Algorithms can help beat cancer in a new approach to personalized medicine.
Researchers at Dana-Farber Cancer Institute have developed a mathematical model to help predict tumor growth and how it’s likely to behave given different treatment options.
The study analyzed breast cancer samples from 47 patients undergoing pre-operative chemotherapy to shrink their tumors. Biopsy samples were taken at diagnosis and again when chemotherapy was completed.
Tumors contain a variety of cancer cells that change over time, known as tumor heterogeneity. The researchers looked at the changing gene information to help pinpoint the best treatments. They found gene location within a tumor and heterogeneity changes affected patient outcomes.
After analyzing the information they found a few trends:
- Cancer treatments with zero or partial response had few changes in how many copies of DNA segment are present.
- Tumors with less genetic diversity respond better to treatment than tumors with more genetic complexity
- Cancer cells that grow more rapidly seem to be easier to kill with treatment
Physicians in the near future can run a gene sample and target the best treatments for their patients, potentially improving patient outcomes.
Oncolytic viruses can get a math assist
Researchers in Ottawa have also been using mathematical models to predict which tumors will respond to “cancer killing viruses.” Oncolytic viruses can be introduced into a cancer patient. The virus will target cancer cells while leaving healthy cells alone. The results can be dramatic but only if the cells are susceptible to the infection.
Dr. Mad Kaern, Canada Research Chair at the University of Ottawa Institute of Systems Biology said, “By using mathematical models to predict how viral modifications would actually impact cancer cells and normal cells, we were able to accelerate the pace of research. It allows us to quickly identify the most promising approaches … something that is usually done through expensive and time-consuming trial and error.”
What do you think about mathematical modeling to improve cancer outcomes? How about the layering of algorithms with genetic testing? We’ll be talking about this in more detail inside Sermo, if you’re an M.D. or D.O. please join us.
Leave a Reply