tissue samples, which provide static snapshots, living cell models allow researchers to observe tumor behavior and evolution over time.
Tumor cells (typically from epithelial tumors found in the skin or the mouth lining) in tissue culture dishes are grown in a controlled environment with fibroblasts, which support epithelial cell proliferation, further enhanced by adding the drug Y-27362. When cells start to fill a dish following repeated division, some are “passaged” into a fresh dish for extended proliferation to produce large quantities for study.
Researchers found that adding a second drug targeting a separate pathway significantly improves the success rate, enabling tumor cell growth indefinitely with an 80% success rate.
However, Dr. Mroz points out that the technique they use wasn’t originally developed in their lab, but was adopted and improved. “We added an important step, a process that made a big difference,” he says.
Researchers also use xenograft models, implanting tumor cells into mice to examine growth patterns and test therapies. This approach has been especially valuable to study rare cancers like adenoid cystic carcinoma, a slow-growing yet fatal tumor.
“The goal is to understand how these cultured cells compare to the original tumor,” Dr. Mroz says. “We’re looking to determine why this happens and how the biology of the tumor cells might change over time.”
Bridging technology and precision for better cancer treatment
The OSUCCC – James’ combination of AI analytics and living tumor models lays the groundwork for personalized, targeted cancer treatments. The researchers are expanding these techniques to other tumor types and leveraging AI to analyze complex genetic and clinical data.
“We’re not just focused on effectiveness, but on finding therapies that are less debilitating for patients,” Dr. Mroz says, emphasizing the need to minimize the long-term side effects of traditional treatments like chemotherapy and radiation.
Dr. Mroz also highlights the importance of collaboration and data sharing through resources like The Cancer Genome Atlas (TCGA). “Access to shared data allows researchers to validate findings and accelerate breakthroughs,” he says.
This is the future of cancer research: a fusion of advanced technology, innovative techniques and extensive collaboration.