The grant will go toward developing a computer model that will consider personal and genetic factors when determining the most effective use of cancer drugs.
The survival rate for those inflicted with more forms of cancer has never been higher than it is today. Thanks to advancement in drug therapy and medical technology, the human race is closer to curing cancer than ever before.
But there is still much work to be done, because cancer hasn't been completely eradicated just yet. A grant from the National Institutes of Health awarded to a Texas Tech University engineering professor could be a huge step in that direction.
Ranadip Pal, an associate professor in the Department of Electrical & Computer Engineering in the Whitacre College of Engineering, and Souparno Ghosh, an assistant professor in the Department of Mathematics and Statistics in the College of Arts and Sciences, received a grant totaling $641,745 over the next three years that will go toward developing a computer model that can predict the effectiveness of a cancer drug or combination of drugs depending on a variety of genetic and cellular factors unique to each individual.
This will allow doctors in the future who treat cancer patients to develop personalized methods of treatment and give the patient the best chance to survive cancer.
“This can be extremely beneficial because we are looking for personalized therapy for cancer, which is still kind of the Holy Grail of cancer research,” Pal said. “We've been searching for two or three decades and haven't found it yet. Yes, we have improved overall cancer treatment and response to different drugs, and with a lot of cancers these days, the survival rates are really good, but there is still a lot that can be done.”
The research will involve taking biological data obtained through biopsies of various cancer patients and seeing how their various forms of cancer reacted to a certain drug or combination of drugs. But because there are so many types of cancer, as well as thousands of drugs used to treat cancer, doing so experimentally would take too much time and be prohibitively expensive.
But by narrowing the number of drugs and using a wide variety of cancer cells, Pal is hoping to get enough of a range where he can develop a computer model that will predict how cancer cells will react to a certain drug or combination of drugs, and then extrapolate responses for other drugs based on the model.
There are other factors, however, that must be taken into consideration, such as a person's DNA or gene expression, protein expression and metabolism. Because each person's genetic factors are different, Pal said it's like considering each patient having his or her own personal type of cancer.
That's where the computer modeling comes into play, giving researchers a response curve that gives them a clearer picture of how a drug reacts, taking into account all medical and biological factors instead of just one particular response along that curve.
“Previously we used to just take the median or the average response,” Pal said. “Now what we can say is we can predict the whole thing. We are coming up with the mathematical framework that will allow us to do this.”
The model also will be able to determine how certain cancer drugs would work based on various genetic mutations, or biomarkers, such as DNA brand and metabolism. Then drugs can be tailored to take those biomarkers into account in order to be most effective.
“If we can integrate all of the biomarkers together, then we get the complete picture rather than just looking at genetics,” Pal said.
Pal feels the initial data set of drugs and types of cancer need to be as widespread as possible to broaden the mathematical predictions. The hope is the model will help doctors not only treat patients with hopes of extending their life but also improve the quality of that life as well in an effort to rid the world of cancer altogether.
“We are primarily working with targeted drugs whose overall side effects are much less compared to chemotherapy where the side effects are so high the quality of life decreases,” Pal said. “If we are successful, we can say that for any person based on the tumor, we will be able to give a much more predictive model which can say what combination of drugs is suitable for that, and we can extrapolate that to predict toxicity. We can give a much better quality of life to that person rather than just trying to save a person or extend life span. The main goal is to extend the life span with a good quality of life.”
Pal is in the stage of taking the initial data and building a model to study its performance and whether what they've seen in the preliminary data holds true for bigger data sets. He also has enlisted the help of a medical consultant who will look at what else is needed from a biological perspective and what needs to be incorporated, and what else may need to be incorporated into the mathematical model that hasn't yet been considered.
He also is hoping the model can prove useful in other areas outside of this particular study.
“This is more of a computational framework that can be applied to any other engineering problem or where you are trying to predict a huge search space and it's not feasible to do so experimentally,” Pal said.
Pal also has used this method of research to publish a book entitled “Predictive Modeling of Drug Sensitivity,” published by Elsevier Academic Press. The book gives an overview of drugs sensitivity modeling for personalized medicine that includes data characterizations, modeling techniques, applications and research challenges.
The book includes mathematical techniques used for modeling drug sensitivity as well as biological knowledge needed to guide a user to apply the mathematical tools in different biological scenarios.
“This book is targeted towards computational biology and medical researchers who are interested in applying mathematical and computational tools to analyze genomic and functional data for personalized therapies,” Pal wrote. “This book can serve as a text or reference book for introductory courses on computational biology for drug sensitivity prediction.”