Computer models recommending drugs for cancer treatments
Posted by evolvingwheel on August 5, 2007
Often, two patients with identical types of cancers respond differently to the same drug or the same treatment. This is caused by different genetic histories of two individuals. However, if the medical professional knew about a predictive system that could take into account the various genetic information, tissue characteristics, and reactions to more than 100,000 cancer treating compounds and accurately design a treatment plan, long-term care for cancer patients would have been so much promising.
Dan Theodorescu and his researchers at the University of Virginia have exactly developed a computer model that addresses such expectation. This one of a kind model has been developed using a database of information on human can-cell lines and their responses to different anti-cancer drugs and compounds. In order to articulate a better treatment path, the model also processes gene-expression analysis of the cancer types. As more and more data gets available about the gene characteristics, the model gets more accurate in predicting the responses to any particular treatment regimen.
One other interesting thing is the use of this model in clinical trials of a specific drug. The model selects patients who could be appropriate for one particular drug. This method will greatly benefit drug development in the coming days. Pharmaceutical companies can now target several subsets of patients and develop customized drugs with much less investments in large-scale clinical analysis. The model will also mitigate investment risks associated with the potency of the drug among a wide range of patients. Furthermore, doctors and researchers, who often sit on a goldmine of gene information, can adequately provide a drug development path along with a treatment path.
Read [article] here.
Picutre: Dan Theodorescu – courtesy UVA