On July 24, Dr. Jae Lee, a bioinformatics statistician, and Dr. Dan Theodorescu, an oncologist, announced a breakthrough algorithm (calculation method) that will predict success rates for cancer patients considering treatment options. C-VILLE spoke with Dr. Lee about this research and how it might change cancer therapy.
 Dr. Jae Lee and a colleague at UVA have figured out a better way to design cancer treatments, possibly saving the lives of critically sick patients. |
C-VILLE: What does this mean for cancer patients?
Dr. Jae Lee: Usually doctors prescribe some combination of chemotherapeutic treatment, but often doctors don’t always know which combination is best for each patient. It’s really like a coin toss or comes down to individual preferences, or—I have to say—bias. We want to use this [algorithm] so that we can nearly tailor those combinations to each individual patient.
Also, we want to test with patients in critical stages. Usually patients who haven’t responded to the first treatment of chemotherapy come into more critical stages of cancer, and usually, the survival rate at that stage really plummets. With all the FDA-approved compounds, our technology can determine the best combination for that patient, and there’ll be much better hope for those in critical stages.
Can you explain the algorithm?
It’s based on several big databases, the first one being the National Cancer Institute (NCI) drug-screening database. They’ve been doing drug-screening experiments on a panel of 60 cancer cell lines, consisting of nine cancer types. They treat each cell with a varying concentration of different compounds to see if the cancer cell doesn’t grow or dies. The second one is called gene-expression profiling. You’ve heard about the Human Genome Project. Thanks to that, we’ve nearly mapped out all human genes, about 40,000.
The question is: How can we predict the chemosensitivity of individual patients? We are trying to make a translation between the two systems, and we now know of the 100,000 compounds which of the 60 cell lines are sensitive, and which are not. Since we have all that information for each compound, we can connect the genome information from the patient. Using this type of Rosetta Stone, we can translate the information from the invitro system to the invivo system.
Does it take a long time to determine which works the best?
That’s really the advantage about our kind of technology. NCI already has screened the 68 FDA-approved compounds, so the only thing we need is the expression profiling from each patient’s tumor. So if the patient comes in, like almost for a regular visit, and the doctor gets the biopsy samples from the patient, then that’s it. Instead of running the traditional lab testing, we will run genomic-expression profiling.
C-VILLE welcomes news tips from readers. Send them to news@c-ville.com.