Abstract
Virtual tumors—computer simulations that incorporate a patient's genomic information—can help personalize cancer care. They have already been used to predict patient responses to drug treatments. In particular, virtual tumors have identified which patients are likely to respond to checkpoint inhibitors, drugs meant to prevent cancers from overriding host immune checkpoints.
Predicting patient responses to checkpoint inhibitors can not only help patients get started on a helpful therapy more quickly, they can avoid unnecessarily subjecting patients to drugs that have toxic effects.
iStock / wildpixel
With these potential benefits in mind, researchers based at The University of Iowa College of Dentistry developed a computational technique that predicted immune checkpoint ligand PD-L1 and other immunosuppressive biomarker responses. The computational technique's predictions were subsequently validated by laboratory experiments.
The computational technique was detailed in a presentation (“Predicting Non-Responders to Immunotherapy Treatments through Simulation of NGS Information”) during the 57th American Hematological Society Annual Meeting and Exposition, which took place last month in Orlando, FL.
The technique involves taking genetic information from a cancer cell, importing it to the simulation model, and predicting the response that the cell would have to a particular treatment. Next, live cancer cells are taken and grown in the laboratory. Then the actual response to the identical treatment is determined.
If researchers get the same results from both experiments, they have a match. The cells growing in the laboratory have verified that the computer model works. If they give different results, then researchers have a mismatch, meaning the simulated model and lab tests are not in agreement and need to be aligned.
“We developed an immunotherapy response phenotype, which is a function of biomarkers representing immune evasion, immune activation, metastasis, and dendritic cell infiltration in cancer cells,” revealed the presentation's abstract. “We then modeled two myeloma cell lines using available genomics information as proxy for myeloma patient cancer. Finally, we predicted these two representative patient classes would vary in response to PD-L1/PD-1 inhibitors. We validated our prediction that biomarkers contribute to the immunotherapy response.”
This presentation resulted from cooperation between the University of Iowa and Cellworks Group, a private company that works to personalize cancer treatment by developing virtual tumors based on a person's genetic profile. The collaborators noted that their technology is timely, particularly to pharmaceutical partners who want to test their cancer drugs using these simulated models.
“Our goal is to develop a very patient-specific workflow that could be used early after cancer diagnosis to aid in the identification of effective cancer treatments,” said Kim Alan Brogden, Ph.D., Director of the Dows Institute for Dental Research at the University of Iowa College of Dentistry.
The simulation and laboratory models also allow for the screening of combination treatments, which could involve more than one immunotherapeutic agent or a combination of immunotherapeutic and chemotherapeutic agents.
The researchers say that they hope their work leads to a personalized medicine approach that will save treatment time, cut costs, and improve long-term prognoses for cancer patients.