Decoding cellular conversations to stop oral cancer before it starts

Researchers collaborate to leverage artificial intelligence to identify the exact moment when cells go wrong, leading to cancer.

Over the course of human history decoders have evolved as invaluable tools for intercepting enemy messages, preventing destructive encounters, deadly disasters and shortening major battles.

Now, an OHSU School of Dentistry researcher is poised to take the decoder concept to the next level. She will use computer models and tissue chips as cellular codebreakers to intercept encrypted messages cells send to each other.

The outcome could revolutionize the way oral cancer is diagnosed and prevented.

“We’re going to stop cancer before it becomes cancer,” says Cristiane Miranda França, D.D.S., M.S., Ph.D., an assistant professor in biomaterial and biomedical sciences in the school.

Ultimately, the goal is to stop the progression of the disease through a combination of artificial intelligence, reverse engineering and laboratory modeling.

Millions in a moment

França secured a two-year pilot grant to team up with Jake Searcy, Ph.D. and Paul Dalton, Ph.D., from the University of Oregon to build artificial intelligence learning models. The initial aim is to increase the accuracy of oral cancer diagnoses.

The model will analyze thousands to millions of genes at once from patient’s biopsies, allowing for the identification of specific cellular parallels and proteins in immune cells that predict whether a lesion is likely to progress to cancer.

“We want AI to help us define hot spots that we don’t see as pathologists or with our human eyes, or the patterns or the correlations,” says França, who is also a researcher in the Knight Cancer Precision Biofabrication Hub.

Medical and dental providers sometimes encounter borderline lesions that are difficult to definitively classify as malignant or benign. Assistance from a machine-learning algorithm will refine this decision-making process, according to França.

The intent isn’t to replace human analysis; but rather to provide a tool that allows providers and researchers to work faster and more accurately.

Personalized medicine

The research signals advances in personalized medicine, where treatments can be tailored to a patient’s specific immunological background.

The potential benefits for patients are:

Early intervention: Move beyond diagnosis to find therapeutic targets allowing for early interception.

Predictive analysis: Provide patients with a reliable timeline and risk assessment for whether a lesion will become malignant.

Non-invasive alternatives: Understand cell-to-cell communication to develop drug-based therapies to stop cancer progression. This will allow a less invasive and more efficient alternative to surgery, which often fails to prevent cancer from reappearing in other areas of the mouth.

Decoding cellular conversations

Franca and her collaborators anticipate diving deeply into the problem; they will come to understand how oral cells and immune cells communicate with one another.

They will decode the cell language to identify the exact moment the conversation goes wrong, allowing cells to escape the body’s natural surveillance and become cancerous. By understanding the signals, specifically the proteins in immune cells and the crosstalk with the immune system, França and her peers will predict the progression of lesions.

Collaboration

The unique project brings together investigators from OHSU and the University of Oregon who have not previously worked together or who have not worked in cancer. The academic researchers will share their knowledge using artificial intelligence, spatial gene analysis and organs-on-a-chip to understand how the immune system influences the earliest stages of oral cancer.

The pilot project is funded by the Center for Biomedical Data Science as part of its CBDS Collaborative Project Awards, providing seed funds for biomedical data science research between UO and OHSU.

---

Story by Rhonda Morin, APR

Photo above, Cristiane Miranda França, D.D.S., M.S., Ph.D., in the Skourtes Tower lab. Photo by OHSU.

Enigma cipher coding machine by Getty Images.

Generative artificial intelligence tools were used to identify themes and summarize the research in this story. Editors reviewed and approved it